From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb  3 15:35:19 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id PAA07721
	for <iptel-archive@odin.ietf.org>; Wed, 3 Feb 1999 15:35:18 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id CC2CE7FB7; Wed,  3 Feb 1999 14:43:27 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DA7AA7FC8; Wed,  3 Feb 1999 14:43:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Wed, 03 Feb 1999 16:43:34 -0300
From: Baisys Sistemas <exin@cvtci.com.ar>
Subject: Bases de emails y hardware
Reply-To: exin@cvtci.com.ar
To: iptel@lists.research.bell-labs.com
X-Mailer: dbMail v1.3b1 by Mach5 Software. Please report spamming to dbspam@mach5.com.
Message-ID: <067f00040190329MAIL2@cvtci.com.ar>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Atención!!! servicios para PYMEs:

Baisys Sistemas ofrece a las pequeñas y medianas empresas una solución
integral en materia de equipamiento en cuanto a instalaciones y
mantenimiento de equipos, redes, Internet y comunicaciones.

Instalación, Configuración y Mantenimiento de Redes

· Cableado alimentación 220/110 V c/tierra
· Cableado par trenzado (RJ45) o coaxil (RG58)
· Canalización con cable-canal y piso-canal
· Instalación de cajas y periscopios
· Instalación de hubs
· Instalación de impresoras y modems de red
· Sistemas de alimentación ininterrumpible (UPS)
· Instalación de sistemas operativos Windows 98/95/NT o Novell
· Configuración de recursos compartidos
· Administración y Políticas de seguridad
· Sistemas de backup (CD, Tape Backup, Zip, Jaz)

Servicio Técnico y Mantenimiento de Equipos

Abonos mensuales de mantenimiento que incluyen:
· Servicio técnico a domicilio sin cargo ante falla de equipos, caídas de
  red y otras eventualidades
· Configuración de software
· Copias de seguridad en CD
· Limpieza periódica de equipos, teclados e impresoras

Servicios de Internet

Asesoramiento en materia de conexión a Internet, correo electrónico,
conexiones y configuración de redes para acceso simultáneo
desde varios equipos.



***********  GRAN OFERTA DE FEBRERO 1999:

OFREZCA SUS PRODUCTOS Y SERVICIOS POR EMAIL!!!!
CD ROM CON 44.000 EMAILS DE ARGENTINA 
para publicidad por email $ 300.00.-

Mouse Genius.............................................. $      4.30.-
Mouse genérico........................................... $      2.90.-
Parlantes 120 watts...................................... $      7.85.-
Micrófono con pié....................................... $      2.50.-
Disco rígido 5.1 GB UDMA........................ $    157.00.-
Módem 56 KB Po......................................... $     33.10.-
Gabinete mini tower fuente 250 w.............. $     23.15.-
Monitor 17 pulgadas Digital....................... $    318.00.-
Placa de red COMBO genérica................... $     12.40.-
Grabadora CD Mitsubishi 6 x 2 IDE........... $    405.00.-
Impresora LEXMARK Chorro de tinta...... $    136.35.-
Scanner FULL PAGE Color......................... $     66.15.-

Por otros artículos, consúltenos!!!!
Precios de contado más IVA.

Baisys Sistemas.
Florida 165 - Piso 14 - Of. 1409 - (1333) Capital Federal.
Teléfono : 4345-3801. Lunes a viernes de 10 a 19 hs.

E-mail : exin@cvtci.com.ar  Web Page : http://members.spree.com/sip.baisys

Este mensaje llega a usted desde una lista de distribución.
Su incorporación a la misma fue sugerida por alguno de nuestros clientes.

Si por cualquier motivo NO DESEA RECIBIR esta lista periódicamente, por
favor háganoslo saber enviando un mensaje con la palabre ELIMINAR.




---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb  5 18:09:29 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id SAA12359
	for <iptel-archive@odin.ietf.org>; Fri, 5 Feb 1999 18:09:28 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7F0F77FB7; Fri,  5 Feb 1999 17:59:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D1E557FB8; Fri,  5 Feb 1999 17:59:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: TERADYNE
From: "Kevin Davis" <Kevin_Davis@notes.teradyne.com>
To: iptel@lists.research.bell-labs.com
Message-ID: <8525670F.007E6B39.00@notes.teradyne.com>
Date: Fri, 5 Feb 1999 17:42:37 -0500
Subject: pre-encoded audio
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

For testing purposes, I wish to pre-encode a speech clip
as G.723  (and maybe G.729) and then I will put it into RTP
payload without the need for real-time encoder.
The typical off-the-shelf audio tools do not support this.
Can anyone point me to a tool that generate VoIP audio ?



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb  5 18:28:36 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id SAA12886
	for <iptel-archive@odin.ietf.org>; Fri, 5 Feb 1999 18:28:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3B4177FB8; Fri,  5 Feb 1999 18:19:39 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 959027FC8; Fri,  5 Feb 1999 18:19:38 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36BB7C89.E2316D85@cs.columbia.edu>
Date: Fri, 05 Feb 1999 18:19:37 -0500
From: Henning Schulzrinne <hgs@cs.columbia.edu>
X-Mailer: Mozilla 4.04 [en] (X11; I; SunOS 5.6 sun4u)
MIME-Version: 1.0
To: Kevin Davis <Kevin_Davis@notes.teradyne.com>
Cc: iptel@lists.research.bell-labs.com
Subject: Re: pre-encoded audio
References: <8525670F.007E6B39.00@notes.teradyne.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Kevin Davis wrote:
> 
> For testing purposes, I wish to pre-encode a speech clip
> as G.723  (and maybe G.729) and then I will put it into RTP
> payload without the need for real-time encoder.
> The typical off-the-shelf audio tools do not support this.
> Can anyone point me to a tool that generate VoIP audio ?

Easy: just use rtpdump to record the output from (say) NetMeeting. See
http://www.cs.columbia.edu/~hgs/rtpdump.

> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Henning Schulzrinne   http://www.cs.columbia.edu/~hgs

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Feb  8 03:55:19 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id DAA15619
	for <iptel-archive@odin.ietf.org>; Mon, 8 Feb 1999 03:55:18 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id AEEA57FB7; Mon,  8 Feb 1999 03:01:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0F5077FB8; Mon,  8 Feb 1999 03:01:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36BE99B3.8981CBE2@cnet.francetelecom.fr>
Date: Mon, 08 Feb 1999 09:00:51 +0100
From: "Stephane GORSE - FT.BD/CNET/DAC/ARP" <stephane.gorse@cnet.francetelecom.fr>
X-Mailer: Mozilla 4.05 [en] (X11; I; SunOS 5.5.1 sun4m)
MIME-Version: 1.0
To: iptel@lists.research.bell-labs.com
Subject: Point-to-point call to  multipoint conference.
Content-Type: multipart/alternative; boundary="------------957CB67FF938C798CEAD58AA"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


--------------957CB67FF938C798CEAD58AA
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi,

I have a question related to the expansion of a point-to-point call to
a multipoint conference in
the gatekeeper call signalling routing in the H323 recommandation.

Make the assumptions that an endpoint A creates a point-to-point call
with an endpoint B. Using the
aforementionned mode, all signalling messages transit through the
gatekeeper, but no MCU is used in
this call.

Now, I want to know if it is possile to extand this call to a
multipoint conference using a MCU by the
gatekeeper initiative. Currently, we see a lot gatekeeper with other
interfaces than H323 ones
(like INAP interface...). So, it is conceivable to think that, in
Internet/IN synergy context, a gatekeeper
can take the initiative in expanding a call.

Following this scenario, what is the message exchange between each
endpoint and the gatekeeper
(including RAS, Q931 and H245) ?

In the Recommandation, the paragraph 8.4.3 explains an extansion but
in this scenario includes a MCU
with the point-to-point call.


--

-----------------------------------------------------------------
        Stephane Gorse - France Telecom BD/CNET/DAC/ARP
            stephane.gorse@cnet.francetelecom.fr
-----------------------------------------------------------------
  _/_/_/_/ Tel: +33 2 96 05 03 31  | Technopole Anticipa
 _/_/_/_/  Fax: +33 2 96 05 37 84  | 2, av. Pierre Marzin
_/_/_/_/                           | 22307 Lannion Cedex - France



--------------957CB67FF938C798CEAD58AA
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<HTML>
Hi,

<P>I have a question related to the expansion of a point-to-point call
to a multipoint conference in
<BR>the gatekeeper call signalling routing in the H323 recommandation.

<P>Make the assumptions that an endpoint A creates a point-to-point call
with an endpoint B. Using the
<BR>aforementionned mode, all signalling messages transit through the gatekeeper,
but no MCU is used in
<BR>this call.

<P>Now, I want to know if it is possile to extand this call to a multipoint
conference using a MCU by the
<BR>gatekeeper initiative. Currently, we see a lot gatekeeper with other
interfaces than H323 ones
<BR>(like INAP interface...). So, it is conceivable to think that, in Internet/IN
synergy context, a gatekeeper
<BR>can take the initiative in expanding a call.

<P>Following this scenario, what is the message exchange between each endpoint
and the gatekeeper
<BR>(including RAS, Q931 and H245) ?

<P>In the Recommandation, the paragraph 8.4.3 explains an extansion but
in this scenario includes a MCU
<BR>with the point-to-point call.
<BR>&nbsp;
<PRE>--&nbsp;

-----------------------------------------------------------------
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Stephane Gorse - France Telecom BD/CNET/DAC/ARP
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; stephane.gorse@cnet.francetelecom.fr
-----------------------------------------------------------------
&nbsp; _/_/_/_/ Tel: +33 2 96 05 03 31&nbsp; | Technopole Anticipa
&nbsp;_/_/_/_/&nbsp; Fax: +33 2 96 05 37 84&nbsp; | 2, av. Pierre Marzin&nbsp;
_/_/_/_/&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; | 22307 Lannion Cedex - France</PRE>
&nbsp;</HTML>

--------------957CB67FF938C798CEAD58AA--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Feb  8 09:53:59 1999
Received: from dirty.research.bell-labs.com (dirty.research.bell-labs.com [204.178.16.6])
	by ietf.org (8.8.5/8.8.7a) with SMTP id JAA18030
	for <iptel-archive@odin.ietf.org>; Mon, 8 Feb 1999 09:53:59 -0500 (EST)
Received: from lists.research.bell-labs.com ([135.180.161.172]) by dirty; Mon Feb  8 09:52:42 EST 1999
Received: by lists.research.bell-labs.com (Postfix)
	id 5E94B7FC8; Mon,  8 Feb 1999 09:31:34 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BCA917FC9; Mon,  8 Feb 1999 09:31:33 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Cc: iptel@lists.research.bell-labs.com
Message-ID: <36BEF4C0.C52E9811@lucent.com>
Date: Mon, 08 Feb 1999 09:29:20 -0500
From: Binshi Cao <bcao@lucent.com>
Organization: Lucent Technologies
X-Mailer: Mozilla 4.05 [en]C-EMS-1.4  (Win95; U)
MIME-Version: 1.0
To: Kevin Davis <Kevin_Davis@notes.teradyne.com>
Original-CC: iptel@lists.research.bell-labs.com
Subject: Re: pre-encoded audio
References: <8525670F.007E6B39.00@notes.teradyne.com>
Content-Type: multipart/mixed; boundary="------------58129083307B6104A5B6AB55"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.
--------------58129083307B6104A5B6AB55
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

You may go to ITU web site to get the test vectors for the original speech,
coded packets, and decoded speech.

Binshi Cao.

Kevin Davis wrote:
> 
> For testing purposes, I wish to pre-encode a speech clip
> as G.723  (and maybe G.729) and then I will put it into RTP
> payload without the need for real-time encoder.
> The typical off-the-shelf audio tools do not support this.
> Can anyone point me to a tool that generate VoIP audio ?
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
--------------58129083307B6104A5B6AB55
Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Binshi Cao
Content-Disposition: attachment; filename="vcard.vcf"

begin:          vcard
fn:             Binshi Cao
n:              Cao;Binshi
org:            jk12eaj00
adr:            Room 2C-122, Lucent Technologies;;600, Mountain Av.;Murray Hill;NJ;07974;USA
email;internet: bcao@lucent.com
title:          Member of Technical Staff
tel;work:       (908) 582 5824
tel;fax:        (908) 582 6729 
tel;home:       (908) 541 9774
x-mozilla-cpt:  ;0
x-mozilla-html: FALSE
version:        2.1
end:            vcard


--------------58129083307B6104A5B6AB55--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Feb  8 10:51:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id KAA19506
	for <iptel-archive@odin.ietf.org>; Mon, 8 Feb 1999 10:51:53 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3B7D97FB8; Mon,  8 Feb 1999 09:11:52 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 043A37FC8; Mon,  8 Feb 1999 09:11:50 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <C51ED84B6F47D211917A0000F8BCBD11BA196F@zcard00g.ca.nortel.com>
From: "Tom-PT Taylor" <taylor@nortelnetworks.com>
To: iptel@lists.research.bell-labs.com
Subject: RE: Point-to-point call to multipoint conference.
Date: Mon, 8 Feb 1999 08:00:58 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.0.1460.8)
Content-Type: text/plain; charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by ietf.org id KAA19506

The Gatekeeper would reroute all the connections through the MCU by using
H.245 to close the existing logical channels and replace them with the new
ones.  Then it could go on to expand the conference.

> -----Original Message-----
> From:	Stephane GORSE - FT.BD/CNET/DAC/ARP
> [SMTP:stephane.gorse@cnet.francetelecom.fr]
> Sent:	Monday, February 08, 1999 3:01 AM
> To:	iptel@lists.research.bell-labs.com
> Subject:	Point-to-point call to multipoint conference.
> 
> Hi, 
> 
> I have a question related to the expansion of a point-to-point call to a
> multipoint conference in 
> the gatekeeper call signalling routing in the H323 recommandation. 
> 
> Make the assumptions that an endpoint A creates a point-to-point call with
> an endpoint B. Using the 
> aforementionned mode, all signalling messages transit through the
> gatekeeper, but no MCU is used in 
> this call. 
> 
> Now, I want to know if it is possile to extand this call to a multipoint
> conference using a MCU by the 
> gatekeeper initiative. Currently, we see a lot gatekeeper with other
> interfaces than H323 ones 
> (like INAP interface...). So, it is conceivable to think that, in
> Internet/IN synergy context, a gatekeeper 
> can take the initiative in expanding a call. 
> 
> Following this scenario, what is the message exchange between each
> endpoint and the gatekeeper 
> (including RAS, Q931 and H245) ? 
> 
> In the Recommandation, the paragraph 8.4.3 explains an extansion but in
> this scenario includes a MCU 
> with the point-to-point call. 
>   
> -- 
> 
> -----------------------------------------------------------------
>         Stephane Gorse - France Telecom BD/CNET/DAC/ARP
>             stephane.gorse@cnet.francetelecom.fr
> -----------------------------------------------------------------
>   _/_/_/_/ Tel: +33 2 96 05 03 31  | Technopole Anticipa
>  _/_/_/_/  Fax: +33 2 96 05 37 84  | 2, av. Pierre Marzin 
> _/_/_/_/                           | 22307 Lannion Cedex - France
>  

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Feb  8 13:37:37 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id NAA26285
	for <iptel-archive@odin.ietf.org>; Mon, 8 Feb 1999 13:37:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 30D3B7FC9; Mon,  8 Feb 1999 13:27:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8957D7FCA; Mon,  8 Feb 1999 13:27:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Ami Amir" <amir@radvision.rad.co.il>, <ming.lu@globalone.net>,
        <epabhm@epa.ericsson.se>
Cc: "Sebastien Ducarne" <Sebastien.Ducarne@globalone.net>,
        "IP Telephony" <iptel@lists.research.bell-labs.com>
Subject: RE: VOIP question ??
Date: Mon, 8 Feb 1999 13:24:25 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFKENJEBAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
In-Reply-To: <99Jan28.124759ist.113795@firewall.radvision.rad.co.il>
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Importance: Normal
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Ami Amir
> Sent: Thursday, January 28, 1999 5:42 AM
> To: ming.lu@globalone.net; epabhm@epa.ericsson.se
> Cc: Sebastien Ducarne; IP Telephony
> Subject: RE: VOIP question ??
>
>
> Hi All
>
> Once you walk through and analyze the situation, you will
> find out that a
> Gatekeeper is really needed.
>
Not really. A GK is just a RAS engine and RAS is just a packaging of
functionality that could just as easily be provided by judicious use of
existing protocols like DNS, LDAP, RADIUS, etc.

> An example of one issue -dialing schemes: If you are willing
> to support
> multiple stage dialing or assign a rigid dialing scheme (e.g.
> prefix 6 for
> PBX A, prefix 7 for PBX B etc) then you can make do w/o
> gatekeepers - and
> sacrifice functionality, mobility and flexibility.
>
Hmm, let's see. Most of this I could get via an LDAP query. The rest I
need a real routing protocol for anyway, and gatekeepers don't run
routing protocols
(yet).

> Another issue is economical: If you have more than 3 PBX's -
> you have two
> options; either create a mesh, and let each PBX support as
> many trunks as
> there are PBX's (= lots of hardware) - or support a flexible
> scheme, where a
> single PBX to GW trunk can support connections to multiple
> GWs. This, less
> expensive, functionality requires a GK.
>
Huh? This is IP. The IP network provides the full mesh.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb  9 10:51:06 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id KAA20500
	for <iptel-archive@odin.ietf.org>; Tue, 9 Feb 1999 10:51:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B8E797FB7; Tue,  9 Feb 1999 10:35:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 19E0C7FC8; Tue,  9 Feb 1999 10:35:24 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <CA6966C24AC6D111B5A100805FEAB7D85773FE@nsrip00208.mcit.com>
From: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
To: confctrl@ISI.EDU, iptel@lists.research.bell-labs.com
Subject: draft-ietf-mmusic-sip-info-method-00.txt
Date: Tue, 9 Feb 1999 15:32:58 -0000 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/mixed;
	boundary="----_=_NextPart_000_01BE5441.78D61FBE"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_000_01BE5441.78D61FBE
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE5441.78D61FBE"


------_=_NextPart_001_01BE5441.78D61FBE
Content-Type: text/plain

This message is notification of the following Internet Draft which has been
submitted to the IETF for posting.

<draft-ietf-mmusic-sip-info-method-00.txt> 

Abstract: 

This document proposes an extension to the Session Initiation Protocol.
This extension adds the INFO method to the SIP protocol.  The intent
of the INFO method is to allow for the carrying of session related 
control information that is generated during a session.  Examples of
such session control information are ISUP/ISDN signaling messages
and DTMF digits used to control telephony services.

The internet draft is attached.

Steve Donovan
MCIWorldcom

 <<draft-ietf-mmusic-sip-info-method-00.txt>> 

------_=_NextPart_001_01BE5441.78D61FBE
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>draft-ietf-mmusic-sip-info-method-00.txt</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">This message is =
notification of the following Internet Draft which has been submitted =
to the IETF for posting.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 =
FACE=3D"Fixedsys">&lt;draft-ietf-mmusic-sip-info-method-00.txt&gt; =
</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">Abstract:<BR>
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">This document =
proposes an extension to the Session Initiation Protocol.</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">This extension =
adds the INFO method to the SIP protocol.&nbsp; The intent</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">of the INFO =
method is to allow for the carrying of session related </FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">control =
information that is generated during a session.&nbsp; Examples =
of</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">such session =
control information are ISUP/ISDN signaling messages</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">and DTMF digits =
used to control telephony services.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">The internet =
draft is attached.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Fixedsys">Steve =
Donovan</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 =
FACE=3D"Fixedsys">MCIWorldcom</FONT>
</P>

<P><FONT FACE=3D"Arial" SIZE=3D2 COLOR=3D"#000000"> =
&lt;&lt;draft-ietf-mmusic-sip-info-method-00.txt&gt;&gt; </FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01BE5441.78D61FBE--

------_=_NextPart_000_01BE5441.78D61FBE
Content-Type: text/plain;
	name="draft-ietf-mmusic-sip-info-method-00.txt"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment;
	filename="draft-ietf-mmusic-sip-info-method-00.txt"
Content-Location: ATT-0-1E1A16A92DC0D211B16D00805F6D8FC3-D
	RAFT-%7E3.TXT

Internet Engineering Task Force                        Steven R. =
Donovan
INTERNET DRAFT                                                       =
MCI
February 8, 1999                                  Expires August 8, =
1999
                              =
<draft-ietf-mmusic-sip-info-method-00.txt>

                        The SIP INFO Method

Status of this document

This document is an Internet-Draft and is in full conformance with all
provisions of Section 10 of RFC 2026.

Internet-Drafts are working documents of the Internet Engineering Task
Force (IETF), its areas, and its working groups.  Note that other =
groups
may also distribute working documents as Internet-Drafts.

Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time.  It is inappropriate to use Internet- Drafts as reference =
material
or to cite them other than as "work in progress."

To view the entire list of current Internet-Drafts, please check the
"1id-abstracts.txt" listing contained in the Internet-Drafts Shadow
Directories on ftp.is.co.za (Africa), ftp.nordu.net (Northern Europe),
ftp.nis.garr.it (Southern Europe), munnari.oz.au (Pacific Rim),
ftp.ietf.org (US East Coast), or ftp.isi.edu (US West Coast).



Abstract

This document proposes an extension to the Session Initiation Protocol.
This extension adds the INFO method to the SIP protocol.  The intent
of the INFO method is to allow for the carrying of session related=20
control information that is generated during a session.  Examples of
such session control information are ISUP/ISDN signaling messages
and DTMF digits used to control telephony services.


















Donovan                                                         [Page =
1]
=0C
Internet draft           The SIP INFO Method            February 8, =
1999

1.0 Introduction

There are situations where session related control information needs to
be sent during a session.  This information is separate from the media=20
that is being exchanged as part of the session.

Two examples of this are motivated by telephony related services:

1 - Mid Call Telephony Signaling Messages
2 - DTMF Digit/Dial Plus Control of Telephony Services

It can also be envisioned that there will be non telephony inspired=20
uses of a mechanism for relaying mid session information between=20
participants of the session and to Proxy Servers interested in the
session.

This document proposes the addition of the INFO Request method to the
SIP specification. =20

1.1 Mid Call Telephony Signaling Messages

The first use for the INFO method is the need to carry mid call=20
signaling information resulting from the interworking between an ISUP=20
or ISDN network/device and a SIP controlled network. =20

One specific example of this interworking is when the SIP controlled
network is used for transport between two PSTN locations.  For this=20
call, there will be a PSTN leg from the calling party to the SIP=20
network, a SIP leg through the SIP network and a PSTN leg from the SIP=20
network to the called party.  There needs to be a method to carry mid-
call PSTN signaling that is originated by the calling party through the
SIP network to the called party.

1.2 DTMF Digit/Dial Plus Control of Telephony Services

The second type of telephony session control information that needs to=20
be carried during a session is DTMF or dial plus (refered to from here
on as DTMF) generated information.  There are various telephony =
services
implemented today which require the use of DTMF digits.  Due to the=20
design of these features, the DTMF information needs to be carried both
as part of the media stream (in the RTP flow) and as part of the=20
signaling or control path.  This is due to the fact that there is an=20
implicit separation of the media and control path in the SIP protocol.
Thus, SIP Proxy Servers that implement services that require DTMF=20
control and that are not in the media path require a mechanism to be=20
notified of the DTMF digits.






Donovan                                                         [Page =
2]
=0C
Internet draft           The SIP INFO Method            February 8, =
1999

2.0 INFO Method

The INFO method is used for communicating mid-session control=20
information along the signaling path for the session.  The signaling
path for the INFO method is the signaling path established as a=20
result of the session setup. =20

The mid-session control information can be communicated in either
an INFO message header or as part of an attachment. =20

If the control information is telephony signaling information than the
signaling message would be carried as part of an ISUP attachment to the
INFO message as described in draft-ietf-sigtran-mime-isup-00.txt.

The method for carrying the DTMF information in the INFO message has=20
not yet been defined and is outside the scope of this document.

2.1 Header Field Support for INFO Method

The following table is an extension of tables 4 and 5 in the SIP
specification.  Refer to the SIP Specification for a description of
the content of the table.

Header                    Where    INFO
------                    -----    ----
Accept                      R       -
Accept-Encoding             R       -
Accept-Language             R       o
Allow                      200      -
Allow                      405      o
Authorization               R       o
Call-ID                    gc       m
Contact                     R       -
Contact                    1xx      -     =20
Contact                    2xx      -
Contact                    3xx      -
Contact                    485      -
Content-Encoding            e       o
Content-Length              e       o
Content-Type                e       *
CSeq                       gc       m
Date                        g       o
Encryption                  g       o
Expires                     g       -
From                       gc       m
Hide                        R       o
Max-Forwards                R       o
Organization                g       o





Donovan                                                         [Page =
3]
=0C
Internet draft           The SIP INFO Method            February 8, =
1999

Header                    Where    INFO
------                    -----    ----
Priority                    R       o
Proxy-Authenticate         407      o
Proxy-Authorization         R       o
Proxy-Require               R       o
Require                     R       o
Retry-After                 R       -
Retry-After            404,480,486  o
Retry-After                503      o
Retry-After              600,603    o
Response-Key                R       o
Record-Route                R       o
Record-Route               2xx      o
Route                       R       o
Server                      r       o
Subject                     R       -
Timestamp                   g       o
To                        gc(1)     m
Unsupported                420      o
User-Agent                  g       o
Via                       gc(2)     m
Warning                     r       o
WWW-Authenticate           401      o

2.2 Responses to the INFO Request Method

A 200 OK response shall be sent if the INFO request was successful.

Request Failure (4xx), Server Failure (5xx) and Global Failure (6xx)=20
responses can also be sent for the INFO Request.

2.3 Message Body Inclusion

The INFO request may contain a message body.

2.4 Behavior of SIP User Agents

The protocol rules applied by the SIP User Agent shall be similar to=20
those applied used for the BYE request.  However, the INFO message =
shall
shall not change the state of the session.

2.5 Behavior of SIP Proxy and Redirect Servers

2.5.1 Proxy Server

The protocol rules applied by the SIP Proxy Server shall be similar to=20
those applied used for the BYE request.  However, the INFO message =
shall
shall not change the state of the session.




Donovan                                                         [Page =
4]
=0C
Internet draft           The SIP INFO Method            February 8, =
1999

2.5.2 Forking Proxy Server

The protocol rules applied by the SIP Forking Proxy Server shall be=20
similar to those applied used for the BYE request.  However, the INFO=20
message shall shall not change the state of the session.

2.5.3 Redirection Server

A redirection server should not receive the INFO method as it is a part
of the signaling path only at the initiation of the session.  As=20
such, a redirection server should send a 403 Forbidden response.

2.6 Security Considerations

There are no security issues specific to the INFO method.  The security
requirements specified in the SIP specification apply to the INFO
method.

3.0 References

[1] M. Handley, H. Schulzrinne, E. Schooler, and J. Rosenberg,=20
    "SIP: Session Initiation Protocol", Internet Draft, Internet=20
    Engineering Task Force, January 15, 1999.  Work in progress.

[2] C. Huitema, "The multipart/sip-id media type", Internet Draft,
    Internet Engineering Task Force, February 5, 1999.  Work in=20
    Progress

4.0 Author's Address

   Steve Donovan
   MCI Worldcom
   1493/678
   901 International Parkway
   Richardson, Texas 75081
   Email: steven.r.donovan@mci.com

















Donovan                                                         [Page =
5]




------_=_NextPart_000_01BE5441.78D61FBE--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 02:25:57 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id CAA10243
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 02:25:57 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C147B7FB7; Thu, 11 Feb 1999 02:15:58 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0F1DC7FC8; Thu, 11 Feb 1999 02:15:57 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36C28375.F73DB78D@dnrc.bell-labs.com>
Date: Thu, 11 Feb 1999 02:15:01 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: list iptel <iptel@lists.research.bell-labs.com>
Subject: Revised framework document
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Folks,

I have submitted a revised version of the GLP framework document. It
should appear in the archives shortly. Until that time, you can find a
copy from our web page:

http://www.bell-labs.com/mailing-lists/iptel/

The changes are:

* discussion of QoS and capacity metrics, as per the Orlando mtg.
* discussion of cost structures, as per Orlando
* discussion of the aggregation/attribute tradeoffs, as per Orlando
* inclusion of minor comments raised on the list

We were supposed to go to last call on this last month, so we are
already behind. I think we are ready for working group last call on this
now. Any objections?

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 11:33:27 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id LAA17192
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 11:33:25 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1FDAA7FB7; Thu, 11 Feb 1999 11:19:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 82C737FC8; Thu, 11 Feb 1999 11:19:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <CA6966C24AC6D111B5A100805FEAB7D8577409@nsrip00208.mcit.com>
From: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
To: "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Thu, 11 Feb 1999 16:16:49 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE55D9.EED8C380"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE55D9.EED8C380
Content-Type: text/plain

> This message is notification of the following Internet Draft which has
> been submitted to the IETF for posting.
> 
> <draft-ietf-mmusic-sip-session-timer-00.txt> 
> 
> Abstract: 
> 
This document proposes an extension to the SIP specification.  This
extension adds a new message header that is used to specify the 
duration of a requested session.

The session timer can be used to limit the total duration of a session
if, for instance, one of the participants in the session wants to
limit the cost of the session.  It can also be used by stateful SIP
Proxy Servers to track the status of sessions for which session state
exists on the servers. Currently a stateful SIP Proxy Server that is 
not handling the media stream(s) for the session has no mechanism to 
definitively determine the state of all sessions for which it has state.
While the SIP Specification does provide the BYE method for terminating
the session, there is no mechanism for a SIP Proxy Server to detect the
end of a session when the BYE message is not sent or is lost due to 
network problems.

> The internet draft is attached.
> 
> Steve Donovan
> MCIWorldcom

------_=_NextPart_001_01BE55D9.EED8C380
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>draft-ietf-mmusic-sip-session-timer-00.txt</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">This message is =
notification of the following Internet Draft which has been submitted =
to the IETF for posting.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 =
FACE=3D"Fixedsys">&lt;draft-ietf-mmusic-sip-</FONT><FONT =
COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">session-timer</FONT><FONT =
COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">-00.txt&gt; </FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">Abstract:<BR>
</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">This document proposes an =
extension to the SIP specification.&nbsp; This</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">extension adds a new message =
header that is used to specify the </FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">duration of a requested =
session.</FONT>
</P>

<P><FONT SIZE=3D2 FACE=3D"Courier New">The session timer can be used to =
limit the total duration of a session</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">if, for instance, one of the =
participants in the session wants to</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">limit the cost of the =
session.&nbsp; It can also be used by stateful SIP</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">Proxy Servers to track the =
status of sessions for which session state</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">exists on the servers. =
Currently a stateful SIP Proxy Server that is </FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">not handling the media =
stream(s) for the session has no mechanism to </FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">definitively determine the =
state of all sessions for which it has state.</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">While the SIP Specification =
does provide the BYE method for terminating</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">the session, there is no =
mechanism for a SIP Proxy Server to detect the</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">end of a session when the BYE =
message is not sent or is lost due to </FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">network problems.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">The internet =
draft is attached.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">Steve =
Donovan</FONT>
<BR><FONT COLOR=3D"#0000FF" SIZE=3D2 =
FACE=3D"Fixedsys">MCIWorldcom</FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01BE55D9.EED8C380--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 12:13:33 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id MAA18197
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 12:13:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 245BE7FC8; Thu, 11 Feb 1999 12:01:21 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 720DF7FC9; Thu, 11 Feb 1999 12:01:20 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <CA6966C24AC6D111B5A100805FEAB7D857740C@nsrip00208.mcit.com>
From: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
To: "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Thu, 11 Feb 1999 09:59:22 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/mixed;
	boundary="----_=_NextPart_000_01BE55DF.DFFC8E72"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_000_01BE55DF.DFFC8E72
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE55DF.DFFC8E72"


------_=_NextPart_001_01BE55DF.DFFC8E72
Content-Type: text/plain

This time with the draft attached.  Sorry for the extra mail.

Steve

 <<draft-ietf-mmusic-sip-session-timer-00.txt>> 

> -----Original Message-----
> From:	Donovan, Steven R. (MCI) 
> Sent:	Thursday, February 11, 1999 10:17 AM
> To:	'confctrl@ISI.EDU'; 'iptel@lists.research.bell-labs.com'
> Subject:	draft-ietf-mmusic-sip-session-timer-00.txt
> 
> This message is notification of the following Internet Draft which has
> been submitted to the IETF for posting.
> 
> <draft-ietf-mmusic-sip-session-timer-00.txt> 
> 
> Abstract: 
> 
> This document proposes an extension to the SIP specification.  This
> extension adds a new message header that is used to specify the 
> duration of a requested session.
> 
> The session timer can be used to limit the total duration of a session
> if, for instance, one of the participants in the session wants to
> limit the cost of the session.  It can also be used by stateful SIP
> Proxy Servers to track the status of sessions for which session state
> exists on the servers. Currently a stateful SIP Proxy Server that is 
> not handling the media stream(s) for the session has no mechanism to 
> definitively determine the state of all sessions for which it has state.
> While the SIP Specification does provide the BYE method for terminating
> the session, there is no mechanism for a SIP Proxy Server to detect the
> end of a session when the BYE message is not sent or is lost due to 
> network problems.
> 
> The internet draft is attached.
> 
> Steve Donovan
> MCIWorldcom

------_=_NextPart_001_01BE55DF.DFFC8E72
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>RE: draft-ietf-mmusic-sip-session-timer-00.txt</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">This time with =
the draft attached.&nbsp; Sorry for the extra mail.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">Steve</FONT>
</P>

<P><FONT FACE=3D"Arial" SIZE=3D2 COLOR=3D"#000000"> =
&lt;&lt;draft-ietf-mmusic-sip-session-timer-00.txt&gt;&gt; </FONT>
</P>
<UL>
<P><FONT SIZE=3D1 FACE=3D"Arial">-----Original Message-----</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">From:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Donovan, Steven R. (MCI) </FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">Sent:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Thursday, February 11, 1999 10:17 AM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">'confctrl@ISI.EDU'; =
'iptel@lists.research.bell-labs.com'</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT>=
</B> <FONT SIZE=3D1 =
FACE=3D"Arial">draft-ietf-mmusic-sip-session-timer-00.txt</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">This message is =
notification of the following Internet Draft which has been submitted =
to the IETF for posting.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 =
FACE=3D"Fixedsys">&lt;draft-ietf-mmusic-sip-</FONT><FONT =
COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">session-timer</FONT><FONT =
COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">-00.txt&gt; </FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">Abstract:<BR>
</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">This document proposes an =
extension to the SIP specification.&nbsp; This</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">extension adds a new message =
header that is used to specify the </FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">duration of a requested =
session.</FONT>
</P>

<P><FONT SIZE=3D2 FACE=3D"Courier New">The session timer can be used to =
limit the total duration of a session</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">if, for instance, one of the =
participants in the session wants to</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">limit the cost of the =
session.&nbsp; It can also be used by stateful SIP</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">Proxy Servers to track the =
status of sessions for which session state</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">exists on the servers. =
Currently a stateful SIP Proxy Server that is </FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">not handling the media =
stream(s) for the session has no mechanism to </FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">definitively determine the =
state of all sessions for which it has state.</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">While the SIP Specification =
does provide the BYE method for terminating</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">the session, there is no =
mechanism for a SIP Proxy Server to detect the</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">end of a session when the BYE =
message is not sent or is lost due to </FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">network problems.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">The internet =
draft is attached.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">Steve =
Donovan</FONT>
<BR><FONT COLOR=3D"#0000FF" SIZE=3D2 =
FACE=3D"Fixedsys">MCIWorldcom</FONT>
</P>
</UL>
</BODY>
</HTML>
------_=_NextPart_001_01BE55DF.DFFC8E72--

------_=_NextPart_000_01BE55DF.DFFC8E72
Content-Type: text/plain;
	name="draft-ietf-mmusic-sip-session-timer-00.txt"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment;
	filename="draft-ietf-mmusic-sip-session-timer-00.txt"
Content-Location: ATT-0-DDA0D810CEC1D211AFA900805FEAAC98-D
	RAFT-%7E6.TXT

Internet Engineering Task Force                        Steven R. =
Donovan
INTERNET DRAFT                                                       =
MCI
February 10, 1999                                Expires August 10, =
1999
                            =
<draft-ietf-mmusic-sip-session-timer-00.txt>

                         SIP Session Timer

Status of this document

This document is an Internet-Draft and is in full conformance with all
provisions of Section 10 of RFC 2026.

Internet-Drafts are working documents of the Internet Engineering Task
Force (IETF), its areas, and its working groups.  Note that other =
groups
may also distribute working documents as Internet-Drafts.

Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time.  It is inappropriate to use Internet- Drafts as reference =
material
or to cite them other than as "work in progress."

To view the entire list of current Internet-Drafts, please check the
"1id-abstracts.txt" listing contained in the Internet-Drafts Shadow
Directories on ftp.is.co.za (Africa), ftp.nordu.net (Northern Europe),
ftp.nis.garr.it (Southern Europe), munnari.oz.au (Pacific Rim),
ftp.ietf.org (US East Coast), or ftp.isi.edu (US West Coast).



Abstract

This document proposes an extension to the SIP specification.  This
extension adds a new message header that is used to specify the=20
duration of a requested session.

The session timer can be used to limit the total duration of a session
if, for instance, one of the participants in the session wants to
limit the cost of the session.  It can also be used by stateful SIP
Proxy Servers to track the status of sessions for which session state
exists on the servers. Currently a stateful SIP Proxy Server that is=20
not handling the media stream(s) for the session has no mechinism to=20
definitively determine the state of all sessions for which it has =
state.
While the SIP Specification does provide the BYE method for terminating
the session, there is no mechinism for a SIP Proxy Server to detect the
end of a session when the BYE message is not sent or is lost due to=20
network problems.








Donovan                                                         [Page =
1]
=0C
Internet draft          The SIP Session Timer          February 10, =
1999

1.0 Introduction

The need for the addition of a session timer was initially motivated by
the realization that stateful proxy servers currently have no method of =

determining the end of a session in certain anomalous situations.  For=20
instance, when a user agent fails to send a BYE message at the end of=20
a session or the BYE message gets lost due to network problems.  In=20
this situation, the stateful proxy will maintain state for the session=20
and will have no deterministic method for determining when the state=20
no longer applies.

With the addition of a session timer to the SIP protocol, the stateful
proxy server will have the option of tearing down the session upon the
expiration of the session timer.

A session timer can also be used for other purposes.  For instance, it
could be used in the implementation of a prepaid service.  In this case
all session related signaling would be routed through a SIP server that
would control the session time based on a subscriber's prepaid account.
The SIP server could use the session timer to force tear down of the=20
session within a specific time.  The session timer could also be used
by the user agent as a trigger to indicate to the subscriber that=20
the prepaid account requires more funds to extend the call.

This document proposes the addition of a the Session-Expires header to=20
the SIP INVITE message.  In addition, two new Request Failure messages
are proposed.  The first to allow a SIP Proxy Server (SPS) server or a
User Agent Server (UAS) to indicate the need for the Session-Expires=20
header in INVITE messages.  The second to allow a SPS or UAS to reject=20
a User Agent Client's (UAC) session timer request.  This can be the=20
UAC's initial INVITE or the UAC's request to extent an existing =
session.

2.0 Session-Expires Header Field Definition

The Session-Expires header will be used by UAC to indicate the maximum=20
duration of the session.  The format of the contents of the=20
Session-Expires header shall be equivalent to the format of the=20
Expires header.  As such, the end of the session can be specified as
either an absolute or delta time.

The Session-Expires header shall be valid only in INVITE requests.

The UAS or any SPS in the path of the session signaling shall have the=20
option of rejecting an INVITE that does not contain the Session-Expires
header if the service context in which the session request is made=20
requires the use of the session timer.

The begin time of the session shall start upon receipt of the ACK by=20
the SPS and UAS message indicating successful setup of the session.




Donovan                                                         [Page =
2]
=0C
Internet draft          The SIP Session Timer          February 10, =
1999

The UAC shall have the ability to request an extension to the duration=20
of the session by sending a re-INVITE message for the session with a=20
new Session-Expires header.  See section 1.4.6, Changing an Existing=20
Session, in [1].

The UAS or any SPS in the path of the session signaling shall have the=20
ability to reject the change in the session duration. =20

Upon expiration of the session timer, the UAS shall have the option of=20
terminating the session using the normal BYE method.

In addition, a SPS shall have the option of tearing down an expired=20
session by sending a BYE to both the UAS and the UAC.

3.0 Behavior of SIP User Agents

3.1 Behavior of User Agent Clients

A User Agent Client shall have the ability to include the Session-
Expires header in an INVITE message. =20

The UAC may need to keep a session based timer in order to determine
the need for extending the session. =20

The UAC shall have the ability to request an extension of the session=20
timer by sending an INVITE message for an existing session with a=20
Session-Expires header. =20

3.2 Behavior of User Agent Servers

The User Agent Server shall have the ability to reject an INVITE=20
message that does not contain a Session-Expires header if the INVITE=20
is received in a service context that requires a session timer.  The=20
UAS shall reject the INVITE using the following response:

     4xx Session-Expires Header Required

If the Session-Expires header contains a delta time then the UAS shall=20
calculate the end of the session based on receipt of the ACK message=20
that completes a successful session initiation:

  End of session time =3D Receipt of ACK time + Session-Expires time

The UAS shall have the ability to reject an INVITE request based on=20
the Session-Expires header.  For instance, the UAS may choose to reject =

the INVITE if the requested session time is longer than the UAS desires =

to participate in the session.  The UAS shall use the following message =

to reject the request:

     4xx Session Time Request Rejected



Donovan                                                         [Page =
3]
=0C
Internet draft          The SIP Session Timer          February 10, =
1999

The UAS shall have the ability to reject a request to extend the length =

of the session.  The UAS shall do so by sending the following response:

     4xx Session Time Request Rejected

If the Session-Expires header in the re-INVITE contains a delta time=20
then the UAS shall calculate the new end of the previous end of session =

time:

  New end of session time =3D Previous end of session time + delta time =


If a request to extend the session time is rejected by the UAS then
the end of session time that existed prior to the extension request
shall be used.

Upon expiration of the session timer, the UAS shall explicitly=20
terminate the session by sending a BYE message.

4.0 Behavior of SIP Proxy Servers

A SIP Proxy Server has the option of tracking the duration of sessions=20
for which it is a proxy.

If the SPS receives an INVITE without a Session-Expires header, it has=20
the option of sending the following message indicating that the=20
Session-Expires header is required:

     4xx Session-Expires Header Required

A SPS has the option of rejecting an INVITE with a Session-Expires=20
header based on the time specified in the header.  For instance, if=20
the time specified is too long. The proxy server UAS shall do by=20
sending the following response:

     4xx Session Time Request Rejected

A SPS has the option of rejecting a request for an extension of the=20
session timer for an existing session by sending the following message:

     4xx Session Time Request Rejected

If the SPS accepts the request to extend the session timer then it=20
shall use the method of calculating the new end of session time
specified in section 3.2.

A SPS shall have the option of tearing down sessions that have expired.
Note that the SPS should wait for the UAS to terminate the call using=20
normal mechanisms.  If the UAS fails to send a BYE as a result of=20
session expiration then the SPS can choose to end the call.

The SPS shall terminate the call by sending BYE requests to the UAC=20
and the UAS.

Donovan                                                         [Page =
4]
=0C
Internet draft          The SIP Session Timer          February 10, =
1999

5.0 Header Field Definitions

The following is an extension of tables 4 and 5 in [1] for the Session-
Expires header:

                  where  enc.  e-e  ACK  BYE  CAN  INV  OPT  REG
 Session-Expires    r           e    -    -    -    o    -    -

5.0 Security Considerations

There are no security considerations unique to the Session-Expires
header.

6.0 Example

The following examples are meant to illustrate the functionality=20
associated with the Session-Expires header.  As such, other headers
are intentionally left out of the SIP messages.

Example One -=20

In this example, the UAC initiates a session directly with the UAS =
using
the Session-Expires header.  The UAC subsequently requests and=20
extension to the session timer which is accepted by the UAS.  The=20
session eventually times out and is terminated by the UAS.

UAC->UAS     INVITE
             Session-Expires: 120
UAS->UAC     200 OK
UAC->UAS     ACK                        UAC starts session timer
             UAS Receipt of ACK         UAS starts session timer
UAC->UAC     INVITE
             Session-Expires: 120
UAC->UAS     200 OK
UAC->UAS     ACK                        UAC resets session end time
             UAS Receipt of ACK         UAS resets session timer
Session timeout
UAS->UAC     BYE
UAC->UAS     200 OK














Donovan                                                         [Page =
5]
=0C
Internet draft          The SIP Session Timer          February 10, =
1999

Example Two=20

In this scenario, the UAC session request without a Session-Expires=20
header.  The SPS rejects the INVITE, indicating that a Session-Expires
header is required.  The UAC re-requests a session with a=20
Session-Expires header.  This session is successfully set up through
the SPS.  The SPS eventually detects session timeout. =20

UAC->SPS     INVITE
SPS->UAC     4xx Session-Expires Header Required
UAC->SPS     INVITE
             Session-Expires: 120
SPS->UAS     INVITE                     To extend existing session
             Session-Expires: 120
UAS->SPS     200 OK
SPS->UAC     200 OK
UAC->SPS     ACK                        UAC starts session timer
             SPS Receipt of ACK         SPS starts session timer
SPS->UAS     ACK                        UAS starts session timer
Session timeout detected by SPS
SPS->UAC     BYE
SPS->UAS     BYE
UAS->SPS     200 OK
UAC->SPS     200 OK

7.0 References

[1] M. Handley, H. Schulzrinne, E. Schooler, and J. Rosenberg,=20
    "SIP: Session Initiation Protocol", Internet Draft, Internet=20
    Engineering Task Force, January 15, 1999.  Work in progress.

8.0 Author's Address

   Steven R. Donovan
   MCI Worldcom
   1493/678
   901 International Parkway
   Richardson, Texas 75081
   Email: steven.r.donovan@mci.com














Donovan                                                         [Page =
6]
------_=_NextPart_000_01BE55DF.DFFC8E72--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 15:15:17 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id PAA21035
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 15:15:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2CE9C7FC9; Thu, 11 Feb 1999 14:07:13 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6CC9E7FCA; Thu, 11 Feb 1999 14:07:12 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199902111906.LAA28606@hsmpka.eng.sun.com>
Date: Thu, 11 Feb 1999 11:02:08 -0800
To: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Pardon the perhaps very naive question, but I suppose I am not as versant in 
SIP as I would like to be.

My understanding is that a SIP Client issues an Invite message to a SIP Server,
which in turn forwards the request to a target SIP Client (or some GW or some
sort). Once the ACK has returned to the originating client, I was under the
impression that both SIP Clients send RTP packets directly to each other,
bypassing the SIP Server (this is where I am unclear).

So if this is true, how will a SIP Server (stateful, or stateless) enforce any
session timeouts? A malicious SIP Client can simply keep the session active
for as long as it wishes, ignoring any messages from the SIP Server that may
tell it to disconnect.

This, as far as I can see, could cause some serious problems if one wishes to
deploy a service and still maintain some control over the sessions.

Again, perhaps I am way off here. Any clarifications would certainly be 
appreciated.

PatC

>This time with the draft attached.  Sorry for the extra mail.
>
>Steve
>
> <<draft-ietf-mmusic-sip-session-timer-00.txt>> 
>
>> -----Original Message-----
>> From:	Donovan, Steven R. (MCI) 
>> Sent:	Thursday, February 11, 1999 10:17 AM
>> To:	'confctrl@ISI.EDU'; 'iptel@lists.research.bell-labs.com'
>> Subject:	draft-ietf-mmusic-sip-session-timer-00.txt
>> 
>> This message is notification of the following Internet Draft which has
>> been submitted to the IETF for posting.
>> 
>> <draft-ietf-mmusic-sip-session-timer-00.txt> 
>> 
>> Abstract: 
>> 
>> This document proposes an extension to the SIP specification.  This
>> extension adds a new message header that is used to specify the 
>> duration of a requested session.
>> 
>> The session timer can be used to limit the total duration of a session
>> if, for instance, one of the participants in the session wants to
>> limit the cost of the session.  It can also be used by stateful SIP
>> Proxy Servers to track the status of sessions for which session state
>> exists on the servers. Currently a stateful SIP Proxy Server that is 
>> not handling the media stream(s) for the session has no mechanism to 
>> definitively determine the state of all sessions for which it has state.
>> While the SIP Specification does provide the BYE method for terminating
>> the session, there is no mechanism for a SIP Proxy Server to detect the
>> end of a session when the BYE message is not sent or is lost due to 
>> network problems.
>> 
>> The internet draft is attached.
>> 
>> Steve Donovan
>> MCIWorldcom



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 17:06:50 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id RAA22570
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 17:06:50 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 980CB7FB8; Thu, 11 Feb 1999 16:50:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E8B7D7FCA; Thu, 11 Feb 1999 16:50:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36C35066.852BC71B@dnrc.bell-labs.com>
Date: Thu, 11 Feb 1999 16:49:26 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Pat.Calhoun@Eng.Sun.COM
Cc: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@isi.edu>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
References: <199902111906.LAA28606@hsmpka.eng.sun.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Patrice Calhoun wrote:
> 
> My understanding is that a SIP Client issues an Invite message to a SIP Server,
> which in turn forwards the request to a target SIP Client (or some GW or some
> sort). Once the ACK has returned to the originating client, I was under the
> impression that both SIP Clients send RTP packets directly to each other,
> bypassing the SIP Server (this is where I am unclear).

Slight correction - ACK goes from caller to callee.

But, in general, you are right. RTP bypasses the SIP servers.

> 
> So if this is true, how will a SIP Server (stateful, or stateless) enforce any
> session timeouts? A malicious SIP Client can simply keep the session active
> for as long as it wishes, ignoring any messages from the SIP Server that may
> tell it to disconnect.
> 
> This, as far as I can see, could cause some serious problems if one wishes to
> deploy a service and still maintain some control over the sessions.

This is an excellent, excellent point, and is something I have been
considering since the whole timeout/DTMF/INFO issues came up. The
question you ask gets back to a more basic problem. Since, unlike
traditional telephony, voice does not flow through the servers
(gatekeepers or SIP servers or MXCP call agents) in Internet telephony,
any sort of "control" a server would like to execute is based on an
assumption that the user agents (the ones sending/receiving the media)
would obey the servers. Consider the example of call transfer. Lets say
we'd like to signal the service using DTMF, so a proxy subscribes to the
DTMF from the client. The client presses the appropriate buttons on the
phone, and the proxy decides to transfer. But, the media doesn't flow
through the proxy, so it has no way of actually transferring the call.
Rather, it will need to go back and tell the user agent to perform the
transfer, since only the user agent can affect where the media goes to.
In that case, why would the UA ask the network for the service in the
first place?  It can simply be invoked directly by the client. In a
network where voice transport is end to end, services based on changing
characteristics of that transport (coding, destination, etc), must also
be end to end. 

Now, if the action the proxy asks to user to take is something the user
wants to do anyway, maybe thats OK. But, if its something less
cooperative, like forcing a hangup, or transferring to an operator,
there is simply no incentive for the user to accept this. 

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 17:22:11 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id RAA23250
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 17:22:11 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3BD737FCA; Thu, 11 Feb 1999 16:57:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 912637FCB; Thu, 11 Feb 1999 16:57:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Thu, 11 Feb 1999 15:55 -0600 (CST)
From: John Hearty <John.H.Hearty@mci.com>
X-Mailer: MailRoom for Internet v2.3g (www.SierraSol.com)
To: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Cc: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Message-Id: <19990211215608.KKFZ16328@localHost>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


 This draft was written in part to address a concern that stateful
SIP proxy servers could have sessions marked active when in fact they
have been terminated, and invalid state data could remain indefinately
in the server.  It is true that BYE messages could be ignored, but
that would have to be coordinated by all parties wishing to remain
in the session, and then they are just using the IP access pipe they
are already paying for.  You could even go one step further in that
line of thinking and have the originator send a Cancel instead of an
ACK once it got the session info back from the terminator in a 200 OK,
thereby never starting the session, but coordination from both ends
would again be required.  So other than setting up the session (which
could be a billable event), coordination of the session from that point
on would be up to the involved clients, no different than what is now
available.

John Hearty
MCI Worldcom
Global Network Evolution



Date: Thu, 11 Feb 1999 13:02 -0600 (CST)
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
To: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
    "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
    "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>
Sender: owner-iptel@lists.research.bell-labs.com
Reply-to: Pat.Calhoun@Eng.Sun.COM
Delivered-to: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Delivered-to: iptel-local@paperless.dnrc.bell-labs.com
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt

Pardon the perhaps very naive question, but I suppose I am not as versant in 
SIP as I would like to be.

My understanding is that a SIP Client issues an Invite message to a SIP Server,
which in turn forwards the request to a target SIP Client (or some GW or some
sort). Once the ACK has returned to the originating client, I was under the
impression that both SIP Clients send RTP packets directly to each other,
bypassing the SIP Server (this is where I am unclear).

So if this is true, how will a SIP Server (stateful, or stateless) enforce any
session timeouts? A malicious SIP Client can simply keep the session active
for as long as it wishes, ignoring any messages from the SIP Server that may
tell it to disconnect.

This, as far as I can see, could cause some serious problems if one wishes to
deploy a service and still maintain some control over the sessions.

Again, perhaps I am way off here. Any clarifications would certainly be 
appreciated.

PatC

>This time with the draft attached.  Sorry for the extra mail.
>
>Steve
>
> <<draft-ietf-mmusic-sip-session-timer-00.txt>> 
>
>> -----Original Message-----
>> From:    Donovan, Steven R. (MCI) 
>> Sent:    Thursday, February 11, 1999 10:17 AM
>> To:  'confctrl@ISI.EDU'; 'iptel@lists.research.bell-labs.com'
>> Subject: draft-ietf-mmusic-sip-session-timer-00.txt
>> 
>> This message is notification of the following Internet Draft which has
>> been submitted to the IETF for posting.
>> 
>> <draft-ietf-mmusic-sip-session-timer-00.txt> 
>> 
>> Abstract: 
>> 
>> This document proposes an extension to the SIP specification.  This
>> extension adds a new message header that is used to specify the 
>> duration of a requested session.
>> 
>> The session timer can be used to limit the total duration of a session
>> if, for instance, one of the participants in the session wants to
>> limit the cost of the session.  It can also be used by stateful SIP
>> Proxy Servers to track the status of sessions for which session state
>> exists on the servers. Currently a stateful SIP Proxy Server that is 
>> not handling the media stream(s) for the session has no mechanism to 
>> definitively determine the state of all sessions for which it has state.
>> While the SIP Specification does provide the BYE method for terminating
>> the session, there is no mechanism for a SIP Proxy Server to detect the
>> end of a session when the BYE message is not sent or is lost due to 
>> network problems.
>> 
>> The internet draft is attached.
>> 
>> Steve Donovan
>> MCIWorldcom



---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 18:08:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id SAA24001
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 18:08:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 9D3107FCB; Thu, 11 Feb 1999 17:57:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E25FA7FCC; Thu, 11 Feb 1999 17:57:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <CA6966C24AC6D111B5A100805FEAB7D8577412@nsrip00208.mcit.com>
From: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
To: "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Thu, 11 Feb 1999 15:56:16 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE5611.BB663324"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE5611.BB663324
Content-Type: text/plain;
	charset="ISO-8859-1"



> -----Original Message-----
> From:	Pat.Calhoun@Eng.Sun.COM [SMTP:Pat.Calhoun@Eng.Sun.COM]
> Sent:	Thursday, February 11, 1999 1:02 PM
> To:	Donovan, Steven R. (MCI); 'iptel@lists.research.bell-labs.com';
> 'confctrl@ISI.EDU'
> Subject:	RE: draft-ietf-mmusic-sip-session-timer-00.txt
> 
> Pardon the perhaps very naive question, but I suppose I am not as versant
> in 
> SIP as I would like to be.
> 
> My understanding is that a SIP Client issues an Invite message to a SIP
> Server,
> which in turn forwards the request to a target SIP Client (or some GW or
> some
> sort). Once the ACK has returned to the originating client, I was under
> the
> impression that both SIP Clients send RTP packets directly to each other,
> bypassing the SIP Server (this is where I am unclear).
> 
	This depends on the implementation of the SIP Server.  There are
valid reasons to have the SIP Server proxy the IP packets.  But in general,
you are correct.  There will be SIP Servers that do not have a view of the
RTP flow.

> So if this is true, how will a SIP Server (stateful, or stateless) enforce
> any
> session timeouts? A malicious SIP Client can simply keep the session
> active
> for as long as it wishes, ignoring any messages from the SIP Server that
> may
> tell it to disconnect.
> 
> This, as far as I can see, could cause some serious problems if one wishes
> to
> deploy a service and still maintain some control over the sessions.
> 
> Again, perhaps I am way off here. Any clarifications would certainly be 
> appreciated.
> 
	You are correct and I personally would not rely on reporting from
the SIP Server to do reliable billing.  However, if a stateful SIP Proxy
server is built, for whatever reason, the current SIP specification does not
have a mechanism for the SIP server to know which of the sessions for which
it has state are valid.  This draft gives the SIP Server a mechanism to
remove those sessions that it feels are no longer valid.

	The amount of control that a SIP server exerts over sessions is a
separate issue.  As John Hearty accurately points out in a separate email,
the user of the SIP device can choose to bypass the SIP server entirely.
The challenge for service providers will be to implement services that make
user's want to access the service provider's SIP server.

> PatC
> 
> >This time with the draft attached.  Sorry for the extra mail.
> >
> >Steve
> >
> > <<draft-ietf-mmusic-sip-session-timer-00.txt>> 
> >
> >> -----Original Message-----
> >> From:	Donovan, Steven R. (MCI) 
> >> Sent:	Thursday, February 11, 1999 10:17 AM
> >> To:	'confctrl@ISI.EDU'; 'iptel@lists.research.bell-labs.com'
> >> Subject:	draft-ietf-mmusic-sip-session-timer-00.txt
> >> 
> >> This message is notification of the following Internet Draft which has
> >> been submitted to the IETF for posting.
> >> 
> >> <draft-ietf-mmusic-sip-session-timer-00.txt> 
> >> 
> >> Abstract: 
> >> 
> >> This document proposes an extension to the SIP specification.  This
> >> extension adds a new message header that is used to specify the 
> >> duration of a requested session.
> >> 
> >> The session timer can be used to limit the total duration of a session
> >> if, for instance, one of the participants in the session wants to
> >> limit the cost of the session.  It can also be used by stateful SIP
> >> Proxy Servers to track the status of sessions for which session state
> >> exists on the servers. Currently a stateful SIP Proxy Server that is 
> >> not handling the media stream(s) for the session has no mechanism to 
> >> definitively determine the state of all sessions for which it has
> state.
> >> While the SIP Specification does provide the BYE method for terminating
> >> the session, there is no mechanism for a SIP Proxy Server to detect the
> >> end of a session when the BYE message is not sent or is lost due to 
> >> network problems.
> >> 
> >> The internet draft is attached.
> >> 
> >> Steve Donovan
> >> MCIWorldcom
> 

------_=_NextPart_001_01BE5611.BB663324
Content-Type: text/html;
	charset="ISO-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>RE: draft-ietf-mmusic-sip-session-timer-00.txt</TITLE>
</HEAD>
<BODY>
<BR>
<BR>
<UL>
<P><FONT SIZE=3D1 FACE=3D"Arial">-----Original Message-----</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">From:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Pat.Calhoun@Eng.Sun.COM =
[SMTP:Pat.Calhoun@Eng.Sun.COM]</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">Sent:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Thursday, February 11, 1999 1:02 PM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">Donovan, Steven R. (MCI); =
'iptel@lists.research.bell-labs.com'; 'confctrl@ISI.EDU'</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT>=
</B> <FONT SIZE=3D1 FACE=3D"Arial">RE: =
draft-ietf-mmusic-sip-session-timer-00.txt</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">Pardon the =
perhaps very naive question, but I suppose I am not as versant in =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">SIP as I =
would like to be.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">My =
understanding is that a SIP Client issues an Invite message to a SIP =
Server,</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">which in turn =
forwards the request to a target SIP Client (or some GW or some</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">sort). Once =
the ACK has returned to the originating client, I was under the</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">impression =
that both SIP Clients send RTP packets directly to each other,</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">bypassing the =
SIP Server (this is where I am unclear).</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">This depends on =
the implementation of the SIP Server.&nbsp; There are valid reasons to =
have the SIP Server proxy the IP packets.&nbsp; But in general, you are =
correct.&nbsp; There will be SIP Servers that do not have a view of the =
RTP flow.</FONT></P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">So if this is =
true, how will a SIP Server (stateful, or stateless) enforce any</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">session =
timeouts? A malicious SIP Client can simply keep the session =
active</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">for as long =
as it wishes, ignoring any messages from the SIP Server that may</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">tell it to =
disconnect.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">This, as far =
as I can see, could cause some serious problems if one wishes to</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">deploy a =
service and still maintain some control over the sessions.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">Again, perhaps =
I am way off here. Any clarifications would certainly be </FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier =
New">appreciated.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">You are correct =
and I personally would not rely on reporting from the SIP Server to do =
reliable billing.&nbsp; However, if a stateful SIP Proxy server is =
built, for whatever reason, the current SIP specification does not have =
a mechanism for the SIP server to know which of the sessions for which =
it has state are valid.&nbsp; This draft gives the SIP Server a =
mechanism to remove those sessions that it feels are no longer =
valid.</FONT></P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">The amount of =
control that a SIP server exerts over sessions is a separate =
issue.&nbsp; As John Hearty accurately points out in a separate email, =
the user of the SIP device can choose to bypass the SIP server =
entirely.&nbsp; The challenge for service providers will be to =
implement services that make user's want to access the service =
provider's SIP server.</FONT></P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">PatC</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;This time =
with the draft attached.&nbsp; Sorry for the extra mail.</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier =
New">&gt;Steve</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt; =
&lt;&lt;draft-ietf-mmusic-sip-session-timer-00.txt&gt;&gt; </FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
-----Original Message-----</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
From:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Donovan, Steven R. =
(MCI) </FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
Sent:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Thursday, February 11, =
1999 10:17 AM</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
To:&nbsp; 'confctrl@ISI.EDU'; =
'iptel@lists.research.bell-labs.com'</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
Subject:&nbsp;&nbsp;&nbsp;&nbsp; =
draft-ietf-mmusic-sip-session-timer-00.txt</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; This =
message is notification of the following Internet Draft which =
has</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; been =
submitted to the IETF for posting.</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
&lt;draft-ietf-mmusic-sip-session-timer-00.txt&gt; </FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
Abstract: </FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; This =
document proposes an extension to the SIP specification.&nbsp; =
This</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
extension adds a new message header that is used to specify the </FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
duration of a requested session.</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; The =
session timer can be used to limit the total duration of a =
session</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; if, =
for instance, one of the participants in the session wants to</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
limit the cost of the session.&nbsp; It can also be used by stateful =
SIP</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
Proxy Servers to track the status of sessions for which session =
state</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
exists on the servers. Currently a stateful SIP Proxy Server that is =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; not =
handling the media stream(s) for the session has no mechanism to =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
definitively determine the state of all sessions for which it has =
state.</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
While the SIP Specification does provide the BYE method for =
terminating</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; the =
session, there is no mechanism for a SIP Proxy Server to detect =
the</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; end =
of a session when the BYE message is not sent or is lost due to </FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
network problems.</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; The =
internet draft is attached.</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
Steve Donovan</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
MCIWorldcom</FONT>
</P>
</UL>
</BODY>
</HTML>
------_=_NextPart_001_01BE5611.BB663324--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 18:39:16 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id SAA24322
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 18:39:16 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1596E7FCC; Thu, 11 Feb 1999 18:27:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5D31C7FCD; Thu, 11 Feb 1999 18:27:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199902112325.PAA14779@hsmpka.eng.sun.com>
Date: Thu, 11 Feb 1999 15:21:44 -0800
To: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>
>
>> -----Original Message-----
>> From:	Pat.Calhoun@Eng.Sun.COM [SMTP:Pat.Calhoun@Eng.Sun.COM]
>> Sent:	Thursday, February 11, 1999 1:02 PM
>> To:	Donovan, Steven R. (MCI); 'iptel@lists.research.bell-labs.com';
>> 'confctrl@ISI.EDU'
>> Subject:	RE: draft-ietf-mmusic-sip-session-timer-00.txt
>> 
>> Pardon the perhaps very naive question, but I suppose I am not as versant
>> in 
>> SIP as I would like to be.
>> 
>> My understanding is that a SIP Client issues an Invite message to a SIP
>> Server,
>> which in turn forwards the request to a target SIP Client (or some GW or
>> some
>> sort). Once the ACK has returned to the originating client, I was under
>> the
>> impression that both SIP Clients send RTP packets directly to each other,
>> bypassing the SIP Server (this is where I am unclear).
>> 
>	This depends on the implementation of the SIP Server.  There are
>valid reasons to have the SIP Server proxy the IP packets.  But in general,
>you are correct.  There will be SIP Servers that do not have a view of the
>RTP flow.
>
>> So if this is true, how will a SIP Server (stateful, or stateless) enforce
>> any
>> session timeouts? A malicious SIP Client can simply keep the session
>> active
>> for as long as it wishes, ignoring any messages from the SIP Server that
>> may
>> tell it to disconnect.
>> 
>> This, as far as I can see, could cause some serious problems if one wishes
>> to
>> deploy a service and still maintain some control over the sessions.
>> 
>> Again, perhaps I am way off here. Any clarifications would certainly be 
>> appreciated.
>> 
>	You are correct and I personally would not rely on reporting from
>the SIP Server to do reliable billing.  However, if a stateful SIP Proxy
>server is built, for whatever reason, the current SIP specification does not
>have a mechanism for the SIP server to know which of the sessions for which
>it has state are valid.  This draft gives the SIP Server a mechanism to
>remove those sessions that it feels are no longer valid.
>
>	The amount of control that a SIP server exerts over sessions is a
>separate issue.  As John Hearty accurately points out in a separate email,
>the user of the SIP device can choose to bypass the SIP server entirely.
>The challenge for service providers will be to implement services that make
>user's want to access the service provider's SIP server.

This is not actually true. It is quite simple for service providers to install
a set of filters in their systems that require that all RTP traffic be directed
to one of their local SIP Servers.

Note that I am not advocating such an idea. I like free service, but one must
realize that free services are rarely every deployed by the services 
providers :(

PatC
>
>> PatC
>> 
>> >This time with the draft attached.  Sorry for the extra mail.
>> >
>> >Steve
>> >
>> > <<draft-ietf-mmusic-sip-session-timer-00.txt>> 
>> >
>> >> -----Original Message-----
>> >> From:	Donovan, Steven R. (MCI) 
>> >> Sent:	Thursday, February 11, 1999 10:17 AM
>> >> To:	'confctrl@ISI.EDU'; 'iptel@lists.research.bell-labs.com'
>> >> Subject:	draft-ietf-mmusic-sip-session-timer-00.txt
>> >> 
>> >> This message is notification of the following Internet Draft which has
>> >> been submitted to the IETF for posting.
>> >> 
>> >> <draft-ietf-mmusic-sip-session-timer-00.txt> 
>> >> 
>> >> Abstract: 
>> >> 
>> >> This document proposes an extension to the SIP specification.  This
>> >> extension adds a new message header that is used to specify the 
>> >> duration of a requested session.
>> >> 
>> >> The session timer can be used to limit the total duration of a session
>> >> if, for instance, one of the participants in the session wants to
>> >> limit the cost of the session.  It can also be used by stateful SIP
>> >> Proxy Servers to track the status of sessions for which session state
>> >> exists on the servers. Currently a stateful SIP Proxy Server that is 
>> >> not handling the media stream(s) for the session has no mechanism to 
>> >> definitively determine the state of all sessions for which it has
>> state.
>> >> While the SIP Specification does provide the BYE method for terminating
>> >> the session, there is no mechanism for a SIP Proxy Server to detect the
>> >> end of a session when the BYE message is not sent or is lost due to 
>> >> network problems.
>> >> 
>> >> The internet draft is attached.
>> >> 
>> >> Steve Donovan
>> >> MCIWorldcom
>> 



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 19:40:03 1999
Received: from dirty.research.bell-labs.com (dirty.research.bell-labs.com [204.178.16.6])
	by ietf.org (8.8.5/8.8.7a) with SMTP id TAA25165
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 19:40:02 -0500 (EST)
Received: from lists.research.bell-labs.com ([135.180.161.172]) by dirty; Thu Feb 11 19:39:14 EST 1999
Received: by lists.research.bell-labs.com (Postfix)
	id 19A767FCE; Thu, 11 Feb 1999 19:11:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4ED957FCF; Thu, 11 Feb 1999 19:11:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199902120009.QAA27818@hsmpka.eng.sun.com>
Date: Thu, 11 Feb 1999 16:05:43 -0800
To: "Douglas Clowes" <dclowes@ozemail.com.au>, <Pat.Calhoun@Eng.Sun.COM>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: "'confctrl@ISI.EDU'" <confctrl@isi.edu>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Well, perhaps I missed something fundamental here, but my understanding is that
what we have here is analogous to two telephones using the PSTN to get the
call endpoint established, and then completely ignores the PSTN. 

If this is the case, certainly the PSTN carrier can charge for the initial 
endpoint resolution information, but is not involved at all afterwards. So
an intelligent user could figure out what physical port on which switch to
contact to get to his mother, and completely bypass the IN.

My only concern here is how does the service provider offer a service *and*
make at least enough money to recover the capital costs (of course, certain
providers may want to cover many additional costs, which is outside the scope
of this discussion :).

PatC

DISCLAIMER: I have a funny feeling that I've missed something fundamental here.
Please feel free to correct my statement above if it is technically incorrect.

>At 16:49 1999-02-11 -0500, Jonathan Rosenberg wrote:
>>Patrice Calhoun wrote:
>>> 
>>> My understanding is that a SIP Client issues an Invite message to a SIP
>Server,
>>> which in turn forwards the request to a target SIP Client (or some GW or
>some
>>> sort). Once the ACK has returned to the originating client, I was under the
>>> impression that both SIP Clients send RTP packets directly to each other,
>>> bypassing the SIP Server (this is where I am unclear).
>>
>>Slight correction - ACK goes from caller to callee.
>>
>>But, in general, you are right. RTP bypasses the SIP servers.
>>
>>> 
>>> So if this is true, how will a SIP Server (stateful, or stateless)
>enforce any
>>> session timeouts? A malicious SIP Client can simply keep the session active
>>> for as long as it wishes, ignoring any messages from the SIP Server that
>may
>>> tell it to disconnect.
>>> 
>>> This, as far as I can see, could cause some serious problems if one
>wishes to
>>> deploy a service and still maintain some control over the sessions.
>>
>
>One wonders about the nature of the service you think to deploy. Where you
>want to maintain session control without control of at least one enpoint.
>
>>This is an excellent, excellent point, and is something I have been
>>considering since the whole timeout/DTMF/INFO issues came up. The
>>question you ask gets back to a more basic problem.
>
>>Since, unlike
>>traditional telephony, voice does not flow through the servers
>>(gatekeepers or SIP servers or MXCP call agents) in Internet telephony,
>>any sort of "control" a server would like to execute is based on an
>>assumption that the user agents (the ones sending/receiving the media)
>>would obey the servers.
>
>Indeed, in traditional telephony (at least the SS7 kind), voice does not
>flow through the control points. Control is maintained by having voice
>nodes that obey control nodes.
>
>In H.323 gatekeepers, the gatekeeper may route the control channels between
>the endpoints, and issue termination commands in each direction, on those
>channels. A gatekeeper may route the media through itself (although this is
>unlikely), or another entity. What you wouldchoose to do would depend on
>your trust of the level of control you exert on the endpoints, and the
>consequences of any lack of control.
>
>If you don't trust either endpoint, but must control the media session, or
>conceal the endpoints from each other, you may have to route the media
>through a Meter Box (tm), that you can trust. There are gateways that
>transcode media (G.711 to G.723.1 for example) and a similar mechanism,
>without transcoding, could be deployed.
>
>> Consider the example of call transfer. Lets say
>>we'd like to signal the service using DTMF, so a proxy subscribes to the
>>DTMF from the client. The client presses the appropriate buttons on the
>>phone, and the proxy decides to transfer. But, the media doesn't flow
>>through the proxy, so it has no way of actually transferring the call.
>>Rather, it will need to go back and tell the user agent to perform the
>>transfer, since only the user agent can affect where the media goes to.
>>In that case, why would the UA ask the network for the service in the
>>first place?  It can simply be invoked directly by the client. In a
>>network where voice transport is end to end, services based on changing
>>characteristics of that transport (coding, destination, etc), must also
>>be end to end. 
>>
>>Now, if the action the proxy asks to user to take is something the user
>>wants to do anyway, maybe thats OK. But, if its something less
>>cooperative, like forcing a hangup, or transferring to an operator,
>>there is simply no incentive for the user to accept this. 
>>
>>-Jonathan R.
>>-- 
>>Jonathan D. Rosenberg                       Lucent Technologies
>>Member of Technical Staff                   101 Crawfords Corner Rd.
>>High Speed Networks Research                Holmdel, NJ 07733
>>FAX: (732) 834-5379                         Rm. 4C-526
>>EMAIL: jdrosen@bell-labs.com
>>URL: http://www.cs.columbia.edu/~jdrosen
>>
>>---------
>>This message came from the IETF IPTEL Working Group Mailing List.
>>
>
>
>Douglas
>



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 20:15:43 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id UAA25404
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 20:15:43 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1AB897FCF; Thu, 11 Feb 1999 20:05:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5FDE77FD0; Thu, 11 Feb 1999 20:05:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <F16674FCE856D211AC0D00E02910AE0A07B517@INTERNAL>
From: Francois Menard Distribution List Management Account <fm-listproc@mediatrix.com>
To: "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Thu, 11 Feb 1999 20:04:36 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

There is no way to properyly time the session unless the end-point is a
controlled network element.  A gatekeeper and/or SIP server cant do anything
unless the media stream gets through it ... of course, a SIP server, and or
H.323 gatekeeper can talk to routers, which control an IP media stream, and
affect the media stream.

Even worse, a sip server / gatekeeper that wishes to track the end of a call
so that it can bill by the second needs to worry about end points that just
go out and die and never send an end of conversation ... this mechanism is
not yet specified ...

-=Francois=-

-----Original Message-----
From: Pat.Calhoun@Eng.Sun.COM [mailto:Pat.Calhoun@Eng.Sun.COM]
Sent: Thursday, February 11, 1999 6:22 PM
To: Donovan, Steven R. (MCI); 'confctrl@ISI.EDU';
'iptel@lists.research.bell-labs.com'; 'Pat.Calhoun@Eng.Sun.COM'
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt



>
>
>> -----Original Message-----
>> From:	Pat.Calhoun@Eng.Sun.COM [SMTP:Pat.Calhoun@Eng.Sun.COM]
>> Sent:	Thursday, February 11, 1999 1:02 PM
>> To:	Donovan, Steven R. (MCI); 'iptel@lists.research.bell-labs.com';
>> 'confctrl@ISI.EDU'
>> Subject:	RE: draft-ietf-mmusic-sip-session-timer-00.txt
>> 
>> Pardon the perhaps very naive question, but I suppose I am not as versant
>> in 
>> SIP as I would like to be.
>> 
>> My understanding is that a SIP Client issues an Invite message to a SIP
>> Server,
>> which in turn forwards the request to a target SIP Client (or some GW or
>> some
>> sort). Once the ACK has returned to the originating client, I was under
>> the
>> impression that both SIP Clients send RTP packets directly to each other,
>> bypassing the SIP Server (this is where I am unclear).
>> 
>	This depends on the implementation of the SIP Server.  There are
>valid reasons to have the SIP Server proxy the IP packets.  But in general,
>you are correct.  There will be SIP Servers that do not have a view of the
>RTP flow.
>
>> So if this is true, how will a SIP Server (stateful, or stateless)
enforce
>> any
>> session timeouts? A malicious SIP Client can simply keep the session
>> active
>> for as long as it wishes, ignoring any messages from the SIP Server that
>> may
>> tell it to disconnect.
>> 
>> This, as far as I can see, could cause some serious problems if one
wishes
>> to
>> deploy a service and still maintain some control over the sessions.
>> 
>> Again, perhaps I am way off here. Any clarifications would certainly be 
>> appreciated.
>> 
>	You are correct and I personally would not rely on reporting from
>the SIP Server to do reliable billing.  However, if a stateful SIP Proxy
>server is built, for whatever reason, the current SIP specification does
not
>have a mechanism for the SIP server to know which of the sessions for which
>it has state are valid.  This draft gives the SIP Server a mechanism to
>remove those sessions that it feels are no longer valid.
>
>	The amount of control that a SIP server exerts over sessions is a
>separate issue.  As John Hearty accurately points out in a separate email,
>the user of the SIP device can choose to bypass the SIP server entirely.
>The challenge for service providers will be to implement services that make
>user's want to access the service provider's SIP server.

This is not actually true. It is quite simple for service providers to
install
a set of filters in their systems that require that all RTP traffic be
directed
to one of their local SIP Servers.

Note that I am not advocating such an idea. I like free service, but one
must
realize that free services are rarely every deployed by the services 
providers :(

PatC
>
>> PatC
>> 
>> >This time with the draft attached.  Sorry for the extra mail.
>> >
>> >Steve
>> >
>> > <<draft-ietf-mmusic-sip-session-timer-00.txt>> 
>> >
>> >> -----Original Message-----
>> >> From:	Donovan, Steven R. (MCI) 
>> >> Sent:	Thursday, February 11, 1999 10:17 AM
>> >> To:	'confctrl@ISI.EDU'; 'iptel@lists.research.bell-labs.com'
>> >> Subject:	draft-ietf-mmusic-sip-session-timer-00.txt
>> >> 
>> >> This message is notification of the following Internet Draft which has
>> >> been submitted to the IETF for posting.
>> >> 
>> >> <draft-ietf-mmusic-sip-session-timer-00.txt> 
>> >> 
>> >> Abstract: 
>> >> 
>> >> This document proposes an extension to the SIP specification.  This
>> >> extension adds a new message header that is used to specify the 
>> >> duration of a requested session.
>> >> 
>> >> The session timer can be used to limit the total duration of a session
>> >> if, for instance, one of the participants in the session wants to
>> >> limit the cost of the session.  It can also be used by stateful SIP
>> >> Proxy Servers to track the status of sessions for which session state
>> >> exists on the servers. Currently a stateful SIP Proxy Server that is 
>> >> not handling the media stream(s) for the session has no mechanism to 
>> >> definitively determine the state of all sessions for which it has
>> state.
>> >> While the SIP Specification does provide the BYE method for
terminating
>> >> the session, there is no mechanism for a SIP Proxy Server to detect
the
>> >> end of a session when the BYE message is not sent or is lost due to 
>> >> network problems.
>> >> 
>> >> The internet draft is attached.
>> >> 
>> >> Steve Donovan
>> >> MCIWorldcom
>> 



---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 20:37:57 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id UAA25611
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 20:37:57 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 440EA7FD0; Thu, 11 Feb 1999 20:27:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 80F5E7FD1; Thu, 11 Feb 1999 20:27:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199902120125.RAA18009@hsmpka.eng.sun.com>
Date: Thu, 11 Feb 1999 17:20:55 -0800
To: "Francois Menard Distribution List Management Account" <fm-listproc@mediatrix.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>There is no way to properyly time the session unless the end-point is a
>controlled network element.  A gatekeeper and/or SIP server cant do anything
>unless the media stream gets through it ... of course, a SIP server, and or
>H.323 gatekeeper can talk to routers, which control an IP media stream, and
>affect the media stream.
>
>Even worse, a sip server / gatekeeper that wishes to track the end of a call
>so that it can bill by the second needs to worry about end points that just
>go out and die and never send an end of conversation ... this mechanism is
>not yet specified ...

Yes the cellular network that the same problem, hence the "end" button
on your phone. This would require a periodical update from the SIP Client 
to the SIP Server (a refresh of some sort).

PatC
>
>-=Francois=-
>
>-----Original Message-----
>From: Pat.Calhoun@Eng.Sun.COM [mailto:Pat.Calhoun@Eng.Sun.COM]
>Sent: Thursday, February 11, 1999 6:22 PM
>To: Donovan, Steven R. (MCI); 'confctrl@ISI.EDU';
>'iptel@lists.research.bell-labs.com'; 'Pat.Calhoun@Eng.Sun.COM'
>Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
>
>
>
>>
>>
>>> -----Original Message-----
>>> From:	Pat.Calhoun@Eng.Sun.COM [SMTP:Pat.Calhoun@Eng.Sun.COM]
>>> Sent:	Thursday, February 11, 1999 1:02 PM
>>> To:	Donovan, Steven R. (MCI); 'iptel@lists.research.bell-labs.com';
>>> 'confctrl@ISI.EDU'
>>> Subject:	RE: draft-ietf-mmusic-sip-session-timer-00.txt
>>> 
>>> Pardon the perhaps very naive question, but I suppose I am not as versant
>>> in 
>>> SIP as I would like to be.
>>> 
>>> My understanding is that a SIP Client issues an Invite message to a SIP
>>> Server,
>>> which in turn forwards the request to a target SIP Client (or some GW or
>>> some
>>> sort). Once the ACK has returned to the originating client, I was under
>>> the
>>> impression that both SIP Clients send RTP packets directly to each other,
>>> bypassing the SIP Server (this is where I am unclear).
>>> 
>>	This depends on the implementation of the SIP Server.  There are
>>valid reasons to have the SIP Server proxy the IP packets.  But in general,
>>you are correct.  There will be SIP Servers that do not have a view of the
>>RTP flow.
>>
>>> So if this is true, how will a SIP Server (stateful, or stateless)
>enforce
>>> any
>>> session timeouts? A malicious SIP Client can simply keep the session
>>> active
>>> for as long as it wishes, ignoring any messages from the SIP Server that
>>> may
>>> tell it to disconnect.
>>> 
>>> This, as far as I can see, could cause some serious problems if one
>wishes
>>> to
>>> deploy a service and still maintain some control over the sessions.
>>> 
>>> Again, perhaps I am way off here. Any clarifications would certainly be 
>>> appreciated.
>>> 
>>	You are correct and I personally would not rely on reporting from
>>the SIP Server to do reliable billing.  However, if a stateful SIP Proxy
>>server is built, for whatever reason, the current SIP specification does
>not
>>have a mechanism for the SIP server to know which of the sessions for which
>>it has state are valid.  This draft gives the SIP Server a mechanism to
>>remove those sessions that it feels are no longer valid.
>>
>>	The amount of control that a SIP server exerts over sessions is a
>>separate issue.  As John Hearty accurately points out in a separate email,
>>the user of the SIP device can choose to bypass the SIP server entirely.
>>The challenge for service providers will be to implement services that make
>>user's want to access the service provider's SIP server.
>
>This is not actually true. It is quite simple for service providers to
>install
>a set of filters in their systems that require that all RTP traffic be
>directed
>to one of their local SIP Servers.
>
>Note that I am not advocating such an idea. I like free service, but one
>must
>realize that free services are rarely every deployed by the services 
>providers :(
>
>PatC
>>
>>> PatC
>>> 
>>> >This time with the draft attached.  Sorry for the extra mail.
>>> >
>>> >Steve
>>> >
>>> > <<draft-ietf-mmusic-sip-session-timer-00.txt>> 
>>> >
>>> >> -----Original Message-----
>>> >> From:	Donovan, Steven R. (MCI) 
>>> >> Sent:	Thursday, February 11, 1999 10:17 AM
>>> >> To:	'confctrl@ISI.EDU'; 'iptel@lists.research.bell-labs.com'
>>> >> Subject:	draft-ietf-mmusic-sip-session-timer-00.txt
>>> >> 
>>> >> This message is notification of the following Internet Draft which has
>>> >> been submitted to the IETF for posting.
>>> >> 
>>> >> <draft-ietf-mmusic-sip-session-timer-00.txt> 
>>> >> 
>>> >> Abstract: 
>>> >> 
>>> >> This document proposes an extension to the SIP specification.  This
>>> >> extension adds a new message header that is used to specify the 
>>> >> duration of a requested session.
>>> >> 
>>> >> The session timer can be used to limit the total duration of a session
>>> >> if, for instance, one of the participants in the session wants to
>>> >> limit the cost of the session.  It can also be used by stateful SIP
>>> >> Proxy Servers to track the status of sessions for which session state
>>> >> exists on the servers. Currently a stateful SIP Proxy Server that is 
>>> >> not handling the media stream(s) for the session has no mechanism to 
>>> >> definitively determine the state of all sessions for which it has
>>> state.
>>> >> While the SIP Specification does provide the BYE method for
>terminating
>>> >> the session, there is no mechanism for a SIP Proxy Server to detect
>the
>>> >> end of a session when the BYE message is not sent or is lost due to 
>>> >> network problems.
>>> >> 
>>> >> The internet draft is attached.
>>> >> 
>>> >> Steve Donovan
>>> >> MCIWorldcom
>>> 
>
>
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 21:01:46 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id VAA25779
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 21:01:45 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 897057FD5; Thu, 11 Feb 1999 20:43:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BB7C67FD1; Thu, 11 Feb 1999 20:43:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990212124447.007603a8@mail.ozemail.com.au>
X-Sender: dclowes@mail.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Fri, 12 Feb 1999 12:44:47 +1100
To: <Pat.Calhoun@Eng.Sun.COM>, <Pat.Calhoun@Eng.Sun.COM>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
Cc: "'confctrl@ISI.EDU'" <confctrl@isi.edu>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
In-Reply-To: <199902120009.QAA27818@hsmpka.eng.sun.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 16:05 1999-02-11 -0800, Patrice Calhoun wrote:

People talking about completely different subjects in the same thread is
possible - been there, done that, got the flame-broiled t-shirt.

>Well, perhaps I missed something fundamental here, but my understanding is
that
>what we have here is analogous to two telephones using the PSTN to get the
>call endpoint established, and then completely ignores the PSTN. 
>

I don't think that the analogy quite fits.

In the PSTN world (which varies from country to country, if not county to
county) you have a {C|I}LEC which controls your access to services. Beyond
them is the long-distance carrier, which the LEC may control, or you may
have choices. Beyond that is the international carrier, and then ... the
terminating LEC. The LEC desires to control your choices all the way, and
only legislation prevents them from doing so, but only to the extent of the
long-distance carrier.

The IP world is different (in case you haven't noticed :), because (I
think) that once you have selected your LEC-equivalent access provider,
your carriers are fixed. You can, however, choose the services you use in
the IP world, just as you choose the numbers you dial in the PSTN world. IP
telephony is a service in the IP world.

The connection-broker and the carrier may be different in the IP world, but
not the PSTN world. You use the connection-broker to get the call
connected, and then the IP-carrier to continue the call. There will
possibly be connection-broker service providers who want to tie you in and
sell you a transport service that they don't provide, and you already pay
for to your ISP. Avoid them.

To extend the model, in cases where you are using a PSTN gateway to egreee
into the POTS world, there may be another service provider involved. This
is a service you may not be able to avoid.

If you use your POTS phone and an ingress gateway, there may be yet another
service provider involved. They will have more success in charging you for
transport, which they do provide, and egress, for which they have to pay.

>If this is the case, certainly the PSTN carrier can charge for the initial 
>endpoint resolution information, but is not involved at all afterwards. So
>an intelligent user could figure out what physical port on which switch to
>contact to get to his mother, and completely bypass the IN.

If your mother has an IP phone, and she tells you her IP address, go for
it. If you have to pay someone for it, keep it. If it's dynamic, you'll
have to get it each time ...

If your mother has a black phone, and her switch provider lets you do that,
go for it. I think they'll want to charge you, or her, for access.

>My only concern here is how does the service provider offer a service *and*
>make at least enough money to recover the capital costs (of course, certain
>providers may want to cover many additional costs, which is outside the scope
>of this discussion :).

Well, I think the terms "what the market will bear", and "supply and
demand" may come into effect here. The term "caveat emptor" should too,
because there are people out there who will ask you to pay for the air you
breath (some people call them "government").

The question I ask, is what service is the provider actually providing, and
what are the capital cost involved. Also, unless there is a healthy profit
margin, most entrepreneurs will take their capital somewhere that there is.

Your access provider (ISP) may charge you by the connect-second, by the
calendar-month, by the byte, or all of the above. They have to provide you
with access, and backbone transport, and that's about it. Is your ISP
charging $19.95/month flat-rate? A cable modem here will cost you $65/month
plus $0.30/Mb. Either way, we already pay for connection and transport.

What is the Internet Telephony Service Provider (ITSP) providing if you and
your mother both have NetMissing, and you use
DNS/LDAP/key-it-in-dotted-decimal as your connection broker? Well, if you
use a $0.20 per lookup LDAP service ...

No, the major service providers will be ingress/egress gateway service
providers. That's where your PSTN analogies become more real, and so does
the service providers control. And the egress GW service provider has to
pay the terminating LEC for access ...

And when the LEC offers egress GW service in their switch, and charge the
ITSP higher access charges ...

>PatC
>
>DISCLAIMER: I have a funny feeling that I've missed something fundamental
here.
>Please feel free to correct my statement above if it is technically
incorrect.
>

Be carefull you correctly identify which services you provide.

Douglas


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 21:39:58 1999
Received: from dirty.research.bell-labs.com (dirty.research.bell-labs.com [204.178.16.6])
	by ietf.org (8.8.5/8.8.7a) with SMTP id VAA26041
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 21:39:57 -0500 (EST)
Received: from lists.research.bell-labs.com ([135.180.161.172]) by dirty; Thu Feb 11 21:38:52 EST 1999
Received: by lists.research.bell-labs.com (Postfix)
	id 606417FD2; Thu, 11 Feb 1999 21:21:40 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 838937FDB; Thu, 11 Feb 1999 21:21:39 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36C38FF9.4D179AF5@dnrc.bell-labs.com>
Date: Thu, 11 Feb 1999 21:20:41 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Pat.Calhoun@Eng.Sun.COM
Cc: Douglas Clowes <dclowes@ozemail.com.au>,
        "'confctrl@ISI.EDU'" <confctrl@isi.edu>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
References: <199902120009.QAA27818@hsmpka.eng.sun.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Patrice Calhoun wrote:
> 
> My only concern here is how does the service provider offer a service *and*
> make at least enough money to recover the capital costs (of course, certain
> providers may want to cover many additional costs, which is outside the scope
> of this discussion :).

It comes back to the same question thats been brought up here on
numerous occassions - what are you billing for? SIP provides the setup
services, not the transport services. So, it makes sense for a SIP
server to charge for what it actually does - set the call up. It really
doesn't make sense, I don't think, to try to bill by the minute for the
SIP part of the call, since its really a transaction service. SIP is
only one part of the picture, though. There is also transport service -
diffserv prioritization, for example. A service provider can happily
charge by the minute,bit, packet, or whatever here.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 22:06:10 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id WAA26204
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 22:06:09 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D7FF87FD3; Thu, 11 Feb 1999 21:55:21 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0E8897FD4; Thu, 11 Feb 1999 21:55:20 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36C397FE.5168D5B8@cs.columbia.edu>
Date: Thu, 11 Feb 1999 21:54:54 -0500
From: Henning Schulzrinne <hgs@cs.columbia.edu>
Organization: Columbia University, Dept. of Computer Science
X-Mailer: Mozilla 4.5 [en] (X11; I; SunOS 5.6 sun4u)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Pat.Calhoun@Eng.Sun.COM, Douglas Clowes <dclowes@ozemail.com.au>,
        "'confctrl@ISI.EDU'" <confctrl@isi.edu>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
References: <199902120009.QAA27818@hsmpka.eng.sun.com> <36C38FF9.4D179AF5@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Jonathan Rosenberg wrote:
> 
> Patrice Calhoun wrote:
> >
> > My only concern here is how does the service provider offer a service *and*
> > make at least enough money to recover the capital costs (of course, certain
> > providers may want to cover many additional costs, which is outside the scope
> > of this discussion :).
> 
> It comes back to the same question thats been brought up here on
> numerous occassions - what are you billing for? SIP provides the setup
> services, not the transport services. So, it makes sense for a SIP
> server to charge for what it actually does - set the call up. It really
> doesn't make sense, I don't think, to try to bill by the minute for the
> SIP part of the call, since its really a transaction service. SIP is
> only one part of the picture, though. There is also transport service -
> diffserv prioritization, for example. A service provider can happily
> charge by the minute,bit, packet, or whatever here.

One argument for a closer relationship between a transport stream and
signaling is that signaling may have access to information not (easily)
available to the lower layers, for example, customer-level strong
authentication. It could be argued that one should fix, say, RSVP to
include this information in the appropriate form since it may well be
needed for other, non-telephony services like media-on-demand.

Another point is that Internet telephony may well fail in more
interesting ways compared to telephones. For example, if I have an
architecture (as in the mbone tools) with separate audio/video media
agents and a signaling application, I may want to terminate the call if
any one of these components dies, as I may loose control over the rest. 
You could easily have the SIP call setup agent stay up (one would hope
so), with the audio agent hung and gone. The customer can't talk and
will demand a refund. Similarly, if the audio agent stays around (in the
background) and merrily sends receiver reports, but the SIP agent has
bid farewell in the French style, the customer also would want to not
pay for this useless session, even if the number called happens to be a
1-900 number with a continuous loop of the daily horoscope.
-- 
Henning Schulzrinne   http://www.cs.columbia.edu/~hgs

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 23:03:20 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id XAA28377
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 23:03:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5351A7FD1; Thu, 11 Feb 1999 21:01:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7CCB17FDB; Thu, 11 Feb 1999 21:01:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990212130140.0148bbf4@mail.ozemail.com.au>
X-Sender: dclowes@mail.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Fri, 12 Feb 1999 13:01:40 +1100
To: Francois Menard Distribution List Management Account <fm-listproc@mediatrix.com>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
In-Reply-To: <F16674FCE856D211AC0D00E02910AE0A07B517@INTERNAL>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Bonjour Francois, Ca va?

It's about the services that you provide, and the services that you like to
charge for.

If you sell the users Mediatrix gateways, then it is their capital
investment, they would like to be in control, and amortize their
investments. They don't want to pay you for each second they use them, no
matter how much you would like them to.

If they are your gateways, that you connect to the PSTN, and provide an
egress service, then you should be in control, and you can control charging
for the service.

At 20:04 1999-02-11 -0500, Francois Menard Distribution List Management
Account wrote:
>There is no way to properyly time the session unless the end-point is a
>controlled network element.  A gatekeeper and/or SIP server cant do anything
>unless the media stream gets through it ... of course, a SIP server, and or
>H.323 gatekeeper can talk to routers, which control an IP media stream, and
>affect the media stream.

If it's your gateway (you bought it) and you need it to obey your
controller (SIP server or gatekeeper) and it doesn't, then send it back and
get one that will. If it's your gateway (you bought it) and you don't want
it to obey someone elses controller and it does, send it back and get one
that does what you want it to.

>Even worse, a sip server / gatekeeper that wishes to track the end of a call
>so that it can bill by the second needs to worry about end points that just
>go out and die and never send an end of conversation ... this mechanism is
>not yet specified ...

And if it wishes to bill by the byte, it needs endpoints that will report
resource usage. It needs endpoints that will report periodically, or store
records in non-volatile storage to report when they come up.

Work is in progress, in both ITU SG-16, and ETSI EP-TIPHON.

>-=Francois=-




---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 11 23:18:51 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id XAA28412
	for <iptel-archive@odin.ietf.org>; Thu, 11 Feb 1999 23:18:51 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 760AE7FCD; Thu, 11 Feb 1999 18:59:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C38E47FCE; Thu, 11 Feb 1999 18:59:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990212110125.006e80bc@mail.ozemail.com.au>
X-Sender: dclowes@mail.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Fri, 12 Feb 1999 11:01:25 +1100
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>, Pat.Calhoun@Eng.Sun.COM
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
Cc: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@isi.edu>
In-Reply-To: <36C35066.852BC71B@dnrc.bell-labs.com>
References: <199902111906.LAA28606@hsmpka.eng.sun.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 16:49 1999-02-11 -0500, Jonathan Rosenberg wrote:
>Patrice Calhoun wrote:
>> 
>> My understanding is that a SIP Client issues an Invite message to a SIP
Server,
>> which in turn forwards the request to a target SIP Client (or some GW or
some
>> sort). Once the ACK has returned to the originating client, I was under the
>> impression that both SIP Clients send RTP packets directly to each other,
>> bypassing the SIP Server (this is where I am unclear).
>
>Slight correction - ACK goes from caller to callee.
>
>But, in general, you are right. RTP bypasses the SIP servers.
>
>> 
>> So if this is true, how will a SIP Server (stateful, or stateless)
enforce any
>> session timeouts? A malicious SIP Client can simply keep the session active
>> for as long as it wishes, ignoring any messages from the SIP Server that
may
>> tell it to disconnect.
>> 
>> This, as far as I can see, could cause some serious problems if one
wishes to
>> deploy a service and still maintain some control over the sessions.
>

One wonders about the nature of the service you think to deploy. Where you
want to maintain session control without control of at least one enpoint.

>This is an excellent, excellent point, and is something I have been
>considering since the whole timeout/DTMF/INFO issues came up. The
>question you ask gets back to a more basic problem.

>Since, unlike
>traditional telephony, voice does not flow through the servers
>(gatekeepers or SIP servers or MXCP call agents) in Internet telephony,
>any sort of "control" a server would like to execute is based on an
>assumption that the user agents (the ones sending/receiving the media)
>would obey the servers.

Indeed, in traditional telephony (at least the SS7 kind), voice does not
flow through the control points. Control is maintained by having voice
nodes that obey control nodes.

In H.323 gatekeepers, the gatekeeper may route the control channels between
the endpoints, and issue termination commands in each direction, on those
channels. A gatekeeper may route the media through itself (although this is
unlikely), or another entity. What you wouldchoose to do would depend on
your trust of the level of control you exert on the endpoints, and the
consequences of any lack of control.

If you don't trust either endpoint, but must control the media session, or
conceal the endpoints from each other, you may have to route the media
through a Meter Box (tm), that you can trust. There are gateways that
transcode media (G.711 to G.723.1 for example) and a similar mechanism,
without transcoding, could be deployed.

> Consider the example of call transfer. Lets say
>we'd like to signal the service using DTMF, so a proxy subscribes to the
>DTMF from the client. The client presses the appropriate buttons on the
>phone, and the proxy decides to transfer. But, the media doesn't flow
>through the proxy, so it has no way of actually transferring the call.
>Rather, it will need to go back and tell the user agent to perform the
>transfer, since only the user agent can affect where the media goes to.
>In that case, why would the UA ask the network for the service in the
>first place?  It can simply be invoked directly by the client. In a
>network where voice transport is end to end, services based on changing
>characteristics of that transport (coding, destination, etc), must also
>be end to end. 
>
>Now, if the action the proxy asks to user to take is something the user
>wants to do anyway, maybe thats OK. But, if its something less
>cooperative, like forcing a hangup, or transferring to an operator,
>there is simply no incentive for the user to accept this. 
>
>-Jonathan R.
>-- 
>Jonathan D. Rosenberg                       Lucent Technologies
>Member of Technical Staff                   101 Crawfords Corner Rd.
>High Speed Networks Research                Holmdel, NJ 07733
>FAX: (732) 834-5379                         Rm. 4C-526
>EMAIL: jdrosen@bell-labs.com
>URL: http://www.cs.columbia.edu/~jdrosen
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.
>


Douglas


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 04:39:59 1999
Received: from dirty.research.bell-labs.com (dirty.research.bell-labs.com [204.178.16.6])
	by ietf.org (8.8.5/8.8.7a) with SMTP id EAA07035
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 04:39:58 -0500 (EST)
Received: from lists.research.bell-labs.com ([135.180.161.172]) by dirty; Fri Feb 12 04:38:49 EST 1999
Received: by lists.research.bell-labs.com (Postfix)
	id 855827FD6; Fri, 12 Feb 1999 04:19:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D78AA7FD7; Fri, 12 Feb 1999 04:19:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Bodo Rueskamp <br@rueskamp.com>
Message-Id: <199902120918.KAA04836@smilla.rueskamp.com>
Subject: Re: Terminals Implementing H.323 RAS Protocol
In-Reply-To: <01BE568B.6C776740@hss056.hss.hns.com> from Alok Mittal at "Feb 12, 1999  1:27:21 pm"
To: amittal@hss.hns.com (Alok Mittal)
Date: Fri, 12 Feb 1999 10:18:40 +0100 (CET)
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
X-Mailer: ELM [version 2.4ME+ PL43 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
X-MIME-Autoconverted: from 8bit to quoted-printable by smilla.rueskamp.com id KAA04836
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by ietf.org id EAA07035


>  If you know of any such terminals, please let us know! I am
>  currently trying out Intel Trial Applet v2.2, but that keeps
>  sending new GRQs even after a GCF is sent (and never
>  sends a RRQ). It is also not clear to me if the "nameserver"
>  mentioned in there is supposed to refer to a H.323 Gatekeeper.

The Intel Trial Applet v2.2 works fine with my gatekeeper.
The "nameserver" refers to a directory server (LDAP?).

; Bodo

-- 
Bodo Rüskamp, br@rueskamp.com, 51°55' N 7°41' E
(1) Die Erde ist flach.
(2) Elvis lebt.
(3) Das neue Jahrtausend beginnt am 1.1.2000.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 04:53:21 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id EAA07097
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 04:53:21 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 422BB7FD4; Fri, 12 Feb 1999 03:53:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 976807FD6; Fri, 12 Feb 1999 03:53:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BE568B.6C776740@hss056.hss.hns.com>
From: Alok Mittal <amittal@hss.hns.com>
To: "h323implementors@imtc.org" <h323implementors@imtc.org>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Terminals Implementing H.323 RAS Protocol
Date: Fri, 12 Feb 1999 13:27:21 -0000
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi

 We are looking at some inter-operability testing on H.323
 RAS (version 1 and/or version 2) protocol -- we need some
 free (preferred :)/paid terminal software which implement 
 this protocol.

 If you know of any such terminals, please let us know! I am
 currently trying out Intel Trial Applet v2.2, but that keeps
 sending new GRQs even after a GCF is sent (and never
 sends a RRQ). It is also not clear to me if the "nameserver"
 mentioned in there is supposed to refer to a H.323 Gatekeeper.

Thanks in Advance!
Alok

-----------------------------------------------------------------
Alok Mittal	
Technical Leader,	Hughes Software Systems
Prestige Opal, 146 Infantry Road, Bangalore - 560001
Tel: +91-80-2286390,91,92	Fax: +91-80-2286393
-----------------------------------------------------------------


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 08:30:01 1999
Received: from dirty.research.bell-labs.com (dirty.research.bell-labs.com [204.178.16.6])
	by ietf.org (8.8.5/8.8.7a) with SMTP id IAA08793
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 08:30:00 -0500 (EST)
Received: from lists.research.bell-labs.com ([135.180.161.172]) by dirty; Fri Feb 12 08:28:15 EST 1999
Received: by lists.research.bell-labs.com (Postfix)
	id 0691B7FD8; Fri, 12 Feb 1999 08:09:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3EFDD7FD9; Fri, 12 Feb 1999 08:09:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: ernst.horvath@siemens.at
X-OpenMail-Hops: 1
Date: Fri, 12 Feb 1999 12:07:36 +0100
Message-Id: <H00011ad0259790c@MHS>
Subject: AW: RE: draft-ietf-mmusic-sip-session-timer-00.txt
MIME-Version: 1.0
To: confctrl@ISI.EDU, fm-listproc@mediatrix.com,
        iptel@lists.research.bell-labs.com, Pat.Calhoun@Eng.Sun.COM,
        Steven.R.Donovan@mci.com
Content-Type: text/plain; charset=ISO-8859-1; name="BDY.TXT"
Content-Disposition: inline; filename="BDY.TXT"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by ietf.org id IAA08793



-----Ursprüngliche Nachricht-----
Von: fm-listproc [SMTP:fm-listproc@mediatrix.com]
Gesendet am: Freitag, 12. Februar 1999 02:05
An: Pat.Calhoun; Steven.R.Donovan; confctrl; iptel
Cc: fm-listproc
Betreff: RE: draft-ietf-mmusic-sip-session-timer-00.txt

There is no way to properyly time the session unless the end-point is a
controlled network element.  A gatekeeper and/or SIP server cant do  
anything
unless the media stream gets through it ... of course, a SIP server,  
and or
H.323 gatekeeper can talk to routers, which control an IP media stream,  
and
affect the media stream.

Even worse, a sip server / gatekeeper that wishes to track the end of a  
call
so that it can bill by the second needs to worry about end points that  
just
go out and die and never send an end of conversation ... this mechanism  
is
not yet specified ...

[Ernst Horvath]  Not quite true - for the gatekeeper, ITU-T  
Recommendation H.225.0 specifies the Disengage message for the regular  
termination of a call (i.e. the gatekeeper is informed immediately upon  
termination), but there is also the Status request/report procedure to  
check the status of an end point. If exchanged at regular intervals  
(say every n seconds, while a call exists), the gatekeeper will notice  
the "death" of an end point with a delay of at most n seconds.

 -=Francois=-

Ernst Horvath
Siemens AG
Emai: ernst.horvath@siemens.at   


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 09:39:00 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id JAA09690
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 09:39:00 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E10627FD7; Fri, 12 Feb 1999 07:57:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4649F7FD8; Fri, 12 Feb 1999 07:57:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <F16674FCE856D211AC0D00E02910AE0A07B518@INTERNAL>
From: Francois Menard Distribution List Management Account <fm-listproc@mediatrix.com>
To: "'Douglas Clowes'" <dclowes@ozemail.com.au>,
        Francois Menard Distribution List Management Account <fm-listproc@mediatrix.com>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Fri, 12 Feb 1999 07:56:19 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi Douglas, how is it in the summer in Australia ?

> If they are your gateways, that you connect to the PSTN, and provide an
egress service, then you should be in control, and you can control charging
for the service.

Yes, its pretty easy to track PSTN resource usage, because the media stream
towards the PSTN must go through them ...

> If it's your gateway (you bought it) and you need it to obey your
controller (SIP server or gatekeeper) and it doesn't, then send it back and
get one that will. If it's your gateway (you bought it) and you don't want

I'm not against this scenario, its proven for militarized zones (i.e.
Intranets in entreprises controlled by MIS). Where I worrry is in DMZs,
where an ISP forces me to use its gatekeeper.  That's where I'm saying that
the end point has to become a trusted network element, in the control of the
ISP, in order to bill by the second/bit.  The control of the session becomes
in practice when the routers of the ISP refuse to send bits on the network
on the behalf of non-trusted (to behave according to the ISP's policy) SIP
end point (I guess that's another form of admission control ...).  This is
the way that the MSO's are gearing up their VoIP network presently
(PacketCable).  I.e., if you want to bill by the minute/second/bit, you
either need to route the media stream through the SIP server and/or make the
SIP server control the admission of the bits to the IP layer ... that is,
there needs to be some form of policy decision flowing between a SIP server
and routers that forwards the media stream bits from SIP endpoints ...

>>Even worse, a sip server / gatekeeper that wishes to track the end of a
call
>>so that it can bill by the second needs to worry about end points that
just
>>go out and die and never send an end of conversation ... this mechanism is
>>not yet specified ...
And if it wishes to bill by the byte, it needs endpoints that will report
resource usage. It needs endpoints that will report periodically, or store
records in non-volatile storage to report when they come up.
Work is in progress, in both ITU SG-16, and ETSI EP-TIPHON.

So I can have the following, when a SIP server never gets a BYE, it doesn't
need to ping the end point to see if its still alive, it can simply wait
until the end-point re-REGISTERs, and then flag the call as terminated, and
then, query the end point to tell it how many seconds were spent on line
before the end-point died ... has anybody thought of extending the REGISTER
message to include the duration of the previous call?  I think that this is
a pretty cool idea ;-)  Of course, if the end-point never re-registers, the
timer will still be running ... indefinitely.  We have a new policy on IP
networks, any call of length greater than 1 million minutes is free ...

BTW, I asked the question today to a supplier of H.323 stacks, as to what
generally did the commercial gatekeepers do in practice to track the end of
calls.  The person told me that the gatekeepers generally wait for the next
registration (i.e. their assumption is that the H.323 terminal will
re-register within hours), to flag the call as terminated.  They also rely
on either one of the two legs of the call to send a termination message, so
the case where the two legs are cut off never happens very frequently.  As
to what is the mechanism that a gatekeeper is supposed to use to "ping" the
terminal, that's not in the standard in their opinion.  Personally, I begin
to like the idea of the end-point presenting the duration of the last call
in the registration message ...

-=Francois=-
fmenard@mediatrix.com

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 09:49:36 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id JAA09845
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 09:49:35 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 172EA7FD9; Fri, 12 Feb 1999 09:26:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 53BEE7FDA; Fri, 12 Feb 1999 09:26:28 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36C4397F.B7E61E7C@dnrc.bell-labs.com>
Date: Fri, 12 Feb 1999 09:23:59 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Francois Menard Distribution List Management Account <fm-listproc@mediatrix.com>
Cc: "'Douglas Clowes'" <dclowes@ozemail.com.au>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'confctrl@ISI.EDU'" <confctrl@isi.edu>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
References: <F16674FCE856D211AC0D00E02910AE0A07B518@INTERNAL>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Francois Menard Distribution List Management Account wrote:
> 
> > If it's your gateway (you bought it) and you need it to obey your
> controller (SIP server or gatekeeper) and it doesn't, then send it back and
> get one that will. If it's your gateway (you bought it) and you don't want
> 
> I'm not against this scenario, its proven for militarized zones (i.e.
> Intranets in entreprises controlled by MIS). Where I worrry is in DMZs,
> where an ISP forces me to use its gatekeeper.  That's where I'm saying that
> the end point has to become a trusted network element, in the control of the
> ISP, in order to bill by the second/bit.  The control of the session becomes
> in practice when the routers of the ISP refuse to send bits on the network
> on the behalf of non-trusted (to behave according to the ISP's policy) SIP
> end point (I guess that's another form of admission control ...).  This is
> the way that the MSO's are gearing up their VoIP network presently
> (PacketCable).  I.e., if you want to bill by the minute/second/bit, you
> either need to route the media stream through the SIP server and/or make the
> SIP server control the admission of the bits to the IP layer ... that is,
> there needs to be some form of policy decision flowing between a SIP server
> and routers that forwards the media stream bits from SIP endpoints ...

Well, bits are bits. Why would you want to control "admission" of bits
for this application, and not for others? If the model is that best
effort traffic is not billed by the bit, then you cannot bill for IP
telephony transport when best effort is used. Now, if you have diffserv,
then it seems quite reasonable to have the diffserv treatments get
"turned off" (i.e., packets go back to best effort treatment) when the
session terminates. In that case, billing would probably best be done
for transport in the COPS server, bandwidth broker, or whatever diffserv
element normally provides this function.


> 
> >>Even worse, a sip server / gatekeeper that wishes to track the end of a
> call
> >>so that it can bill by the second needs to worry about end points that
> just
> >>go out and die and never send an end of conversation ... this mechanism is
> >>not yet specified ...
> And if it wishes to bill by the byte, it needs endpoints that will report
> resource usage. It needs endpoints that will report periodically, or store
> records in non-volatile storage to report when they come up.
> Work is in progress, in both ITU SG-16, and ETSI EP-TIPHON.
> 
> So I can have the following, when a SIP server never gets a BYE, it doesn't
> need to ping the end point to see if its still alive, it can simply wait
> until the end-point re-REGISTERs, and then flag the call as terminated, and
> then, query the end point to tell it how many seconds were spent on line
> before the end-point died 

Registration is separate from in-progress calls. A client periodically
re-registers independently from whether its in a call or not. A client
may have been in many calls between two registers. Would you include
durations for all of these? Why would the user agent report them anyway,
being it is just going to get billed (again, keeping the DMZ model in
mind). I don't see the benefit of tying these mechanisms together.

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 10:59:48 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id KAA11004
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 10:59:48 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E0D567FDA; Fri, 12 Feb 1999 10:45:25 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 373C17FDB; Fri, 12 Feb 1999 10:45:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199902121543.HAA11937@hsmpka.eng.sun.com>
Date: Fri, 12 Feb 1999 07:39:20 -0800
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        <Pat.Calhoun@Eng.Sun.COM>
Cc: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@isi.edu>,
        "Douglas Clowes" <dclowes@ozemail.com.au>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>Patrice Calhoun wrote:
>> 
>> My only concern here is how does the service provider offer a service *and*
>> make at least enough money to recover the capital costs (of course, certain
>> providers may want to cover many additional costs, which is outside the scope
>> of this discussion :).
>
>It comes back to the same question thats been brought up here on
>numerous occassions - what are you billing for? SIP provides the setup
>services, not the transport services. So, it makes sense for a SIP
>server to charge for what it actually does - set the call up. It really
>doesn't make sense, I don't think, to try to bill by the minute for the
>SIP part of the call, since its really a transaction service. SIP is
>only one part of the picture, though. There is also transport service -
>diffserv prioritization, for example. A service provider can happily
>charge by the minute,bit, packet, or whatever here.
>

This is perhaps my mis-understanding. I have been working with the TIA recently
and they have adopted MobileIP/DIAMETER for the next generation cellular
data networks. When one considers voice over IP using IP cellular phones, one
quickly wonders how the charging will really work. Yes, it is true that one
will charge for network access, but conceivably, one could also charge for
calls as well. This is especially true if these calls span providers.

Perhaps I have all of this wrong. I understand that charging for an LDAP
query cannot amount to much, and perhaps that is all bundled as part of the
access charge. Are there any providers on this list that have an idea on how
they will actually bill for these services?

PatC


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 12:02:37 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id MAA12052
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 12:02:37 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 96D0A7FB7; Fri, 12 Feb 1999 11:49:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E3B7E7FC9; Fri, 12 Feb 1999 11:49:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199902121647.IAA25261@hsmpka.eng.sun.com>
Date: Fri, 12 Feb 1999 08:43:02 -0800
To: "Francois Menard Distribution List Management Account" <fm-listproc@mediatrix.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        "Francois Menard Distribution List Management Account" <fm-listproc@mediatrix.com>,
        "'Douglas Clowes'" <dclowes@ozemail.com.au>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>Hi Douglas, how is it in the summer in Australia ?
>
>> If they are your gateways, that you connect to the PSTN, and provide an
>egress service, then you should be in control, and you can control charging
>for the service.
>
>Yes, its pretty easy to track PSTN resource usage, because the media stream
>towards the PSTN must go through them ...
>
>> If it's your gateway (you bought it) and you need it to obey your
>controller (SIP server or gatekeeper) and it doesn't, then send it back and
>get one that will. If it's your gateway (you bought it) and you don't want
>
>I'm not against this scenario, its proven for militarized zones (i.e.
>Intranets in entreprises controlled by MIS). Where I worrry is in DMZs,
>where an ISP forces me to use its gatekeeper.  That's where I'm saying that
>the end point has to become a trusted network element, in the control of the
>ISP, in order to bill by the second/bit.  The control of the session becomes
>in practice when the routers of the ISP refuse to send bits on the network
>on the behalf of non-trusted (to behave according to the ISP's policy) SIP
>end point (I guess that's another form of admission control ...).  This is
>the way that the MSO's are gearing up their VoIP network presently
>(PacketCable).  I.e., if you want to bill by the minute/second/bit, you
>either need to route the media stream through the SIP server and/or make the
>SIP server control the admission of the bits to the IP layer ... that is,
>there needs to be some form of policy decision flowing between a SIP server
>and routers that forwards the media stream bits from SIP endpoints ...
>
>>>Even worse, a sip server / gatekeeper that wishes to track the end of a
>call
>>>so that it can bill by the second needs to worry about end points that
>just
>>>go out and die and never send an end of conversation ... this mechanism is
>>>not yet specified ...
>And if it wishes to bill by the byte, it needs endpoints that will report
>resource usage. It needs endpoints that will report periodically, or store
>records in non-volatile storage to report when they come up.
>Work is in progress, in both ITU SG-16, and ETSI EP-TIPHON.
>
>So I can have the following, when a SIP server never gets a BYE, it doesn't
>need to ping the end point to see if its still alive, it can simply wait
>until the end-point re-REGISTERs, and then flag the call as terminated, and
>then, query the end point to tell it how many seconds were spent on line
>before the end-point died ... has anybody thought of extending the REGISTER
>message to include the duration of the previous call?  I think that this is
>a pretty cool idea ;-)  Of course, if the end-point never re-registers, the
>timer will still be running ... indefinitely.  We have a new policy on IP
>networks, any call of length greater than 1 million minutes is free ...
>
>BTW, I asked the question today to a supplier of H.323 stacks, as to what
>generally did the commercial gatekeepers do in practice to track the end of
>calls.  The person told me that the gatekeepers generally wait for the next
>registration (i.e. their assumption is that the H.323 terminal will
>re-register within hours), to flag the call as terminated.  They also rely
>on either one of the two legs of the call to send a termination message, so
>the case where the two legs are cut off never happens very frequently.  As
>to what is the mechanism that a gatekeeper is supposed to use to "ping" the
>terminal, that's not in the standard in their opinion.  Personally, I begin
>to like the idea of the end-point presenting the duration of the last call
>in the registration message ...

Having built PPP routers in the past, I would never trust the PPP client to
report usage, and generate accounting information from it. Actually, the only
boxes that I ever would really trust are the ones that I have control over,
which in this case seems to be the SIP Servers.

PatC
>
>-=Francois=-
>fmenard@mediatrix.com



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 12:23:38 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id MAA12222
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 12:23:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 31A3C7FC8; Fri, 12 Feb 1999 12:11:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 86F5E7FCB; Fri, 12 Feb 1999 12:11:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <F16674FCE856D211AC0D00E02910AE0A07B51A@INTERNAL>
From: Francois Menard <fmenard@mediatrix.com>
To: "'ernst.horvath@siemens.at'" <ernst.horvath@siemens.at>
Cc: "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        Francois Menard Distribution List Management Account <fm-listproc@mediatrix.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        "'Steven.R.Donovan@mci.com'" <Steven.R.Donovan@mci.com>
Subject: RE: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Fri, 12 Feb 1999 12:10:33 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

> [Ernst Horvath]  Not quite true - for the gatekeeper, ITU-T  
> Recommendation H.225.0 specifies the Disengage message for the regular  
> termination of a call (i.e. the gatekeeper is informed immediately upon  
> termination), but there is also the Status request/report procedure to  
> check the status of an end point. If exchanged at regular intervals  
> (say every n seconds, while a call exists), the gatekeeper will notice  
> the "death" of an end point with a delay of at most n seconds.
> 

Which means that if you want to maintain the billing accuracy that is
popular in today's PSTN, you'd have to set this frequency to one second...
Has anyone analyzed the scalability of this ? How big is the PDU of the
status message ?

-=Francois=-

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 15:04:24 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id PAA14517
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 15:04:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BB1157FB8; Fri, 12 Feb 1999 14:51:26 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2190A7FCB; Fri, 12 Feb 1999 14:51:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Sean Olson <eussean@exu.ericsson.se>
Date: Fri, 12 Feb 1999 13:49:33 -0600 (CST)
Message-Id: <199902121949.NAA03123@b04a42.exu.ericsson.se>
To: ernst.horvath@siemens.at, fmenard@mediatrix.com
Subject: RE: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Cc: confctrl@ISI.EDU, fm-listproc@mediatrix.com,
        iptel@lists.research.bell-labs.com, Pat.Calhoun@Eng.Sun.COM,
        Steven.R.Donovan@mci.com
X-Sun-Charset: US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


> 
> > [Ernst Horvath]  Not quite true - for the gatekeeper, ITU-T  
> > Recommendation H.225.0 specifies the Disengage message for the regular  
> > termination of a call (i.e. the gatekeeper is informed immediately upon  
> > termination), but there is also the Status request/report procedure to  
> > check the status of an end point. If exchanged at regular intervals  
> > (say every n seconds, while a call exists), the gatekeeper will notice  
> > the "death" of an end point with a delay of at most n seconds.
> > 
> 
> Which means that if you want to maintain the billing accuracy that is
> popular in today's PSTN, you'd have to set this frequency to one second...
> Has anyone analyzed the scalability of this ? How big is the PDU of the
> status message ?

This assumes the best case scenario w/no re-transmissions and perfect
reliability of the status requests (which does not exist for UDP SIP 
communication). The bottom line seems to be that if you want to charge
on a time usage basis, you MUST have control of the media stream which is
outside the scope of SIP itself.

Sean Olson

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 15:40:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id PAA14922
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 15:40:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7A3757FC9; Fri, 12 Feb 1999 15:27:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C0B3E7FCA; Fri, 12 Feb 1999 15:27:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <CA6966C24AC6D111B5A100805FEAB7D8577416@nsrip00208.mcit.com>
From: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
To: "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@isi.edu>,
        Douglas Clowes <dclowes@ozemail.com.au>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Fri, 12 Feb 1999 20:25:06 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE56C5.CB3AE254"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE56C5.CB3AE254
Content-Type: text/plain

This is an interesting discussion, however I'm not sure that we are going to
get answers to the questions that Pat is asking.  Only time will tell how
services providers make money off of IP based telephony services.

However, this is not the purpose of proposing the addition of the session
timer to the SIP protocol.  The primary purpose is to give the stateful SIP
Proxy Server a method of determining the validity of the individual sessions
for which it is keeping state.

Are there comments on the mechanisms proposed in the session-timer draft?

Steve

> -----Original Message-----
> From:	Pat.Calhoun@Eng.Sun.COM [SMTP:Pat.Calhoun@Eng.Sun.COM]
> Sent:	Friday, February 12, 1999 9:39 AM
> To:	Jonathan Rosenberg; Pat.Calhoun@Eng.Sun.COM
> Cc:	Donovan, Steven R. (MCI); 'iptel@lists.research.bell-labs.com';
> 'confctrl@ISI.EDU'; Douglas Clowes
> Subject:	Re: draft-ietf-mmusic-sip-session-timer-00.txt
> 
> 
> >Patrice Calhoun wrote:
> >> 
> >> My only concern here is how does the service provider offer a service
> *and*
> >> make at least enough money to recover the capital costs (of course,
> certain
> >> providers may want to cover many additional costs, which is outside the
> scope
> >> of this discussion :).
> >
> >It comes back to the same question thats been brought up here on
> >numerous occassions - what are you billing for? SIP provides the setup
> >services, not the transport services. So, it makes sense for a SIP
> >server to charge for what it actually does - set the call up. It really
> >doesn't make sense, I don't think, to try to bill by the minute for the
> >SIP part of the call, since its really a transaction service. SIP is
> >only one part of the picture, though. There is also transport service -
> >diffserv prioritization, for example. A service provider can happily
> >charge by the minute,bit, packet, or whatever here.
> >
> 
> This is perhaps my mis-understanding. I have been working with the TIA
> recently
> and they have adopted MobileIP/DIAMETER for the next generation cellular
> data networks. When one considers voice over IP using IP cellular phones,
> one
> quickly wonders how the charging will really work. Yes, it is true that
> one
> will charge for network access, but conceivably, one could also charge for
> calls as well. This is especially true if these calls span providers.
> 
> Perhaps I have all of this wrong. I understand that charging for an LDAP
> query cannot amount to much, and perhaps that is all bundled as part of
> the
> access charge. Are there any providers on this list that have an idea on
> how
> they will actually bill for these services?
> 
> PatC
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

------_=_NextPart_001_01BE56C5.CB3AE254
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>RE: draft-ietf-mmusic-sip-session-timer-00.txt</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">This is an =
interesting discussion, however I'm not sure that we are going to get =
answers to the questions that Pat is asking.&nbsp; Only time will tell =
how services providers make money off of IP based telephony =
services.</FONT></P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">However, this is =
not the purpose of proposing the addition of the session timer to the =
SIP protocol.&nbsp; The primary purpose is to give the stateful SIP =
Proxy Server a method of determining the validity of the individual =
sessions for which it is keeping state.</FONT></P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">Are there =
comments on the mechanisms proposed in the session-timer draft?</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">Steve</FONT>
</P>
<UL>
<P><FONT SIZE=3D1 FACE=3D"Arial">-----Original Message-----</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">From:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Pat.Calhoun@Eng.Sun.COM =
[SMTP:Pat.Calhoun@Eng.Sun.COM]</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">Sent:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Friday, February 12, 1999 9:39 AM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">Jonathan Rosenberg; Pat.Calhoun@Eng.Sun.COM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Cc:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">Donovan, Steven R. (MCI); =
'iptel@lists.research.bell-labs.com'; 'confctrl@ISI.EDU'; Douglas =
Clowes</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT>=
</B> <FONT SIZE=3D1 FACE=3D"Arial">Re: =
draft-ietf-mmusic-sip-session-timer-00.txt</FONT>
</P>
<BR>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;Patrice =
Calhoun wrote:</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; My =
only concern here is how does the service provider offer a service =
*and*</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; make =
at least enough money to recover the capital costs (of course, =
certain</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; =
providers may want to cover many additional costs, which is outside the =
scope</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;&gt; of =
this discussion :).</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;It comes =
back to the same question thats been brought up here on</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;numerous =
occassions - what are you billing for? SIP provides the setup</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;services, =
not the transport services. So, it makes sense for a SIP</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;server to =
charge for what it actually does - set the call up. It really</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;doesn't =
make sense, I don't think, to try to bill by the minute for the</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;SIP part =
of the call, since its really a transaction service. SIP is</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;only one =
part of the picture, though. There is also transport service -</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;diffserv =
prioritization, for example. A service provider can happily</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;charge by =
the minute,bit, packet, or whatever here.</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt;</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">This is =
perhaps my mis-understanding. I have been working with the TIA =
recently</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">and they have =
adopted MobileIP/DIAMETER for the next generation cellular</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">data =
networks. When one considers voice over IP using IP cellular phones, =
one</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">quickly =
wonders how the charging will really work. Yes, it is true that =
one</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">will charge =
for network access, but conceivably, one could also charge for</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">calls as =
well. This is especially true if these calls span providers.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">Perhaps I have =
all of this wrong. I understand that charging for an LDAP</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">query cannot =
amount to much, and perhaps that is all bundled as part of the</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">access =
charge. Are there any providers on this list that have an idea on =
how</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">they will =
actually bill for these services?</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">PatC</FONT>
</P>
<BR>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier =
New">---------</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">This message =
came from the IETF IPTEL Working Group Mailing List.</FONT>
</P>
</UL>
</BODY>
</HTML>
------_=_NextPart_001_01BE56C5.CB3AE254--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 16:02:36 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id QAA15551
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 16:02:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2FB0C7FCA; Fri, 12 Feb 1999 15:51:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 710BF7FCB; Fri, 12 Feb 1999 15:51:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199902122049.MAA04666@hsmpka.eng.sun.com>
Date: Fri, 12 Feb 1999 12:45:31 -0800
To: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>
Cc: "Douglas Clowes" <dclowes@ozemail.com.au>,
        "'confctrl@ISI.EDU'" <confctrl@isi.edu>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>This is an interesting discussion, however I'm not sure that we are going to
>get answers to the questions that Pat is asking.  Only time will tell how
>services providers make money off of IP based telephony services.
>
>However, this is not the purpose of proposing the addition of the session
>timer to the SIP protocol.  The primary purpose is to give the stateful SIP
>Proxy Server a method of determining the validity of the individual sessions
>for which it is keeping state.
>
>Are there comments on the mechanisms proposed in the session-timer draft?

Yes, and I appologize for side-tracking the whole thread. Is there anyway
that we could have shorter INVITEs (perhaps the SIP Server can determine
the length of a session), and have the SIP Client re-register with the 
server periodically, such as:

   SIP Client --------- INVITE ------------> SIP Server
               (some stuff happens here)
   SIP Client < ------- ACK (3 mins) ------- SIP Server
   (2 mins 45 seconds pass, enough to allow for retransmissions)
   SIP Client ---------- FOOBAR -----------> SIP Server
   (where FOOBAR is some message that refreshes the session)

This would allow the SIP Server to have soft-state and time out any sessions
that have not been re-registered (or foobar'ed in the above example).

PatC
>
>Steve
>
>> -----Original Message-----
>> From:	Pat.Calhoun@Eng.Sun.COM [SMTP:Pat.Calhoun@Eng.Sun.COM]
>> Sent:	Friday, February 12, 1999 9:39 AM
>> To:	Jonathan Rosenberg; Pat.Calhoun@Eng.Sun.COM
>> Cc:	Donovan, Steven R. (MCI); 'iptel@lists.research.bell-labs.com';
>> 'confctrl@ISI.EDU'; Douglas Clowes
>> Subject:	Re: draft-ietf-mmusic-sip-session-timer-00.txt
>> 
>> 
>> >Patrice Calhoun wrote:
>> >> 
>> >> My only concern here is how does the service provider offer a service
>> *and*
>> >> make at least enough money to recover the capital costs (of course,
>> certain
>> >> providers may want to cover many additional costs, which is outside the
>> scope
>> >> of this discussion :).
>> >
>> >It comes back to the same question thats been brought up here on
>> >numerous occassions - what are you billing for? SIP provides the setup
>> >services, not the transport services. So, it makes sense for a SIP
>> >server to charge for what it actually does - set the call up. It really
>> >doesn't make sense, I don't think, to try to bill by the minute for the
>> >SIP part of the call, since its really a transaction service. SIP is
>> >only one part of the picture, though. There is also transport service -
>> >diffserv prioritization, for example. A service provider can happily
>> >charge by the minute,bit, packet, or whatever here.
>> >
>> 
>> This is perhaps my mis-understanding. I have been working with the TIA
>> recently
>> and they have adopted MobileIP/DIAMETER for the next generation cellular
>> data networks. When one considers voice over IP using IP cellular phones,
>> one
>> quickly wonders how the charging will really work. Yes, it is true that
>> one
>> will charge for network access, but conceivably, one could also charge for
>> calls as well. This is especially true if these calls span providers.
>> 
>> Perhaps I have all of this wrong. I understand that charging for an LDAP
>> query cannot amount to much, and perhaps that is all bundled as part of
>> the
>> access charge. Are there any providers on this list that have an idea on
>> how
>> they will actually bill for these services?
>> 
>> PatC
>> 
>> 
>> ---------
>> This message came from the IETF IPTEL Working Group Mailing List.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 12 23:33:20 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id XAA22084
	for <iptel-archive@odin.ietf.org>; Fri, 12 Feb 1999 23:33:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 4EAE97FCB; Fri, 12 Feb 1999 23:23:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 808D27FCC; Fri, 12 Feb 1999 23:23:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Dean Willis" <Dean.Willis@MCI.COM>
To: <Pat.Calhoun@Eng.Sun.COM>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@MCI.COM>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: "Douglas Clowes" <dclowes@ozemail.com.au>, <confctrl@ISI.EDU>,
        <iptel@lists.research.bell-labs.com>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Fri, 12 Feb 1999 22:22:01 -0600
Message-ID: <000101be5708$66f2dde0$a9a22ca6@dwillispc3.mcit.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0
In-reply-to: <199902122049.MAA04666@hsmpka.eng.sun.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Ahah! Finally back on track.

Pat Calhoun asked
> Steve Donovan said:
> >Are there comments on the mechanisms proposed in the
> session-timer draft?
>
> Yes, and I appologize for side-tracking the whole thread. Is
> there anyway
> that we could have shorter INVITEs (perhaps the SIP Server
> can determine
> the length of a session), and have the SIP Client re-register
> with the
> server periodically, such as:
>
>    SIP Client --------- INVITE ------------> SIP Server
>                (some stuff happens here)
>    SIP Client < ------- ACK (3 mins) ------- SIP Server
>    (2 mins 45 seconds pass, enough to allow for retransmissions)
>    SIP Client ---------- FOOBAR -----------> SIP Server
>    (where FOOBAR is some message that refreshes the session)
>

The problem with this approach is that the SIP server (read, proxy) to
which the client REGISTERs may well not be the one with the
state-keeping probem.

Some calls might be signalled:

C-P1-P2-P3-P4-P5-P6-S

Maybe C registers with P1 and S registers with P6. REGISTERS are
targeted at the proxy acting as a location server -- they don't get
proxied along the chain.

Assume P4 has to keep state, because it is a firewall and needs to know
when to tear the ports down. P4 never sees a register, just the
INVITE-OK-ACK sequence. Maybe P4 sees a BYE at the end of the call, or
maybe C and S just get turned off . . .

How does P4 know when to close the ports? It can't. The timer parameter
gives it a way to expire the sessions . . . alternative to probing the
endpoints.

I had earlier discussed an approach method using SUBSCRIBE/NOTIFY (P4
SUBSCRIBEs to time-mark NOTIFY events on call X from C and S when it
sees the final ACK of INVITE/OK/ACK for call X). Steve Donovan
counterproposed the INFO method and was kind enough to write an ID on
it.

Different solution, but the problem is the same . . .

Can we at least get some consensus that it IS a problem?

Thanks,

--
Dean Willis


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Feb 13 18:32:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id SAA03967
	for <iptel-archive@odin.ietf.org>; Sat, 13 Feb 1999 18:32:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id EB2C67FCC; Sat, 13 Feb 1999 06:41:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4DB907FCD; Sat, 13 Feb 1999 06:41:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990213223827.0076ae50@pop.ozemail.com.au>
X-Sender: dclowes@pop.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Sat, 13 Feb 1999 22:38:27 +1100
To: Francois Menard Distribution List Management Account <fm-listproc@mediatrix.com>,
        Francois Menard Distribution List Management Account <fm-listproc@mediatrix.com>,
        "'Pat.Calhoun@Eng.Sun.COM'" <Pat.Calhoun@Eng.Sun.COM>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
In-Reply-To: <F16674FCE856D211AC0D00E02910AE0A07B518@INTERNAL>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 07:56 1999-02-12 -0500, Francois Menard Distribution List Management
Account wrote:
>Hi Douglas, how is it in the summer in Australia ?
>

It's warm and wet. Warmer that Sherbrooke, but I'll be in sunny California
in a couple of days.

>I'm not against this scenario, its proven for militarized zones (i.e.
>Intranets in entreprises controlled by MIS). Where I worrry is in DMZs,
>where an ISP forces me to use its gatekeeper.  That's where I'm saying that
>the end point has to become a trusted network element, in the control of the
>ISP, in order to bill by the second/bit.  The control of the session becomes
>in practice when the routers of the ISP refuse to send bits on the network
>on the behalf of non-trusted (to behave according to the ISP's policy) SIP
>end point (I guess that's another form of admission control ...).  This is
>the way that the MSO's are gearing up their VoIP network presently
>(PacketCable).  I.e., if you want to bill by the minute/second/bit, you
>either need to route the media stream through the SIP server and/or make the
>SIP server control the admission of the bits to the IP layer ... that is,
>there needs to be some form of policy decision flowing between a SIP server
>and routers that forwards the media stream bits from SIP endpoints ...

What's an MSO?

The ISP, in this case is trying to be like the LEC. It's more difficult for
them to do so in a deregulated market, with enforced competition. Is that
an oxymoron?

>
>
>So I can have the following, when a SIP server never gets a BYE, it doesn't
>need to ping the end point to see if its still alive, it can simply wait
>until the end-point re-REGISTERs, and then flag the call as terminated, and
>then, query the end point to tell it how many seconds were spent on line
>before the end-point died ... has anybody thought of extending the REGISTER
>message to include the duration of the previous call?  I think that this is
>a pretty cool idea ;-)  Of course, if the end-point never re-registers, the
>timer will still be running ... indefinitely.  We have a new policy on IP
>networks, any call of length greater than 1 million minutes is free ...
>

Well, perhaps less than that.

>BTW, I asked the question today to a supplier of H.323 stacks, as to what
>generally did the commercial gatekeepers do in practice to track the end of
>calls.  The person told me that the gatekeepers generally wait for the next
>registration (i.e. their assumption is that the H.323 terminal will
>re-register within hours), to flag the call as terminated.  They also rely
>on either one of the two legs of the call to send a termination message, so
>the case where the two legs are cut off never happens very frequently.  As
>to what is the mechanism that a gatekeeper is supposed to use to "ping" the
>terminal, that's not in the standard in their opinion.  Personally, I begin
>to like the idea of the end-point presenting the duration of the last call
>in the registration message ...
>

IRR is being used in H.323 to determine if the endpoint is still alive, but
the preference seems to be heading toward gatekeeper routed calls. Here,
you have a TCP connection through the gatekeeper for session control, but
media still goes direct.

>-=Francois=-
>fmenard@mediatrix.com
>

Douglas


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Feb 14 13:39:44 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id NAA03217
	for <iptel-archive@odin.ietf.org>; Sun, 14 Feb 1999 13:39:43 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 58C727FB7; Sun, 14 Feb 1999 12:49:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B9CDB7FB8; Sun, 14 Feb 1999 12:49:34 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <C51ED84B6F47D211917A0000F8BCBD11C2AB18@zcard00g.ca.nortel.com>
From: "Tom-PT Taylor" <taylor@nortelnetworks.com>
To: "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Fri, 12 Feb 1999 14:45:05 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.0.1460.8)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

The conversation is interesting, but is there any chance it could be
directed to one list (say confctrl) rather than two? My mail client is
ending up putting two copies of each message into each of the two
sub-folders.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 16 09:45:09 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id JAA27390
	for <iptel-archive@odin.ietf.org>; Tue, 16 Feb 1999 09:45:09 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B03077FE5; Mon, 15 Feb 1999 10:51:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 228AF7FE9; Mon, 15 Feb 1999 10:51:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199902151550.KAA20047@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: iptel@lists.research.bell-labs.com
From: Internet-Drafts@ietf.org
Reply-To: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-iptel-gwloc-framework-02.txt
Date: Mon, 15 Feb 1999 10:50:31 -0500
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the IP Telephony Working Group of the IETF.

	Title		: A Framework for a Gateway Location Protocol
	Author(s)	: J. Rosenberg, H. Schulzrinne
	Filename	: draft-ietf-iptel-gwloc-framework-02.txt
	Pages		: 13
	Date		: 12-Feb-99
	
   This document serves as a framework for a protocol for locating an IP
   telephony gateway. It defines terminology, specifies the various
   architectural elements and their functions, overviews the services
   provided by the protocol, and discusses how it fits into the broader
   context of Internet telephony.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-iptel-gwloc-framework-02.txt

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-iptel-gwloc-framework-02.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-iptel-gwloc-framework-02.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<19990212153543.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-gwloc-framework-02.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-gwloc-framework-02.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19990212153543.I-D@ietf.org>

--OtherAccess--

--NextPart--



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 16 10:06:17 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id KAA28241
	for <iptel-archive@odin.ietf.org>; Tue, 16 Feb 1999 10:06:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A9E0B8009; Mon, 15 Feb 1999 21:39:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 14578800B; Mon, 15 Feb 1999 21:39:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <CA6966C24AC6D111B5A100805FEAB7D8577419@nsrip00208.mcit.com>
From: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>
To: "'Tom-PT Taylor'" <taylor@nortelnetworks.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: RE: draft-ietf-mmusic-sip-session-timer-00.txt
Date: Mon, 15 Feb 1999 19:38:16 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE5955.6AD00524"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE5955.6AD00524
Content-Type: text/plain

My apologies, I copied the announcement to both lists because I though there
would be interest on both lists.  Unless anyone objects, I will send all of
my responses only to the confctrl list after this message.

Steve

> -----Original Message-----
> From:	Tom-PT Taylor [SMTP:taylor@nortelnetworks.com]
> Sent:	Friday, February 12, 1999 1:45 PM
> To:	'confctrl@ISI.EDU'; 'iptel@lists.research.bell-labs.com'
> Subject:	RE: RE: draft-ietf-mmusic-sip-session-timer-00.txt
> 
> The conversation is interesting, but is there any chance it could be
> directed to one list (say confctrl) rather than two? My mail client is
> ending up putting two copies of each message into each of the two
> sub-folders.
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

------_=_NextPart_001_01BE5955.6AD00524
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>RE: RE: draft-ietf-mmusic-sip-session-timer-00.txt</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">My apologies, I =
copied the announcement to both lists because I though there would be =
interest on both lists.&nbsp; Unless anyone objects, I will send all of =
my responses only to the confctrl list after this message.</FONT></P>

<P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Fixedsys">Steve</FONT>
</P>
<UL>
<P><FONT SIZE=3D1 FACE=3D"Arial">-----Original Message-----</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">From:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Tom-PT Taylor =
[SMTP:taylor@nortelnetworks.com]</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">Sent:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Friday, February 12, 1999 1:45 PM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">'confctrl@ISI.EDU'; =
'iptel@lists.research.bell-labs.com'</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT>=
</B> <FONT SIZE=3D1 FACE=3D"Arial">RE: RE: =
draft-ietf-mmusic-sip-session-timer-00.txt</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">The =
conversation is interesting, but is there any chance it could be</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">directed to =
one list (say confctrl) rather than two? My mail client is</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">ending up =
putting two copies of each message into each of the two</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier =
New">sub-folders.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier =
New">---------</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">This message =
came from the IETF IPTEL Working Group Mailing List.</FONT>
</P>
</UL>
</BODY>
</HTML>
------_=_NextPart_001_01BE5955.6AD00524--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 16 10:27:06 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id KAA29284
	for <iptel-archive@odin.ietf.org>; Tue, 16 Feb 1999 10:27:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6AC208007; Mon, 15 Feb 1999 19:23:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id CBC608009; Mon, 15 Feb 1999 19:23:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: 16 Feb 1999 00:17:13 -0000 (added by MTA mailnew.fibertel.com.ar)
X-Internal-ID: 36C1984F0001C6FF
Message-ID: <36C1984F0001C6FF@mailnew.fibertel.com.ar> (added by mailnew.fibertel.com.ar)
From: IIBA <iiba@bitsmart.com>
To: iptel@lists.research.bell-labs.com
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

iptel@lists.research.bell-labs.com>
Subject: Cursos Intensivos de Informática de Sábados y Domingos.
Reply-To: iiba@bitsmart.com
X-Priority: 3
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7Bit

ATENCION: Cursos Intensivos de Informática de fin de semana.

NO DEJE PASAR ESTA OPORTUNIDAD DE CAPACITARSE YA!

OBJETO: Los cursos apuntan a resolver la problemática de las personas que no 
disponen del tiempo necesario para capacitarse durante la semana. 

Se entregará certificado de realización del curso.

Práctica intensiva personalizada sobre computadoras (1 PC por alumno, 100% de 
práctica).

CURSOS DISPONIBLES:

Windows 95 (2 días - 16 hs.).
Windows 98 - Operación (2 días 16 hs.).
Windows 98 - Comunicaciones e Internet (1 día 8 hs.).
Internet (1 día 8 hs.).
Word 97 - Básico (2 días 16 hs.).
Word 97 - Avanzado (1 día 8 hs.).
Excel 97 - Básico (2 días 16 hs.).
Excel 97 - Avanzado (2 días 16 hs.).
AutoCAD 14 - Dibujo en 2 dimensiones (2 días 16 hs.).
AutoCAD 14 - Dibujo en 3D y Modelado de Sólidos (2 días 16 hs.).

OTROS CURSOS:

Access 95
Access 97
Corel Draw 6.0
Corel Draw 8.0
Excel 95
Introducción al PC
Office 95
Office 97
PowerPoint 95
PowerPoint 97
Visual Basic 4.0
Windows NT 4 - Workstation
Word 95

Descuentos por más de un curso o a grupos de personas (consultar).

Puede consultar los contenidos de los cursos, sus duraciones y costos en:

http://www.bitsmart.com/iiba 

o enviarnos un e-mail a:     iiba@bitsmart.com

FECHAS Y HORARIOS DISPONIBLES (Febrero de 1999):

Sábado 20/02
Domingo 21/02
Sábado 27/02
Domingo 28/02

En horarios de 9 a 13 hs. y  15 a 19 hs (consultar por horarios especiales).

LUGAR:

Los cursos de realizarán en Capital Federal (también se realizan cursos a distancia).

RESERVAR VACANTE (Cupos limitados).

A los interesados rogamos enviar los datos siguientes y nos pondremos en contacto.

NOMBRE Y APELLIDO:
EMPRESA y CARGO (Si corresponde):
TELEFONO:
CURSO(S) DE INTERES:
FECHAS DE INTERES:


Cordialmente,

Depto. de Capacitación
Instituto Informático Buenos Aires
iiba@bitsmart.com

Para mayor información sobre los cursos visite nuestra página en:

http://www.bitsmart.com/iiba



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 16 10:49:11 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id KAA00507
	for <iptel-archive@odin.ietf.org>; Tue, 16 Feb 1999 10:49:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A4E597FC9; Tue, 16 Feb 1999 10:35:34 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9E0F47FEC; Tue, 16 Feb 1999 10:35:33 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199902161534.HAA26159@hsmpka.eng.sun.com>
Date: Tue, 16 Feb 1999 07:29:46 -0800
To: "Dean Willis" <Dean.Willis@MCI.COM>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@MCI.COM>,
        <Pat.Calhoun@Eng.Sun.COM>
Cc: <iptel@lists.research.bell-labs.com>, <confctrl@ISI.EDU>,
        "Douglas Clowes" <dclowes@ozemail.com.au>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>Ahah! Finally back on track.
>
>Pat Calhoun asked
>> Steve Donovan said:
>> >Are there comments on the mechanisms proposed in the
>> session-timer draft?
>>
>> Yes, and I appologize for side-tracking the whole thread. Is
>> there anyway
>> that we could have shorter INVITEs (perhaps the SIP Server
>> can determine
>> the length of a session), and have the SIP Client re-register
>> with the
>> server periodically, such as:
>>
>>    SIP Client --------- INVITE ------------> SIP Server
>>                (some stuff happens here)
>>    SIP Client < ------- ACK (3 mins) ------- SIP Server
>>    (2 mins 45 seconds pass, enough to allow for retransmissions)
>>    SIP Client ---------- FOOBAR -----------> SIP Server
>>    (where FOOBAR is some message that refreshes the session)
>>
>
>The problem with this approach is that the SIP server (read, proxy) to
>which the client REGISTERs may well not be the one with the
>state-keeping probem.
>
>Some calls might be signalled:
>
>C-P1-P2-P3-P4-P5-P6-S
>
>Maybe C registers with P1 and S registers with P6. REGISTERS are
>targeted at the proxy acting as a location server -- they don't get
>proxied along the chain.
>
>Assume P4 has to keep state, because it is a firewall and needs to know
>when to tear the ports down. P4 never sees a register, just the
>INVITE-OK-ACK sequence. Maybe P4 sees a BYE at the end of the call, or
>maybe C and S just get turned off . . .
>
>How does P4 know when to close the ports? It can't. The timer parameter
>gives it a way to expire the sessions . . . alternative to probing the
>endpoints.
>
>I had earlier discussed an approach method using SUBSCRIBE/NOTIFY (P4
>SUBSCRIBEs to time-mark NOTIFY events on call X from C and S when it
>sees the final ACK of INVITE/OK/ACK for call X). Steve Donovan
>counterproposed the INFO method and was kind enough to write an ID on
>it.
>
>Different solution, but the problem is the same . . .
>
>Can we at least get some consensus that it IS a problem?

Oh, I agree that there is a problem. I am just not sure that the solution 
proposed is the best one. I suppose I need to really understand why the INVITE
approach cannot work to setup soft-state along the proxy chain. you state that
P4 never sees the REGISTER, but I think that the INVITE/ACK is sufficient
for the proxies to build state. Am I wrong?

PatC
>
>Thanks,
>
>--
>Dean Willis
>



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 16 13:32:05 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id NAA06178
	for <iptel-archive@odin.ietf.org>; Tue, 16 Feb 1999 13:32:05 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8CC727FB7; Tue, 16 Feb 1999 13:19:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D9B3C7FC8; Tue, 16 Feb 1999 13:19:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36C9B605.636365F6@dnrc.bell-labs.com>
Date: Tue, 16 Feb 1999 13:16:37 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: list iptel <iptel@lists.research.bell-labs.com>
Subject: Draft of glp protocol
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

iptel participants,

Matt Squire has submitted an Internet Draft which contains a proposal
for the GLP. Its an excellent piece of work, and proposes the use of
SCSP (RFC2334) as as baseline. Until it appears in the archives, you can
find a copy at:

http://www.bell-labs.com/mailing-lists/iptel/draft-ietf-iptel-glp-00.txt

which is off of our homepage.

I'd like to make this draft the main topic of conversation at the
upcoming meeting (I have reserved a slot, but don't know when it is yet.
When I know, I'll send an announcement). So, please take a look at this
and lets get some discussion going on the list ahead of time. Matt has
identified many open issues in the editorial comments within the draft.
Please take a look and speak up if you have any thoughts.

Thanks,
Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 17 04:53:35 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id EAA29899
	for <iptel-archive@odin.ietf.org>; Wed, 17 Feb 1999 04:53:35 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1B1467FB7; Wed, 17 Feb 1999 04:43:21 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 762F97FC8; Wed, 17 Feb 1999 04:43:20 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <000701be5a5a$add7a560$b15195c1@agloverhome>
Reply-To: "Ashley Glover" <nr51@dial.pipex.com>
From: "Ashley Glover" <nr51@dial.pipex.com>
To: <iptel@lists.research.bell-labs.com>
Subject: list of main players in UK and European IP Telephony
Date: Wed, 17 Feb 1999 09:48:30 -0000
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_0004_01BE5A5A.AD1762A0"
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.

------=_NextPart_000_0004_01BE5A5A.AD1762A0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

anyone know where I'd get a list ?

Ashley

------=_NextPart_000_0004_01BE5A5A.AD1762A0
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">
<HTML>
<HEAD>

<META content=3Dtext/html;charset=3Diso-8859-1 =
http-equiv=3DContent-Type>
<META content=3D'"MSHTML 4.72.2106.6"' name=3DGENERATOR>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT color=3D#000000 size=3D2>anyone know where I'd get a list =
?</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#000000 size=3D2>Ashley</FONT></DIV></BODY></HTML>

------=_NextPart_000_0004_01BE5A5A.AD1762A0--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 17 09:44:50 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id JAA03435
	for <iptel-archive@odin.ietf.org>; Wed, 17 Feb 1999 09:44:50 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 12A267FCA; Wed, 17 Feb 1999 09:33:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 731547FCB; Wed, 17 Feb 1999 09:33:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Jiri Kuthan <kuthan@fokus.gmd.de>
Message-Id: <199902171431.PAA15730@jobe.fokus.gmd.de>
Subject: Re: FW: list of main players in UK and European IP Telephony
To: bjarne.nordberg@ausys.se (Bjarne Nordberg)
Date: Wed, 17 Feb 1999 15:31:55 +0100 (MET)
Cc: iptel@lists.research.bell-labs.com
In-Reply-To: <D1E5CE9B4067D111A012002048112CECCC93DD@aunt9.ausys.se> from "Bjarne Nordberg" at Feb 17, 99 02:53:18 pm
X-Mailer: ELM [version 2.4 PL25]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Neither this site is UK/EU specific but is is definitly worth trying:

http://www.fokus.gmd.de/research/cc/glone/projects/ipt/

A list of players (section "players") and companies (section "products") is 
provided. Other websites relevant to the IP Telephony are refered (section
"References-Weblinks-Resources"). This web-site is free of commercial ads 
and is updated daily.


Regards, Jiri

> 
> Good places to start looking for "players", both operators and =
> suppliers,
> not specifically for Europe and UK though:
> =A0
> http://www.computertelephony.org/ <http://www.computertelephony.org/>=20
> http://www.pulver.com/ <http://www.pulver.com/>=20
> =A0
> 
> Bjarne Nordberg=20
> AU-System Network=20
> Box 47612, Liljeholmsstranden 3, S-117 94 Stockholm, SWEDEN=20
> email:=A0 bjarne.nordberg@ausys.se=A0=A0=A0=A0=A0=A0 www:=A0 =
> http://www.ausys.com
> <http://www.ausys.com/> =20
> 
> =A0
> -----Original Message-----
> From: Ashley Glover [mailto:nr51@dial.pipex.com]=20
> Sent: 17 February 1999 10:49
> To: iptel@lists.research.bell-labs.com
> <mailto:iptel@lists.research.bell-labs.com>=20
> Subject: list of main players in UK and European IP Telephony
> 
> 
> anyone know where I'd get a list ?
> =A0
> Ashley
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 


-- 
Jiri Kuthan		  Phone: ++49 / (0)30 / 3463 - 7 271
GMD FOKUS                 Fax  : ++49 / (0)30 / 3463 - 8 271
Kaiserin-Augusta-Allee 31 E-Mail : kuthan@fokus.gmd.de
D-10589 Berlin, Germany   WWW    : http://www.fokus.gmd.de/usr/kuthan

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 17 10:01:33 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id KAA04064
	for <iptel-archive@odin.ietf.org>; Wed, 17 Feb 1999 10:01:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C9CBB7FC8; Wed, 17 Feb 1999 08:55:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 353BA7FCA; Wed, 17 Feb 1999 08:55:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <D1E5CE9B4067D111A012002048112CECCC93DD@aunt9.ausys.se>
From: Bjarne Nordberg <bjarne.nordberg@ausys.se>
To: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: FW: list of main players in UK and European IP Telephony
Date: Wed, 17 Feb 1999 14:53:18 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.1960.3)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by ietf.org id KAA04064

Good places to start looking for "players", both operators and suppliers,
not specifically for Europe and UK though:
 
http://www.computertelephony.org/ <http://www.computertelephony.org/> 
http://www.pulver.com/ <http://www.pulver.com/> 
 

Bjarne Nordberg 
AU-System Network 
Box 47612, Liljeholmsstranden 3, S-117 94 Stockholm, SWEDEN 
email:  bjarne.nordberg@ausys.se       www:  http://www.ausys.com
<http://www.ausys.com/>  

 
-----Original Message-----
From: Ashley Glover [mailto:nr51@dial.pipex.com] 
Sent: 17 February 1999 10:49
To: iptel@lists.research.bell-labs.com
<mailto:iptel@lists.research.bell-labs.com> 
Subject: list of main players in UK and European IP Telephony


anyone know where I'd get a list ?
 
Ashley

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 17 18:36:56 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id SAA13370
	for <iptel-archive@odin.ietf.org>; Wed, 17 Feb 1999 18:36:56 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D7F577FC9; Wed, 17 Feb 1999 18:27:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3E2C27FCB; Wed, 17 Feb 1999 18:27:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36CB5012.51160924@ics.uci.edu>
Date: Wed, 17 Feb 1999 15:26:10 -0800
From: Sachin Vasudeva <sachin@ics.uci.edu>
Organization: UCI
X-Mailer: Mozilla 4.05 [en] (X11; I; SunOS 5.5.1 sun4u)
MIME-Version: 1.0
To: IPTEL <iptel@lists.research.bell-labs.com>
Subject: Information required
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi
I wanted to know exactly which protocols are (eg. GLP etc.)
are being implemented  at the software level , which protocols
at the firmware and hardware levels.
thanx
Sachin Vasudeva
ICS (CSN) dept UCI.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 17 21:23:39 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id VAA14935
	for <iptel-archive@odin.ietf.org>; Wed, 17 Feb 1999 21:23:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8C6B87FCB; Wed, 17 Feb 1999 20:27:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E44C97FCC; Wed, 17 Feb 1999 20:27:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <008f01be5add$8fed04e0$98a1b3c7@jtk-laptop>
From: "John Kristoff" <jkristof@interaccess.com>
To: <iptel@lists.research.bell-labs.com>
Subject: comp.dcom.voice-over-ip vote now in progress
Date: Wed, 17 Feb 1999 19:25:25 -0600
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.5
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

The vote to decide whether to create a new "comp.dcom.voice-over-ip"
newsgroup is now in progress.  Anyone with an opinion is entitled and
encouraged to vote; there are no special qualifications, it's not difficult
and you don't have to be a computer expert.

If you have an opinion on this issue, please find the article with "Subject:
CFV: comp.dcom.voice-over-ip" in comp.dcom.telecom.tech,
comp.dcom.sys.bay-networks, comp.dcom.sys.cisco, comp.dcom.sys.nortel,
comp.protocols.tcp-ip, news.announce.newgroups, news.groups or
alt.dcom.telecom.ip.

Then follow the directions in that article to cast your vote.

If you have trouble finding that article, e-mail me and I'll gladly tell you
how to get a ballot by e-mail.

ONLY votes which use the official ballot and follow the directions count.

In order to be created, a new group needs at least 100 more YES votes than
NO votes and at least 2/3 of the votes cast must be YES votes.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 18 06:35:01 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id GAA27534
	for <iptel-archive@odin.ietf.org>; Thu, 18 Feb 1999 06:35:01 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BE0317FCF; Thu, 18 Feb 1999 05:19:30 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E04477FCD; Thu, 18 Feb 1999 05:19:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X400-Received: by /PRMD=INTERNET/ADMD=ATLAS/C=FR/; Relayed;
               Thu, 18 Feb 1999 11:09:30 +0100
X400-Received: by mta xn1-gw.atlas.fr in /PRMD=INTERNET/ADMD=ATLAS/C=FR/;
               Relayed; Thu, 18 Feb 1999 11:09:30 +0100
X400-Received: by /ADMD=ATLAS/C=FR/; Relayed; Thu, 18 Feb 1999 11:08:37 +0100
X400-Received: by /PRMD=art-telecom/ADMD=atlas/C=fr/; Relayed;
               Thu, 18 Feb 1999 11:11:32 +0100
Date: Thu, 18 Feb 1999 11:11:32 +0100
X400-Originator: Didier.CHAUVEAU@art-telecom.fr
X400-Recipients: non-disclosure:;
X400-MTS-Identifier: [/PRMD=art-telecom/ADMD=atlas/C=fr/;XGW-990218110221+0100-23318]
Original-Encoded-Information-Types: undefined
X400-Content-Type: P2-1988 (22)
Content-Identifier: Voice over IP - 
From: Didier.CHAUVEAU@art-telecom.fr
To: iptel@lists.research.bell-labs.com, pint@lists.research.bell-labs.com,
        ss7-internet@bayNetworks.com, e164-to-ip@lserv.vocaltec.com
Subject:  Voice over IP - public call for comments
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Message-Id: <19990218101905.4BE097FCC@lists.research.bell-labs.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by ietf.org id GAA27534

 Dear IETF members,
 
 Please note that ART, the telecommunications regulatory authority in 
 France, launched a public call for comments this week concerning Voice 
 over Ip.
 
 ART invites users and all market parties to contribute to this study. The 
 public call for comments is available on the following web site: 
 http://www.art-telecom.fr
 Responses should reach ART by noon on 13 March 1999 (see details on 
 public call for comments).
 
 Do not hesitate to share this information with interested parties.
 
 Regards,
 
 
 didier chauveau
 ART
 didier.chauveau@art-telecom.fr 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 18 14:03:33 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id OAA07409
	for <iptel-archive@odin.ietf.org>; Thu, 18 Feb 1999 14:03:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C3D3A7FB7; Thu, 18 Feb 1999 12:33:13 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2460B7FC8; Thu, 18 Feb 1999 12:33:13 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199902181731.MAA02440@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: iptel@lists.research.bell-labs.com
From: Internet-Drafts@ietf.org
Reply-To: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-iptel-glp-00.txt
Date: Thu, 18 Feb 1999 12:31:31 -0500
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the IP Telephony Working Group of the IETF.

	Title		: A Gateway Location Protocol
	Author(s)	: M. Squire
	Filename	: draft-ietf-iptel-glp-00.txt
	Pages		: 22
	Date		: 17-Feb-99
	
   Given a gateway and a (possibly empty) list  of  gateway  attributes,
   the 'gateway location problem' is to find a gateway serving the given
   phone number that satisfies the attribute set.  This problem has also
   been referred to as the 'call routing problem' as the answer returned
   may not be an IP-PSTN gateway but an intermediate signaling server.

   Part  of  the  solution  to  this  problem  is  the  maintenance  and
   distribution  of  the call routing tables.  This document describes a
   protocol for maintaining a distributed call routing  database  across
   multiple  administrative domains.  The protocol uses the Server Cache
   Synchronization Protocol (SCSP) to maintain the distributed database.
   This  document  describes  the  problem  of  gateway  location  and a
   potential solution that uses SCSP as the foundation for  distributing
   the call routing tables.


A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-iptel-glp-00.txt

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-iptel-glp-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-iptel-glp-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<19990217150337.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-glp-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-glp-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19990217150337.I-D@ietf.org>

--OtherAccess--

--NextPart--



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 18 17:38:15 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.8.5/8.8.7a) with ESMTP id RAA23131
	for <iptel-archive@odin.ietf.org>; Thu, 18 Feb 1999 17:38:15 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 468677FCA; Thu, 18 Feb 1999 17:27:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 968CF7FCB; Thu, 18 Feb 1999 17:27:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <m10DbqX-0003IrC@fwd08.btx.dtag.de>
Date: Thu, 18 Feb 1999 23:23:01 +0100
To: fhollatz@compuserve.com
Cc: iptel@lists.research.bell-labs.com
References: <199902181731.MAA02440@ietf.org>
Subject: I-D ACTION:draft-ietf-iptel-glp-00.txt (fwd)
X-Mailer: T-Online eMail 2.0
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8BIT
X-Sender: 0897912484-0005@t-online.de
From: bettina.gier@t-online.de (Bettina Gier)
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Internet-Drafts@ietf.org schrieb:
> 	access-type="mail-server";
> 	server="mailserv@ietf.org"
>
>
> Content-Type: text/plain
> Content-ID:	<19990217150337.I-D@ietf.org>
>
> ENCODING mime
> FILE /internet-drafts/draft-ietf-iptel-glp-00.txt
>
> --OtherAccess
> Content-Type: Message/External-body;
> 	name="draft-ietf-iptel-glp-00.txt";
> 	site="ftp.ietf.org";
> 	access-type="anon-ftp";
> 	directory="internet-drafts"
>
> Content-Type: text/plain
> Content-ID:	<19990217150337.I-D@ietf.org>
>
> --OtherAccess--
>
>
>
>
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

**************************************************
Bettina Gier
Phone: ++49 89 664283
Email/mailto:Bettina.Gier@t-online.de
**************************************************



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Feb 21 23:10:51 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA08035
	for <iptel-archive@odin.ietf.org>; Sun, 21 Feb 1999 23:10:51 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C7A0F7FB7; Sun, 21 Feb 1999 23:01:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3B31C7FB8; Sun, 21 Feb 1999 23:01:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BE5E45.A9271DE0@hss056.hss.hns.com>
From: Alok Mittal <amittal@hss.hns.com>
To: "'msquire@nortelnetworks.com'" <msquire@nortelnetworks.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Comments on draft-ietf-iptel-glp-00.txt
Date: Mon, 22 Feb 1999 09:28:08 -0000
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hi Matt,

  Some comments/suggestions:

5.1.2 Hello Phase -- the heading number should be 5.1.1 -- It might be 
	useful to have a version number, may be to indicate what kind
	of attributes might be expected of the other end -- this also brings
	me to another point, whether we actually need to make some of
	the attributes mandatory in the sense of being able to make some
	call routing decisions.
5.2 -- Using a network layer identifier (IP address) for an essentially 
	application level protocol might cause problems in other networks.
	Can we just define the semantics of the id as being locally unique,
	and suggest that one possible way of doing it on IP networks is to
	use the IP address? Or use some other network-independent
	formation?
5.2.2 -- For the Telephone number range, another useful way to express
	it might be a prefix -- so if it says +9191, then the node handles
	all numbers starting from +9191 -- I understand that this can be
	expressed in current range format, but that will unnecessarily 
	increase message size, and comparison complexity -- especially
	since in this case the high end number has to be completely
	specified (such as +919199999999999) owing to inclusive limits
	on both lower and upper bounds. This could be done also by 
	just using len(Upper Bound) to be 0.
5.2.3.1 -- In the sig-proto-id, it is useful to have a version number along with
	protocol name, as this is likely to be important in selection decisions.
5.2.3.2 -- This is kinda generic -- Having pricing information in glp is 
	encapsulating not the application requirements, but the business
	requirements that probably form another higher layer from a 
	protocol standpoint. Business rules are likely to be based on more
	than just the price, and even the price may be different for each
	business relationship. As such, if we can replace this by some
	cost measure, then that might serve the application protocol 
	semantics better.

Best Regards
Alok
-----------------------------------------------------------------
Alok Mittal	
Technical Leader,	Hughes Software Systems
Prestige Opal, 146 Infantry Road, Bangalore - 560001
Tel: +91-80-2286390,91,92	Fax: +91-80-2286393
-----------------------------------------------------------------


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Feb 22 09:15:55 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA17779
	for <iptel-archive@odin.ietf.org>; Mon, 22 Feb 1999 09:15:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 188417FB8; Mon, 22 Feb 1999 09:03:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 74D7D7FCA; Mon, 22 Feb 1999 09:03:34 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990222085041.00a4a320@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Mon, 22 Feb 1999 08:50:42 -0500
To: Alok Mittal <amittal@hss.hns.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


Alok - 

Thanks for all of the comments.  My responses are included below.  

At 09:28 AM 2/22/99 -0000, Alok Mittal wrote:
>Hi Matt,
>
>  Some comments/suggestions:
>
>5.1.2 Hello Phase -- the heading number should be 5.1.1 -- It might be 
>	useful to have a version number, may be to indicate what kind
>	of attributes might be expected of the other end -- this also brings
>	me to another point, whether we actually need to make some of
>	the attributes mandatory in the sense of being able to make some
>	call routing decisions.

I agree that some kind of version/feature exchange is desireable.  I'm
currently working with one of the SCSP authors to determine the best way to
address this.  

In Section 5 I somewhat address the mandatory nature of attributes, but not
completely, and probably not clearly enough. 

Within the individual sections describing each attribute, I tried to say if
particular attribute was required or had some limitation on the number of
times it could appear within a single call routing object.  Currently, the
only ones which I say 'must' be included are the TTL (to prevent unlimited
propagation) and the next hop provider (seemed logical that you would want
to know to whom you are sending your signalling messages).  So these are
'mandatory' in the sense that the must be included in each call routing
object.  

Now there's another interpretation of mandatory, and that is that the
receiver must be able to understand the attribute.  We could provide this
by having a flag in the attribute signify that if the receiver does not
understand the attribute, then the receiver must ignore the entire call
routing object.  We could take a bit from the type field as follows.  
  C - Compulsory.  Set to indicate that the receiver of this call routing
object should ignore the entire object if they do not understand this
attribute.  

There are alternate methods to get similar effects to the compulsory flag.
Specifically, we could use feature negotiation in the Hello Phase to
determine the set of attributes understood by each peer.  Using a
compulsory bit also implies that we'd need an error indication to let the
sender know that some attribute type is not understood.  

So I agree that determining which attributes are understood by each peer is
important.  I haven't yet decided if this function is better provided by
negotiations during hello phase or by flags in the attribute.  Opinions?

>5.2 -- Using a network layer identifier (IP address) for an essentially 
>	application level protocol might cause problems in other networks.
>	Can we just define the semantics of the id as being locally unique,
>	and suggest that one possible way of doing it on IP networks is to
>	use the IP address? Or use some other network-independent
>	formation?

I think that other applications of SCSP use similar information as their
identifiers (I can't get to the IETF server to confirm right now).  The
semantics of the IDs are that the sender/receiver/originator ID must be
unique with the server group.  It seems logical that we should define what
goes in the field in a way that guarantees this uniqueness criteria is
satisfied.  The IP address seemed like the easiest way.  This information
is certainly available to the application layer, so I didn't consider it
problematic (yet! see below).  

One of the Editor's notes in the draft asks if we need to apply GLP to
non-IPv4 networks (ie use more general address arguments).  I think we need
such generality.  If this is the consensus, then the use and specification
of the network address as the SCSP sender/receiver ID is not as clear.  In
this case, I agree that we should choose some other method of specifying
the IDs.  To guarantee uniqueness, I think we'd need a format like
ITAD:Server.  Since any provider controls only their machines, we'd need to
scope each server identifier with the provider identifier.  Each provider
can ensure uniqueness within their domain. 

But I can't specify this format until we figure out how we want to
represent the Internet Telephony Administrative Domain.  Another Editor's
note asks for opinions of how an ITAD should be represented (url, string,
decimal, etc).  After we answer this question, I'll go back and re-specify
what the SCSP identifiers should be.

Does that sound reasonable?  

>5.2.2 -- For the Telephone number range, another useful way to express
>	it might be a prefix -- so if it says +9191, then the node handles
>	all numbers starting from +9191 -- I understand that this can be
>	expressed in current range format, but that will unnecessarily 
>	increase message size, and comparison complexity -- especially
>	since in this case the high end number has to be completely
>	specified (such as +919199999999999) owing to inclusive limits
>	on both lower and upper bounds. This could be done also by 
>	just using len(Upper Bound) to be 0.

The first time I wrote this up, I used prefixes.  But then I generalized it
to a range, and for the life of me, I can't remember why.  

The range is certainly more general.  I wasn't worried about the comparison
complexity because in many (most?) cases, the decision will be
multi-attributed, so the complexity of a simple string comparison seemed
small compared to determining a server using the right protocols by the
right provider within a given price range ...  

I think the real question is whether a prefix is sufficient to specify the
likely sets of target telephone numbers.  In the US, I think that a prefix
is sufficient.  The proximity of area/city codes does not mean the
proximity of areas/cities, which seems a likely criteria for a gateway
serving 'range' but not a 'prefix'.  Is this common of phone numbering
systems throughout the world?  If so, then I agree we should go with
prefixes and not ranges. 


>5.2.3.1 -- In the sig-proto-id, it is useful to have a version number
along with
>	protocol name, as this is likely to be important in selection decisions.

Good point.  I'll fix it.  

>5.2.3.2 -- This is kinda generic -- Having pricing information in glp is 
>	encapsulating not the application requirements, but the business
>	requirements that probably form another higher layer from a 
>	protocol standpoint. Business rules are likely to be based on more
>	than just the price, and even the price may be different for each
>	business relationship. As such, if we can replace this by some
>	cost measure, then that might serve the application protocol 
>	semantics better.
>
>

The pricing attribute does encapsulate business and not application
requirements, but I think that's what people are after.  The pricing
attribute is a big black box as far as the protocol is concerned.  Its
specified that way because we can't set any limits on business
rules/practices.  The peer-to-peer nature of GLP would allow one to have
different pricing structures for different business relationships (ie peers). 

Its not clear to me what you mean by cost. Cost, to most routing protocols,
is the number of hops.  But we don't really care about the number of
signalling hops.  A couple weeks ago, there was something on the mailing
list which enumerated many variables affecting the cost/price of an
internet phone call.  So the meaning of cost is ambiguous at best.  

If you have something specific that you mean by cost then let me know, but
my view is that we can't do anything with real cost, and that we should
stick to a black-box structure to express business practices and
relationships.

- Matt  

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Feb 22 23:41:03 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA29223
	for <iptel-archive@odin.ietf.org>; Mon, 22 Feb 1999 23:41:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A827D7FB7; Mon, 22 Feb 1999 23:31:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 040757FB8; Mon, 22 Feb 1999 23:31:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <4.1.19990222202949.00a5bc80@infoserv.infoserv.com>
X-Sender: pjdurand@infoserv.infoserv.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1 
Date: Mon, 22 Feb 1999 20:29:54 -0800
To: iptel@lists.research.bell-labs.com
From: Jerry Durand <jdurand@interstellar.com>
Subject: Fwd: Re: pre-encoded audio
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

>X-From_: infoserv!ihgw1.lucent.com!bcao Mon Feb 22 20:26:25 1999
>Date: Mon, 22 Feb 1999 23:05:14 -0500
>From: Binshi Cao <bcao@lucent.com>
>Reply-To: bcao@lucent.com
>Organization: Consumer Products
>X-Mailer: Mozilla 4.0 [en] (Win95; I)
>To: Jerry Durand <jdurand@interstellar.com>
>Subject: Re: pre-encoded audio
>
>Hi,
>
>I am not a person who takes care of this list.
>
>Please ask :
>iptel@lists.research.bell-labs.com
>to do so.
>
>Thanks!
>
>Binshi Cao.
>
>
>
>
>Jerry Durand wrote:
>
>> Please remove me from your mailing list.  Somebody else was using my
>> account.
>>
>> ----- Original Message -----
>> From: Binshi Cao <bcao@lucent.com>
>> To: Kevin Davis <Kevin_Davis@notes.teradyne.com>
>> Cc: <iptel@lists.research.bell-labs.com>
>> Sent: Monday, February 08, 1999 6:29 AM
>> Subject: Re: pre-encoded audio
>>
>> >You may go to ITU web site to get the test vectors for the original
>> speech,
>> >coded packets, and decoded speech.
>> >
>> >Binshi Cao.
>> >
>> >Kevin Davis wrote:
>> >>
>> >> For testing purposes, I wish to pre-encode a speech clip
>> >> as G.723  (and maybe G.729) and then I will put it into RTP
>> >> payload without the need for real-time encoder.
>> >> The typical off-the-shelf audio tools do not support this.
>> >> Can anyone point me to a tool that generate VoIP audio ?
>> >>
>> >> ---------
>> >> This message came from the IETF IPTEL Working Group Mailing List.
>
>
>

----------
Jerry Durand
Durand Interstellar, Inc.
Los Gatos, California  USA
tel:  +1 408 356-3886
fax:  +1 408 356-4659
net:  jdurand@interstellar.com


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 02:54:52 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA06852
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 02:54:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 044997FB8; Tue, 23 Feb 1999 02:45:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4E3B07FCE; Tue, 23 Feb 1999 02:45:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BE5F2E.21F60C20@hss056.hss.hns.com>
From: Alok Mittal <amittal@hss.hns.com>
To: "'Matt Squire'" <msquire@nortelnetworks.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Date: Tue, 23 Feb 1999 13:12:13 -0000
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hi Matt,

So I agree that determining which attributes are understood by each peer is
important.  I haven't yet decided if this function is better provided by
negotiations during hello phase or by flags in the attribute.  Opinions?

To me it seems that this activity is best done in the negotiation so that a lot
of overhead is avoided if the peers are not compatible. The degree of control
offered by per-attribute control does not seem to be necessary. 

I think that other applications of SCSP use similar information as their
identifiers (I can't get to the IETF server to confirm right now).  
<snip>
But I can't specify this format until we figure out how we want to
represent the Internet Telephony Administrative Domain.  Another Editor's
note asks for opinions of how an ITAD should be represented (url, string,
decimal, etc).  After we answer this question, I'll go back and re-specify
what the SCSP identifiers should be.

Does that sound reasonable?  

Yes, we could go with that approach. Or just to use what H.323 does for Call
Identifiers -- it specifies a 6 byte field within Call ID for "node" specifying that
it has to be spatially unique. The recommendation then suggests that ethernet
card address should be used, BUT if there is no netowork card, the system
may use other means to generate globally unique node identifier -- such as 
make least significant bit of first octet to 1 (refers to multicast so doesn't conflict
with any ethernet card addresses) and fill a 47 bit pseudo random number
elsewhere.

I think the real question is whether a prefix is sufficient to specify the
likely sets of target telephone numbers.  In the US, I think that a prefix
is sufficient.  The proximity of area/city codes does not mean the
proximity of areas/cities, which seems a likely criteria for a gateway
serving 'range' but not a 'prefix'.  Is this common of phone numbering
systems throughout the world?  If so, then I agree we should go with
prefixes and not ranges. 

In India, again there is no strong correlation between the two, though in many
cases such a property may be satisfied.

Its not clear to me what you mean by cost. Cost, to most routing protocols,
is the number of hops.  But we don't really care about the number of
signalling hops.  A couple weeks ago, there was something on the mailing
list which enumerated many variables affecting the cost/price of an
internet phone call.  So the meaning of cost is ambiguous at best.  

If you have something specific that you mean by cost then let me know, but
my view is that we can't do anything with real cost, and that we should
stick to a black-box structure to express business practices and
relationships.

Yes, I agree that this is a rather vague thing. To think of a cost metric at an 
application level, we could use the number of administrative domains through
which the shortest path to destination lies (from this gateway/signaling node).
This is based on an assumption that every administrative domain would charge
*some* fees for routing packets through its gateways. But this will lead to the 
problem of propagating this information, and it also makes a simplifying
assumption that cost of passing through each administrative domain is more
or less the same (or at least the difference is insignificant as compared to 
passing through an additional domain). Will think more and get back to you
if I have better suggestions!

Best Regards
Alok



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 03:35:01 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA07157
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 03:35:00 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8DBB67FB7; Tue, 23 Feb 1999 03:23:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E4CE77FCE; Tue, 23 Feb 1999 03:23:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BE5F33.97C7FF30@hss056.hss.hns.com>
From: Alok Mittal <amittal@hss.hns.com>
To: "'Matt Squire'" <msquire@nortelnetworks.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Date: Tue, 23 Feb 1999 13:51:17 -0000
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Sorry for the lost indentations.
Hi Matt,

>So I agree that determining which attributes are understood by each peer is
>important.  I haven't yet decided if this function is better provided by
>negotiations during hello phase or by flags in the attribute.  Opinions?

To me it seems that this activity is best done in the negotiation so that a lot
of overhead is avoided if the peers are not compatible. The degree of control
offered by per-attribute control does not seem to be necessary. 

>I think that other applications of SCSP use similar information as their
>identifiers (I can't get to the IETF server to confirm right now).  
> <snip>
>But I can't specify this format until we figure out how we want to
>represent the Internet Telephony Administrative Domain.  Another Editor's
>note asks for opinions of how an ITAD should be represented (url, string,
>decimal, etc).  After we answer this question, I'll go back and re-specify
>what the SCSP identifiers should be.
>
>Does that sound reasonable?  

Yes, we could go with that approach. Or just to use what H.323 does for Call
Identifiers -- it specifies a 6 byte field within Call ID for "node" specifying that
it has to be spatially unique. The recommendation then suggests that ethernet
card address should be used, BUT if there is no netowork card, the system
may use other means to generate globally unique node identifier -- such as 
make least significant bit of first octet to 1 (refers to multicast so doesn't conflict
with any ethernet card addresses) and fill a 47 bit pseudo random number
elsewhere.

>I think the real question is whether a prefix is sufficient to specify the
>likely sets of target telephone numbers.  In the US, I think that a prefix
>is sufficient.  The proximity of area/city codes does not mean the
>proximity of areas/cities, which seems a likely criteria for a gateway
>serving 'range' but not a 'prefix'.  Is this common of phone numbering
>systems throughout the world?  If so, then I agree we should go with
>prefixes and not ranges. 

In India, again there is no strong correlation between the two, though in many
cases such a property may be satisfied.

>Its not clear to me what you mean by cost. Cost, to most routing protocols,
>is the number of hops.  But we don't really care about the number of
>signalling hops.  A couple weeks ago, there was something on the mailing
>list which enumerated many variables affecting the cost/price of an
>internet phone call.  So the meaning of cost is ambiguous at best.  
>
>If you have something specific that you mean by cost then let me know, but
>my view is that we can't do anything with real cost, and that we should
>stick to a black-box structure to express business practices and
>relationships.

Yes, I agree that this is a rather vague thing. To think of a cost metric at an 
application level, we could use the number of administrative domains through
which the shortest path to destination lies (from this gateway/signaling node).
This is based on an assumption that every administrative domain would charge
*some* fees for routing packets through its gateways. But this will lead to the 
problem of propagating this information, and it also makes a simplifying
assumption that cost of passing through each administrative domain is more
or less the same (or at least the difference is insignificant as compared to 
passing through an additional domain). Will think more and get back to you
if I have better suggestions!

Best Regards
Alok



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 08:04:37 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA08354
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 08:04:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 235EB7FB7; Tue, 23 Feb 1999 07:53:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7B14E7FB8; Tue, 23 Feb 1999 07:53:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Tue, 23 Feb 1999 12:50:18 +0000 (GMT)
From: Aishling <irelanda@tcd.ie>
To: Patrice Calhoun <Pat.Calhoun@Eng.Sun.COM>
Cc: "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>
Subject: RE: draft-ietf-mmusic-sip-session-timer-00.txt
In-Reply-To: <199902111906.LAA28606@hsmpka.eng.sun.com>
Message-ID: <Pine.OSF.3.96.990223124818.32086A-100000@alf2.tcd.ie>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



> 
> My understanding is that a SIP Client issues an Invite message to a SIP Server,
> which in turn forwards the request to a target SIP Client (or some GW or some

Do clients not only issue requests, I didn't realise they could accept
requests???

> sort). Once the ACK has returned to the originating client, I was under the
> impression that both SIP Clients send RTP packets directly to each other,
> bypassing the SIP Server (this is where I am unclear).



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 09:19:32 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA08833
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 09:19:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 4911F7FB8; Tue, 23 Feb 1999 09:09:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9DE637FC9; Tue, 23 Feb 1999 09:09:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D2B626.DE1022AB@cs.columbia.edu>
Date: Tue, 23 Feb 1999 09:07:34 -0500
From: Henning Schulzrinne <hgs@cs.columbia.edu>
Organization: Columbia University, Dept. of Computer Science
X-Mailer: Mozilla 4.5 [en] (X11; I; SunOS 5.6 sun4u)
X-Accept-Language: en
MIME-Version: 1.0
To: Aishling <irelanda@tcd.ie>
Cc: Patrice Calhoun <Pat.Calhoun@Eng.Sun.COM>,
        "Donovan, Steven R. (MCI)" <Steven.R.Donovan@mci.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'confctrl@ISI.EDU'" <confctrl@ISI.EDU>
Subject: Re: draft-ietf-mmusic-sip-session-timer-00.txt
References: <Pine.OSF.3.96.990223124818.32086A-100000@alf2.tcd.ie>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Aishling wrote:
> 
> >
> > My understanding is that a SIP Client issues an Invite message to a SIP Server,
> > which in turn forwards the request to a target SIP Client (or some GW or some
> 
> Do clients not only issue requests, I didn't realise they could accept
> requests???

Because it sounds strange to some to talk about clients accepting
requests, the SIP spec uses the terms UAC and UAS (user agent client and
server). However, this is obviously more awkward than just saying
"client". Just do the mental substitution...



-- 
Henning Schulzrinne   http://www.cs.columbia.edu/~hgs

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 12:56:46 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA12350
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 12:56:46 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 66A927FC9; Tue, 23 Feb 1999 12:45:26 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C8FC87FCA; Tue, 23 Feb 1999 12:45:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <001301be5f54$907a5360$d2b8c281@td.unige.ch>
From: "Marc hauswirth" <hauswirth@netserver1.td.unige.ch>
To: "IPTEL" <iptel@lists.research.bell-labs.com>
Subject: H.323 - H.225 ... analyser
Date: Tue, 23 Feb 1999 18:47:21 +0100
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.1012.300
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.1012.300
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hello,

Did somebody hear about an ethernet analyser that support the IP telehony
protocols.

Tank's

Marc hauswirth


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 13:28:54 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA12793
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 13:28:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 90F907FCA; Tue, 23 Feb 1999 13:17:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D8AC67FD1; Tue, 23 Feb 1999 13:17:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D2EF49.A057A6EB@americasm01.nt.com>
Date: Tue, 23 Feb 1999 10:11:21 -0800
From: "Mo Zonoun" <mzonoun@nortelnetworks.com>
Organization: Nortel
X-Mailer: Mozilla 4.5 [en] (Win95; U)
X-Accept-Language: en
MIME-Version: 1.0
Cc: IPTEL <iptel@lists.research.bell-labs.com>
Subject: IP Telephone Standards
References: <001301be5f54$907a5360$d2b8c281@td.unige.ch>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

ANSI TIA TR-41.3.4 has recently started a new project "Performance and
Interoperability Requirements for Voice-over-IP Telephone Terminals", to produce
a standard for design of IP phones. This standard is to include all aspects of
such designs, from physical connectors, to speech coding, audio quality and echo
management, to control protocols.  Initial focus has been on use of H.323 and
Megaco. The kickoff meeting was held on 15th Feb 1999 in Sarasota FL, with 21
attendees including Lucent, Microsoft, Nortel Networks, Siemens, Cisco,
Dialogic, NEC, Bell Canada, and a number of small companies.  The next official
TIA 41 meeting is slated for the week of May 17th, with the TR-41.3.4 working
group meeting to finalize as much of the spec as possible.  An interim
face-to-face meeting of the TR-41.3.4 working group was also proposed,
tentatively mid-April.  The spec is scheduled for completion by Sept 1999, using
a TIA fast track process. For more info go to www.tiaonline.org

Mo Zonoun



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 13:50:55 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA13213
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 13:50:55 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E97557FC8; Tue, 23 Feb 1999 13:41:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 425BE7FD1; Tue, 23 Feb 1999 13:41:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Jiri Kuthan <kuthan@fokus.gmd.de>
Message-Id: <199902231838.TAA01815@jobe.fokus.gmd.de>
Subject: Re: H.323 - H.225 ... analyser
To: hauswirth@netserver1.td.unige.ch (Marc hauswirth)
Date: Tue, 23 Feb 1999 19:38:36 +0100 (MET)
Cc: iptel@lists.research.bell-labs.com
In-Reply-To: <001301be5f54$907a5360$d2b8c281@td.unige.ch> from "Marc hauswirth" at Feb 23, 99 06:47:21 pm
X-Mailer: ELM [version 2.4 PL25]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Try the HP and RadCom solutions listed in the section "References-
weblinks-Testing'n'measurement" of my iptel website:
   http://www.fokus.gmd.de/cc/research/glone/projects/ipt

Regards, Jiri

> 
> Hello,
> 
> Did somebody hear about an ethernet analyser that support the IP telehony
> protocols.
> 
> Tank's
> 
> Marc hauswirth
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 


-- 
Jiri Kuthan		  Phone: ++49 / (0)30 / 3463 - 7 271
GMD FOKUS                 Fax  : ++49 / (0)30 / 3463 - 8 271
Kaiserin-Augusta-Allee 31 E-Mail : kuthan@fokus.gmd.de
D-10589 Berlin, Germany   WWW    : http://www.fokus.gmd.de/usr/kuthan

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 14:34:56 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA13969
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 14:34:56 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 219ED7FD1; Tue, 23 Feb 1999 14:25:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 66DDD7FD4; Tue, 23 Feb 1999 14:25:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Jiri Kuthan <kuthan@fokus.gmd.de>
Message-Id: <199902231923.UAA01867@jobe.fokus.gmd.de>
Subject: Re: H.323 - H.225 ... analyser
To: aanaya@infosel.com.mx (Alejandro Anaya Valverde)
Date: Tue, 23 Feb 1999 20:23:32 +0100 (MET)
Cc: kuthan@fokus.gmd.de, iptel@lists.research.bell-labs.com,
        hauswirth@netserver1.td.unige.ch
In-Reply-To: <91521560BC07D111B33F006097BC38AF01DF7088@CIDMSEXCHANGE> from "Alejandro Anaya Valverde" at Feb 23, 99 01:07:12 pm
X-Mailer: ELM [version 2.4 PL25]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

I am very sorry. The correct URL of the iptel
webpages is
 
 http://www.fokus.gmd.de/research/cc/glone/projects/ipt/

Thanks, Jiri

> 
> Apparently thats url is incorrect. See the message:
> Not Found
> The requested URL /cc/research/glone/projects/ipt was not found on this
> server.
> 
>> Try the HP and RadCom solutions listed in the section "References-
>> weblinks-Testing'n'measurement" of my iptel website:
>>    http://www.fokus.gmd.de/cc/research/glone/projects/ipt
>>> 
>>> Did somebody hear about an ethernet analyser that support the IP telehony
>>> protocols.

-- 
Jiri Kuthan		  Phone: ++49 / (0)30 / 3463 - 7 271
GMD FOKUS                 Fax  : ++49 / (0)30 / 3463 - 8 271
Kaiserin-Augusta-Allee 31 E-Mail : kuthan@fokus.gmd.de
D-10589 Berlin, Germany   WWW    : http://www.fokus.gmd.de/usr/kuthan

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 18:09:26 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA19125
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 18:09:26 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E6ED77FD4; Tue, 23 Feb 1999 17:59:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3F8CB7FD5; Tue, 23 Feb 1999 17:59:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Hanlin_Fang@3com.com
X-Lotus-FromDomain: 3COM
To: list iptel <iptel@lists.research.bell-labs.com>
Message-ID: <88256721.007E1A4C.00@hqoutbound.ops.3com.com>
Date: Tue, 23 Feb 1999 14:57:09 -0800
Subject: Looking for papers
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



Hello, All:

Does any one know where I can find the IETF Internet Draft  " An Inter-Gateway
Signalling Protocol for VoIP Networks" , by K, Melkild,

Thanks for any information

Hanlin



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 19:53:22 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA23008
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 19:53:22 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1B1E17FD5; Tue, 23 Feb 1999 19:43:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6528C7FD9; Tue, 23 Feb 1999 19:43:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <E32BD2062DAFD1119CFD0000F804F31702C0478E@cosexch003.mcit.com>
From: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>
To: "'Marc hauswirth'" <hauswirth@netserver1.td.unige.ch>,
        IPTEL <iptel@lists.research.bell-labs.com>
Subject: RE: H.323 - H.225 ... analyser
Date: Wed, 24 Feb 1999 00:42:43 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE5F8E.9A42F34A"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE5F8E.9A42F34A
Content-Type: text/plain

See Hewlett Packard ... don't recall the model number, but I've seen it
work.
Ralph

> -----Original Message-----
> From:	Marc hauswirth [SMTP:hauswirth@mail.td.unige.ch]
> Sent:	Tuesday, February 23, 1999 10:47 AM
> To:	IPTEL
> Subject:	H.323 - H.225 ... analyser
> 
> Hello,
> 
> Did somebody hear about an ethernet analyser that support the IP telehony
> protocols.
> 
> Tank's
> 
> Marc hauswirth
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

------_=_NextPart_001_01BE5F8E.9A42F34A
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>RE: H.323 - H.225 ... analyser</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">See Hewlett Packard ... don't =
recall the model number, but I've seen it work.</FONT>
<BR><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Ralph</FONT>
</P>
<UL>
<P><FONT SIZE=3D1 FACE=3D"Arial">-----Original Message-----</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">From:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Marc hauswirth =
[SMTP:hauswirth@mail.td.unige.ch]</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">Sent:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Tuesday, February 23, 1999 10:47 AM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">IPTEL</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT>=
</B> <FONT SIZE=3D1 FACE=3D"Arial">H.323 - H.225 ... analyser</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">Hello,</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">Did somebody =
hear about an ethernet analyser that support the IP telehony</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier =
New">protocols.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">Tank's</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">Marc =
hauswirth</FONT>
</P>
<BR>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier =
New">---------</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">This message =
came from the IETF IPTEL Working Group Mailing List.</FONT>
</P>
</UL>
</BODY>
</HTML>
------_=_NextPart_001_01BE5F8E.9A42F34A--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 20:23:57 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24454
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 20:23:57 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A84CB7FD9; Tue, 23 Feb 1999 20:09:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id EF37C7FDA; Tue, 23 Feb 1999 20:09:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: crk@research.att.com (Chuck Kalmanek)
Message-Id: <9902232006.ZM3703316@corona.research.att.com>
Date: Tue, 23 Feb 1999 20:06:38 -0500
X-Mailer: Z-Mail (3.2.1 10oct95)
To: iptel@lists.research.bell-labs.com
Subject: NOSSDAV99 Call-for-Papers
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


	The 9th International Workshop on Network and Operating Systems
             Support for Digital Audio and Video (NOSSDAV 99)

    *** NOTE: CHANGE IN SUBMISSION DEADLINE AND PROCEDURES BELOW. ***

 	        http://weblab.research.att.com/edas/nossdav99.html

                           23-25 June, 1999
                              Hosted By:
                       	       AT&T Labs
	   Networking and Distributed Systems Research Laboratory

                          CALL FOR PAPERS


Objectives

The 9th International Workshop on Network and Operating Systems
Support for Digital Audio and Video (NOSSDAV 99) is the international
workshop concerned with state of the art technology in networking and
operating system support for multimedia systems.  For eight years,
NOSSDAV has proven to be an outstanding forum for researchers involved
in building innovative multimedia systems, networks and applications
in both industry and academia.

A key aspect of the workshop is that it provides extensive discussion
periods during which attendees can informally discuss their current
work and future research directions. Traditionally, NOSSDAV has
emphasized high quality experimental research based on prototype or
real systems.  NOSSDAV99 will continue this tradition.

Submissions

Submissions are sought in any area related to Network and Operating
Systems Support for Digital Audio and Video. Two types of submissions
are solicited:  position statements or work in progress reports and full
papers.  These will be reviewed in entirely different ways and have
different deadlines.

Full papers will be subject to a normal review process and are
restricted to fifteen pages.  Accepted papers will be presented at the
Workshop and will appear in the published proceedings.  The best papers
will be forwarded to selected journals for publication.

Position statements and work in progress reports of no more than three
pages, will be reviewed by the program committee only.  Papers will be
chosen on the basis of their potential to provoke discussion.  Roughly
half of the workshop presentations will be allocated to this category.

To electronically register and submit a paper, please go to:

	http://weblab.research.att.com/edas/

Important Dates

Please note that the submission deadline for full papers has been
extended by two weeks.  However, this is a hard deadline.

Submission Deadline for Full Papers:            	15 March 1999
Submission Deadline for Position Statements:    	30 March 1999
Acceptance Notification:				 8 May 1999
Final Full Papers Due:                          	 5 June 1999

Workshop:                                       	23-25 June  1999

Program Chair:

Chuck Kalmanek
AT&T Labs
Shannon Laboratory
Room A113
180 Park Avenue
Florham Park NJ 07032
crk@research.att.com

Other Correspondence:

nossdav99@research.att.com

Program Committee

     Ian Leslie, Cambridge University, UK
     Charles Kalmanek, AT&T Labs Research
     Jim Kurose, University of Masachusetts
     Tom Little, Boston University
     Derek McAuley, Microsoft Research, UK
     Henning Schulzrinne, Columbia University
     Hide Tokuda, Keio University
     Domenico Ferrari, Universita Cattolica, Italy
     Kevin Jeffay, University of North Carolina
     Mike Jones, Microsoft
     Doug Shepherd, Lancaster University, UK
     Richard Black, University of Glasgow, UK
     Jason Nieh, Columbia University
     Partho Mishra, AT&T Labs Research
     S. Keshav, Cornell University
     H. Zhang, Carnegie Mellon University
     Raj Yavatkar, Intel
     Duane Northcutt, Sun Microsystems

LOCATION

AT&T Learning Center, Basking Ridge NJ

-- 
Chuck Kalmanek					Email: crk@research.att.com
AT&T Labs - Research 				Phone: 973-360-8720
Shannon Laboratory, Room A113			Fax:   973-360-8871
180 Park Avenue, Florham Park, N.J. 07932	

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Feb 23 22:05:56 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA00264
	for <iptel-archive@odin.ietf.org>; Tue, 23 Feb 1999 22:05:56 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8A4D17FDB; Tue, 23 Feb 1999 21:53:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D83647FDC; Tue, 23 Feb 1999 21:53:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D36924.B930D010@mail.e-comtech.com>
Date: Wed, 24 Feb 1999 10:51:16 +0800
From: Thad KP Fong <tfong@mail.e-comtech.com>
X-Mailer: Mozilla 4.05 [en] (Win95; I)
MIME-Version: 1.0
To: IP Mo Zonoun <mzonoun@nortelnetworks.com>,
        IP IPTEL <iptel@lists.research.bell-labs.com>
Subject: Re: [Fwd: IP Telephone Standards]
References: <36D3557B.4ECEB79B@mail.e-comtech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

To Mr.Mo Zonoun

Can you informed us how to be "listed", thus we can
receive steady information of IP Telephone Standards progress

If we cannot, may be not even qualified, to participate
IP technology progressing, at least we can be sideline LISTENER


Regards



> Subject: IP Telephone Standards
> Date: Tue, 23 Feb 1999 10:11:21 -0800
> From: "Mo Zonoun" <mzonoun@nortelnetworks.com>
> Organization: Nortel
> CC: IPTEL <iptel@lists.research.bell-labs.com>
> References: <001301be5f54$907a5360$d2b8c281@td.unige.ch>
> 
> ANSI TIA TR-41.3.4 has recently started a new project "Performance and
> Interoperability Requirements for Voice-over-IP Telephone Terminals", to produce
> a standard for design of IP phones. This standard is to include all aspects of
> such designs, from physical connectors, to speech coding, audio quality and echo
> management, to control protocols.  Initial focus has been on use of H.323 and
> Megaco. The kickoff meeting was held on 15th Feb 1999 in Sarasota FL, with 21
> attendees including Lucent, Microsoft, Nortel Networks, Siemens, Cisco,
> Dialogic, NEC, Bell Canada, and a number of small companies.  The next official
> TIA 41 meeting is slated for the week of May 17th, with the TR-41.3.4 working
> group meeting to finalize as much of the spec as possible.  An interim
> face-to-face meeting of the TR-41.3.4 working group was also proposed,
> tentatively mid-April.  The spec is scheduled for completion by Sept 1999, using
> a TIA fast track process. For more info go to www.tiaonline.org
> 
> Mo Zonoun
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 24 06:10:26 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA15371
	for <iptel-archive@odin.ietf.org>; Wed, 24 Feb 1999 06:10:26 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 78AC57FDC; Wed, 24 Feb 1999 05:03:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D371E7FDE; Wed, 24 Feb 1999 05:03:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <1.5.4.32.19990224100509.02699790@steptoe.hpl.hp.com>
X-Sender: hmd@steptoe.hpl.hp.com
X-Mailer: Windows Eudora Light Version 1.5.4 (32)
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Wed, 24 Feb 1999 10:05:09 +0000
To: IPTEL@lists.research.bell-labs.com,
        " <iptel@lists.research.bell-labs.com>"@steptoe.hpl.hp.com
From: Hans Daanen <hmd@hplb.hpl.hp.com>
Subject: Re: H.323 - H.225 ... analyser
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 18:47 23/02/99 +0100, Marc hauswirth wrote:
>Hello,
>
>Did somebody hear about an ethernet analyser that support the IP telehony
>protocols.

See:
http://www.hp.com/go/internetadvisor

Hans.
------------------------------------------------------------------------

Hans Daanen             e-mail: hmd@hplb.hpl.hp.com    #######   #######
Hewlett Packard Labs    phone:  +44 117 3128221        ####  /_  _  ####
Filton Road             fax:    +44 117 3128003        ###  / / /_/  ###
Stoke Gifford                                          ####    /    ####
Bristol BS34 8QZ        Location: building 3, EEL      #######   #######
United Kingdom


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 24 09:37:38 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA17019
	for <iptel-archive@odin.ietf.org>; Wed, 24 Feb 1999 09:37:34 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 03E497FB8; Wed, 24 Feb 1999 09:27:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 563357FC8; Wed, 24 Feb 1999 09:27:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <C51ED84B6F47D211917A0000F8BCBD11D15A86@zcard00g.ca.nortel.com>
From: "Tom-PT Taylor" <taylor@nortelnetworks.com>
To: Hanlin_Fang@3com.com, list iptel <iptel@lists.research.bell-labs.com>
Subject: RE: Looking for papers
Date: Wed, 24 Feb 1999 08:24:16 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Yopu can find it at
http://search.ietf.org/internet-drafts/draft-thomas-sigtran-igsp-00.txt

> -----Original Message-----
> From:	Hanlin_Fang@3com.com [SMTP:Hanlin_Fang@3com.com]
> Sent:	Tuesday, February 23, 1999 5:57 PM
> To:	list iptel
> Subject:	Looking for papers
> 
> 
> 
> Hello, All:
> 
> Does any one know where I can find the IETF Internet Draft  " An
> Inter-Gateway
> Signalling Protocol for VoIP Networks" , by K, Melkild,
> 
> Thanks for any information
> 
> Hanlin
> 
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 24 13:08:19 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA21150
	for <iptel-archive@odin.ietf.org>; Wed, 24 Feb 1999 13:08:19 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 4A6DD7FC8; Wed, 24 Feb 1999 12:09:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9ADFC7FC9; Wed, 24 Feb 1999 12:09:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Reply-To: <zehl@berkom.de>
From: "Andre' Zehl" <zehl@berkom.de>
To: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>,
        "'Marc hauswirth'" <hauswirth@netserver1.td.unige.ch>,
        "IPTEL" <iptel@lists.research.bell-labs.com>
Subject: RE: H.323 - H.225 ... analyser
Date: Wed, 24 Feb 1999 18:02:52 +0100
Message-Id: <000a01be6017$841f2320$270c278d@quake.berkom.de>
Mime-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_000B_01BE601F.E5E38B20"
X-Priority: 3 (Normal)
X-Msmail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Importance: Normal
X-Mimeole: Produced By Microsoft MimeOLE V4.72.2106.4
In-Reply-To: <E32BD2062DAFD1119CFD0000F804F31702C0478E@cosexch003.mcit.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.

------=_NextPart_000_000B_01BE601F.E5E38B20
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit

We are using an HP Internet Advisor currently, it does basic packet decode
of H.323 packets, fancy stuff like keeping a state machine, intelligent
error detection of signaling problems etc. is not (yet) there. Radcom also
has a decoder as well, AFAIK, which compares.


Andre'

zehl@berkom.de


    -----Original Message-----
    From: owner-iptel@lists.research.bell-labs.com
[mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Hoefelmeyer,
Ralph (MCI)
    Sent: Mittwoch, 24. Februar 1999 01:43
    To: 'Marc hauswirth'; IPTEL
    Subject: RE: H.323 - H.225 ... analyser


    See Hewlett Packard ... don't recall the model number, but I've seen it
work.
    Ralph

        -----Original Message-----
        From:   Marc hauswirth [SMTP:hauswirth@mail.td.unige.ch]
        Sent:   Tuesday, February 23, 1999 10:47 AM
        To:     IPTEL
        Subject:        H.323 - H.225 ... analyser

        Hello,

        Did somebody hear about an ethernet analyser that support the IP
telehony
        protocols.

        Tank's

        Marc hauswirth



        ---------
        This message came from the IETF IPTEL Working Group Mailing List.


------=_NextPart_000_000B_01BE601F.E5E38B20
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">
<HTML>
<HEAD>

<META content=3D"text/html; charset=3Dus-ascii" =
http-equiv=3DContent-Type><TITLE>RE: H.323 - H.225 ... analyser</TITLE>
<META content=3D'"MSHTML 4.72.2106.6"' name=3DGENERATOR>
</HEAD>
<BODY>
<DIV><SPAN class=3D110345216-24021999><FONT color=3D#0000ff face=3DArial =
size=3D2>We are=20
using an HP Internet Advisor currently, it does basic packet decode of =
H.323=20
packets, fancy stuff like keeping a state machine, intelligent error =
detection=20
of signaling problems etc. is not (yet) there. Radcom also has a decoder =
as=20
well, AFAIK, which compares.</FONT></SPAN></DIV><BR>
<P><FONT size=3D2>Andre'</FONT></P>
<P><FONT size=3D2></FONT><FONT size=3D2><A=20
href=3D"mailto:zehl@berkom.de">zehl@berkom.de</A></FONT> </P>
<DIV>&nbsp;</DIV>
<BLOCKQUOTE=20
style=3D"BORDER-LEFT: #0000ff solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: =
5px">
    <DIV class=3DOutlookMessageHeader><FONT face=3D"Times New Roman"=20
    size=3D2>-----Original Message-----<BR><B>From:</B>=20
    owner-iptel@lists.research.bell-labs.com=20
    [mailto:owner-iptel@lists.research.bell-labs.com]<B>On Behalf Of</B> =

    Hoefelmeyer, Ralph (MCI)<BR><B>Sent:</B> Mittwoch, 24. Februar 1999=20
    01:43<BR><B>To:</B> 'Marc hauswirth'; IPTEL<BR><B>Subject:</B> RE: =
H.323 -=20
    H.225 ... analyser<BR><BR></FONT></DIV>
    <P><FONT color=3D#0000ff face=3DArial>See Hewlett Packard ... don't =
recall the=20
    model number, but I've seen it work.</FONT> <BR><FONT =
color=3D#0000ff=20
    face=3DArial>Ralph</FONT> </P>
    <UL>
        <P><FONT face=3DArial size=3D1>-----Original Message-----</FONT> =

        <BR><B><FONT face=3DArial size=3D1>From:&nbsp;&nbsp;</FONT></B> =
<FONT=20
        face=3DArial size=3D1>Marc hauswirth=20
        [SMTP:hauswirth@mail.td.unige.ch]</FONT> <BR><B><FONT =
face=3DArial=20
        size=3D1>Sent:&nbsp;&nbsp;</FONT></B> <FONT face=3DArial =
size=3D1>Tuesday,=20
        February 23, 1999 10:47 AM</FONT> <BR><B><FONT face=3DArial=20
        size=3D1>To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT =
face=3DArial=20
        size=3D1>IPTEL</FONT> <BR><B><FONT face=3DArial=20
        =
size=3D1>Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B>=20
        <FONT face=3DArial size=3D1>H.323 - H.225 ... analyser</FONT> =
</P>
        <P><FONT color=3D#000000 face=3D"Courier New" =
size=3D2>Hello,</FONT> </P>
        <P><FONT color=3D#000000 face=3D"Courier New" size=3D2>Did =
somebody hear about=20
        an ethernet analyser that support the IP telehony</FONT> =
<BR><FONT=20
        color=3D#000000 face=3D"Courier New" size=3D2>protocols.</FONT> =
</P>
        <P><FONT color=3D#000000 face=3D"Courier New" =
size=3D2>Tank's</FONT> </P>
        <P><FONT color=3D#000000 face=3D"Courier New" size=3D2>Marc =
hauswirth</FONT>=20
        </P><BR>
        <P><FONT color=3D#000000 face=3D"Courier New" =
size=3D2>---------</FONT>=20
        <BR><FONT color=3D#000000 face=3D"Courier New" size=3D2>This =
message came from=20
        the IETF IPTEL Working Group Mailing List.</FONT>=20
</P></UL></BLOCKQUOTE></BODY></HTML>

------=_NextPart_000_000B_01BE601F.E5E38B20--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 24 13:28:15 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA21366
	for <iptel-archive@odin.ietf.org>; Wed, 24 Feb 1999 13:28:14 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 200847FC9; Wed, 24 Feb 1999 13:07:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7D5A17FCA; Wed, 24 Feb 1999 13:07:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D43E26.ECA5533F@americasm01.nt.com>
Date: Wed, 24 Feb 1999 10:00:07 -0800
From: "Mo Zonoun" <mzonoun@nortelnetworks.com>
Organization: Nortel
X-Mailer: Mozilla 4.5 [en] (Win95; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Thad KP Fong <tfong@mail.e-comtech.com>
Cc: IP IPTEL <iptel@lists.research.bell-labs.com>
Subject: Re: [Fwd: IP Telephone Standards]
References: <36D3557B.4ECEB79B@mail.e-comtech.com> <36D36924.B930D010@mail.e-comtech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

TIA TR-41 does not have a listing for non-member. If IpTel is interested in the work
that is going on in TIA TR-41, there are several ways we can exchange information with
them. Establishing liaison with them for formal document exchange or may be a
presentation by TIA TR-41 to IpTel (which probably I can arrange).

Mo Zonoun



Thad KP Fong wrote:

> To Mr.Mo Zonoun
>
> Can you informed us how to be "listed", thus we can
> receive steady information of IP Telephone Standards progress
>
> If we cannot, may be not even qualified, to participate
> IP technology progressing, at least we can be sideline LISTENER
>
> Regards
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 24 14:54:43 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA22999
	for <iptel-archive@odin.ietf.org>; Wed, 24 Feb 1999 14:54:43 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id CD98B7FCA; Wed, 24 Feb 1999 14:45:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 396D97FD1; Wed, 24 Feb 1999 14:45:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Dave Winton" <dwinton@viennasys.com>
To: <iptel@lists.research.bell-labs.com>
Subject: Nortel's IPS7
Date: Wed, 24 Feb 1999 14:54:49 -0500
Message-ID: <001601be602f$89b7b500$64c9a8c0@baltic.viennasys.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

All;

  Any ideas where can I find Nortel's IPS7 specification???? Any links????





Thanks;


Dave Winton 
Nokia IP Telepony Business Unit
555 legget Drive, Suite 400
Kanata, Ontario
Canada, K2K 2X3

Tel: (613) 591-7450 Ext. 2688
E-mail: dwinton@viennasys.com


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 24 15:35:03 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA23791
	for <iptel-archive@odin.ietf.org>; Wed, 24 Feb 1999 15:35:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A9E637FD1; Wed, 24 Feb 1999 15:23:50 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 148AC7FD4; Wed, 24 Feb 1999 15:23:49 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D45F47.636C8C0B@dnrc.bell-labs.com>
Date: Wed, 24 Feb 1999 15:21:27 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Mo Zonoun <mzonoun@nortelnetworks.com>
Cc: Thad KP Fong <tfong@mail.e-comtech.com>,
        IP IPTEL <iptel@lists.research.bell-labs.com>
Subject: Re: [Fwd: IP Telephone Standards]
References: <36D3557B.4ECEB79B@mail.e-comtech.com> <36D36924.B930D010@mail.e-comtech.com> <36D43E26.ECA5533F@americasm01.nt.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Thad, Mo,

I believe Thad's request was for his company specifically, not for iptel
in general. But, if either of you want specifically to establish a
liaison between TIA and iptel, please approach me privately along those
lines.

Thanks,
Jonathan R.

Mo Zonoun wrote:
> 
> TIA TR-41 does not have a listing for non-member. If IpTel is interested in the work
> that is going on in TIA TR-41, there are several ways we can exchange information with
> them. Establishing liaison with them for formal document exchange or may be a
> presentation by TIA TR-41 to IpTel (which probably I can arrange).
> 
> Mo Zonoun
> 
> Thad KP Fong wrote:
> 
> > To Mr.Mo Zonoun
> >
> > Can you informed us how to be "listed", thus we can
> > receive steady information of IP Telephone Standards progress
> >
> > If we cannot, may be not even qualified, to participate
> > IP technology progressing, at least we can be sideline LISTENER
> >
> > Regards
> >
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 24 15:54:43 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA24109
	for <iptel-archive@odin.ietf.org>; Wed, 24 Feb 1999 15:54:43 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id AD0BF7FD4; Wed, 24 Feb 1999 15:43:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F3F4A7FDD; Wed, 24 Feb 1999 15:43:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D46426.D3011447@cs.columbia.edu>
Date: Wed, 24 Feb 1999 15:42:14 -0500
From: Henning Schulzrinne <hgs@cs.columbia.edu>
Organization: Columbia University, Dept. of Computer Science
X-Mailer: Mozilla 4.5 [en] (X11; I; SunOS 5.6 sun4u)
X-Accept-Language: en
MIME-Version: 1.0
To: confctrl@ISI.EDU, iptel@lists.research.bell-labs.com, rem-conf@es.net,
        voip-tech@vocaltec.com
Cc: Jeff Pulver <jeff@pulver.com>
Subject: First SIP Bake-Off at Columbia University, April 8/9, 1999
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

After an informal survey, a number of people have expressed interest in
participating in a first ``SIP Bake-Off''. The purpose of the bake-off
is to test for interoperability of finished and in in-progress SIP
implementations.

The bake-off will take place at Columbia University, New York, NY on
Thursday, April 8th and Friday, April 9th.

Details can be found at http://www.cs.columbia.edu/~hgs/sip/bakeoff.html

-- 
Henning Schulzrinne   http://www.cs.columbia.edu/~hgs

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Feb 24 17:19:00 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA25585
	for <iptel-archive@odin.ietf.org>; Wed, 24 Feb 1999 17:19:00 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3AF9C7FDC; Wed, 24 Feb 1999 17:09:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 940F57FDD; Wed, 24 Feb 1999 17:09:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <F033F6FEF3F1D111BD150000F8CD143101AAC3A2@zcard007.ca.nortel.com>
From: "Nancy-M Greene" <ngreene@nortelnetworks.com>
To: iptel <iptel@lists.research.bell-labs.com>,
        "'Dave Winton'" <dwinton@viennasys.com>
Subject: RE: Nortel's IPS7
Date: Wed, 24 Feb 1999 15:25:29 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

> A more relevant mailing list for this question is the sigtran one. But in
> fact this kind of question doesn't really belong on the IETF mailing
> lists... In any case, I'll get back to you directly with an answer.
> 
Nancy
--------------------------------------------------------------------------
Nancy M. Greene 
Internet & Service Provider Networks, Nortel Networks
T:514-271-7221 (internal:ESN853-1077) E:ngreene@nortelnetworks.com

> ----------
> From: 	Dave Winton[SMTP:dwinton@viennasys.com]
> Sent: 	Wednesday, February 24, 1999 2:54 PM
> To: 	iptel
> Subject: 	Nortel's IPS7
> 
> All;
> 
>   Any ideas where can I find Nortel's IPS7 specification???? Any links????
> 
> Thanks;
> 
> 
> Dave Winton 
> Nokia IP Telepony Business Unit
> 555 legget Drive, Suite 400
> Kanata, Ontario
> Canada, K2K 2X3
> 
> Tel: (613) 591-7450 Ext. 2688
> E-mail: dwinton@viennasys.com
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 25 00:12:06 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA23983
	for <iptel-archive@odin.ietf.org>; Thu, 25 Feb 1999 00:12:05 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id F38F97FB7; Thu, 25 Feb 1999 00:01:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4813D7FD9; Thu, 25 Feb 1999 00:01:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D4D92C.6A2F1972@mail.e-comtech.com>
Date: Thu, 25 Feb 1999 13:01:32 +0800
From: Thad KP Fong <tfong@mail.e-comtech.com>
X-Mailer: Mozilla 4.05 [en] (Win95; I)
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Mo Zonoun <mzonoun@nortelnetworks.com>,
        IP IPTEL <iptel@lists.research.bell-labs.com>,
        E-Com pengcong <pengcong@mail.e-comtech.com>
Subject: Re: [Fwd: IP Telephone Standards]
References: <36D3557B.4ECEB79B@mail.e-comtech.com> <36D36924.B930D010@mail.e-comtech.com> <36D43E26.ECA5533F@americasm01.nt.com> <36D45F47.636C8C0B@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

To Jonathan Rosenberg,  Mo, 

Yes.  The request is for our company.
Then again, we are in this business, thus vital to follow 
closely.

+++
Some Background of E-Com Technology Ltd :
E-Com Technology is software company in China, 100% privated 
ownership 2 years old (established 1996 December).  
Still small, but we are into expansion phase getting more, 
more engineers.  Several person in the group are USA educated 
PH.D. mainland  Chinese.
Our products are two area
   a> VOIP.  This is Internet IP technology, we did/are doing
      develop all the core technology. one of our major
      customer is IBM Asia Pacific (VOIP gateway under OS Warp
      platform)  + others
   b> Medical Imaging : currently we are working with Fuji Film
      and CR Corporation (formerly Sterling Diagnostic, ex-DuPont
      Medicals) on their CR & DR.  As well in contact with Kodak
      All these are for Greater China Market.  We are working
      on Mini-Pacs & DICOM etc, etc, etc
      We have members with Medical background (Mainland
     China hospitals Radiologies),  + super Computer background

Lastly I am not technical side, I am just taking care of money
& international side

Our address as follows:
    Thad Fong
    E-Com Technology Ltd
    6/F  Glass Building
    Nanhai Oil Silver Bay Resort
    Zhuhai, GD519015
    China

    tel : 86 756 3367066   and or 86 756 3367076(direct line)
    fax : 86 756 3363065

    em : tfong@mail.e-comtech.com  or zuwinltd@akkord.com



Regards


Jonathan Rosenberg wrote:
> 
> Thad, Mo,
> 
> I believe Thad's request was for his company specifically, not for iptel
> in general. But, if either of you want specifically to establish a
> liaison between TIA and iptel, please approach me privately along those
> lines.
> 
> Thanks,
> Jonathan R.
> 
> Mo Zonoun wrote:
> >
> > TIA TR-41 does not have a listing for non-member. If IpTel is interested in the work
> > that is going on in TIA TR-41, there are several ways we can exchange information with
> > them. Establishing liaison with them for formal document exchange or may be a
> > presentation by TIA TR-41 to IpTel (which probably I can arrange).
> >
> > Mo Zonoun
> >
> > Thad KP Fong wrote:
> >
> > > To Mr.Mo Zonoun
> > >
> > > Can you informed us how to be "listed", thus we can
> > > receive steady information of IP Telephone Standards progress
> > >
> > > If we cannot, may be not even qualified, to participate
> > > IP technology progressing, at least we can be sideline LISTENER
> > >
> > > Regards
> > >
> >
> > ---------
> > This message came from the IETF IPTEL Working Group Mailing List.
> 
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 25 10:56:34 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA12037
	for <iptel-archive@odin.ietf.org>; Thu, 25 Feb 1999 10:56:33 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6277A7FB8; Thu, 25 Feb 1999 10:45:38 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BF5887FD1; Thu, 25 Feb 1999 10:45:37 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D56F8D.C1CCB388@dnrc.bell-labs.com>
Date: Thu, 25 Feb 1999 10:43:09 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Alok Mittal <amittal@hss.hns.com>
Cc: "'Matt Squire'" <msquire@nortelnetworks.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
References: <01BE5F33.97C7FF30@hss056.hss.hns.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Alok Mittal wrote:
> 
> Sorry for the lost indentations.
> Hi Matt,
> 
> >So I agree that determining which attributes are understood by each peer is
> >important.  I haven't yet decided if this function is better provided by
> >negotiations during hello phase or by flags in the attribute.  Opinions?
> 
> To me it seems that this activity is best done in the negotiation so that a lot
> of overhead is avoided if the peers are not compatible. The degree of control
> offered by per-attribute control does not seem to be necessary.

On the other hand, I think marking the attributes is simpler, avoiding
the need to define a specific capabilities exchange. 

> 
> >I think that other applications of SCSP use similar information as their
> >identifiers (I can't get to the IETF server to confirm right now).
> > <snip>
> >But I can't specify this format until we figure out how we want to
> >represent the Internet Telephony Administrative Domain.  Another Editor's
> >note asks for opinions of how an ITAD should be represented (url, string,
> >decimal, etc).  After we answer this question, I'll go back and re-specify
> >what the SCSP identifiers should be.
> >
> >Does that sound reasonable?
> 
> Yes, we could go with that approach. Or just to use what H.323 does for Call
> Identifiers -- it specifies a 6 byte field within Call ID for "node" specifying that
> it has to be spatially unique. The recommendation then suggests that ethernet
> card address should be used, BUT if there is no netowork card, the system
> may use other means to generate globally unique node identifier -- such as
> make least significant bit of first octet to 1 (refers to multicast so doesn't conflict
> with any ethernet card addresses) and fill a 47 bit pseudo random number
> elsewhere.

I agree. Let implementations determine how to guarantee uniqueness, and
make some reccomendations in the draft.

> 
> >I think the real question is whether a prefix is sufficient to specify the
> >likely sets of target telephone numbers.  In the US, I think that a prefix
> >is sufficient.  The proximity of area/city codes does not mean the
> >proximity of areas/cities, which seems a likely criteria for a gateway
> >serving 'range' but not a 'prefix'.  Is this common of phone numbering
> >systems throughout the world?  If so, then I agree we should go with
> >prefixes and not ranges.
> 
> In India, again there is no strong correlation between the two, though in many
> cases such a property may be satisfied.

The problem is that there is really no other scalable way to do it. In
an environment where the correlation between prefixes and geographical
regions (and more specifically, cost) is lost, there really is no easy
way to specify phone number ranges.



> 
> >Its not clear to me what you mean by cost. Cost, to most routing protocols,
> >is the number of hops.  But we don't really care about the number of
> >signalling hops.  A couple weeks ago, there was something on the mailing
> >list which enumerated many variables affecting the cost/price of an
> >internet phone call.  So the meaning of cost is ambiguous at best.
> >
> >If you have something specific that you mean by cost then let me know, but
> >my view is that we can't do anything with real cost, and that we should
> >stick to a black-box structure to express business practices and
> >relationships.
> 
> Yes, I agree that this is a rather vague thing. To think of a cost metric at an
> application level, we could use the number of administrative domains through
> which the shortest path to destination lies (from this gateway/signaling node).
> This is based on an assumption that every administrative domain would charge
> *some* fees for routing packets through its gateways. But this will lead to the
> problem of propagating this information, and it also makes a simplifying
> assumption that cost of passing through each administrative domain is more
> or less the same (or at least the difference is insignificant as compared to
> passing through an additional domain). Will think more and get back to you
> if I have better suggestions!

I think we still need a pricing structure, opaque to GLP. We need a hop
count for TTL purposes already, and we have also discussed a capacity
metric. This would seem to provide a reasonably large enough set of
information for an LS to decide what to do with the routing object,
based on whatever total measure of "cost" it decides to define.

If there is additional cost for transit service, shouldn't this be
wrapped into the pricing attribute? An ITAD would simply modify the
pricing to include its own transit costs before propagating the
attribute, if it so chooses.

-Jonathan R.


-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 25 15:47:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA18894
	for <iptel-archive@odin.ietf.org>; Thu, 25 Feb 1999 15:47:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D32137FC8; Thu, 25 Feb 1999 13:41:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1B3327FD4; Thu, 25 Feb 1999 13:41:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X400-Received: by /PRMD=INTERNET/ADMD=ATLAS/C=FR/; Relayed;
               Thu, 25 Feb 1999 18:25:22 +0100
X400-Received: by mta xn1-gw.atlas.fr in /PRMD=INTERNET/ADMD=ATLAS/C=FR/;
               Relayed; Thu, 25 Feb 1999 18:25:22 +0100
X400-Received: by /ADMD=ATLAS/C=FR/; Relayed; Thu, 25 Feb 1999 18:25:23 +0100
X400-Received: by /PRMD=art-telecom/ADMD=atlas/C=fr/; Relayed;
               Thu, 25 Feb 1999 18:16:04 +0100
Date: Thu, 25 Feb 1999 18:16:04 +0100
X400-Originator: Didier.CHAUVEAU@art-telecom.fr
X400-Recipients: non-disclosure:;
X400-MTS-Identifier: [/PRMD=art-telecom/ADMD=atlas/C=fr/;XGW-990225180637+0100-28773]
Original-Encoded-Information-Types: undefined
X400-Content-Type: P2-1988 (22)
Content-Identifier: Voice over IP: p
From: Didier.CHAUVEAU@art-telecom.fr
To: iptel@lists.research.bell-labs.com, pint@lists.research.bell-labs.com,
        ss7-internet@BayNetworks.com, e164-to-Ip@lserv.vocaltec.com
Subject:  Voice over IP: public call for comments
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Message-Id: <19990225184106.45A947FC8@lists.research.bell-labs.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by ietf.org id PAA18894

 Dear IETF members,
 
 ART launched a public call for comments concerning voice over IP (see 
 http://www.art-telecom.fr).
 
 Please note that the deadline to submit your contribution is postponed: 
 Responses should reach ART by noon on 23 april 1999 (see details on public 
 call for comments).
 
 Do not hesitate to share this information with interested parties.
 
 Regards,
 
 
 didier chauveau
 ART
 didier.chauveau@art-telecom.fr 
 regards 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Feb 25 18:18:35 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA25695
	for <iptel-archive@odin.ietf.org>; Thu, 25 Feb 1999 18:18:34 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 67B8F7FC9; Thu, 25 Feb 1999 17:11:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id AF5CA7FD1; Thu, 25 Feb 1999 17:11:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D5C8E7.879391DE@americasm01.nt.com>
Date: Thu, 25 Feb 1999 14:04:24 -0800
From: "Mo Zonoun" <mzonoun@nortelnetworks.com>
Organization: Nortel
X-Mailer: Mozilla 4.5 [en] (Win95; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Thad KP Fong <tfong@mail.e-comtech.com>,
        IP IPTEL <iptel@lists.research.bell-labs.com>
Subject: Re: [Fwd: IP Telephone Standards]
References: <36D3557B.4ECEB79B@mail.e-comtech.com> <36D36924.B930D010@mail.e-comtech.com> <36D43E26.ECA5533F@americasm01.nt.com> <36D45F47.636C8C0B@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Orig: <mzonoun@americasm01.nt.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Jonathan,
I understand.  I wanted to see if others in iptel are also interested. If that is the case
then certainly it is up to the chair to initiate the liaison.

Mo Zonoun

Jonathan Rosenberg wrote:

> Thad, Mo,
>
> I believe Thad's request was for his company specifically, not for iptel
> in general. But, if either of you want specifically to establish a
> liaison between TIA and iptel, please approach me privately along those
> lines.
>
> Thanks,
> Jonathan R.
>
> Mo Zonoun wrote:
> >
> > TIA TR-41 does not have a listing for non-member. If IpTel is interested in the work
> > that is going on in TIA TR-41, there are several ways we can exchange information with
> > them. Establishing liaison with them for formal document exchange or may be a
> > presentation by TIA TR-41 to IpTel (which probably I can arrange).
> >
> > Mo Zonoun
> >
> > Thad KP Fong wrote:
> >
> > > To Mr.Mo Zonoun
> > >
> > > Can you informed us how to be "listed", thus we can
> > > receive steady information of IP Telephone Standards progress
> > >
> > > If we cannot, may be not even qualified, to participate
> > > IP technology progressing, at least we can be sideline LISTENER
> > >
> > > Regards
> > >
> >
> > ---------
> > This message came from the IETF IPTEL Working Group Mailing List.
>
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 26 10:09:26 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA20241
	for <iptel-archive@odin.ietf.org>; Fri, 26 Feb 1999 10:09:26 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A4D3B7FD1; Fri, 26 Feb 1999 09:57:50 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 06D7B7FD4; Fri, 26 Feb 1999 09:57:49 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990226094434.00a67310@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Fri, 26 Feb 1999 09:44:35 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        Alok Mittal <amittal@hss.hns.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 10:43 AM 2/25/99 -0500, Jonathan Rosenberg wrote:
>Alok Mittal wrote:
>> 
>> Sorry for the lost indentations.
>> Hi Matt,
>> 
>> >So I agree that determining which attributes are understood by each
peer is
>> >important.  I haven't yet decided if this function is better provided by
>> >negotiations during hello phase or by flags in the attribute.  Opinions?
>> 
>> To me it seems that this activity is best done in the negotiation so
that a lot
>> of overhead is avoided if the peers are not compatible. The degree of
control
>> offered by per-attribute control does not seem to be necessary.
>
>On the other hand, I think marking the attributes is simpler, avoiding
>the need to define a specific capabilities exchange. 
>

The two approaches are not mutually exclusive.  

I think there should be version negotiation in the hello phase.  Other
features/attributes which must be agreed upon before two peers synchronize
should also be negotiated in the hello phase.  BGP has an I-D where
capabilities exhange is added to the OPEN message, and I think we should
have similar flexibility.  

Attributes could also be marked 'required', and if the attribute is not
understood by the peer then the entire routing object is ignored by the
peer.  

Including a list of attributes that an LS understands in the hello could be
used to reduce/eliminate the sending of 'required' attributes to a peer
that it does not understand.   This is just one of the ways to use hello
negotiation.  I think the up-front negotiation could be used to optimize,
but not replace, flagging individual attributes as 'required'.  
  

>> 
>> >I think that other applications of SCSP use similar information as their
>> >identifiers (I can't get to the IETF server to confirm right now).
>> > <snip>
>> >But I can't specify this format until we figure out how we want to
>> >represent the Internet Telephony Administrative Domain.  Another Editor's
>> >note asks for opinions of how an ITAD should be represented (url, string,
>> >decimal, etc).  After we answer this question, I'll go back and re-specify
>> >what the SCSP identifiers should be.
>> >
>> >Does that sound reasonable?
>> 
>> Yes, we could go with that approach. Or just to use what H.323 does for
Call
>> Identifiers -- it specifies a 6 byte field within Call ID for "node"
specifying that
>> it has to be spatially unique. The recommendation then suggests that
ethernet
>> card address should be used, BUT if there is no netowork card, the system
>> may use other means to generate globally unique node identifier -- such as
>> make least significant bit of first octet to 1 (refers to multicast so
doesn't conflict
>> with any ethernet card addresses) and fill a 47 bit pseudo random number
>> elsewhere.
>
>I agree. Let implementations determine how to guarantee uniqueness, and
>make some reccomendations in the draft.
>

Although its not required for SCSP, I think we should at least pick a size
for the identifier in order to simplify packet processing.  Picking a
meaningful size requires some assumption about the content.  Larger
identifiers imply more overhead because every CSA/CallRoutingObject is
tagged with an identifier.  

Should the size/recommendation be based on:
  (a) internet address
  (b) mac address
  (c) domain:server
  (d) all/any of the above
  (e) others
  (f) variable length, unspecified content

Personally, I'd prefer to specify a 4-byte identifier with a recommendation
to use an IPv4 address or to set it administratively.  This ID is just has
to be unique withing the server group(s) that the LS is in, it doesn't have
to be globally unique.  

A more general possibility is to specify a 16-byte identifier.  A
recommendation could be to use 
  - ipv6 unicast addr
  - 0xff leading byte, 0x01 second byte, then all zeros, with ipv4 address
at end
  - 0xff leading byte, 0x02 second byte, then all zeros followed by mac
address at end
  - 0xff leading byte, 0x03 second byte, then some kind of ITAD:server
representation (yet to be determined)
  - 0xff leading byte, 0xff second byte, followed by anything you want (a
local problem to guarantee uniqueness)
But I think this is overkill.  


>
>
>> 
>> >Its not clear to me what you mean by cost. Cost, to most routing
protocols,
>> >is the number of hops.  But we don't really care about the number of
>> >signalling hops.  A couple weeks ago, there was something on the mailing
>> >list which enumerated many variables affecting the cost/price of an
>> >internet phone call.  So the meaning of cost is ambiguous at best.
>> >
>> >If you have something specific that you mean by cost then let me know, but
>> >my view is that we can't do anything with real cost, and that we should
>> >stick to a black-box structure to express business practices and
>> >relationships.
>> 
>> Yes, I agree that this is a rather vague thing. To think of a cost
metric at an
>> application level, we could use the number of administrative domains
through
>> which the shortest path to destination lies (from this gateway/signaling
node).
>> This is based on an assumption that every administrative domain would
charge
>> *some* fees for routing packets through its gateways. But this will lead
to the
>> problem of propagating this information, and it also makes a simplifying
>> assumption that cost of passing through each administrative domain is more
>> or less the same (or at least the difference is insignificant as
compared to
>> passing through an additional domain). Will think more and get back to you
>> if I have better suggestions!
>
>I think we still need a pricing structure, opaque to GLP. We need a hop
>count for TTL purposes already, and we have also discussed a capacity
>metric. This would seem to provide a reasonably large enough set of
>information for an LS to decide what to do with the routing object,
>based on whatever total measure of "cost" it decides to define.
>
>If there is additional cost for transit service, shouldn't this be
>wrapped into the pricing attribute? An ITAD would simply modify the
>pricing to include its own transit costs before propagating the
>attribute, if it so chooses.
>

I agree.

- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 26 10:24:39 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA20443
	for <iptel-archive@odin.ietf.org>; Fri, 26 Feb 1999 10:24:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A26F17FD4; Fri, 26 Feb 1999 10:12:33 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 02F0F7FDA; Fri, 26 Feb 1999 10:12:32 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D6B949.4D224D2C@dnrc.bell-labs.com>
Date: Fri, 26 Feb 1999 10:10:01 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: list iptel <iptel@lists.research.bell-labs.com>
Subject: New draft on glp
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

members of iptel,

Dave Hampton, Dave Oran, Hussein Salama, and Dhaval Shah have just
submitted a draft on using BGP4 as a basis for the GLP. Until the draft
appears in the archives, it is available from our webpage:

http://www.bell-labs.com/mailing-lists/iptel/

Please note that Cisco has indicated it may have IPR related to this
draft.

It seems we now have two competing proposals (one based on SCSP, and one
on BGP) for GLP. I'd like to begin some discussion on the lists here
debating the relative merits of the two approaches. Fire away....

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 26 14:50:31 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA25600
	for <iptel-archive@odin.ietf.org>; Fri, 26 Feb 1999 14:50:30 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DBE297FDA; Fri, 26 Feb 1999 13:56:25 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 422447FDB; Fri, 26 Feb 1999 13:56:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D6EDC2.9EF4AEA0@dnrc.bell-labs.com>
Date: Fri, 26 Feb 1999 13:53:54 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: list iptel <iptel@lists.research.bell-labs.com>
Subject: iptel meeting time
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Folks,

We have been assigned our slot at Minneapolis. It is: 

 Tuesday, March 16 at 1300-1515 (two one-hour sessions combined)
        other groups scheduled at that time: hfn,
ediint,                                 calsch, vrrp, dnsind
        usv area open meeting, idwg, aaa

I expect to have a packed agenda. Anyone who has something they'd like
to discuss, please let me know.

Thanks,
Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 26 17:00:48 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA28914
	for <iptel-archive@odin.ietf.org>; Fri, 26 Feb 1999 17:00:48 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DF22C7FDB; Fri, 26 Feb 1999 16:49:00 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3EDDC7FDC; Fri, 26 Feb 1999 16:49:00 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36D71636.18EFA219@dnrc.bell-labs.com>
Date: Fri, 26 Feb 1999 16:46:30 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: list iptel <iptel@lists.research.bell-labs.com>
Subject: Two more drafts
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

members of iptel,

Two additional drafts are now available from the iptel web page
(http://www.bell-labs.com/mailing-lists/iptel/), both of which should
appear in the archives shortly. 

The first is draft-ietf-iptel-cpl-00.{txt,ps,pdf}, which is a draft of
the CPL syntax itself. Please take a look. There are a number of open
issues there, and we need to discuss whether the set of tags described
in the document is sufficient or not to meet peoples desires.

The second is draft-ietf-iptel-sip-reg-payload-00.{txt,ps,pdf}, which
discusses how to transport both CPL's and other user logic (such as a
CGI script) in a SIP register message. We'll need to decide if the
approach is good, and also what to do with this draft - should we have a
new document on transporting CPL's, should this be folded into the
framework, or should we forget about it for now?

Thanks,
Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Feb 26 17:23:58 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA00483
	for <iptel-archive@odin.ietf.org>; Fri, 26 Feb 1999 17:23:57 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1E7CD7FCA; Fri, 26 Feb 1999 17:13:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 649237FD5; Fri, 26 Feb 1999 17:13:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "list iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: Two more drafts
Date: Fri, 26 Feb 1999 17:10:38 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFIEHKEDAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
In-Reply-To: <36D71636.18EFA219@dnrc.bell-labs.com>
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Note that there's a proposal into MMUSIC to allow multiple body parts in
SIP - look for *sip-multipart*. This may have some bearing on the
encoding.

> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Jonathan
> Rosenberg
> Sent: Friday, February 26, 1999 4:47 PM
> To: list iptel
> Subject: Two more drafts
>
>
> members of iptel,
>
> Two additional drafts are now available from the iptel web page
> (http://www.bell-labs.com/mailing-lists/iptel/), both of which should
> appear in the archives shortly.
>
> The first is draft-ietf-iptel-cpl-00.{txt,ps,pdf}, which is a draft of
> the CPL syntax itself. Please take a look. There are a number of open
> issues there, and we need to discuss whether the set of tags described
> in the document is sufficient or not to meet peoples desires.
>
> The second is draft-ietf-iptel-sip-reg-payload-00.{txt,ps,pdf}, which
> discusses how to transport both CPL's and other user logic (such as a
> CGI script) in a SIP register message. We'll need to decide if the
> approach is good, and also what to do with this draft -
> should we have a
> new document on transporting CPL's, should this be folded into the
> framework, or should we forget about it for now?
>
> Thanks,
> Jonathan R.
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen
>
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  1 19:17:40 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA06556
	for <iptel-archive@odin.ietf.org>; Mon, 1 Mar 1999 19:17:40 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BE18A7FB7; Mon,  1 Mar 1999 19:07:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 26A627FB8; Mon,  1 Mar 1999 19:07:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903020006.TAA06197@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: iptel@lists.research.bell-labs.com
From: Internet-Drafts@ietf.org
Reply-To: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-iptel-glp-tbgp-00.txt
Date: Mon, 01 Mar 1999 19:06:43 -0500
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the IP Telephony Working Group of the IETF.

	Title		: The IP Telephony Border Gateway Protocol Architecture
	Author(s)	: D. Hampton, D. Oran, H. Salama, D. Shah
	Filename	: draft-ietf-iptel-glp-tbgp-00.txt
	Pages		: 12
	Date		: 26-Feb-99
	
   This document presents the architecture of the IP Telephony Border
   Gateway Protocol (TBGP). TBGP is an inter-domain protocol, for
   routing voice calls through the IP network towards their
   destinations, which may be inside the IP network, IP destinations, or
   outside it, PSTN destinations. TBGP's operation is independent of any
   VoIP call signaling protocols (H.323, SIP, MGCP, ...), but it can
   serve as the call routing protocol for any of these signaling
   protocols. TBGP is based on the Border Gateway Protocol 4 (BGP-4).
 

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-iptel-glp-tbgp-00.txt

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-iptel-glp-tbgp-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-iptel-glp-tbgp-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<19990226154216.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-glp-tbgp-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-glp-tbgp-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19990226154216.I-D@ietf.org>

--OtherAccess--

--NextPart--



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  1 23:49:01 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA25101
	for <iptel-archive@odin.ietf.org>; Mon, 1 Mar 1999 23:48:57 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3564F7FB8; Mon,  1 Mar 1999 23:37:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 898E07FC9; Mon,  1 Mar 1999 23:37:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <008101be6466$390abf40$b18cd0cf@jtk-laptop>
From: "John Kristoff" <jkristof@interaccess.com>
To: <iptel@lists.research.bell-labs.com>
Subject: comp.dcom.voice-over-ip vote now in progress Last Call  for Votes
Date: Mon, 1 Mar 1999 22:36:20 -0600
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.5
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

The vote to decide whether to create a new "comp.dcom.voice-over-ip"
newsgroup is now in progress.  Anyone with an opinion is entitled and
encouraged to vote; there are no special qualifications, it's not difficult
and you don't have to be a computer expert.

If you have an opinion on this issue, please find the article with "Subject:
2nd CFV: comp.dcom.voice-over-ip" in comp.dcom.telecom.tech,
comp.dcom.sys.bay-networks, comp.dcom.sys.cisco, comp.dcom.sys.nortel,
comp.protocols.tcp-ip, news.announce.newgroups, news.groups or
alt.dcom.telecom.ip.

Then follow the directions in that article to cast your vote.

If you have trouble finding that article, e-mail me and I'll gladly tell you
how to get a ballot by e-mail.

ONLY votes which use the official ballot and follow the directions count.

In order to be created, a new group needs at least 100 more YES votes than
NO votes and at least 2/3 of the votes cast must be YES votes.




---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar  3 00:24:52 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA13620
	for <iptel-archive@odin.ietf.org>; Wed, 3 Mar 1999 00:24:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id CF0FB7FB7; Wed,  3 Mar 1999 00:13:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2617C7FB8; Wed,  3 Mar 1999 00:13:24 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903020006.TAA06197@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: iptel@lists.research.bell-labs.com
From: Internet-Drafts@ietf.org
Reply-To: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-iptel-glp-tbgp-00.txt
Date: Mon, 01 Mar 1999 19:06:43 -0500
X-Rcpt-To: aalok@bisquare.com
X-UIDL: ca5b4e52d587095b4e33b2a32569d696
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the IP Telephony Working Group of the IETF.

	Title		: The IP Telephony Border Gateway Protocol Architecture
	Author(s)	: D. Hampton, D. Oran, H. Salama, D. Shah
	Filename	: draft-ietf-iptel-glp-tbgp-00.txt
	Pages		: 12
	Date		: 26-Feb-99
	
   This document presents the architecture of the IP Telephony Border
   Gateway Protocol (TBGP). TBGP is an inter-domain protocol, for
   routing voice calls through the IP network towards their
   destinations, which may be inside the IP network, IP destinations, or
   outside it, PSTN destinations. TBGP's operation is independent of any
   VoIP call signaling protocols (H.323, SIP, MGCP, ...), but it can
   serve as the call routing protocol for any of these signaling
   protocols. TBGP is based on the Border Gateway Protocol 4 (BGP-4).
 

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-iptel-glp-tbgp-00.txt

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-iptel-glp-tbgp-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-iptel-glp-tbgp-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<19990226154216.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-glp-tbgp-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-glp-tbgp-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19990226154216.I-D@ietf.org>

--OtherAccess--

--NextPart--





---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar  4 02:11:39 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA23079
	for <iptel-archive@odin.ietf.org>; Thu, 4 Mar 1999 02:11:37 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DDFC37FB7; Thu,  4 Mar 1999 02:01:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3E7647FB8; Thu,  4 Mar 1999 02:01:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BE663A.B6B0A9C0@hss056.hss.hns.com>
From: Alok Mittal <amittal@hss.hns.com>
To: "'hampton@cisco.com'" <hampton@cisco.com>
Cc: list iptel <iptel@lists.research.bell-labs.com>
Subject: RE: glp draft based on BGP
Date: Thu, 4 Mar 1999 12:29:55 -0000
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hi

 Some comments:

1. IT_SIG_PROTO should specify H.323 instead of Q.931/RAS -- a version
    number should also be supported here.
2. Some fields present in the SCSP based draft, such as lifetime and support 
    for range of numbers (for E.164) are useful -- Should we include those?
3. One drawback that I find in this approach is that the policy for forwarding 
    route information is present at every TBGP speaker (section 7.3) and hence
    when the information arrives at a particular speaker, it is already subjected 
    to a cascade of policies, which it doesn't even know. Support for exchange
    of full database (as in SCSP based draft) seems to be useful, if the policy
    leading to decision, should be applied independently in each domain. Of
    course, changes are exchanged incrementally so the network wouldn't be flooded.
4. In absence of a precise semantic definition of the IT_PATH_COST attribute,
    it seems premature to define an aggregation rule for it (section 7.4) -- we 
    should define what it really means, and then the aggregation rule would follow.
 
Best Regards
Alok

-----------------------------------------------------------------
Alok Mittal	
Technical Leader,	Hughes Software Systems
Prestige Opal, 146 Infantry Road, Bangalore - 560001
Tel: +91-80-2286390,91,92	Fax: +91-80-2286393
-----------------------------------------------------------------

-----Original Message-----
From:	Jonathan Rosenberg [SMTP:jdrosen@dnrc.bell-labs.com]
Sent:	Friday, February 26, 1999 8:40 PM
To:	list iptel
Subject:	New draft on glp

members of iptel,

Dave Hampton, Dave Oran, Hussein Salama, and Dhaval Shah have just
submitted a draft on using BGP4 as a basis for the GLP. Until the draft
appears in the archives, it is available from our webpage:

http://www.bell-labs.com/mailing-lists/iptel/

Please note that Cisco has indicated it may have IPR related to this
draft.

It seems we now have two competing proposals (one based on SCSP, and one
on BGP) for GLP. I'd like to begin some discussion on the lists here
debating the relative merits of the two approaches. Fire away....

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar  4 09:55:15 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA04754
	for <iptel-archive@odin.ietf.org>; Thu, 4 Mar 1999 09:55:15 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 54B437FD0; Thu,  4 Mar 1999 09:45:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id AC9D37FD1; Thu,  4 Mar 1999 09:45:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <4.1.19990303181622.009db100@pop3.metatel.com>
X-Sender: scott.petrack@metatel.com@pop3.metatel.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1 
Date: Wed, 03 Mar 1999 18:20:20 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        Alok Mittal <amittal@hss.hns.com>
From: Scott Petrack <scott.petrack@metatel.com>
Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
Cc: "'Matt Squire'" <msquire@nortelnetworks.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
In-Reply-To: <36D56F8D.C1CCB388@dnrc.bell-labs.com>
References: <01BE5F33.97C7FF30@hss056.hss.hns.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 10:43 AM 2/25/99 -0500, Jonathan Rosenberg wrote:
>Alok Mittal wrote:
>> 
>> >I think the real question is whether a prefix is sufficient to specify the
>> >likely sets of target telephone numbers.  In the US, I think that a prefix
>> >is sufficient.  The proximity of area/city codes does not mean the
>> >proximity of areas/cities, which seems a likely criteria for a gateway
>> >serving 'range' but not a 'prefix'.  Is this common of phone numbering
>> >systems throughout the world?  If so, then I agree we should go with
>> >prefixes and not ranges.
>> 
>> In India, again there is no strong correlation between the two, though in 
>many
>> cases such a property may be satisfied.
>
>The problem is that there is really no other scalable way to do it. In
>an environment where the correlation between prefixes and geographical
>regions (and more specifically, cost) is lost, there really is no easy
>way to specify phone number ranges.
>

Is an MGCP-style "digit map" string a practical matching pattern which is
more general than just a prefix? (I guess that I can see the problem with
using a general regular expression....).



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar  4 14:07:54 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10958
	for <iptel-archive@odin.ietf.org>; Thu, 4 Mar 1999 14:07:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 37E467FB7; Thu,  4 Mar 1999 13:57:25 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9114F7FC8; Thu,  4 Mar 1999 13:57:24 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <5854E0C731A0D211B30400805FFED541599E46@MTL_EXCHANGE>
From: Youcef Dahmane <youcefd@EICON.COM>
To: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Date: Thu, 4 Mar 1999 13:52:31 -0500 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.0.1460.8)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

iptel-request@lists.research.bell-labs.com 


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar  4 15:21:02 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA11995
	for <iptel-archive@odin.ietf.org>; Thu, 4 Mar 1999 15:21:01 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 63DA47FC8; Thu,  4 Mar 1999 15:11:13 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B4FFB7FC9; Thu,  4 Mar 1999 15:11:12 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36DEE85A.25189194@cisco.com>
Date: Thu, 04 Mar 1999 12:08:58 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Alok Mittal <amittal@hss.hns.com>
Cc: list iptel <iptel@lists.research.bell-labs.com>
Subject: Re: glp draft based on BGP
References: <01BE663A.B6B0A9C0@hss056.hss.hns.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit


Alok,

Thanks for your comments. Some responses below.

Hussein

Alok Mittal wrote:
> 
> Hi
> 
>  Some comments:
> 
> 1. IT_SIG_PROTO should specify H.323 instead of Q.931/RAS -- a version
>     number should also be supported here.

An administrative domain, AD1, may wish that all incoming H.323 calls
from neighboring domains be routing by its gatekeeper first. The
gatekeeper then directs the incoming call to the appropriate gateway or
proxy. In this case, AD1, specifies the address of a gatekeeper as the
next hop and RAS as the signaling protocol. Alternatively, AD1, may
specify an H.323 gateway or proxy as the next hop and Q.931 as the call
signaling protocol. In the first case the gatekeeper is involved in both
call routing and admission control. In the second case, the gatekeeper
does admission control only.

I don't see the need for including version numbers. They are negotiated
within the protocols themselves, and I don't think there is a danger of
backward incompatibility in any of the protocols we're considering.

> 2. Some fields present in the SCSP based draft, such as lifetime and support
>     for range of numbers (for E.164) are useful -- Should we include those?

Yes, lifetime could be useful, especially if cost (pricing) information,
which change over time, are included in an Update. However, each new
attribute or field poses additional challenges and hazards to route
aggregation.

Regarding the use of ranges instead of prefixes, as you mentioned in an
earlier message, prefixes take less space than ranges. In addition, I
think, with proper use of aggregates and more specific call routes,
prefixes will be able to achieve the same degree of flexibility ranges
provide.

> 3. One drawback that I find in this approach is that the policy for forwarding
>     route information is present at every TBGP speaker (section 7.3) and hence
>     when the information arrives at a particular speaker, it is already subjected
>     to a cascade of policies, which it doesn't even know. Support for exchange
>     of full database (as in SCSP based draft) seems to be useful, if the policy
>     leading to decision, should be applied independently in each domain. Of
>     course, changes are exchanged incrementally so the network wouldn't be flooded.

If an administrative domain happens to lie on a specific call route, the
TBGP speakers representing that domain definitely have the right to
enforce their own policies on that route before advertising it further.
If an administrative domain doesn't have any policies to enforce, that's
fine.

Even the SCSP-based proposal permits intermediate LSes to enforce their
own policies (Matt, please correct me if I am wrong?). In the
inter-domain case, intermediate LSes may modify attributes and may
perform aggregation. They're not just synchronizing their databases.

> 4. In absence of a precise semantic definition of the IT_PATH_COST attribute,
>     it seems premature to define an aggregation rule for it (section 7.4) -- we
>     should define what it really means, and then the aggregation rule would follow.

Good point.

> 
> Best Regards
> Alok
> 
> -----------------------------------------------------------------
> Alok Mittal
> Technical Leader,       Hughes Software Systems
> Prestige Opal, 146 Infantry Road, Bangalore - 560001
> Tel: +91-80-2286390,91,92       Fax: +91-80-2286393
> -----------------------------------------------------------------
> 
> -----Original Message-----
> From:   Jonathan Rosenberg [SMTP:jdrosen@dnrc.bell-labs.com]
> Sent:   Friday, February 26, 1999 8:40 PM
> To:     list iptel
> Subject:        New draft on glp
> 
> members of iptel,
> 
> Dave Hampton, Dave Oran, Hussein Salama, and Dhaval Shah have just
> submitted a draft on using BGP4 as a basis for the GLP. Until the draft
> appears in the archives, it is available from our webpage:
> 
> http://www.bell-labs.com/mailing-lists/iptel/
> 
> Please note that Cisco has indicated it may have IPR related to this
> draft.
> 
> It seems we now have two competing proposals (one based on SCSP, and one
> on BGP) for GLP. I'd like to begin some discussion on the lists here
> debating the relative merits of the two approaches. Fire away....
> 
> -Jonathan R.
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-2499

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar  4 17:15:42 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA14988
	for <iptel-archive@odin.ietf.org>; Thu, 4 Mar 1999 17:15:42 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 61B027FC9; Thu,  4 Mar 1999 17:06:06 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C49247FCA; Thu,  4 Mar 1999 17:06:05 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36DF0318.6E806C3E@dnrc.bell-labs.com>
Date: Thu, 04 Mar 1999 17:03:04 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: Alok Mittal <amittal@hss.hns.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: Re: glp draft based on BGP
References: <01BE663A.B6B0A9C0@hss056.hss.hns.com> <36DEE85A.25189194@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> > 1. IT_SIG_PROTO should specify H.323 instead of Q.931/RAS -- a version
> >     number should also be supported here.
> 
> An administrative domain, AD1, may wish that all incoming H.323 calls
> from neighboring domains be routing by its gatekeeper first. The
> gatekeeper then directs the incoming call to the appropriate gateway or
> proxy. In this case, AD1, specifies the address of a gatekeeper as the
> next hop and RAS as the signaling protocol. Alternatively, AD1, may
> specify an H.323 gateway or proxy as the next hop and Q.931 as the call
> signaling protocol. In the first case the gatekeeper is involved in both
> call routing and admission control. In the second case, the gatekeeper
> does admission control only.
> 
> I don't see the need for including version numbers. They are negotiated
> within the protocols themselves, and I don't think there is a danger of
> backward incompatibility in any of the protocols we're considering.

Thats true so long as the protocol remains backwards compatible. If it
doesn't, we can always assign a new signaling protocol name. In other
words, we can group compatible versions together into a single name, and
define new names for incompatible versions. So, should H.323v4 be not
backwards compatible (just as an example), we would define two names:
H323 and H323V4.

> 
> > 2. Some fields present in the SCSP based draft, such as lifetime and support
> >     for range of numbers (for E.164) are useful -- Should we include those?
> 
> Yes, lifetime could be useful, especially if cost (pricing) information,
> which change over time, are included in an Update. However, each new
> attribute or field poses additional challenges and hazards to route
> aggregation.
> 
> Regarding the use of ranges instead of prefixes, as you mentioned in an
> earlier message, prefixes take less space than ranges. In addition, I
> think, with proper use of aggregates and more specific call routes,
> prefixes will be able to achieve the same degree of flexibility ranges
> provide.

This alludes to an important point. It seems we have three possible ways
to represent number ranges: (1) a prefix, (2) a range, and (3) an MGCP
style digit map. When evaluating the choices, we need to consider the
utility of the approach (do we need ranges, or is a prefix sufficient?),
and the aggregatability of the approach. I suspect prefixes are easier
to aggregate than MGCP digit maps. The digit maps were not designed for
aggregation, but we will need to do it.

I'd really love to hear from some experts in PSTN numbering here as
well.


> 
> > 3. One drawback that I find in this approach is that the policy for forwarding
> >     route information is present at every TBGP speaker (section 7.3) and hence
> >     when the information arrives at a particular speaker, it is already subjected
> >     to a cascade of policies, which it doesn't even know. Support for exchange
> >     of full database (as in SCSP based draft) seems to be useful, if the policy
> >     leading to decision, should be applied independently in each domain. Of
> >     course, changes are exchanged incrementally so the network wouldn't be flooded.
> 
> If an administrative domain happens to lie on a specific call route, the
> TBGP speakers representing that domain definitely have the right to
> enforce their own policies on that route before advertising it further.
> If an administrative domain doesn't have any policies to enforce, that's
> fine.
> 
> Even the SCSP-based proposal permits intermediate LSes to enforce their
> own policies (Matt, please correct me if I am wrong?). In the
> inter-domain case, intermediate LSes may modify attributes and may
> perform aggregation. They're not just synchronizing their databases.

I agree that the SCSP proposal also allows the enforcement of policies,
and thus with that approach routing objects will be received that have
had lots of cascaded policies applied. Thats the nature of policy.

The advantage of the synchronization approach in SCSP is for an I-BGP
equivalent. All of the LS's in an AD can sync their databases, and use
any sort of interconnection topology. This is as opposed to BGP, which
uses full mesh, or specific things like communities (I believe). SCSP
would allow arbitrary connection graphs.



There are really two separate issues we need to discuss. The first is
BGP vs. SCSP as a framework, and the second is structure of the
attributes, which is likely to be the same between the two. So, I would
propose that we try and discuss these separately. I'd love to see mails
from people about the relative pros and cons of BGP vs. SCSP as a
framework, without regard to the attributes. We can carry separate
threads to argue the attribute issues. The agenda for the upcoming
meeting will follow this structure as well.

-jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar  4 17:48:37 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA16000
	for <iptel-archive@odin.ietf.org>; Thu, 4 Mar 1999 17:48:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5E94A7FCA; Thu,  4 Mar 1999 17:39:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A49FB7FCB; Thu,  4 Mar 1999 17:39:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990305093825.013d6ba4@pop.ozemail.com.au>
X-Sender: dclowes@pop.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Fri, 05 Mar 1999 09:38:25 +1100
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
Cc: Alok Mittal <amittal@hss.hns.com>,
        "'Matt Squire'" <msquire@nortelnetworks.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
In-Reply-To: <36D56F8D.C1CCB388@dnrc.bell-labs.com>
References: <01BE5F33.97C7FF30@hss056.hss.hns.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 10:43 1999-02-25 -0500, Jonathan Rosenberg wrote:
>> 
>> >I think the real question is whether a prefix is sufficient to specify the
>> >likely sets of target telephone numbers.  In the US, I think that a prefix
>> >is sufficient.  The proximity of area/city codes does not mean the
>> >proximity of areas/cities, which seems a likely criteria for a gateway
>> >serving 'range' but not a 'prefix'.  Is this common of phone numbering
>> >systems throughout the world?  If so, then I agree we should go with
>> >prefixes and not ranges.
>> 
>> In India, again there is no strong correlation between the two, though
in many
>> cases such a property may be satisfied.
>
>The problem is that there is really no other scalable way to do it. In
>an environment where the correlation between prefixes and geographical
>regions (and more specifically, cost) is lost, there really is no easy
>way to specify phone number ranges.

Number prefixes is the way to go. However, there are a few things to note.

Cost is a significant factor to be considered, and that can vary in
pseudo-absurd ways. Number ranges can have holes in them, where cost varies
significantly. Take for example the US number range, prefixed by "1", where
"1800" and "1900" are significantly different to the broader range in which
they are contained.

Similar characteristics exist even within an area code where, in my
experience and opinion, strange thing can occur. Here, I am assuming that a
gateway is connected to a regular phone line, at a particular point in
space. Within one are code, calls to some exchanges are free, or fixed
price, while calls to other exchanges are toll calls. Similarly, sometimes
you can dial the seven digit number with or without the 1-area-code prefix,
while sometimes you must, and other times you must not dial it, when you
are in that area code.

Consider, for example, a corporation that has a directory number (say
1-616-555-1212) and have local operations in NYC, LAX and SFO. They want to
carry calls to that number on their private network, from their gateways in
those cities, and offer it for free. The gateways in SFO will offer "1-415"
for a price and "1-616-555-1212" for free, while the Michigan gateways
offer "1-616" for a price and "1-616-555-1212" for free.

A similar situation exists for 1-800 numbers, where the callee pays by the
distance of the caller from the destination. While all gateways will be
able to access the 1-800 number, there will be a preference for a gateway
closer to the point of termination. What should happen is that all the
gateways offer "1-800", while the gateway closest to the destination offers
"1-800-555-5000".

It can be suggested, that part of the requirement is to match the longest
prefix, at the lowest cost.

>> 
>> >Its not clear to me what you mean by cost. Cost, to most routing
protocols,
>> >is the number of hops.  But we don't really care about the number of
>> >signalling hops.  A couple weeks ago, there was something on the mailing
>> >list which enumerated many variables affecting the cost/price of an
>> >internet phone call.  So the meaning of cost is ambiguous at best.
>> >
>> >If you have something specific that you mean by cost then let me know, but
>> >my view is that we can't do anything with real cost, and that we should
>> >stick to a black-box structure to express business practices and
>> >relationships.

One mas price is another man's cost.
 
>> Yes, I agree that this is a rather vague thing. To think of a cost
metric at an
>> application level, we could use the number of administrative domains
through
>> which the shortest path to destination lies (from this gateway/signaling
node).
>> This is based on an assumption that every administrative domain would
charge
>> *some* fees for routing packets through its gateways. But this will lead
to the
>> problem of propagating this information, and it also makes a simplifying
>> assumption that cost of passing through each administrative domain is more
>> or less the same (or at least the difference is insignificant as
compared to
>> passing through an additional domain). Will think more and get back to you
>> if I have better suggestions!
>
>I think we still need a pricing structure, opaque to GLP. We need a hop
>count for TTL purposes already, and we have also discussed a capacity
>metric. This would seem to provide a reasonably large enough set of
>information for an LS to decide what to do with the routing object,
>based on whatever total measure of "cost" it decides to define.
>
>If there is additional cost for transit service, shouldn't this be
>wrapped into the pricing attribute? An ITAD would simply modify the
>pricing to include its own transit costs before propagating the
>attribute, if it so chooses.

What cost is, depends on who you are. If you are the egress gateway
operator, your costs would include things like fixed costs (interest,
depreciation, rental of premises, lines and Internet connection) and
variable costs would include things like calls into the PSTN, bandwidth
charges on your Internet connection. The egress gateway operator would then
set a price, to cover costs and make a profit (unless they are buying
market share ...).

For the ingress gateway operator, or private caller, their costs include
the price charged by the egress gateway operator, their own equipment and
rental costs, and their bandwidth charges. Plus tax.

When I go looking for a gateway to terminate my call, all of my costs are
fixed, except the price charged by the egress gateway operator. Unless I
pay hugely different backbone bandwidth charges to different gateways,
egress gateway price is going to be the determinant of cost, and dwarf any
other cost variables.

The other issues come down to reachability, capacity, and QoS. In a "well
engineered" network, these will not matter. If we assume that they do
matter, then it is something that the call originator must determine.

I think that pricing must *not* be opaque to GLP, because it must aggregate
routes based on cost. Gateways will act as secondary alternates (and
tertiary, and quaternary, ...). Typically, they will reflect increasing
prices by their operators, because of increasing geographic distance
between the gateway and callee being reflected in increasing PSTN charges.

An LS outside your network can never know the reachability and QoS for a
gateway that is also outside your network. How can it possibly know that
your router is misconfigured and that gateway is unreachable from your
network, or that the link over the shortest path is overloaded - it can't.
Only the caller can ascertain the QoS on the link, and decide it doesn't
meet its requirements.

As for gateway capacity, it could be useful in weighting the decision, but
it will not be absolute. If there will be other LS giving out the same
information, an LS can't know about traffic from other LS. Since calls
don't all have the same statistical characteristics, load will not be
balanced by weighted round robin.

Again, in the end, the caller must make the determination. Take a list of
possible gateways, ordered by price (and then capacity), add your own
costs, go down the list trying each one. If you're not prepared to pay the
price/cost combination give up. If you are, check the QoS (ping, RSVP,
whatever), if that passes, try to connect, etc..

Douglas


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar  4 21:41:26 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA24971
	for <iptel-archive@odin.ietf.org>; Thu, 4 Mar 1999 21:41:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BBD837FCB; Thu,  4 Mar 1999 21:31:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2CDBE7FCC; Thu,  4 Mar 1999 21:31:28 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990304213006.00a96e70@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 04 Mar 1999 21:30:08 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: glp draft based on BGP
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>> 3. One drawback that I find in this approach is that the policy for
forwarding
>>     route information is present at every TBGP speaker (section 7.3) and
hence
>>     when the information arrives at a particular speaker, it is already
subjected
>>     to a cascade of policies, which it doesn't even know. Support for
exchange
>>     of full database (as in SCSP based draft) seems to be useful, if the
policy
>>     leading to decision, should be applied independently in each domain. Of
>>     course, changes are exchanged incrementally so the network wouldn't
be flooded.
>
>If an administrative domain happens to lie on a specific call route, the
>TBGP speakers representing that domain definitely have the right to
>enforce their own policies on that route before advertising it further.
>If an administrative domain doesn't have any policies to enforce, that's
>fine.
>
>Even the SCSP-based proposal permits intermediate LSes to enforce their
>own policies (Matt, please correct me if I am wrong?). In the
>inter-domain case, intermediate LSes may modify attributes and may
>perform aggregation. They're not just synchronizing their databases.
>

Right.  The SCSP/BGP question addresses probably the simplest part of the
GLP problem, how to keep peer LSs synchronized.  It may turn out to be a
controversial issue, but at least the problem is simple, and could be
solved by a number of other protocols as well.  There are some subtle
underlying issues to the debate, but the basic problem is just a two-party
database synchronization.  

IMHO, the tougher questions deal with the information being synchronized,
what is it, how does it get represented, how does it get aggregated, etc.
These are issues that both approaches must eventually address better than
they currently do.  

- Matt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar  5 10:33:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA15249
	for <iptel-archive@odin.ietf.org>; Fri, 5 Mar 1999 10:33:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6C6797FB7; Fri,  5 Mar 1999 10:21:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D1D067FC8; Fri,  5 Mar 1999 10:21:34 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36DFF5CA.3519848E@dnrc.bell-labs.com>
Date: Fri, 05 Mar 1999 10:18:34 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: list iptel <iptel@lists.research.bell-labs.com>
Subject: Drafts for consideration
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Folks,

As our meeting approaches, I wanted to make sure it was clear what
drafts were to be discussed. The drafts we will be discussing are:

CPL efforts:
draft-ietf-iptel-cpl-00
draft-ietf-iptel-sig-reg-payload-00

GLP efforts:
draft-ietf-iptel-glp-00
draft-ietf-iptel-glp-tbgp-00


Attendees should have also read the two framework documents, although
they won't be discussed:

draft-ietf-iptel-gwloc-framework-02
draft-ietf-iptel-cpl-requirements-00

Thanks,
Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar  5 16:10:35 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA22691
	for <iptel-archive@odin.ietf.org>; Fri, 5 Mar 1999 16:10:34 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0719A7FC8; Fri,  5 Mar 1999 15:59:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 618D27FC9; Fri,  5 Mar 1999 15:59:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Scott Petrack" <scott.petrack@metatel.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Alok Mittal" <amittal@hss.hns.com>
Cc: "'Matt Squire'" <msquire@nortelnetworks.com>,
        <iptel@lists.research.bell-labs.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Date: Fri, 5 Mar 1999 15:58:27 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFCEDMEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
In-Reply-To: <4.1.19990303181622.009db100@pop3.metatel.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

The choice of prefixes and the destination descriptor was guided
primarily by two considerations:
 - The ability to use longest-match easily as a route discriminator.
Other destination descriptors don't necessarily have well defined
partial orders.
- The ability to do fairly strightforward route aggregation (of course
modulo the problems of multiple metrics and the lack of a totaly order
among them).

We seriously considered using regexps (some existing Cisco products do
just that) because of their power and generality, but it's quite
difficult to achieve the two characteristics above using them.

Dave.

> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Scott
> Petrack
> Sent: Wednesday, March 03, 1999 6:20 PM
> To: Jonathan Rosenberg; Alok Mittal
> Cc: 'Matt Squire'; 'iptel@lists.research.bell-labs.com'
> Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
>
>
> At 10:43 AM 2/25/99 -0500, Jonathan Rosenberg wrote:
> >Alok Mittal wrote:
> >>
> >> >I think the real question is whether a prefix is
> sufficient to specify the
> >> >likely sets of target telephone numbers.  In the US, I
> think that a prefix
> >> >is sufficient.  The proximity of area/city codes does not mean the
> >> >proximity of areas/cities, which seems a likely criteria
> for a gateway
> >> >serving 'range' but not a 'prefix'.  Is this common of
> phone numbering
> >> >systems throughout the world?  If so, then I agree we
> should go with
> >> >prefixes and not ranges.
> >>
> >> In India, again there is no strong correlation between the
> two, though in
> >many
> >> cases such a property may be satisfied.
> >
> >The problem is that there is really no other scalable way to
> do it. In
> >an environment where the correlation between prefixes and
> geographical
> >regions (and more specifically, cost) is lost, there really
> is no easy
> >way to specify phone number ranges.
> >
>
> Is an MGCP-style "digit map" string a practical matching
> pattern which is
> more general than just a prefix? (I guess that I can see the
> problem with
> using a general regular expression....).
>
>
>
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar  5 16:25:45 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA23343
	for <iptel-archive@odin.ietf.org>; Fri, 5 Mar 1999 16:25:45 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E08F87FCA; Fri,  5 Mar 1999 16:18:11 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 363C37FCE; Fri,  5 Mar 1999 16:18:10 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E0495C.F9933385@dnrc.bell-labs.com>
Date: Fri, 05 Mar 1999 16:15:08 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: "David R. Oran" <oran@cisco.com>
Cc: Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        "'Matt Squire'" <msquire@nortelnetworks.com>,
        iptel@lists.research.bell-labs.com
Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
References: <NBBBIFCAKKOPNMMKHHEFCEDMEEAA.oran@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

David R. Oran wrote:
> 
> The choice of prefixes and the destination descriptor was guided
> primarily by two considerations:
>  - The ability to use longest-match easily as a route discriminator.
> Other destination descriptors don't necessarily have well defined
> partial orders.

Good point. 

> - The ability to do fairly strightforward route aggregation (of course
> modulo the problems of multiple metrics and the lack of a totaly order
> among them).

Actually even with prefixes, it seems hard. Basically - are the prefixes
expressed in terms of bits, or in terms of decimal numbers. Decimal is
more logical, but then to aggregate to +1-555, you must collect
+1-555-1, +1-555-2, +1-555-3, etc., which is 10 routes. If its done
bitwise, it only requires 2.

Thoughts?

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar  5 16:38:45 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA23691
	for <iptel-archive@odin.ietf.org>; Fri, 5 Mar 1999 16:38:45 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 97E727FC9; Fri,  5 Mar 1999 16:11:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0502B7FCA; Fri,  5 Mar 1999 16:11:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Matt Squire" <msquire@nortelnetworks.com>,
        "IPTEL IETF Mailing List" <iptel@lists.research.bell-labs.com>
Subject: RE: glp draft based on BGP
Date: Fri, 5 Mar 1999 16:08:59 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFKEDMEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
In-Reply-To: <3.0.32.19990304213006.00a96e70@bl-mail1.corpeast.baynetworks.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

> Right.  The SCSP/BGP question addresses probably the simplest
> part of the
> GLP problem, how to keep peer LSs synchronized.  It may turn
> out to be a
> controversial issue, but at least the problem is simple, and could be
> solved by a number of other protocols as well.  There are some subtle
> underlying issues to the debate, but the basic problem is
> just a two-party
> database synchronization.
>
I guess routing has finally made the big time, when people consider the
routing convergence, scalability, policy satisfiability, and robustness
problems "simple". :-)

Oh, and yes, what makes you think this is a two-party problem? If you
can convince me that it is, I'll retract what I said above and stop
worrying about all the usual things I worry about when designing a
routing scheme.

> IMHO, the tougher questions deal with the information being
> synchronized,
> what is it, how does it get represented, how does it get
> aggregated, etc.
> These are issues that both approaches must eventually address
> better than
> they currently do.
>
While I agree in principle, there are certain mathematical/physical
constraints which will so constrain the solution space that we'd better
face that up front. For example, the general multi-metric optimization
problem has exponential time complexity. Information that gets
aggregated needs to have at least a globally-defined partial order. Etc.
Etc.

> - Matt
>
This is going to be fun!


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar  5 17:22:41 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA25279
	for <iptel-archive@odin.ietf.org>; Fri, 5 Mar 1999 17:22:40 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8622B7FCF; Fri,  5 Mar 1999 17:11:36 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 73C6F7FCC; Fri,  5 Mar 1999 17:11:31 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903052209.RAA24423@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: iptel@lists.research.bell-labs.com
From: Internet-Drafts@ietf.org
Reply-To: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-iptel-sip-reg-payload-00.txt,.ps
Date: Fri, 05 Mar 1999 17:09:05 -0500
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the IP Telephony Working Group of the IETF.

	Title		: Transporting User Control Information in 
                          SIP REGISTER Payloads
	Author(s)	: J. Lennox, H. Schulzrinne
	Filename	: draft-ietf-iptel-sip-reg-payload-00.txt,.ps
	Pages		: 9
	Date		: 04-Mar-99
	
   Several newly developed languages and interfaces, such as the CPL and
   SIP CGI, allow users or administrators to specify how Internet
   telephony servers should process calls. There needs to be a method of
   transporting scripts for such languages between a client and a
   server. This document proposes using the payload of SIP REGISTER
   messages, and their responses, as one method to transport them.


A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-iptel-sip-reg-payload-00.txt

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-iptel-sip-reg-payload-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-iptel-sip-reg-payload-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<19990304092749.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-sip-reg-payload-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-sip-reg-payload-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19990304092749.I-D@ietf.org>

--OtherAccess--

--NextPart--



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar  5 17:39:27 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA25907
	for <iptel-archive@odin.ietf.org>; Fri, 5 Mar 1999 17:39:26 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id EFFD27FCD; Fri,  5 Mar 1999 17:11:33 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6E00B7FCE; Fri,  5 Mar 1999 17:11:31 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903052209.RAA24412@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: iptel@lists.research.bell-labs.com
From: Internet-Drafts@ietf.org
Reply-To: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-iptel-cpl-00.txt,.ps
Date: Fri, 05 Mar 1999 17:09:00 -0500
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the IP Telephony Working Group of the IETF.

	Title		: CPL: A Language for User Control of Internet 
                          Telephony Services
	Author(s)	: J. Lennox, H. Schulzrinne
	Filename	: draft-ietf-iptel-cpl-00.txt,.ps
	Pages		: 23
	Date		: 04-Mar-99
	
   The Call Processing Language (CPL) is a language that can be used to
   describe and control Internet telephony services. It is designed to
   be implementable on either network servers or user agent servers. It
   is meant to be simple, extensible, easily edited by graphical
   clients, and independent of operating system or signalling protocol.
   It is suitable for running on a server where users may not be allowed
   to execute arbitrary programs, as it has no variables, loops, or
   ability to run external programs.
 
   This document is a product of the IP Telephony (IPTEL) working group
   of the Internet Engineering Task Force. Comments are solicited and
   should be addressed to the working group's mailing list at
   iptel@lists.research.bell-labs.com and/or the authors.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-iptel-cpl-00.txt

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-iptel-cpl-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-iptel-cpl-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<19990304092449.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-cpl-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-cpl-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19990304092449.I-D@ietf.org>

--OtherAccess--

--NextPart--



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar  5 22:11:10 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA15147
	for <iptel-archive@odin.ietf.org>; Fri, 5 Mar 1999 22:11:10 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 985837FCC; Fri,  5 Mar 1999 22:01:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F0A787FCE; Fri,  5 Mar 1999 22:01:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <003001be677d$657cc040$9c4566ca@xmzhao>
From: "Xiaoming Zhao" <xmzhao@990.net>
To: <iptel@lists.research.bell-labs.com>
Subject: A question about CPL
Date: Sat, 6 Mar 1999 10:59:47 +0800
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_002D_01BE67C0.7322E100"
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.5
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.

------=_NextPart_000_002D_01BE67C0.7322E100
Content-Type: text/plain;
	charset="gb2312"
Content-Transfer-Encoding: quoted-printable

Hi,
  I'm a freshman on IP telephony, especially CPL. After I read the CPL =
draft, I have a question.

What's the relationship between several scripts for one user?
    For example, Jones is an IPtel user. Of course a script for what =
actions should be taken when calls for Jones come is needed. However, =
another script used when Jones calls out for someone else is also =
needed. How can these two scripts work correctly together?

 =20
--------------------------------------------
Xiaoming Zhao.
xmzhao@990.net

Institute of Communications Engineering.=20
210016, Nanjing, Jiangsu, P.R. of China.
--------------------------------------------

------=_NextPart_000_002D_01BE67C0.7322E100
Content-Type: text/html;
	charset="gb2312"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">
<HTML>
<HEAD>

<META content=3Dtext/html;charset=3Dgb2312 http-equiv=3DContent-Type>
<META content=3D'"MSHTML 4.72.3110.7"' name=3DGENERATOR>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT color=3D#000000 size=3D2>Hi,</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT><FONT size=3D2>&nbsp; I'm a =
freshman on IP=20
telephony, especially CPL. After I read the CPL draft, I have a=20
question.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>What's the relationship between several scripts for =
one=20
user?</FONT></DIV>
<DIV><FONT size=3D2></FONT><FONT color=3D#000000 =
size=3D2>&nbsp;&nbsp;&nbsp; For=20
example, Jones is an IPtel user. Of course a script for what actions =
should be=20
taken when calls for Jones come is needed. However, another script used =
when=20
Jones calls out for someone else is also needed. How can these two =
scripts work=20
correctly together?</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#000000 size=3D2>&nbsp; </FONT></DIV>
<DIV><FONT color=3D#000000=20
size=3D2>--------------------------------------------<BR>Xiaoming =
Zhao.<BR><A=20
href=3D"mailto:xmzhao@990.net">xmzhao@990.net</A></FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#000000 size=3D2>Institute of Communications =
Engineering.=20
<BR>210016, Nanjing, Jiangsu, P.R. of=20
China.<BR>--------------------------------------------</FONT></DIV></BODY=
></HTML>

------=_NextPart_000_002D_01BE67C0.7322E100--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar  5 23:56:20 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA20959
	for <iptel-archive@odin.ietf.org>; Fri, 5 Mar 1999 23:56:19 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id CBFF87FD2; Fri,  5 Mar 1999 23:45:38 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 932F97FD0; Fri,  5 Mar 1999 23:45:32 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903052209.RAA24412@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: iptel@lists.research.bell-labs.com
From: Internet-Drafts@ietf.org
Reply-To: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-iptel-cpl-00.txt,.ps
Date: Fri, 05 Mar 1999 17:09:00 -0500
X-Rcpt-To: aalok@bisquare.com
X-UIDL: ef9fc48843d067b9231fb49246c70e8d
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the IP Telephony Working Group of the IETF.

	Title		: CPL: A Language for User Control of Internet 
                          Telephony Services
	Author(s)	: J. Lennox, H. Schulzrinne
	Filename	: draft-ietf-iptel-cpl-00.txt,.ps
	Pages		: 23
	Date		: 04-Mar-99
	
   The Call Processing Language (CPL) is a language that can be used to
   describe and control Internet telephony services. It is designed to
   be implementable on either network servers or user agent servers. It
   is meant to be simple, extensible, easily edited by graphical
   clients, and independent of operating system or signalling protocol.
   It is suitable for running on a server where users may not be allowed
   to execute arbitrary programs, as it has no variables, loops, or
   ability to run external programs.
 
   This document is a product of the IP Telephony (IPTEL) working group
   of the Internet Engineering Task Force. Comments are solicited and
   should be addressed to the working group's mailing list at
   iptel@lists.research.bell-labs.com and/or the authors.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-iptel-cpl-00.txt

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-iptel-cpl-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-iptel-cpl-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<19990304092449.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-cpl-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-cpl-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19990304092449.I-D@ietf.org>

--OtherAccess--

--NextPart--





---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar  6 00:14:25 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA21460
	for <iptel-archive@odin.ietf.org>; Sat, 6 Mar 1999 00:14:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 42AE37FCE; Fri,  5 Mar 1999 23:45:27 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 90C6F7FD1; Fri,  5 Mar 1999 23:45:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903052209.RAA24423@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: iptel@lists.research.bell-labs.com
From: Internet-Drafts@ietf.org
Reply-To: Internet-Drafts@ietf.org
Subject: I-D ACTION:draft-ietf-iptel-sip-reg-payload-00.txt,.ps
Date: Fri, 05 Mar 1999 17:09:05 -0500
X-Rcpt-To: aalok@bisquare.com
X-UIDL: de646ac46213646dcf244d91b27b3bc2
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the IP Telephony Working Group of the IETF.

	Title		: Transporting User Control Information in 
                          SIP REGISTER Payloads
	Author(s)	: J. Lennox, H. Schulzrinne
	Filename	: draft-ietf-iptel-sip-reg-payload-00.txt,.ps
	Pages		: 9
	Date		: 04-Mar-99
	
   Several newly developed languages and interfaces, such as the CPL and
   SIP CGI, allow users or administrators to specify how Internet
   telephony servers should process calls. There needs to be a method of
   transporting scripts for such languages between a client and a
   server. This document proposes using the payload of SIP REGISTER
   messages, and their responses, as one method to transport them.


A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-iptel-sip-reg-payload-00.txt

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-iptel-sip-reg-payload-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-iptel-sip-reg-payload-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<19990304092749.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-sip-reg-payload-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-sip-reg-payload-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19990304092749.I-D@ietf.org>

--OtherAccess--

--NextPart--





---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 06:37:30 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA26394
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 06:37:29 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 427AB7FD3; Mon,  8 Mar 1999 06:27:21 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 996767FD4; Mon,  8 Mar 1999 06:27:20 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <6525672E.0037D5CD.00@sampark.hss.hns.com>
Date: Mon, 8 Mar 1999 16:07:25 +0530
Subject: Syntax of h.323, version 1 and version 2 endpoints.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi All,
    I have a question regarding the syntax of version 1 and 2, h.323
endpoints.

For h.225.0 call signalling messages, ASN syntax given in ITU-T h.225.0
specifications includes only UUIE for each message. This has an extension
marker to show newly added fields in version 2. Syntax does not say
anything about other information elements of the message. Does that mean
that both version 1 and 2 will have all other IEs, as given in Q.931
message details, in section 7.3? Can we assume that both version 1 and
version 2 endpoints have all the IEs same, other than UUIE?

Any help or pointers to more information will be highly appreciated.

Thanks,
Ritu Gupta.
_____________________________________________________________________
Ritu Gupta                               |       email: rtgupta@hss.hns.com
Senior Software Engineer,          |                   rtgupta@hotmail.com
Hughes Software Systems,           |       phone: 91-80-2286390/91/92(W)
Banglore, Karnataka                |                    91-80-3374075(H)
INDIA.                                   |       Fax    : 91-80-2286393
_____________________________________________________________________
Restlessness and discontent are the first necessities of progress.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 08:40:08 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA28193
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 08:40:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 33A1B7FD4; Mon,  8 Mar 1999 08:29:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 83A4C7FD5; Mon,  8 Mar 1999 08:29:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "'Matt Squire'" <msquire@nortelnetworks.com>,
        <iptel@lists.research.bell-labs.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Date: Mon, 8 Mar 1999 08:27:07 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFKEFEEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
In-Reply-To: <36E0495C.F9933385@dnrc.bell-labs.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

> > - The ability to do fairly strightforward route aggregation
> (of course
> > modulo the problems of multiple metrics and the lack of a
> totaly order
> > among them).
>
> Actually even with prefixes, it seems hard. Basically - are
> the prefixes
> expressed in terms of bits, or in terms of decimal numbers. Decimal is
> more logical, but then to aggregate to +1-555, you must collect
> +1-555-1, +1-555-2, +1-555-3, etc., which is 10 routes. If its done
> bitwise, it only requires 2.
>
> Thoughts?
>
Well, if all phone numbers can be normalized to a "natural" length or
canonicalized into full 15 digit E.164 numbers, then you just convert
the number to binary and the bitwise scheme will work fine. Good idea.
One interesting question is wherhter to actually encode the numbers in
binary in the protocol. Whether this saves or wastes space depends on
whether you expect long or short prefixes. From a protocol debugging
point of view, sending strings wins.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 13:31:23 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA06216
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 13:31:22 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D17597FD5; Mon,  8 Mar 1999 13:21:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2DE1B7FD6; Mon,  8 Mar 1999 13:21:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Mon, 8 Mar 1999 12:21:22 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: "David R. Oran" <oran@cisco.com>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        Alok Mittal <amittal@hss.hns.com>,
        "'Matt Squire'" <msquire@nortelnetworks.com>,
        iptel@lists.research.bell-labs.com
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
In-Reply-To: <NBBBIFCAKKOPNMMKHHEFKEFEEEAA.oran@cisco.com>
Message-ID: <Pine.LNX.4.04.9903081218190.3058-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



On Mon, 8 Mar 1999, David R. Oran wrote:

> One interesting question is wherhter to actually encode the numbers in
> binary in the protocol. Whether this saves or wastes space depends on
> whether you expect long or short prefixes. From a protocol debugging
> point of view, sending strings wins.

Surely the processing time is more important than the space issue. I
realize that the packets don't need to be processed anywhere near line
speed, but a server that needs to process 1000s of phone calls will surely
benefit from not having to process the strings. 

Anyone who needs to debug can easily write a filter to translates the bits
into decimal.

Scott


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 15:18:47 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09077
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 15:18:47 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 668597FD7; Mon,  8 Mar 1999 15:09:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BCC197FD8; Mon,  8 Mar 1999 15:09:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Scott Petrack" <scott.petrack@metatel.com>
Cc: <iptel@lists.research.bell-labs.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Date: Mon, 8 Mar 1999 15:07:25 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFKEGKEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
In-Reply-To: <Pine.LNX.4.04.9903081218190.3058-100000@petrack.metatel.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

>
> > One interesting question is wherhter to actually encode the
> numbers in
> > binary in the protocol. Whether this saves or wastes space
> depends on
> > whether you expect long or short prefixes. From a protocol debugging
> > point of view, sending strings wins.
>
> Surely the processing time is more important than the space issue. I
> realize that the packets don't need to be processed anywhere near line
> speed, but a server that needs to process 1000s of phone
> calls will surely
> benefit from not having to process the strings.
>
These are in ROUTING messages, not SIGNALING messages. If we're sending
routing updates at anywhere near the signaling rate we're toast no
matter how things are encoded.

> Anyone who needs to debug can easily write a filter to
> translates the bits
> into decimal.
>
True, but this is a swamp I'm not going to wade into...


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 15:35:18 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09532
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 15:35:18 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B4F447FD6; Mon,  8 Mar 1999 15:01:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 013177FD7; Mon,  8 Mar 1999 15:01:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990308145943.007e32a0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Mon, 08 Mar 1999 14:59:44 -0500
To: "David R. Oran" <oran@cisco.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Cc: Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 08:27 AM 3/8/99 -0500, David R. Oran wrote:
>> > - The ability to do fairly strightforward route aggregation
>> (of course
>> > modulo the problems of multiple metrics and the lack of a
>> totaly order
>> > among them).
>>
>> Actually even with prefixes, it seems hard. Basically - are
>> the prefixes
>> expressed in terms of bits, or in terms of decimal numbers. Decimal is
>> more logical, but then to aggregate to +1-555, you must collect
>> +1-555-1, +1-555-2, +1-555-3, etc., which is 10 routes. If its done
>> bitwise, it only requires 2.
>>
>> Thoughts?
>>

Planning for general aggregation is probably wasteful.  I'm in area code
919 in North Carolina.  Area code 920 is in Wisconsin, and 918 is in
Oklahoma.  I don't see that I'd be aggregating these together unless I'm
aggregating all +1 numbers together.  I don't predict many prefixes of the
form +1-9 or +1-91 being of any use, nor do I think many like +1-919-9 are
of great value.  Writing code that goes and checks for possible aggregation
at arbitrary bit/byte boundaries seems like a waste of time.  Phone numbers
are heirachical and based on groups of digits, we should plan for
aggregation along the same lines.  

So I don't think we should worry about how to make aggregation easier when
we choose representations, we should optimize representations for processing.

- Matt



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 15:46:56 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09928
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 15:46:55 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 632187FDD; Mon,  8 Mar 1999 15:23:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B57617FDC; Mon,  8 Mar 1999 15:23:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Matt Squire" <msquire@nortelnetworks.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Date: Mon, 8 Mar 1999 15:19:45 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFCEGLEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
In-Reply-To: <3.0.32.19990308145943.007e32a0@bl-mail1.corpeast.baynetworks.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

The major reason I would argue for general aggregation is so that the
aggregator doesn't need to understand the semantics or assignment
politics of telephone numbers. We once had such an aggregation scheme in
the IP world (class-based addressing) and we didn't like where we wound
up.

Furthermore, semantics-blind aggregation is likely to wind up being
simpler to both implement and administer (IMHO). The performance is
likely to be as good or better as well, with data structures like
level-compressed tries and the like.

I'd make one exception to this principle. If there's a set of cases
where the aggregator can be told a purported "hole" in a set of
otherwise aggregatable prefixes isn't actually present somewhere, then
the aggregator can do a much better job of coalescing parts of the
address space. For example, if there's block of numbers like "+1-600"
which are known not to exist then telling the aggregator this will allow
for "+1-***" to be aggregated when otherwise it would be missed due to
the hole.

Dave.

> -----Original Message-----
> From: Matt Squire [mailto:msquire@nortelnetworks.com]
> Sent: Monday, March 08, 1999 3:00 PM
> To: David R. Oran; Jonathan Rosenberg
> Cc: Scott Petrack; Alok Mittal; iptel
> Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
>
>
> At 08:27 AM 3/8/99 -0500, David R. Oran wrote:
> >> > - The ability to do fairly strightforward route aggregation
> >> (of course
> >> > modulo the problems of multiple metrics and the lack of a
> >> totaly order
> >> > among them).
> >>
> >> Actually even with prefixes, it seems hard. Basically - are
> >> the prefixes
> >> expressed in terms of bits, or in terms of decimal
> numbers. Decimal is
> >> more logical, but then to aggregate to +1-555, you must collect
> >> +1-555-1, +1-555-2, +1-555-3, etc., which is 10 routes. If its done
> >> bitwise, it only requires 2.
> >>
> >> Thoughts?
> >>
>
> Planning for general aggregation is probably wasteful.  I'm
> in area code
> 919 in North Carolina.  Area code 920 is in Wisconsin, and 918 is in
> Oklahoma.  I don't see that I'd be aggregating these together
> unless I'm
> aggregating all +1 numbers together.  I don't predict many
> prefixes of the
> form +1-9 or +1-91 being of any use, nor do I think many like
> +1-919-9 are
> of great value.  Writing code that goes and checks for
> possible aggregation
> at arbitrary bit/byte boundaries seems like a waste of time.
> Phone numbers
> are heirachical and based on groups of digits, we should plan for
> aggregation along the same lines.
>
> So I don't think we should worry about how to make
> aggregation easier when
> we choose representations, we should optimize representations
> for processing.
>
> - Matt
>
>
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 16:39:12 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA11346
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 16:39:12 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 869507FD8; Mon,  8 Mar 1999 16:29:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DB45F7FD9; Mon,  8 Mar 1999 16:29:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990308162829.00abfa80@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Mon, 08 Mar 1999 16:28:30 -0500
To: "David R. Oran" <oran@cisco.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Cc: Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


I don't think we disagree (much).  My main point was that I think
aggregation is going to a royal pain, and that our choice of
representations shouldn't let 'easing aggregation algorithms' be a factor.
With multiple, onordered attributes, whether we use ascii or binary isn't
going make aggregation fun or easy, and probably won't have an impact on
the performance of aggregation.  

I also agree that when one writes an aggregation algorithm, it should be
general, ie apply at any reasonable boundry.  I'm not convinced that we
need to get into the bits, but I agree that any digit boundry is
reasonable.  A bit boundry would permit pairwise aggregation, which is
generally easier than performing aggregation over larger sets, but it just
seems so "wrong" to be looking at bits in phone numbers.  It might be
better, but it just seems weird.  

The efficiency of relative aggregation algorithms depends on many things,
but with the kinds of attributes that we're talking about, its not clear
(at least to me) that the cost of determining whether two objects are
capable of being aggregated is free.  As the cost of making this
determination goes up and the probability of its success goes down, it
becomes more attractive to shortcut the decision process by whether the
result is a reasonable prefix.  I'd consider this check to be an
optimization in a more general aggregation algorithm.  It could be an quick
way to cut-off potential wasted effort.  

But my main point was that aggregation isn't going to be optimal no matter
how we represent numbers, so let's try to optimize call routing decisions,
not aggregation algorithms.   

Didn't mean to get off on the wrong tangent. 

- Matt

At 03:19 PM 3/8/99 -0500, David R. Oran wrote:
>The major reason I would argue for general aggregation is so that the
>aggregator doesn't need to understand the semantics or assignment
>politics of telephone numbers. We once had such an aggregation scheme in
>the IP world (class-based addressing) and we didn't like where we wound
>up.
>
>Furthermore, semantics-blind aggregation is likely to wind up being
>simpler to both implement and administer (IMHO). The performance is
>likely to be as good or better as well, with data structures like
>level-compressed tries and the like.
>
>I'd make one exception to this principle. If there's a set of cases
>where the aggregator can be told a purported "hole" in a set of
>otherwise aggregatable prefixes isn't actually present somewhere, then
>the aggregator can do a much better job of coalescing parts of the
>address space. For example, if there's block of numbers like "+1-600"
>which are known not to exist then telling the aggregator this will allow
>for "+1-***" to be aggregated when otherwise it would be missed due to
>the hole.
>
>Dave.
>
>> -----Original Message-----
>> From: Matt Squire [mailto:msquire@nortelnetworks.com]
>> Sent: Monday, March 08, 1999 3:00 PM
>> To: David R. Oran; Jonathan Rosenberg
>> Cc: Scott Petrack; Alok Mittal; iptel
>> Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
>>
>>
>> At 08:27 AM 3/8/99 -0500, David R. Oran wrote:
>> >> > - The ability to do fairly strightforward route aggregation
>> >> (of course
>> >> > modulo the problems of multiple metrics and the lack of a
>> >> totaly order
>> >> > among them).
>> >>
>> >> Actually even with prefixes, it seems hard. Basically - are
>> >> the prefixes
>> >> expressed in terms of bits, or in terms of decimal
>> numbers. Decimal is
>> >> more logical, but then to aggregate to +1-555, you must collect
>> >> +1-555-1, +1-555-2, +1-555-3, etc., which is 10 routes. If its done
>> >> bitwise, it only requires 2.
>> >>
>> >> Thoughts?
>> >>
>>
>> Planning for general aggregation is probably wasteful.  I'm
>> in area code
>> 919 in North Carolina.  Area code 920 is in Wisconsin, and 918 is in
>> Oklahoma.  I don't see that I'd be aggregating these together
>> unless I'm
>> aggregating all +1 numbers together.  I don't predict many
>> prefixes of the
>> form +1-9 or +1-91 being of any use, nor do I think many like
>> +1-919-9 are
>> of great value.  Writing code that goes and checks for
>> possible aggregation
>> at arbitrary bit/byte boundaries seems like a waste of time.
>> Phone numbers
>> are heirachical and based on groups of digits, we should plan for
>> aggregation along the same lines.
>>
>> So I don't think we should worry about how to make
>> aggregation easier when
>> we choose representations, we should optimize representations
>> for processing.
>>
>> - Matt
>>
>>
>>
>
>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 16:51:18 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA11709
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 16:51:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2145D7FDA; Mon,  8 Mar 1999 16:41:26 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id ACA127FDF; Mon,  8 Mar 1999 16:41:24 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <010501be69aa$79336240$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "David R. Oran" <oran@cisco.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Matt Squire" <msquire@nortelnetworks.com>
Cc: "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: GLP, number portability and wireless subscribers
Date: Mon, 8 Mar 1999 13:26:56 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hello,

One issue needs to be considered is number portability. Number portability
(NP) is reality for wired POTS lines for some time now. Wireless Number
Portability will be a reality in a few years' time. The portable numbers
cannot
be routed by strictly following the hierarchy derived from the digits of
called
number. When the number of portable numbers is small compared to all
telephone numbers in the network, it can be handled on an exception bases
meaning route the call following the hierarchy then back track when hitting
a dead end. This is approach used in the first generation systems that
supports number portability. However, when the number of portable numbers
gets higher, the inefficiency in the approach becomes more pronounced.
Differentiation of portable numbers or the correct lookup of the
switch/gateway
where the called number resides needs to be done much earlier in the
SETUP forwarding path.

With the competition in local access market heating up and new area code
added using overlay instead of split, aggregation based on the number
hierarchy starts to make less and less sense. On top of this, with cell
phones
becomes cheaper and cheaper to have, and the wireless bandwidth grows
higher, to support roaming/mobility of wireless subscribers both within
wireless networks and across wireless and wireline based networks,
phone numbers will need to be handled more and more on a individual
bases.

To make GLP for IP based telephony networks future proof, the protocol
needs to support unaggregated  location information for telephone numbers.
Of course, this will increase the volume of traffic. But if the information
is
basically unaggregable, it is just the nature of the problem itself. It will
probably requires a different solution than it is currently outlined in the
two proposals that we currently have.

I am in the process of writing up a design that is used for propagating
location information among mobile telephony gateways. Although the
original design target is a IP based carrier class wireless telephone
network. It can be used to solve the generic gateway location problem
in a IP telephony network. If there is enough interest, I can outline the
design or send my document to the group for discussion.

Best regards,

- Greg



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 17:18:51 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA12688
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 17:18:50 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E28C47FD9; Mon,  8 Mar 1999 17:09:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 353797FDC; Mon,  8 Mar 1999 17:09:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Pat.Calhoun@Eng.Sun.COM (Patrice Calhoun)
Message-Id: <199903082207.OAA08912@hsmpka.eng.sun.com>
Date: Mon, 8 Mar 1999 14:00:19 -0800
To: "Greg Zhang" <gzhang@utstar.com>,
        "Matt Squire" <msquire@nortelnetworks.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "David R. Oran" <oran@cisco.com>
Cc: "iptel" <iptel@lists.research.bell-labs.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "Scott Petrack" <scott.petrack@metatel.com>
Reply-To: <Pat.Calhoun@Eng.Sun.COM>
Subject: Re: GLP, number portability and wireless subscribers
X-Mailer: Sun NetMail 2.2.5
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Greg,

Some of us have been addressing this problem, but not necessarily in the IPTel
space. Alot of work was done in TR45.6 (3G cellular systems), and much of this
was brought forth at the Mobile IP WG. Perhaps the work there applies here 
as well. I would be interested in seeing your requirements/architecture.

For a start you can look at draft-hiller-3gwireless-00.txt. For protocol 
information you can look at draft-calhoun-diameter-mobileip-03.txt. Again,
perhaps the problem is not exactly the same, but we may be able to re-use
some of the work already done.


PatC

>Hello,
>
>One issue needs to be considered is number portability. Number portability
>(NP) is reality for wired POTS lines for some time now. Wireless Number
>Portability will be a reality in a few years' time. The portable numbers
>cannot
>be routed by strictly following the hierarchy derived from the digits of
>called
>number. When the number of portable numbers is small compared to all
>telephone numbers in the network, it can be handled on an exception bases
>meaning route the call following the hierarchy then back track when hitting
>a dead end. This is approach used in the first generation systems that
>supports number portability. However, when the number of portable numbers
>gets higher, the inefficiency in the approach becomes more pronounced.
>Differentiation of portable numbers or the correct lookup of the
>switch/gateway
>where the called number resides needs to be done much earlier in the
>SETUP forwarding path.
>
>With the competition in local access market heating up and new area code
>added using overlay instead of split, aggregation based on the number
>hierarchy starts to make less and less sense. On top of this, with cell
>phones
>becomes cheaper and cheaper to have, and the wireless bandwidth grows
>higher, to support roaming/mobility of wireless subscribers both within
>wireless networks and across wireless and wireline based networks,
>phone numbers will need to be handled more and more on a individual
>bases.
>
>To make GLP for IP based telephony networks future proof, the protocol
>needs to support unaggregated  location information for telephone numbers.
>Of course, this will increase the volume of traffic. But if the information
>is
>basically unaggregable, it is just the nature of the problem itself. It will
>probably requires a different solution than it is currently outlined in the
>two proposals that we currently have.
>
>I am in the process of writing up a design that is used for propagating
>location information among mobile telephony gateways. Although the
>original design target is a IP based carrier class wireless telephone
>network. It can be used to solve the generic gateway location problem
>in a IP telephony network. If there is enough interest, I can outline the
>design or send my document to the group for discussion.
>
>Best regards,
>
>- Greg
>
>
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar  8 18:42:48 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA16140
	for <iptel-archive@odin.ietf.org>; Mon, 8 Mar 1999 18:42:48 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 30F557FDC; Mon,  8 Mar 1999 18:33:14 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7704C7FDE; Mon,  8 Mar 1999 18:33:13 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990309103311.00707944@pop.ozemail.com.au>
X-Sender: dclowes@pop.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Tue, 09 Mar 1999 10:33:11 +1100
To: Scott Petrack <scott.petrack@metatel.com>
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Cc: "David R. Oran" <oran@cisco.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        Alok Mittal <amittal@hss.hns.com>,
        "'Matt Squire'" <msquire@nortelnetworks.com>,
        iptel@lists.research.bell-labs.com
In-Reply-To: <Pine.LNX.4.04.9903081218190.3058-100000@petrack.metatel.co
 m>
References: <NBBBIFCAKKOPNMMKHHEFKEFEEEAA.oran@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Scott,

The big issue of text versus binary usually arises where the processing
domain and the transmission domain are potentially different. This is,
IMHO, not such a case.

The problem domain of mapping dialed digits, is not a numeric one, but a
string one. The processing needs to be done "in string", so the
transmission may as well be, to save the conversions that text transmission
usually involves.

While 49 and 50 are quite close in binary, they are very different in terms
of the digits dialed. Extending that thought, 599-9999 (5B8D7FH) might
represent something very close to 600-0000, in hex/ (5B8D80H) binary, but
they are very different as dialing strings.

Additionally, numerics are "right-centric", while strings (dialed digits)
are "left-centric". Here, 600 and 6000000 are very different numerically,
but map the same in a three digit string compare. That is, if 600 is the
template, and 6000000 is the dialed digits, they match.

You have to store them and process them as strings, you should transmit
them as strings, even in a binary protocol.

Regards,

Douglas

At 12:21 1999-03-08 -0500, Scott Petrack wrote:
>
>
>On Mon, 8 Mar 1999, David R. Oran wrote:
>
>> One interesting question is wherhter to actually encode the numbers in
>> binary in the protocol. Whether this saves or wastes space depends on
>> whether you expect long or short prefixes. From a protocol debugging
>> point of view, sending strings wins.
>
>Surely the processing time is more important than the space issue. I
>realize that the packets don't need to be processed anywhere near line
>speed, but a server that needs to process 1000s of phone calls will surely
>benefit from not having to process the strings. 
>
>Anyone who needs to debug can easily write a filter to translates the bits
>into decimal.
>
>Scott
>
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.
>
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 03:21:47 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA08153
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 03:21:47 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A56767FB7; Tue,  9 Mar 1999 02:29:21 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 03CD57FC8; Tue,  9 Mar 1999 02:29:20 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <6525672F.00261979.00@sampark.hss.hns.com>
Date: Tue, 9 Mar 1999 13:00:47 +0530
Subject: Handling of Setup message by Gatekeeper.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi All,
       Here is a query regarding handling of setup message received by
Gatekeeper.

In normal case, any endpoint in the zone of a gatekeeper will send ARQ
message before sending Setup message. On receiving ARQ, Gatekeeper will
check for registration, allocate bandwidth and also decides the callmodel.

But there are cases when Gatekeeper receives Setup without receiving ARQ
from an endpoint. Scenarios include:
1. Called party Gatekeeper decides for GKrouted call model and sends ARJ.
Called endpoint sends facility message with reason = routeCallToGatekeeper
to the calling party. In this case, calling party releases the previous
call and sends a new Setup message to the called endpoint's Gatekeeper.
2. If calling and called endpoints are registered with different
gatekeepers and called party's gatekeeper decides for GKrouted call model.
Calling endpoint, registered with gatekeeper1, will send Setup to
gatekeeper2 while it has performed ARQ/ACF transaction with gatekeeper1.
3. GK receives Setup message from another gatekeeper.
4. GK receives a Setup message from a gateway in preGrantedArq scenario.

In all these cases, gatekeeper will receive Setup message without receiving
ARQ message. So how can gatekeeper check whether ARQ is expected from the
endpoint, which sent the Setup or not? No field in the Setup message seem
to give this information. If gatekeeper doesn't check for this, that means
even the endpoints in its own zone can bypass gatekeeper for RAS
signalling?

Any help/information on this issue will be appreciated.

Thanks,
Ritu.
_____________________________________________________________________
Ritu Gupta                                   |       email:
rtgupta@hss.hns.com
Senior Software Engineer,       |                   rtgupta@hotmail.com
Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
Banglore, Karnataka                  |                    91-80-3374075(H)
INDIA.                                           |       Fax  :
91-80-2286393
_____________________________________________________________________
Restlessness and discontent are the first necessities of progress.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 06:02:54 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA09454
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 06:02:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8641B7FC9; Tue,  9 Mar 1999 05:51:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DD6AD7FC8; Tue,  9 Mar 1999 05:51:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BE6A2B.8F25A6E0.amir@radvision.rad.co.il>
From: Ami Amir <amir@radvision.rad.co.il>
To: iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 02:44:23 +0200
Organization: RADVision
X-Mailer: Microsoft Internet E-mail/MAPI - 8.0.0.4211
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Greg,

I, for one, would be very interested in your design.

Ami Amir

-----Original Message-----
From:	Patrice Calhoun [SMTP:Pat.Calhoun@Eng.Sun.COM]
Sent:	Tuesday, March 09, 1999 12:00 AM
To:	Greg Zhang; Matt Squire; Jonathan Rosenberg; David R. Oran
Cc:	iptel; Alok Mittal; Scott Petrack
Subject:	Re: GLP, number portability and wireless subscribers

Greg,

Some of us have been addressing this problem, but not necessarily in the IPTel
space. Alot of work was done in TR45.6 (3G cellular systems), and much of this
was brought forth at the Mobile IP WG. Perhaps the work there applies here 
as well. I would be interested in seeing your requirements/architecture.

For a start you can look at draft-hiller-3gwireless-00.txt. For protocol 
information you can look at draft-calhoun-diameter-mobileip-03.txt. Again,
perhaps the problem is not exactly the same, but we may be able to re-use
some of the work already done.


PatC

>Hello,
>
>One issue needs to be considered is number portability. Number portability
>(NP) is reality for wired POTS lines for some time now. Wireless Number
>Portability will be a reality in a few years' time. The portable numbers
>cannot
>be routed by strictly following the hierarchy derived from the digits of
>called
>number. When the number of portable numbers is small compared to all
>telephone numbers in the network, it can be handled on an exception bases
>meaning route the call following the hierarchy then back track when hitting
>a dead end. This is approach used in the first generation systems that
>supports number portability. However, when the number of portable numbers
>gets higher, the inefficiency in the approach becomes more pronounced.
>Differentiation of portable numbers or the correct lookup of the
>switch/gateway
>where the called number resides needs to be done much earlier in the
>SETUP forwarding path.
>
>With the competition in local access market heating up and new area code
>added using overlay instead of split, aggregation based on the number
>hierarchy starts to make less and less sense. On top of this, with cell
>phones
>becomes cheaper and cheaper to have, and the wireless bandwidth grows
>higher, to support roaming/mobility of wireless subscribers both within
>wireless networks and across wireless and wireline based networks,
>phone numbers will need to be handled more and more on a individual
>bases.
>
>To make GLP for IP based telephony networks future proof, the protocol
>needs to support unaggregated  location information for telephone numbers.
>Of course, this will increase the volume of traffic. But if the information
>is
>basically unaggregable, it is just the nature of the problem itself. It will
>probably requires a different solution than it is currently outlined in the
>two proposals that we currently have.
>
>I am in the process of writing up a design that is used for propagating
>location information among mobile telephony gateways. Although the
>original design target is a IP based carrier class wireless telephone
>network. It can be used to solve the generic gateway location problem
>in a IP telephony network. If there is enough interest, I can outline the
>design or send my document to the group for discussion.
>
>Best regards,
>
>- Greg
>
>
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.



---------
This message came from the IETF IPTEL Working Group Mailing List.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 06:17:58 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA09493
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 06:17:58 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B02DF7FB8; Tue,  9 Mar 1999 05:49:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1D7557FC8; Tue,  9 Mar 1999 05:49:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: "Ramesh B S" <bsr@netlab.hcltech.com>
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <6525672F.003A1326.00@sampark.hss.hns.com>
Date: Tue, 9 Mar 1999 16:18:48 +0530
Subject: Re: Handling of Setup message by Gatekeeper.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi Ramesh,
    Thanks for your input. But still things are not clear. It's not only
registration which matters but presence of endpoint in the zone also
affects this case. What if endpoint is in the zone but not registered? An
endpoint just got power on and directly sent Setup message. Therefore three
checks are required at the Gatekeeper.
1. whether or not endpoint is in the zone of this GK.
2. If yes, then whether it's registered or not.
3. If it is registered, whether ARQ was sent or not.

May be GK needs to maintain a static database of all the endpoints in its
zone along with the registration table. But as Anant pointed out, there is
no mandatory field in the Setup message, which provides source information.
So even if both the databases are available with the GK, how to perform the
checks? As getting a setup message without an ARQ is a valid case (as given
in first mail), we can not just discard the message.

Any inputs?

Thanks,
Ritu.





"Ramesh B S" <bsr@netlab.hcltech.com> on 03/09/99 02:49:44 PM

To:   rtgupta@hss.hns.com
cc:
Subject:  Re: Handling of Setup message by Gatekeeper.







>Hi All,
>       Here is a query regarding handling of setup message received by
>Gatekeeper.
>
>In normal case, any endpoint in the zone of a gatekeeper will send ARQ
>message before sending Setup message. On receiving ARQ, Gatekeeper will
>check for registration, allocate bandwidth and also decides the callmodel.
>
>But there are cases when Gatekeeper receives Setup without receiving ARQ
>from an endpoint. Scenarios include:
>1. Called party Gatekeeper decides for GKrouted call model and sends ARJ.
>Called endpoint sends facility message with reason = routeCallToGatekeeper
>to the calling party. In this case, calling party releases the previous
>call and sends a new Setup message to the called endpoint's Gatekeeper.
>2. If calling and called endpoints are registered with different
>gatekeepers and called party's gatekeeper decides for GKrouted call model.
>Calling endpoint, registered with gatekeeper1, will send Setup to
>gatekeeper2 while it has performed ARQ/ACF transaction with gatekeeper1.
>3. GK receives Setup message from another gatekeeper.
>4. GK receives a Setup message from a gateway in preGrantedArq scenario.
>
>In all these cases, gatekeeper will receive Setup message without
receiving
>ARQ message. So how can gatekeeper check whether ARQ is expected from the
>endpoint, which sent the Setup or not? No field in the Setup message seem
>to give this information. If gatekeeper doesn't check for this, that means
>even the endpoints in its own zone can bypass gatekeeper for RAS
>signalling?
>

If there is a gatekeeper in the system, then, according to the spec.
(H.323,
v2, section 7.2.2),
all endpoints shall register with it. It is possible to bypass a gatekeeper
but
that would be a protocol violation.
To answer the question as to how a gatekeeper checks for ARQs:
A gatekeeper will have a database of all endpoints that
have registered with it. A gatekeeper should also maintain states
associated
with such endpoints. So when a Setup message arrives at the GK, it can
tell if the endpoint is from the same zone and if it is, whether it has
sent
an ARQ (if !preGrantedArq) before this request.







---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 06:44:20 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA09642
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 06:44:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2322A7FC8; Tue,  9 Mar 1999 06:33:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7711E7FCA; Tue,  9 Mar 1999 06:33:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <4110BC526A58D211B56C0000F6B9636DA851A9@ws-exch01.dev.madge.com>
From: Chris Purvis <CPurvis@Madge.com>
To: "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>,
        Ramesh B S <bsr@netlab.hcltech.com>
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject: RE: Handling of Setup message by Gatekeeper.
Date: Tue, 9 Mar 1999 11:28:33 -0000 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain;
	charset="windows-1252"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

All,

A few options:

1. Let the endpoint get away with it.
2. Use some other form of security, such as H.235.
3. Do everything Ramesh suggests AND have a proxy on the edge of the zone so
that all calls into the zone come via the proxy (which is registered and
will do ARQ).
4. Within an administrative domain, it is possible to have a restricted
database of entities (such as other gatekeepers) which are permitted to
setup without ARQ.

Maintaining a static database of all endpoints in a GK's zone is not by
itself a good solution to this particular problem: it seems unreasonable to
give MORE rights to things outside of one's zone than things within it.

A few other pointers:

A gatekeeper knows whether it has pregranted an ARQ to a particular entity.
It is up to the entity to fill in the setup message such that the gatekeeper
knows who it is - sourceCallSignalAddress and SourceAddress fields are the
starter for this, if you want security use H.235.

Regards,
Chris

> -----Original Message-----
> From: rtgupta@hss.hns.com [mailto:rtgupta@hss.hns.com]
> Sent: 09 March 1999 10:49
> To: Ramesh B S
> Cc: h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject: Re: Handling of Setup message by Gatekeeper.
> 
> 
> Hi Ramesh,
>     Thanks for your input. But still things are not clear. 
> It's not only
> registration which matters but presence of endpoint in the zone also
> affects this case. What if endpoint is in the zone but not 
> registered? An
> endpoint just got power on and directly sent Setup message. 
> Therefore three
> checks are required at the Gatekeeper.
> 1. whether or not endpoint is in the zone of this GK.
> 2. If yes, then whether it's registered or not.
> 3. If it is registered, whether ARQ was sent or not.
> 
> May be GK needs to maintain a static database of all the 
> endpoints in its
> zone along with the registration table. But as Anant pointed 
> out, there is
> no mandatory field in the Setup message, which provides 
> source information.
> So even if both the databases are available with the GK, how 
> to perform the
> checks? As getting a setup message without an ARQ is a valid 
> case (as given
> in first mail), we can not just discard the message.
> 
> Any inputs?
> 
> Thanks,
> Ritu.
> 
> 
> 
> 
> 
> "Ramesh B S" <bsr@netlab.hcltech.com> on 03/09/99 02:49:44 PM
> 
> To:   rtgupta@hss.hns.com
> cc:
> Subject:  Re: Handling of Setup message by Gatekeeper.
> 
> 
> 
> 
> 
> 
> 
> >Hi All,
> >       Here is a query regarding handling of setup message 
> received by
> >Gatekeeper.
> >
> >In normal case, any endpoint in the zone of a gatekeeper 
> will send ARQ
> >message before sending Setup message. On receiving ARQ, 
> Gatekeeper will
> >check for registration, allocate bandwidth and also decides 
> the callmodel.
> >
> >But there are cases when Gatekeeper receives Setup without 
> receiving ARQ
> >from an endpoint. Scenarios include:
> >1. Called party Gatekeeper decides for GKrouted call model 
> and sends ARJ.
> >Called endpoint sends facility message with reason = 
> routeCallToGatekeeper
> >to the calling party. In this case, calling party releases 
> the previous
> >call and sends a new Setup message to the called endpoint's 
> Gatekeeper.
> >2. If calling and called endpoints are registered with different
> >gatekeepers and called party's gatekeeper decides for 
> GKrouted call model.
> >Calling endpoint, registered with gatekeeper1, will send Setup to
> >gatekeeper2 while it has performed ARQ/ACF transaction with 
> gatekeeper1.
> >3. GK receives Setup message from another gatekeeper.
> >4. GK receives a Setup message from a gateway in 
> preGrantedArq scenario.
> >
> >In all these cases, gatekeeper will receive Setup message without
> receiving
> >ARQ message. So how can gatekeeper check whether ARQ is 
> expected from the
> >endpoint, which sent the Setup or not? No field in the Setup 
> message seem
> >to give this information. If gatekeeper doesn't check for 
> this, that means
> >even the endpoints in its own zone can bypass gatekeeper for RAS
> >signalling?
> >
> 
> If there is a gatekeeper in the system, then, according to the spec.
> (H.323,
> v2, section 7.2.2),
> all endpoints shall register with it. It is possible to 
> bypass a gatekeeper
> but
> that would be a protocol violation.
> To answer the question as to how a gatekeeper checks for ARQs:
> A gatekeeper will have a database of all endpoints that
> have registered with it. A gatekeeper should also maintain states
> associated
> with such endpoints. So when a Setup message arrives at the GK, it can
> tell if the endpoint is from the same zone and if it is, 
> whether it has
> sent
> an ARQ (if !preGrantedArq) before this request.
> 
> 
> 
> 
> 
> 
> --------------------------------------------------------------
> -----------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"
> 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 08:25:09 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA12115
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 08:25:09 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 35AB47FCA; Tue,  9 Mar 1999 08:13:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7F3907FCB; Tue,  9 Mar 1999 08:13:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990309080950.00acd850@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Tue, 09 Mar 1999 08:09:51 -0500
To: Greg Zhang <gzhang@utstar.com>, "David R. Oran" <oran@cisco.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


I agree with the importance of number portability.   But we are just
talking about the signalling messages here.  Having a user register with
with a 'home server' which serves as a proxy signalling server for that
user (as he/she roams about) doesn't seem unreasonable.  Data would go a
more direct route via RTP.   Unlike the PSTN, the data is not tied to the
signalling path.  

The problem of users roaming after the data path is established is an
interesting problem, but I think its a problem for a different working
group.  In this case, you'd want the RTP stream to switch from one
destination to another in mid conversation.  My personal opinion is that
the 'inter-domain' signalling messages should still go to a home server
which then has worry about maintaining current status on its users.  Having
the inter-domain world-wide call routing protocol update every time any
user changes cell towers would be impractical.  

I think that most people agree that aggregation is going to be difficult.
Exactly how difficult is hard to say.  If it can be done, then it should
be, so that the database size is reduced.  This leads to smaller table
transfers, fewer updates, and quicker lookups.  If it can't be done, then
it can't be done.  How much effort should be spent attempting aggregation
was one of the points of our previous thread.  I think that Dave feels
aggregation will be more successful than I think it will be.  

I'm not sure that this is the right group for your proposal, but I'd be
interested in seeing it.

- Matt

At 01:26 PM 3/8/99 -0800, Greg Zhang wrote:
>Hello,
>
>One issue needs to be considered is number portability. Number portability
>(NP) is reality for wired POTS lines for some time now. Wireless Number
>Portability will be a reality in a few years' time. The portable numbers
>cannot
>be routed by strictly following the hierarchy derived from the digits of
>called
>number. When the number of portable numbers is small compared to all
>telephone numbers in the network, it can be handled on an exception bases
>meaning route the call following the hierarchy then back track when hitting
>a dead end. This is approach used in the first generation systems that
>supports number portability. However, when the number of portable numbers
>gets higher, the inefficiency in the approach becomes more pronounced.
>Differentiation of portable numbers or the correct lookup of the
>switch/gateway
>where the called number resides needs to be done much earlier in the
>SETUP forwarding path.
>
>With the competition in local access market heating up and new area code
>added using overlay instead of split, aggregation based on the number
>hierarchy starts to make less and less sense. On top of this, with cell
>phones
>becomes cheaper and cheaper to have, and the wireless bandwidth grows
>higher, to support roaming/mobility of wireless subscribers both within
>wireless networks and across wireless and wireline based networks,
>phone numbers will need to be handled more and more on a individual
>bases.
>
>To make GLP for IP based telephony networks future proof, the protocol
>needs to support unaggregated  location information for telephone numbers.
>Of course, this will increase the volume of traffic. But if the information
>is
>basically unaggregable, it is just the nature of the problem itself. It will
>probably requires a different solution than it is currently outlined in the
>two proposals that we currently have.
>
>I am in the process of writing up a design that is used for propagating
>location information among mobile telephony gateways. Although the
>original design target is a IP based carrier class wireless telephone
>network. It can be used to solve the generic gateway location problem
>in a IP telephony network. If there is enough interest, I can outline the
>design or send my document to the group for discussion.
>
>Best regards,
>
>- Greg
>
>
>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 10:40:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA14701
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 10:40:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 73CAD7FCB; Tue,  9 Mar 1999 10:29:54 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B7CB37FCC; Tue,  9 Mar 1999 10:29:53 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E53DE3.694C7D69@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 10:27:31 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Greg Zhang <gzhang@utstar.com>
Cc: "David R. Oran" <oran@cisco.com>, Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <010501be69aa$79336240$2821a8c0@greg_zhang.wacos.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Greg Zhang wrote:
> 
> Hello,
> 
> One issue needs to be considered is number portability. Number portability
> (NP) is reality for wired POTS lines for some time now. Wireless Number
> Portability will be a reality in a few years' time. The portable numbers
> cannot
> be routed by strictly following the hierarchy derived from the digits of
> called
> number. 

I agree LNP makes the problem harder. I have two responses here:

1. Simply "not aggregating" is always an option for a protocol which
provides aggregation. I don't want to not support aggregation because of
LNP. If LNP gets in the way, it can just be turned off if needed.

2. I'm not yet convinced LNP actually prevents aggregation. Lets say
gateway X can reach ten numbers, +1 607-1* (using glob matching), +1
607-2*, etc. so these could be aggregated to +1 607*. Lets further
assume all ten numbers are LNP, and actually terminate all over the
country. These numbers can still all be reached through the gateway.
They cannot be aggregated only if the attributes for each are different.
The only attribute likely to differ is cost. So, the question is, does
the cost (to the gateway provider) for terminating calls to each of the
numbers differ? Even if they do, is the cost the gateway provider wants
to charge to its users different for each number? If the answer to
either is no, then they can still be aggregated. With long distance
rates at 10 cents a minute *anywhere* in the U.S., this seems not so
unlikely.


There is a related issue of 800 numbers, and any of the other
non-routing numbers. Should these be advertised by GLP? Or, should they
be translated first, and then the translated numbers advertised? (Please
correct me if I have misunderstood some aspect of how this works in the
PSTN).

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 11:13:14 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA15349
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 11:13:13 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1162F7FCC; Tue,  9 Mar 1999 11:03:14 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 565417FCD; Tue,  9 Mar 1999 11:03:13 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Greg Zhang" <gzhang@utstar.com>
Cc: "Matt Squire" <msquire@nortelnetworks.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 11:02:00 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFOEIJEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
In-Reply-To: <36E53DE3.694C7D69@dnrc.bell-labs.com>
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Jonathan R. wrote...
> There is a related issue of 800 numbers, and any of the other
> non-routing numbers. Should these be advertised by GLP? Or,
> should they
> be translated first, and then the translated numbers
> advertised? (Please
> correct me if I have misunderstood some aspect of how this
> works in the
> PSTN).
>
We may not be sunk if we try to handle "non-routing numbers" in a
routing protocol, but it sure would be nice to minimize or eliminate the
need to. It would save a lot of bandwidth and memory.

One idea I've kicked around is to have an explicit syntactical marker on
an E.164 number so anybody processing it knows whether that number
represents a destination rather than a "directory number". Then we try
to only feed final destinations into the routing protocol. For the LNP
case, it may mean finding a way to explicitly carry "office codes" in
the protocol. For the 8xx/700 number cases we may wish to simply declare
these to be names and ban them from the routing advertisements.

Dave.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 11:39:16 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA15728
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 11:39:15 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2EFD27FCD; Tue,  9 Mar 1999 11:29:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 82D8F7FCE; Tue,  9 Mar 1999 11:29:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903091625.IAA01401@mailman.cisco.com>
X-Sender: jschnizl@diablo.cisco.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.0.2 
Date: Tue, 09 Mar 1999 11:26:10 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        Greg Zhang <gzhang@utstar.com>
From: John Schnizlein <jschnizl@cisco.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: "David R. Oran" <oran@cisco.com>, Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
In-Reply-To: <36E53DE3.694C7D69@dnrc.bell-labs.com>
References: <010501be69aa$79336240$2821a8c0@greg_zhang.wacos.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 10:27 AM 3/9/99 -0500, Jonathan Rosenberg wrote:
>
>I don't want to not support aggregation because of
>LNP. If LNP gets in the way, it can just be turned off if needed.

Which "it" do you mean can be turned off? 

>2. I'm not yet convinced LNP actually prevents aggregation. Lets say
>gateway X can reach ten numbers, +1 607-1* (using glob matching), +1
>607-2*, etc. so these could be aggregated to +1 607*. 

Aggregation coexists with a moderate amount of "ported numbers" because
most numbers are still covered by the aggregations.
You can keep +1 607-555-1212 in the same database with +1 607-444*
and the ones you listed if you use longest-match lookup.

>There is a related issue of 800 numbers, and any of the other
>non-routing numbers. Should these be advertised by GLP? Or, should they
>be translated first, and then the translated numbers advertised? (Please
>correct me if I have misunderstood some aspect of how this works in the
>PSTN).

I see 800 number portability as the leading use of LNP, not just related.
The PSTN assumes that ported number will be the exception.
Aggregation by <area><exchage> is used as before LNP unless any number
in an exchange has been ported. If an exchange included a ported number,
the "right" switch looks up the exception in a process called "Global
Title Translation" and forwards the call to a guaranteed-not-ported number
of the switch to which the number was ported.

Longest match enables more flexibility than the fixed fields of area and
exchange as partial addresses, and facilitates keeping the routing for
aggregated destinations in the same database as the (LNP) exceptions.

John

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 11:49:37 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA15845
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 11:49:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B5BD17FCE; Tue,  9 Mar 1999 11:39:37 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E997F7FCF; Tue,  9 Mar 1999 11:39:36 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E54E32.C8931AB2@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 11:37:06 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: John Schnizlein <jschnizl@cisco.com>
Cc: Greg Zhang <gzhang@utstar.com>, "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <010501be69aa$79336240$2821a8c0@greg_zhang.wacos.com> <199903091625.IAA01401@mailman.cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

John Schnizlein wrote:
> 
> At 10:27 AM 3/9/99 -0500, Jonathan Rosenberg wrote:
> >
> >I don't want to not support aggregation because of
> >LNP. If LNP gets in the way, it can just be turned off if needed.
> 
> Which "it" do you mean can be turned off?

Aggregation.

> 
> >2. I'm not yet convinced LNP actually prevents aggregation. Lets say
> >gateway X can reach ten numbers, +1 607-1* (using glob matching), +1
> >607-2*, etc. so these could be aggregated to +1 607*.
> 
> Aggregation coexists with a moderate amount of "ported numbers" because
> most numbers are still covered by the aggregations.
> You can keep +1 607-555-1212 in the same database with +1 607-444*
> and the ones you listed if you use longest-match lookup.

Agree.

> 
> >There is a related issue of 800 numbers, and any of the other
> >non-routing numbers. Should these be advertised by GLP? Or, should they
> >be translated first, and then the translated numbers advertised? (Please
> >correct me if I have misunderstood some aspect of how this works in the
> >PSTN).
> 
> I see 800 number portability as the leading use of LNP, not just related.
> The PSTN assumes that ported number will be the exception.
> Aggregation by <area><exchage> is used as before LNP unless any number
> in an exchange has been ported. If an exchange included a ported number,
> the "right" switch looks up the exception in a process called "Global
> Title Translation" and forwards the call to a guaranteed-not-ported number
> of the switch to which the number was ported.

This begs the question, then: does the translation of the 800 number get
done before the gateway selection process, or after? By doing it before,
a gateway can be chosen which is, in fact, close to the real number
being dialed. With 800 numbers, all gateways will be able to reach it.
It seems that this opens the possibility of very non-optimal routing. A
caller in NY that dials an 800 number, goes to a gateway in CA. The
phone network there determines that the 800 number translates to a
number in NY, and routes the call through the PSTN back to NY. However,
if the translation occurs before gateway selection, the caller can
choose a local NY gateway and do much better.

-Jonathan R.


-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 12:05:56 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA16099
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 12:05:56 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 9A62B7FCF; Tue,  9 Mar 1999 11:55:38 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A9E587FD2; Tue,  9 Mar 1999 11:55:37 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Tue, 9 Mar 1999 08:54:30 -0800 (PST)
From: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>
Reply-To: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>
Subject: Re: GLP, number portability and wireless subscribers
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
In-Reply-To: "Your message with ID" <36E54E32.C8931AB2@dnrc.bell-labs.com>
Message-ID: <Roam.SIMCSD.2.0.4.920998470.9397.pcalhoun@hsmpka>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

> John Schnizlein wrote:
> > 
> > At 10:27 AM 3/9/99 -0500, Jonathan Rosenberg wrote:
> > >
> > >I don't want to not support aggregation because of
> > >LNP. If LNP gets in the way, it can just be turned off if needed.
> > 
> > Which "it" do you mean can be turned off?
> 
> Aggregation.
> 
> > 
> > >2. I'm not yet convinced LNP actually prevents aggregation. Lets say
> > >gateway X can reach ten numbers, +1 607-1* (using glob matching), +1
> > >607-2*, etc. so these could be aggregated to +1 607*.
> > 
> > Aggregation coexists with a moderate amount of "ported numbers" because
> > most numbers are still covered by the aggregations.
> > You can keep +1 607-555-1212 in the same database with +1 607-444*
> > and the ones you listed if you use longest-match lookup.
> 
> Agree.
> 
> > 
> > >There is a related issue of 800 numbers, and any of the other
> > >non-routing numbers. Should these be advertised by GLP? Or, should they
> > >be translated first, and then the translated numbers advertised? (Please
> > >correct me if I have misunderstood some aspect of how this works in the
> > >PSTN).
> > 
> > I see 800 number portability as the leading use of LNP, not just related.
> > The PSTN assumes that ported number will be the exception.
> > Aggregation by <area><exchage> is used as before LNP unless any number
> > in an exchange has been ported. If an exchange included a ported number,
> > the "right" switch looks up the exception in a process called "Global
> > Title Translation" and forwards the call to a guaranteed-not-ported number
> > of the switch to which the number was ported.
> 
> This begs the question, then: does the translation of the 800 number get
> done before the gateway selection process, or after? By doing it before,
> a gateway can be chosen which is, in fact, close to the real number
> being dialed. With 800 numbers, all gateways will be able to reach it.
> It seems that this opens the possibility of very non-optimal routing. A
> caller in NY that dials an 800 number, goes to a gateway in CA. The
> phone network there determines that the 800 number translates to a
> number in NY, and routes the call through the PSTN back to NY. However,
> if the translation occurs before gateway selection, the caller can
> choose a local NY gateway and do much better.

Translation before the gateway selection is *obviously* the correct model. We
should strive towards this goal.

PatC


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar  9 12:24:04 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA16342
	for <iptel-archive@odin.ietf.org>; Tue, 9 Mar 1999 12:24:04 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 891F37FD4; Tue,  9 Mar 1999 12:17:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7B3BD7FD6; Tue,  9 Mar 1999 12:17:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Tue, 9 Mar 1999 09:14:40 -0800 (PST)
From: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>
Reply-To: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>
Subject: Re: GLP, number portability and wireless subscribers
To: Matt Squire <msquire@nortelnetworks.com>
Cc: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
In-Reply-To: "Your message with ID" <3.0.32.19990309120938.00ad15c0@bl-mail1.corpeast.baynetworks.com>
Message-ID: <Roam.SIMCSD.2.0.4.920999680.16383.pcalhoun@hsmpka>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

> 
> >> This begs the question, then: does the translation of the 800 number get
> >> done before the gateway selection process, or after? By doing it before,
> >> a gateway can be chosen which is, in fact, close to the real number
> >> being dialed. With 800 numbers, all gateways will be able to reach it.
> >> It seems that this opens the possibility of very non-optimal routing. A
> >> caller in NY that dials an 800 number, goes to a gateway in CA. The
> >> phone network there determines that the 800 number translates to a
> >> number in NY, and routes the call through the PSTN back to NY. However,
> >> if the translation occurs before gateway selection, the caller can
> >> choose a local NY gateway and do much better.
> >
> >Translation before the gateway selection is *obviously* the correct model.
> We >should strive towards this goal.
> >
> >PatC
> >
> >
> 
> The next question is when before gateway selection.  Translation could be
> done as a pre-cursor to call-routing by the ingress, could be done by
> signalling itself via things like SIP redirections/proxies, or could be
> done within the routing infrastructure itself.  
> 
> I'd prefer one of the first two alternatives.  
> 
That seems to be the one that is least disruptive to the Internet (this is my
perception). We are after all simply creating a new service that runs on top
of the Internet, so it does not seem wise to have to play with the routing
infrastructure.

PatC


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 11:36:46 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA05930
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 11:36:46 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1D3A28013; Thu, 11 Mar 1999 11:29:55 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5E96F8092; Thu, 11 Mar 1999 11:22:04 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <36E6A165.62F81438@lucent.com>
Date: Wed, 10 Mar 1999 10:44:21 -0600
From: "ronald h. davis" <ronaldd@lucent.com>
Organization: Lucent Technologies
X-Mailer: Mozilla 4.05 [en]C-EMS-1.4  (WinNT; U)
MIME-Version: 1.0
To: rtgupta@hss.hns.com
Original-CC: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject: Re: Handling of Setup message by Gatekeeper.
References: <65256730.001BC1A0.00@sampark.hss.hns.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

rtgupta@hss.hns.com wrote:
> 
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
> 

i think that i don't understand what you're trying to say.  the setup
message is an h.225 call signalling message, not a ras message.

-- 
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
"if you ain't the lead dog, the scenery never changes"  -- edmund wilson

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 11:57:43 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA06180
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 11:57:43 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 948E37FDE; Tue,  9 Mar 1999 12:53:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DBD5F7FDF; Tue,  9 Mar 1999 12:53:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903091749.JAA01215@mailman.cisco.com>
X-Sender: jschnizl@diablo.cisco.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.0.2 
Date: Tue, 09 Mar 1999 12:49:19 -0500
To: Matt Squire <msquire@nortelnetworks.com>,
        "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: John Schnizlein <jschnizl@cisco.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: Greg Zhang <gzhang@utstar.com>, "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
In-Reply-To: <3.0.32.19990309120938.00ad15c0@bl-mail1.corpeast.baynetwor
 ks.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Notice that names, aka "directory numbers", don't look different from 
addresses, aka "routing numbers" in the PSTN as they do on the Internet.
The tradeoff is that an early lookup can avoid extra call routing,
but that cost is saved only for the relatively few numbers that are ported.

If most numbers really needed lookup, a (LDAP) directory with its schema
reflecting the hierarchy of delegation of ported numbers would be better
than the tedious process used in the PSTN to manage the LPN database.
It might require extensions to LDAP because the parsing of a phone number
into the delegation hierarchy is different in different parts of the world.

The other obstacle to this early-translation approach is the administration
of this directory. While the rudiments of distributing subtrees of the
directory, referrals to other authorities, access control for changing 
pointers, and replication to avoid the chain of queries are specified;
interoperability of these features are still progressing through LDAPext
and LDUP working groups.

One way to deal with it is for gateways to perform only as much lookup
as they desire to avoid poor routes until the directory standards settle.
Even then, would is be OK for some gateways to use poor routes as a local
option to using the global directory?

Even though standardizing these issues is outside the scope of IPTEL,
it is important that the (external) solution is feasible withing the
scope of existing IETF working group charters.

John

At 12:09 PM 3/9/99 -0500, Matt Squire wrote:

>>Translation before the gateway selection is *obviously* the correct model. 
>>We should strive towards this goal.
>>
>>PatC
>
>The next question is when before gateway selection.  Translation 
>could be done as a pre-cursor to call-routing by the ingress, 
>could be done by signalling itself via things like SIP redirections/proxies, 
>or could be done within the routing infrastructure itself.  
>
>I'd prefer one of the first two alternatives.  
>
>- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 12:18:06 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA06523
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 12:18:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E4BC4802D; Thu, 11 Mar 1999 11:10:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 54E408013; Thu, 11 Mar 1999 11:05:01 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <36E5622B.647C0FB@lucent.com>
Date: Tue, 09 Mar 1999 12:02:19 -0600
From: "ronald h. davis" <ronaldd@lucent.com>
Organization: Lucent Technologies
X-Mailer: Mozilla 4.05 [en]C-EMS-1.4  (WinNT; U)
MIME-Version: 1.0
To: rtgupta@hss.hns.com
Original-CC: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject: Re: Handling of Setup message by Gatekeeper.
References: <6525672F.00261979.00@sampark.hss.hns.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

rtgupta@hss.hns.com wrote:
> 
			[scenarios deleted]
> In all these cases, gatekeeper will receive Setup message without receiving
> ARQ message. So how can gatekeeper check whether ARQ is expected from the
> endpoint, which sent the Setup or not? No field in the Setup message seem
> to give this information. If gatekeeper doesn't check for this, that means
> even the endpoints in its own zone can bypass gatekeeper for RAS
> signalling?
> 
> Any help/information on this issue will be appreciated.
> 

the gatekeeper is only responsible for arq/acf/arj with endpoints within 
its zone.  all endpoints within the zone establish a ras channel (over
which these messages are sent) with their gatekeeper.

endpoints in other zones don't establish a ras channel with this
gatekeeper.  thus, when calls are received from outside of the zone
the recipient (be through the direct or gatekeeper routed call model)
has no idea how the calling endpoint came upon sending the setup
message.

this statement may be controversial but it is my opinion that h.323
is lan-oriented, so if you think of gatekeepers as administering
bandwidth allocation, &c for endpoints on different lans it would
be difficult for a gatekeeper to perform this administration for
a different lan.  and it would be even harder for a gatekeeper to
perform such administration over a wide area network consisting of
a number of zones.

this fact leads me to believe that the ras channel should be *optional*
for h.323 gatekeepers.  to abstract your scenarios further it is
possible that gatekeepers could act as application layer routers, in
which case they are terminating h.225 signalling and h.245 media control
channels for endpoints with which they have no ras channel.

-- 
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
"if you ain't the lead dog, the scenery never changes"  -- edmund wilson

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 12:37:26 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA06741
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 12:37:26 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 133568037; Thu, 11 Mar 1999 11:23:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 274E37FDA; Thu, 11 Mar 1999 11:19:32 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E6C8D0.BA795CF2@lucent.com>
Date: Wed, 10 Mar 1999 14:32:32 -0500
From: Alec Brusilovsky <abrusilovsky@lucent.com>
Reply-To: abrusilovsky@lucent.com
Organization: Lucent Technologies, Bell Laboratories Innovation
X-Mailer: Mozilla 4.5 [en] (Win95; U)
X-Accept-Language: en,ru
MIME-Version: 1.0
To: iptel@lists.research.bell-labs.com
Subject: PIN BOF Announcement
Content-Type: multipart/mixed;
 boundary="------------32C77E3858257B4CB7FC4132"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.
--------------32C77E3858257B4CB7FC4132
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

On the advice of the Transport Area Directorate I would like to publish
this announcement regarding PIN BOF on March 15 from 3:30 through 5:30
p.m.

PIN: PSTN - IP Notification

With the proliferation and wide acceptance of the Internet, and more so
with the convergence of the Internet and PSTN, there is an increasing
desire for events occurring on the PSTN domain to be propagated to the
Internet domain.

The protocol of the PINT working group is used to pass requests for
service from an IP host to a telephone network, and to receive responses
back from the telephone network. But some Interworking services require
that requests for services go in the opposite direction: from a
telephone network to an IP host.

The Internet Draft "Need for PSTN Internet Notification (PIN) Services"
is available from the official IETF directory and explains the need and
gives some requirements for such requests for service and for
notification.

The PIN BOF will examine the sorts of Telephone Services that present
requirements for such requests from a telephone network to a set of IP
hosts. It is important that we agree on the scope of requests for
service and notification that such services might generate. There are
various working groups and protocols (most notably IPTEL, PINT, and
MEGACO) which might have relevance for the sort of services we want to
enable.  In addition to a discussion of the scope and requirements of an
eventual PIN working group, the PIN BOF will need to understand if those
requirements are being fulfilled by these or other working groups.

At the BOF there will be a series of short presentations of various
solutions in industry use.  These will be examined only to help us
understand the scope of the services we want to support.  Our goal is to
decide whether or not a new protocol or protocol family is
needed, and, if so, what its scope and requirements should be.

The PIN mailing list is located at
ietf-pin@lists.research.bell-labs.com.

To subscribe or unsubscribe yourself to the mailing list, send E-mail to
ietf-pin-request@lists.research.bell-labs.com with the single word
"subscribe" or "unsubscribe" (without the quotes) in the body of the
message. Please do not send subscription requests to
ietf-pin@lists.research.bell-labs.com.

Regards,
Alec Brusilovsky

--------------32C77E3858257B4CB7FC4132
Content-Type: text/x-vcard; charset=us-ascii;
 name="abrusilovsky.vcf"
Content-Description: Card for Alec Brusilovsky
Content-Disposition: attachment;
 filename="abrusilovsky.vcf"
Content-Transfer-Encoding: 7bit

begin:vcard 
n:Brusilovsky;Alec 
x-mozilla-html:FALSE
org:Lucent (jg7290000)
version:2.1
email;internet:abrusilovsky@lucent.com
title:MTS 
tel;fax:+1 630 713 5840 
tel;work:+1 630 713 8401
adr;quoted-printable:;;IHP 1A-423=0D=0A263 Shuman Blvd,P O Box 3050;Naperville;IL;60566-7050;U S
x-mozilla-cpt:;0
fn:Alec  Brusilovsky
end:vcard

--------------32C77E3858257B4CB7FC4132--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 14:45:22 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09269
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 14:45:22 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 64A147FE9; Thu, 11 Mar 1999 12:01:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 452E27FEC; Thu, 11 Mar 1999 11:47:38 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01a301be6a6e$2ff450c0$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Matt Squire" <msquire@nortelnetworks.com>
Cc: <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 12:46:15 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Matt Squire wrote:
>I've received similar comments from several providers.  This is one of
>things I was going to bring up when discussing general data representation
>issues.  Scope has applicability to 1-800 numbers, but it also has
>applicability to private/corporate numbers.

< some comments cut>

>What do folks think about the necessity of a 'scope' attribute for
> (a) 1-800
> (b) corporate/private addresses
>(ie would I be wasting my time trying to address this concern?)


I think as long as there is a real world requirement for it, the protocol
would need to support it.

With the peer group and a hierarchy scheme I outlined earlier, a scope
becomes a proper definition/configuration of a peer group. A gateway
can belong to more than one peer group on one level. The members
in a peer group does not need to be geographically/topologically
adjacent. Corporate/private numbers can be supported by defining
a peer group for the corporate VPN and having the gateways in the
peer group exchanging their number advertisement privately within
the peer group.

- Greg


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 15:03:18 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09597
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 15:03:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 773C87FE0; Thu, 11 Mar 1999 14:31:34 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 14BD780A9; Thu, 11 Mar 1999 11:45:12 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01c201be6a77$ac69cd20$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        "Matt Squire" <msquire@nortelnetworks.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 13:55:49 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit


-----Original Message-----
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
To: Greg Zhang <gzhang@utstar.com>
Cc: hsalama@cisco.com <hsalama@cisco.com>; John Schnizlein
<jschnizl@cisco.com>; David R. Oran <oran@cisco.com>; Matt Squire
<msquire@nortelnetworks.com>; Scott Petrack <scott.petrack@metatel.com>;
Alok Mittal <amittal@hss.hns.com>; iptel
<iptel@lists.research.bell-labs.com>
Date: Tuesday, March 09, 1999 1:08 PM
Subject: Re: GLP, number portability and wireless subscribers


Jonathan Rosenberg wrote:
>Greg Zhang wrote:
>>
>> One approach is to introduce hierarchy among the GLP speaking Internet
>> telephony gateways. Instead of having a flat network of gateways, a
>> hierarchy is built among the gateways. In a way that is similar to PNNI,
>> the gateways becomes members of a peer group (PG). A peer group
>> leader (PGL) is elected in a group. PGLs in one level becomes peers
>> on the next level. And the process of PGL election process repeats
>> on the new level. The hierarchy building is finished when all gateways
>> are tied into one peer group at the highest global level.
>
>The use of a hierarchical arrangement of LS's seems orthogonal to the
>scope issue. We would still need to specify the encoding of scopes,
>interpretation of scopes, and be able to configure LS's about whether to
>propagate routes with particular scopes or not.

I agree with the spirit of the above statement. However, if the scope can
be expressed in simply the number of level or a list of peer group ids
that a advertisement gets to propagate and aggregated to, we would
not need to administrate, propagate and manage a separate sets of
scope code throughout the network. We can then have a simple solution
that serves our needs. Each gateway receives its nodal hierarchy
information from the peer group leaders in the hierarchy.

In a sense, the administration/configuration of peer group ids and the
hierarchy becomes the administration and interpretation of the scope
ids.

As I mentioned earlier, a gateway can be made to belong to more than
one peer group on one particular level with some of its advertisement
goes to one peer group while others goes to another peer group. The
peer group list that is contained in the original advertisement determines
where the advertisement gets sent to. The list can contain both vertical
and lateral peer groups in the hierarchy.

>In a sense,
>the current model contains a two-level hierachy. How does expanding this
>hierarchy make the problem easier? It seems to complicate the protocol,
>if anything.

Adding more levels to the hierarchy provides better scalability by limiting
the synchronization of distributed caches or databases to a smaller
scope with fewer members on each level. By having one protocol with
multiple levels instead of two protocols one for intra domain, one for
inter domain lets you avoid having to develop two separate pieces of
protocols and code. And it makes the administration and maintenance
of the gateways a little easier by only requires the network administrator
to deal with one protocol for configuration and debug.

Having said that, the implementation of the protocol can be gradual.
In the first release, one can choose to only support the lowest level with
future releases supporting more or all levels. This allows us to have
something that is reflexible, capable of dealing with topological changes,
and is scalable to a certain degree up and running relatively quickly. We
can also learn from the experience when adding the support for the full
hierarchy.

I understand that the approach is somewhat different than the traditional
approach used in the TCP/IP world. I would appreciate it if it can
commented and debated on its merits in the group.

- Greg





---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 15:20:45 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09952
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 15:20:44 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A26107FF8; Thu, 11 Mar 1999 12:04:54 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3BE9B7FF9; Thu, 11 Mar 1999 11:45:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990309170219.00aa9dc0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Tue, 09 Mar 1999 17:02:29 -0500
To: "David R. Oran" <oran@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: RE: GLP, number portability and wireless subscribers
Cc: hsalama <hsalama@cisco.com>, John Schnizlein <jschnizl@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 04:34 PM 3/9/99 -0500, David R. Oran wrote:
>Or, alternatively, project all numbers into a global, unambiguous space.
>This seems cleaner and much less error prone than topologically
>established scope boundaries, and less administratively burdensome than
>solving the problem of scope name collisions and information propagation
>boundaries. It also avoids having to talk about translations between
>scopes, which I'm sure will happen when the other shoe drops :-).
>

I'm a little confused by this comment. To me, scopes provide the
unambiguous global number space.  If there is a scope registration
authority (maybe this wasn't clear before), then the combination of
scope+number is globally unambiguous, with the default scope being global
e164.  If someone registers a scope and they wish it to have topological
significance, then that's an administrative matter and they should be able
to set up their peer association and policy accordingly.  

I'd probably be against scopes of the form "+1-405" because that would just
lead to too many headaches in interpreting, aggregating, etc.   

An opaque scope using a equal/!equal test might provide what we desire.  I
haven't convinced myself completely, but I think it has the potential.

- Matt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 15:38:13 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA10350
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 15:38:13 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C13CD8002; Thu, 11 Mar 1999 11:58:07 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 194887FF0; Thu, 11 Mar 1999 11:51:37 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01b801be6a72$9dfef3a0$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "Scott Petrack" <scott.petrack@metatel.com>,
        "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>
Cc: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "John Schnizlein" <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        "Matt Squire" <msquire@nortelnetworks.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 13:19:37 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Scott Petrack wrote:
>Well, yes, but there are lots of 800 numbers whose translations change on
>time scales MUCH slower than the GLP advertisement changes. In fact, most
>800 numbers are this way -- they are used by small-to-medium
>businesses, point to the 'real' telephone number, with no
>logic at all, and don't change at all.
>
>If a GW wants to advertise that it can service such a number, I don't see
>why it should be forbidden from doing so.

I agree.


>The general problem of for how long the advertisements are valid, and how
>quickly change can be propogated, is just that general -- there is nothing
>special about 800 numbers or other directory numbers.


The duration of the validity of a advertisement should be decision of the
originator. One approach to have a time to live field in the advertisement
and letting it to expire if it has not been refreshed. Another approach
is to let the advertisement lives to perpetuity unless it specifically
updated or canceled by the originator. A Hello mechanisms is employed
to maintain the status of connectivity to the originator. When the
connection
to the originator is lost, all its entries will be removed from the
database.

The second approach will limit the updates to only the changes that have
occurred. There is no need to send refreshes periodically. The overhead
of hello messages are very low comparing with the amount of data in
the actual advertisement.

Aging management of this potentially large database is also greatly
simplified. There is no need for a aging timer that goes off periodically
and a aging handler to go through the entire database to decrement time
to live field, remove the expired ones, and possibly call a handler to
notify
the expiration. Instead, the aging is done on a per originator bases. You
only need to maintain a time or counter to indicate when the last hello was
received. When it exceeds the hello interval * dead factor, the originator
is
declared dead. All its entries get removed. Otherwise, they are all valid.

- Greg



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 15:56:17 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA10729
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 15:56:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7171280A5; Thu, 11 Mar 1999 11:23:59 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C8A118096; Thu, 11 Mar 1999 11:16:09 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <009301be6b48$0bbdad70$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "David R. Oran" <oran@cisco.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Matt Squire" <msquire@nortelnetworks.com>
Cc: "hsalama" <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Dynamic Cache Synchronization Protocol, was Re: GPL, number portability and wireless subscribers
Date: Wed, 10 Mar 1999 14:47:16 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hello, everyone!

Matt Squire wrote:
>It seems like a heirarchy could be built now (using either scsp or bgp
>approach) if each peer group/level was considered a domain, and we just
>defined filters so that at B, information coming from A didn't get sent
>down to C.   So I think one could deploy a heirarchy using static
>port-based filters, which requires some administrative work.  But I think
>Greg was after something more dynamic, where the attribute itself defined
>which ports it would go out on based on the peer/server group on each port.

Let me take the liberty to outline my ideas in this email.

What I was trying to suggest is a different way of defining domain and doing
cache synchronization in and among these domains. Rather than the
traditional AS approach as in BGP and using two different protocols in
intra domain and inter domain, we could define a multi level domain
or peer group and build a hierarchy upon them. The domains or peer
groups are identified by a 32 bit domain or peer group ids. The lowest level
nodes in the hierarchy are configured with a list peer group ids that it
belongs to. i.e.

1st level peer group ID
2nd level peer group ID
3rd level peer group ID
...

On this list, the levels are counted from bottom up. Each node will use
a neighbour discovery hello message sent out using flooding on all its
ports to discovery its neighbours. It is going to carry a small TTL value
that is greater than one, perhaps 5. This is to get pass the routers that
do not speak the protocol in between protocol speaking nodes.

The hello message will include a nodal hierarachy list with contains:

1st level peer group ID
1st level peer group leader
2nd level peer group ID
2nd level peer group leader
3rd level peer group ID
3rd level peer group leader
...

When a peer group leader is not known, its field is filled with a blank
value. On receipt of such hello message, a node is suppose to send
a hello message back using unicast. If a node does not hear back from
any neighbour after sending out its neighbour discovery hello a couple
of times, it will increase the TTL value of the message to try to reach
the neighbours in a wider area. A neighbour list can also be configured
into a node to let it comes in contact with none topologically adjacent
nodes.

Nodes belong to the same peer group on the 1st level will elect a peer
group leader (PGL). Whoever has the highest PGL priority value and
a highest IP address will become the leader. The use of priority value will
ease the job of configuration of peer group list to nodes. You only need
to configure a few node with a more complete peer group list on each
level for the whole hierarchy to build properly. The boundary nodes will
exchange hello messages. Through the hello messages, they will
discovery whether they belong to a same peer group on the 2nd level.
If so, they will each notify its peer group leader. The PGL in turn will
start to exchange hello messages to establish the group and elect its
leader. This process continues until the whole hierarchy is established.

The hello message exchange and the peer group election process is
an on going process. The formation of the peer groups and the hierarchy
will respond to network topology change such as partition or merge.

The nodal hierarchy list sent by peer group leaders on each level
will include a more complete information of the hierarchy. It gets
propagated downward at each level until all node in the hierarchy gets
a complete list of all its parents. Note that only the lowest level nodes
physical node in the network. The nodes on the upper levels are logical
nodes resides in a physical box. In theory, there is no limit on the
number of levels that a hierarchy can have.

With formation of peer groups, members can exchange packets to
synchronize their database or cache. This is done in a very SCSP
fashion. The type of information to synchronize is outside the scope
of the cache synchronization protocol. They are identified by a
protocol ids as in SCSP. The same synchronization protocol can be
used to do cache synchronization for mutiple protocols at the same
time. The difference with SCSP is that the cache synchronization
is among a hierarchy of peer groups rather than between two servers.
The protocol used among the members could be multicast, a flooding
scheme, or a broadcast within a domain. It is up to the peer group
leader to decide which mechanism to use.

Cache entries gets propagated up through the hierarchy by the peer
group leaders on each level. At PGL, the entries can get
aggregated, filtered, or just passed up with minor or no change
depend on the policies installed at PGL and some propagation
control field carried in the cache entry.

The cache entries are only propagated up through the hierarchy.
They are never propagated downward. When a look up fails in a
peer group, the request gets forwarded to its peer group leader.
The PGL will perform the look up on the next level up. The process
continues until a match is found or it fails after reaches the top level
of the hierarchy.

Each cache entry will have a originator field. At the lowest level, the
originator is a physical node identified by its IP address. At upper
levels is the lower level peer group identified by the peer group id.
It is originator's responsibility to update the entry content to keep it
valid. A peer group hello mechanism is used to monitor the
connectivity status among the members of a peer group. As long as the
connection to the originator is alive, its entries is considered valid.
This way, cache entries does not need to be refreshed periodically.
This cut down the update traffic considerably. The cache entries are
sent over the wire only when a change occurred or a new entry get
created.

To have scoping control on cache entry, the originator will
put a list of peer group ids in the cache entry. The list can
contain both vertical and horizontal peer groups in the
hierarchy. The entry only get propagated to the peer
groups in the list. A simpler but less expressive approach is
to have a simple number of levels that the entry gets
propagated up to from its current level.

The peer group is a very arbitrary collection of nodes. It is
not limited to geographical/topological adjacency. On one
level a node can belong to more than one peer group. The
policy inside the node and the originator of the cache
entries determine what information gets propagated to
which peer group. In effect, there could be multiple
hierarchies build with the protocol each carrying out
its cache synchronization in the network.

There are some nice properties with this approach including
the dynamic formation of the peer group and hierarchy, the
scalability provided for propagating both aggregatable and
unaggregatable information, and a easy mechanism to
control scope of information propagation.

I am still working on the details of the protocol. I have two immediate
applications of the protocol. First is to propagate location update
information for mobile telephone subscriber in a IP based mobile
telephone network. The second is to do IP mobility for extending
IP all the way to handset in such a network.

Since there is fair amount of overlap between my location update
need and GLP, I am presenting the design to the group. A lot of
concept here is borrowed from PNNI and SCSP. All I did was adapt
them to the IP world, add some flexibility into the design to make the
association of servers dynamic, and to make the whole thing more
scalable. I would appreciate greatly if the ideas I presented here
can be evaluated and possibly debated on its merit. I would be
grateful to any comment and input.

Best regards,

- Greg



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 16:14:29 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA11148
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 16:14:28 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id EB1F67FE3; Thu, 11 Mar 1999 12:04:51 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 90B758077; Thu, 11 Mar 1999 11:43:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E5A6C0.ED7501A1@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 17:54:56 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Greg Zhang <gzhang@utstar.com>
Cc: hsalama@cisco.com, John Schnizlein <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <01c201be6a77$ac69cd20$2821a8c0@greg_zhang.wacos.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Greg Zhang wrote:
> 
> >In a sense,
> >the current model contains a two-level hierachy. How does expanding this
> >hierarchy make the problem easier? It seems to complicate the protocol,
> >if anything.
> 
> Adding more levels to the hierarchy provides better scalability by limiting
> the synchronization of distributed caches or databases to a smaller
> scope with fewer members on each level. By having one protocol with
> multiple levels instead of two protocols one for intra domain, one for
> inter domain lets you avoid having to develop two separate pieces of
> protocols and code. And it makes the administration and maintenance
> of the gateways a little easier by only requires the network administrator
> to deal with one protocol for configuration and debug.

In the current model, synchronization of the databases occurs from a
point outward until an aggregation point, at which a route change is no
longer propagated. This effectively creates a hierarchy along the
boundaries of the addressing hierarchy. Of course, if the LS topology
doesn't follow the topology of the phone number allocations in the
gateways, aggregation will be more limited, and the system appears more
flat. However, a system with explicitly defined boundaries, as you
propose, will do no better in this scenario either. As such, I don't see
much advantage to explicitly defining hierarchies.

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 16:31:49 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA11693
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 16:31:49 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 09CDA803C; Thu, 11 Mar 1999 12:03:10 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id EC36E7FE3; Thu, 11 Mar 1999 11:47:36 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E58845.BA234BEB@cisco.com>
Date: Tue, 09 Mar 1999 12:44:53 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Greg Zhang <gzhang@utstar.com>, John Schnizlein <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <010901be6a62$ea0ea480$2821a8c0@greg_zhang.wacos.com> <36E57B5C.E357450@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Jonathan Rosenberg wrote:
> 
> Greg Zhang wrote:
> >
> > >This is another good example of why translation before gateway selection
> > >is a good idea. However, I'm not sure there is much we can do from a
> > >protocol perspective to prevent it. There may be valid reasons why a
> > >company would like to advertise an 800 number (only they own it, for
> > >example), and I'd rather not exclude this possibility. It seems like a
> > >sensible approach is to allow non-routing numbers in GLP, but mention
> > >that there are issues with them, and translation before routing is a
> > >good idea. Does this sound sensible?
> >
> > I think to handle this type of cases correctly, we need to introduce a
> > concept of scope for the advertisement. Gateways in CA will advertise
> > the 800 number for company X in CA while gateways in NY will
> > advertise the same 800 for company Y in NY. The advertisement from
> > the CA domain does not get propagated to the NY domain and vice
> > versa. This way the numbers that are locally significant and are being
> > administrated locally stays local. This however does not prevent the
> > same 800 number being used by upper level for a third customer.
> 
> A sensible idea, but seems hard to do. How does one represent scope? How
> does one encode the notion that "this number is valid within the greater
> new york metropolitan area", or "northern california" into a protocol?
> 
> Perhaps one solution is to allow scopes to be represented as opaque
> tokens. These tokens can be registered, perhaps with IANA, along with a
> desription of the what the scope means. This would allow peers to
> program their LS's to recognize certain scope tokens, and do something
> appropriate with them. Prefixes with unrecognized scope tokens probably
> shouldn't be propagated.
> 
> As an alternative, scope could be represented by a separate phone number
> prefix. This second prefix indicates the set of numbers you need to call
> from for this scope to be defined. For example:
> 
> phone numbers: +1-800-55*
> scope: +1-415*
> 
> means that 800-55* is only defined within parts of California.
> 

Jonathan,

I prefer the second alternative. The scope attribute shouldn't be opaque
to intermediate LSs. I think the scope will factor into aggregation and
route selection. For example, an LS shouldn't aggregate two
advertisements with different scopes. And if an LS receives multiple
advertisements for the same 800 number, each with a different scope, it
shouldn't apply route selection to these advertisements, but rather
continue advertising them all. I think that's necessary for GLP to work
in a topology where non-contiguous administrative domains can accept
calls from the same scope?

Sounds like source+destination based routing with the scope representing
the source prefix. Are there any prior experiences with such a scheme?

Hussein



> Any other ideas? Or do people think that this isn't needed?
> 
> -Jonathan R.
> 
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 17:17:28 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA13013
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 17:17:27 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6B8947FD2; Tue,  9 Mar 1999 12:07:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A76887FD3; Tue,  9 Mar 1999 12:07:28 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Tue, 9 Mar 1999 12:07:18 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: 800 numbers
In-Reply-To: <36E54E32.C8931AB2@dnrc.bell-labs.com>
Message-ID: <Pine.LNX.4.04.9903091159110.3156-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


The translation of "directory numbers" to "network numbers" can have a lot
of dependencies that are unrelated to GLP -- the calling party id, the
time of day, some ACD info. 

I suppose that GLP advertisements are meant to be valid for some time
which is related to an estimate of how long the particular gateway will be
in service, rather than (say) that it can service a number until close
of business at 5pm EST.

That being said, of course Dave is right that there are a lot of 800
numbers which are perfectly static, and it would be a very useful
optimization to be able to advertise those numbers. 

Bottom line: I don't see why we should forbid a gateway from advertising
an 800 number if it wishes to, but we still need a separate "directory
number to network number" lookup protocol. 

Scott 



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 17:59:00 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA14300
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 17:58:59 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1A7C18088; Thu, 11 Mar 1999 11:27:32 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id AD0768078; Thu, 11 Mar 1999 11:15:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D183@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'pcalhoun@eng.sun.com'" <Pat.Calhoun@Eng.Sun.COM>,
        Matt Squire <msquire@nortelnetworks.com>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 13:10:34 -0600 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hello,

I am in favor of Pat's point but I am still hoping that we can
achieve a sound solution in order for us to handle number portability
environment in the future, including wireless.

The issues I see is how to synchronize two database between GLP and SS7
world. In AIN, 800 number information are kept in not only Signaling 
Control Point(SCP) but also Service Switching Point(SSP), normally in 
local DMS-100 or 5ESS switches. The 800 calling process had to go through 
a conversation session between SSP and SCP to figure out the final
routing to the destination. There are two information databases been
kept on both SCP and local SSP. Do not forget that there are many SSPs
in the network, not like SCP, we have to deal with.
The major problem occur when a new portable number been provisioned,
into both SSP and SCP but there is no information on GLP side.
GLP and SS7 database is out of sync basically.

How to solve this database synchronization problem, I do not have
a solution for you but I hope that we can see the nature of the
problem first. 

Henry W. Lu
925-598-1245
SBC, Technology Resources, Inc.
4698 Willow Road
Pleasanton, CA 94588

-----Original Message-----
From: pcalhoun@eng.sun.com [mailto:Pat.Calhoun@Eng.Sun.COM]
Sent: Tuesday, March 09, 1999 9:15 AM
To: Matt Squire
Cc: pcalhoun@eng.sun.com; Jonathan Rosenberg; John Schnizlein; Greg
Zhang; David R. Oran; Scott Petrack; Alok Mittal; iptel
Subject: Re: GLP, number portability and wireless subscribers


> 
> >> This begs the question, then: does the translation of the 800 number
get
> >> done before the gateway selection process, or after? By doing it
before,
> >> a gateway can be chosen which is, in fact, close to the real number
> >> being dialed. With 800 numbers, all gateways will be able to reach it.
> >> It seems that this opens the possibility of very non-optimal routing. A
> >> caller in NY that dials an 800 number, goes to a gateway in CA. The
> >> phone network there determines that the 800 number translates to a
> >> number in NY, and routes the call through the PSTN back to NY. However,
> >> if the translation occurs before gateway selection, the caller can
> >> choose a local NY gateway and do much better.
> >
> >Translation before the gateway selection is *obviously* the correct
model.
> We >should strive towards this goal.
> >
> >PatC
> >
> >
> 
> The next question is when before gateway selection.  Translation could be
> done as a pre-cursor to call-routing by the ingress, could be done by
> signalling itself via things like SIP redirections/proxies, or could be
> done within the routing infrastructure itself.  
> 
> I'd prefer one of the first two alternatives.  
> 
That seems to be the one that is least disruptive to the Internet (this is
my
perception). We are after all simply creating a new service that runs on top
of the Internet, so it does not seem wise to have to play with the routing
infrastructure.

PatC


---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 18:27:17 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA14900
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 18:27:16 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2A70D8003; Thu, 11 Mar 1999 11:56:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 324568002; Thu, 11 Mar 1999 11:50:41 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>, <hsalama@cisco.com>
Cc: "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 16:12:08 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFEEJNEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Importance: Normal
In-Reply-To: <36E56BB7.E1A0DB91@dnrc.bell-labs.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

I'm reluctant to bring this up, because some would view it as a corner
case or just a "feeeeture" of the PSTN, but when the PSTN does number
translation (or call forwarding in the network as well), the resulting
number is not revealed to the calling party. The pure directory scheme
would reveal the translated number since that's what would be put in the
end-to-end signaling (e.g. SIP, H.323). Do we care about this?

> -----Original Message-----
> From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
> Sent: Tuesday, March 09, 1999 1:43 PM
> To: hsalama@cisco.com
> Cc: John Schnizlein; Greg Zhang; David R. Oran; Matt Squire; Scott
> Petrack; Alok Mittal; iptel
> Subject: Re: GLP, number portability and wireless subscribers
>
>
> Hussein F. Salama wrote:
> >
> > Jonathan Rosenberg wrote:
> >
> > >
> > > This begs the question, then: does the translation of the
> 800 number get
> > > done before the gateway selection process, or after? By
> doing it before,
> > > a gateway can be chosen which is, in fact, close to the
> real number
> > > being dialed. With 800 numbers, all gateways will be able
> to reach it.
> > > It seems that this opens the possibility of very
> non-optimal routing. A
> > > caller in NY that dials an 800 number, goes to a gateway
> in CA. The
> > > phone network there determines that the 800 number translates to a
> > > number in NY, and routes the call through the PSTN back
> to NY. However,
> > > if the translation occurs before gateway selection, the caller can
> > > choose a local NY gateway and do much better.
> >
> > Because of the above example, I think all types of translation (800
> > numbers or LNP) and mapping should occur prior to call
> routing. This is
> > particularly important for 800 numbers, because they have a regional
> > scope in many cases. The same 800 number may owned by
> company X in CA
> > and company Y in NY. In the above example, the call will be
> incorrectly
> > routed to company X instead of going to the originally intended
> > destination, company Y.
>
> This is another good example of why translation before
> gateway selection
> is a good idea. However, I'm not sure there is much we can do from a
> protocol perspective to prevent it. There may be valid reasons why a
> company would like to advertise an 800 number (only they own it, for
> example), and I'd rather not exclude this possibility. It seems like a
> sensible approach is to allow non-routing numbers in GLP, but mention
> that there are issues with them, and translation before routing is a
> good idea. Does this sound sensible?
>
> Scott P., of course, is right, that the translation service itself is
> outside the scope of iptel.
>
> -Jonathan R.
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 18:50:03 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA15464
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 18:50:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 37E3E80EF; Thu, 11 Mar 1999 18:38:21 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6FB078100; Thu, 11 Mar 1999 18:38:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D18A@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'John Schnizlein'" <jschnizl@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Greg Zhang <gzhang@utstar.com>, "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Thu, 11 Mar 1999 16:48:33 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear John,

>One way to deal with it is for gateways to perform only as much lookup
>as they desire to avoid poor routes until the directory standards settle.
>Even then, would is be OK for some gateways to use poor routes as a local
>option to using the global directory?

The answer to your question will really depending on how gateway getting 
its response from SS7 network. If the gateway just query SCP, then we may
have problem. The normal process works in PSTN is that when a call initiated

from local switch, it consult SCP when destination is not the same prefix. 
I do not see there is a reverse query, from SCP to consult local switches, 
because SCP then has to have all switches information to route this
signaling request. It's not a good news but I hope it will help. Thanks.

Henry W. Lu
SBC, Technology Resources, Inc.
4698 Willow Road,
Pleasanton, CA 94588

-----Original Message-----
From: John Schnizlein [mailto:jschnizl@cisco.com]
Sent: Tuesday, March 09, 1999 9:49 AM
To: Matt Squire; pcalhoun@eng.sun.com; Jonathan Rosenberg
Cc: Greg Zhang; David R. Oran; Scott Petrack; Alok Mittal; iptel
Subject: Re: GLP, number portability and wireless subscribers


Notice that names, aka "directory numbers", don't look different from 
addresses, aka "routing numbers" in the PSTN as they do on the Internet.
The tradeoff is that an early lookup can avoid extra call routing,
but that cost is saved only for the relatively few numbers that are ported.

If most numbers really needed lookup, a (LDAP) directory with its schema
reflecting the hierarchy of delegation of ported numbers would be better
than the tedious process used in the PSTN to manage the LPN database.
It might require extensions to LDAP because the parsing of a phone number
into the delegation hierarchy is different in different parts of the world.

The other obstacle to this early-translation approach is the administration
of this directory. While the rudiments of distributing subtrees of the
directory, referrals to other authorities, access control for changing 
pointers, and replication to avoid the chain of queries are specified;
interoperability of these features are still progressing through LDAPext
and LDUP working groups.

One way to deal with it is for gateways to perform only as much lookup
as they desire to avoid poor routes until the directory standards settle.
Even then, would is be OK for some gateways to use poor routes as a local
option to using the global directory?

Even though standardizing these issues is outside the scope of IPTEL,
it is important that the (external) solution is feasible withing the
scope of existing IETF working group charters.

John

At 12:09 PM 3/9/99 -0500, Matt Squire wrote:

>>Translation before the gateway selection is *obviously* the correct model.

>>We should strive towards this goal.
>>
>>PatC
>
>The next question is when before gateway selection.  Translation 
>could be done as a pre-cursor to call-routing by the ingress, 
>could be done by signalling itself via things like SIP
redirections/proxies, 
>or could be done within the routing infrastructure itself.  
>
>I'd prefer one of the first two alternatives.  
>
>- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 19:03:21 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA15793
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 19:03:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 95F408027; Thu, 11 Mar 1999 17:48:39 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D491380E2; Thu, 11 Mar 1999 17:32:30 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <36E6A165.62F81438@lucent.com>
Date: Wed, 10 Mar 1999 10:44:21 -0600
From: "ronald h. davis" <ronaldd@lucent.com>
Organization: Lucent Technologies
X-Mailer: Mozilla 4.05 [en]C-EMS-1.4  (WinNT; U)
MIME-Version: 1.0
To: rtgupta@hss.hns.com
Original-CC: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject: Re: Handling of Setup message by Gatekeeper.
References: <65256730.001BC1A0.00@sampark.hss.hns.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

rtgupta@hss.hns.com wrote:
> 
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
> 

i think that i don't understand what you're trying to say.  the setup
message is an h.225 call signalling message, not a ras message.

-- 
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
"if you ain't the lead dog, the scenery never changes"  -- edmund wilson

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 19:13:33 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA16026
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 19:13:33 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id AE9D6806D; Thu, 11 Mar 1999 18:51:47 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 70CF2813B; Thu, 11 Mar 1999 18:51:29 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E8558E.268D6203@dnrc.bell-labs.com>
Date: Thu, 11 Mar 1999 18:45:18 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: Christian Huitema <huitema@research.bellcore.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        "David R. Oran" <oran@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        John Schnizlein <jschnizl@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <3.0.32.19990309170219.00aa9dc0@bl-mail1.corpeast.baynetworks.com> <990311181245.ZM1353@seawind.research.bellcore.com> <36E8543D.240CFE28@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> Christian Huitema wrote:
> >
> > Well, it seems that we are back again to the distinction between naming
> > and routing. With GLP, the phone number in the E164 space is really an
> > alias to a "local" phone number.  Basically, the translation is:
> >
> > a) use the number to obtain the id of the local provider,
> >
> > b) route to the local provider
> >
> 
> But there may be multiple local providers (multiple egress gateways
> owned by different providers) servicing the same number. Therefore, I
> think we should route to the destination number. Then local provider
> selection will automatically occur as part of route selection, and it
> can still be controlled by injecting appropriate policy at the various
> GLP speakers.

You took the words right out of my mouth (well, keyboard perhaps).

-Jonathan R.


-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 20:43:40 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA17032
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 20:43:39 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8B59D8067; Thu, 11 Mar 1999 16:59:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4D0CD7FFA; Thu, 11 Mar 1999 16:13:39 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E82200.B3CDC3E1@dnrc.bell-labs.com>
Date: Thu, 11 Mar 1999 15:05:20 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: list iptel <iptel@lists.research.bell-labs.com>
Subject: TEST - please ignore
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

test

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 20:48:17 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA17055
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 20:48:16 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C2E8F7FEE; Thu, 11 Mar 1999 11:53:04 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7B0FD804E; Thu, 11 Mar 1999 11:49:34 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E58D66.E0CD077A@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 16:06:46 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Greg Zhang <gzhang@utstar.com>
Cc: hsalama@cisco.com, John Schnizlein <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <013b01be6a69$403925f0$2821a8c0@greg_zhang.wacos.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Greg Zhang wrote:
> 
> One approach is to introduce hierarchy among the GLP speaking Internet
> telephony gateways. Instead of having a flat network of gateways, a
> hierarchy is built among the gateways. In a way that is similar to PNNI,
> the gateways becomes members of a peer group (PG). A peer group
> leader (PGL) is elected in a group. PGLs in one level becomes peers
> on the next level. And the process of PGL election process repeats
> on the new level. The hierarchy building is finished when all gateways
> are tied into one peer group at the highest global level.

The use of a hierarchical arrangement of LS's seems orthogonal to the
scope issue. We would still need to specify the encoding of scopes,
interpretation of scopes, and be able to configure LS's about whether to
propagate routes with particular scopes or not. We'll need to do this
with the current intra-domain and inter-domain model also. In a sense,
the current model contains a two-level hierachy. How does expanding this
hierarchy make the problem easier? It seems to complicate the protocol,
if anything. 

-Jonathan R.


-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 21:04:27 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA17205
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 21:04:27 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0E4F48005; Thu, 11 Mar 1999 20:57:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9CDF48140; Thu, 11 Mar 1999 19:57:13 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Thu, 11 Mar 1999 19:22:17 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: "David R. Oran" <oran@cisco.com>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>, hsalama@cisco.com,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
In-Reply-To: <NBBBIFCAKKOPNMMKHHEFEEJNEEAA.oran@cisco.com>
Message-ID: <Pine.LNX.4.04.9903111917110.8137-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

On Tue, 9 Mar 1999, David R. Oran wrote:

> I'm reluctant to bring this up, because some would view it as a corner
> case or just a "feeeeture" of the PSTN, but when the PSTN does number
> translation (or call forwarding in the network as well), the resulting
> number is not revealed to the calling party. The pure directory scheme
> would reveal the translated number since that's what would be put in the
> end-to-end signaling (e.g. SIP, H.323). Do we care about this?

This question has nothing to do with PSTN gateways or GLP, right? It has
to do with a system based on end-to-end signalling, like SIP and H.323.

If a called party doesn't want to reveal his/her "real" address, then s/he
must set up a SIP proxy server which will get the invitation to the "fake"
address, and the proxy will pass the invitation to the real address. 

In other words, if you need this feeeeeture within a SIP/H.323 system,
then you need to use a SIP proxy server or H.323 GK routed signalling, and
then the "translation" service will return the "fake" routing number.

Scott


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 21:13:40 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA17238
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 21:13:39 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 54F36815C; Thu, 11 Mar 1999 20:20:24 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9801E8161; Thu, 11 Mar 1999 20:20:03 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Hanlin_Fang@3com.com
X-Lotus-FromDomain: 3COM
To: Deepak Kumar <deepak@microsoft.com>
Cc: "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>,
        "ronald h. davis" <ronaldd@LUCENT.COM>, h323implementors@imtc.org,
        iptel@lists.research.bell-labs.com
Message-ID: <88256732.000726A2.00@hqoutbound.ops.3com.com>
Date: Thu, 11 Mar 1999 17:15:37 -0800
Subject: RE: Handling of Setup message by Gatekeeper.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



How can Gatekeepers know that there are some endpoints in its zone not
registered? If GKs don't know their existence, then how can those endpoints be
managed?
Thanks for any information

Regards!

Hanlin




Deepak Kumar <deepak@microsoft.com> on 03/09/99 10:26:38 PM

To:   "'rtgupta @hss.hns.com'" <rtgupta@hss.hns.com>, "ronald h. davis"
      <ronaldd@LUCENT.COM>
cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com (Hanlin
      Fang/HQ/3Com)
Subject:  RE: Handling of Setup message by Gatekeeper.




Q.931 could originate at a rogue (ie. in the physical zone, but
unregistered, and doesn't send ARQ) end-point or an end-point/GK outside the
zone. Why is a distinction between these two scenarios meaningful?

An end-point that chooses (illegally) not to register with its Gatekeeper
could be treated exactly like an end-point outside the zone, and be denied
the benefits of Gatekeeper support - such as the ability to recv incoming
calls, QoS reservations, and the ability to place and recv calls using
network resources such as Gateways and proxies.

Deepak Kumar,

Program Manager, IP Telephony,
Windows Networking and Communications,
Microsoft Corp.

> -----Original Message-----
> From:   rtgupta@hss.hns.com [SMTP:rtgupta@hss.hns.com]
> Sent:   Tuesday, March 09, 1999 9:19 PM
> To:     ronald h. davis
> Cc:     h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject:     Re: Handling of Setup message by Gatekeeper.
>
> Hi Davis,
>     Thanks for your inputs.
>
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But
> the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
>
> In contrast to what you said, I think RAS signalling is must for all the
> endpoints in any GK's zone. RAS signaling is not required for endpoints,
> outside the zone. Now when GK receives a Setup message, it doesn't know
> whether it is from its own zone or from outside. Before doing any further
> checks, it needs to check that. And as I mentioned in my previous mails,
> to
> do that, it needs to maintain a static databse of its zone. But there is
> no
> mandatory field in Setup message,giving source info. So basic question is
> how to perform this check.
>
> regards,
> Ritu.
>
> _____________________________________________________________________
> Ritu Gupta                                   |       email:
> rtgupta@hss.hns.com
> Senior Software Engineer,       |                   rtgupta@hotmail.com
> Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
> Banglore, Karnataka                  |                    91-80-3374075(H)
> INDIA.                                            |       Fax    :
> 91-80-2286393
> _____________________________________________________________________
> Restlessness and discontent are the first necessities of progress.
>
>
>
>
>
> "ronald h. davis" <ronaldd@lucent.com> on 03/09/99 11:32:19 PM
>
> To:   rtgupta@hss.hns.com
> cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
> Subject:  Re: Handling of Setup message by Gatekeeper.
>
>
>
>
> rtgupta@hss.hns.com wrote:
> >
>                [scenarios deleted]
> > In all these cases, gatekeeper will receive Setup message without
> receiving
> > ARQ message. So how can gatekeeper check whether ARQ is expected from
> the
> > endpoint, which sent the Setup or not? No field in the Setup message
> seem
> > to give this information. If gatekeeper doesn't check for this, that
> means
> > even the endpoints in its own zone can bypass gatekeeper for RAS
> > signalling?
> >
> > Any help/information on this issue will be appreciated.
> >
> the gatekeeper is only responsible for arq/acf/arj with endpoints within
> its zone.  all endpoints within the zone establish a ras channel (over
> which these messages are sent) with their gatekeeper.
> endpoints in other zones don't establish a ras channel with this
> gatekeeper.  thus, when calls are received from outside of the zone
> the recipient (be through the direct or gatekeeper routed call model)
> has no idea how the calling endpoint came upon sending the setup
> message.
> this statement may be controversial but it is my opinion that h.323
> is lan-oriented, so if you think of gatekeepers as administering
> bandwidth allocation, &c for endpoints on different lans it would
> be difficult for a gatekeeper to perform this administration for
> a different lan.  and it would be even harder for a gatekeeper to
> perform such administration over a wide area network consisting of
> a number of zones.
> this fact leads me to believe that the ras channel should be *optional*
> for h.323 gatekeepers.  to abstract your scenarios further it is
> possible that gatekeepers could act as application layer routers, in
> which case they are terminating h.225 signalling and h.245 media control
> channels for endpoints with which they have no ras channel.
> --
>   __  ______  __  / __/ |       lucent technologies, naperville il, usa
> _/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
> "if you ain't the lead dog, the scenery never changes"  -- edmund wilson
>
>
>
>
> -------------------------------------------------------------------------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"

---------
This message came from the IETF IPTEL Working Group Mailing List.







---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 21:19:19 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA17274
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 21:19:18 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 36D6180EE; Thu, 11 Mar 1999 20:45:34 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C6D9A8026; Thu, 11 Mar 1999 20:40:52 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <003801be6a9f$6d922070$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "David R. Oran" <oran@cisco.com>
Cc: "Matt Squire" <msquire@nortelnetworks.com>, "hsalama" <hsalama@cisco.com>,
        "John Schnizlein" <jschnizl@cisco.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 18:40:08 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Jonathan Rosenberg wrote:
>For example, for corporate addresses, I would say scopes are not needed.

In the case of VPN overlaid on top of the public network, a scope that is
not geographically or topologically based provides a easy solution to the
problem.

>For some local address, don't propagate it beyond the boundaries of the
>enterprise.

>No scope is needed since if you've received the prefix,
>you're allowed to call it.

Correct me if I have misunderstood the statement. You are allowed to
call any number you want at any given time. A subscriber normally
acquires a phone number to call out of band of any call routing or
location update protocol. Our job is to determine how to route the
call after the subcriber has dialed the digits. In the process, we need
to avoid route the call to a wrong location and to determine if a call
is misdialed and cannot be routed quickly.

>Same idea as enterprise IP networks with
>internal IP addressing that never goes outside. No special routing
>protocols needed,or extra attributes. Just a policy at the periphery
>which causes nothing to be propagated further.

Skip the following in case I am just stating the obvious here. There is
difference between a enterprise IP network implemented using NAT
on firewalls and the IP telephony network that we are discussing here.
In IP telephony network, the gateways that are within a operator's network
is exposed to the outside world. You need to propagate the local addresses
beyond the boundaries of your network to allow other people to call your
subscriber. And the local address is not translated by a firewall to some
globally significant address. The local address is the global address.
With number portability, the local address in one operator's network does
not need to fall within a predefined prefix. To allow some numbers to
be only locally meaningful, you need to block them from being advertised
beyond its scope so that when people outside the scope dial the number,
they are not routed to the local number by accident. Whether the scope
itself has a hierarchy and be aggregatable is another matter though.

- Greg



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 21:27:30 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA17342
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 21:27:29 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 460F97FE5; Thu, 11 Mar 1999 11:51:33 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DA0577FFD; Thu, 11 Mar 1999 11:42:34 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E57B5C.E357450@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 14:49:48 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Greg Zhang <gzhang@utstar.com>
Cc: hsalama@cisco.com, John Schnizlein <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <010901be6a62$ea0ea480$2821a8c0@greg_zhang.wacos.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Greg Zhang wrote:
> 
> >This is another good example of why translation before gateway selection
> >is a good idea. However, I'm not sure there is much we can do from a
> >protocol perspective to prevent it. There may be valid reasons why a
> >company would like to advertise an 800 number (only they own it, for
> >example), and I'd rather not exclude this possibility. It seems like a
> >sensible approach is to allow non-routing numbers in GLP, but mention
> >that there are issues with them, and translation before routing is a
> >good idea. Does this sound sensible?
> 
> I think to handle this type of cases correctly, we need to introduce a
> concept of scope for the advertisement. Gateways in CA will advertise
> the 800 number for company X in CA while gateways in NY will
> advertise the same 800 for company Y in NY. The advertisement from
> the CA domain does not get propagated to the NY domain and vice
> versa. This way the numbers that are locally significant and are being
> administrated locally stays local. This however does not prevent the
> same 800 number being used by upper level for a third customer.

A sensible idea, but seems hard to do. How does one represent scope? How
does one encode the notion that "this number is valid within the greater
new york metropolitan area", or "northern california" into a protocol? 

Perhaps one solution is to allow scopes to be represented as opaque
tokens. These tokens can be registered, perhaps with IANA, along with a
desription of the what the scope means. This would allow peers to
program their LS's to recognize certain scope tokens, and do something
appropriate with them. Prefixes with unrecognized scope tokens probably
shouldn't be propagated.

As an alternative, scope could be represented by a separate phone number
prefix. This second prefix indicates the set of numbers you need to call
from for this scope to be defined. For example:

phone numbers: +1-800-55*
scope: +1-415*

means that 800-55* is only defined within parts of California.

Any other ideas? Or do people think that this isn't needed?

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 21:33:36 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA17506
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 21:33:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BD5148022; Thu, 11 Mar 1999 17:38:52 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4DF537FB7; Thu, 11 Mar 1999 17:26:06 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E83998.395D893C@dnrc.bell-labs.com>
Date: Thu, 11 Mar 1999 16:46:00 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Matthew Squire <Matthew_Squire@BayNetworks.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: Mail backlog
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Folks,

Unfortunately, our mail server crashed on Tuesday, and has only now come
back up. There is a large mail backlog which is clearing. Things should
be back up to speed by Thursday evening. 

Sorry about that,
Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 21:41:50 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA17609
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 21:41:50 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0687780F2; Thu, 11 Mar 1999 17:44:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 63AFF801C; Thu, 11 Mar 1999 17:34:04 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990310080504.00aaed00@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Wed, 10 Mar 1999 08:05:05 -0500
To: "David R. Oran" <oran@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: RE: GLP, number portability and wireless subscribers
Cc: hsalama <hsalama@cisco.com>, John Schnizlein <jschnizl@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 06:14 PM 3/9/99 -0500, David R. Oran wrote:
>Then scopes are identical to a deeper hierarchy and splitting them out
>is simply a syntactical, not a semantic distinction, right?
>
>Do you think we would need hierarchical scopes? If so, do you see
>aggregation of hierarchical scopes? How would this differ from
>hierarchical aggregation of addresses? Are scopes always more global
>than addresses?
>
>Are we in sync here or am I being obtuse?
>

I think we just have many different ideas of what 'scope' means, and I
offered a poor explanation.

When this whole discussion started, I was in the middle of writing
something on the possibility of a 'vpn-scope' (I was unfortunately using
that word) to satisfy a requirement that I received via email from a couple
of providers.  I was going to suggest an attribute whose basic purpose was
to map overlapping address spaces into a globally unique number (ie
vpn-scope+number is globally unique).  

When I read Greg's ideas on heirarchical arrangements, it seemed like what
he wanted was a basic filtering mechanism so that if we have A as a parent
of B and B as a parent of C, then advertisments get propagated up, but not
down.   From one of Greg's notes:

>I agree with the spirit of the above statement. However, if the scope can
>be expressed in simply the number of level or a list of peer group ids
>that a advertisement gets to propagate and aggregated to, we would
>not need to administrate, propagate and manage a separate sets of
>scope code throughout the network. We can then have a simple solution
>that serves our needs. Each gateway receives its nodal hierarchy
>information from the peer group leaders in the hierarchy.


It seems like a heirarchy could be built now (using either scsp or bgp
approach) if each peer group/level was considered a domain, and we just
defined filters so that at B, information coming from A didn't get sent
down to C.   So I think one could deploy a heirarchy using static
port-based filters, which requires some administrative work.  But I think
Greg was after something more dynamic, where the attribute itself defined
which ports it would go out on based on the peer/server group on each port.
 I was hoping that, if we decide that that this support is necessary, a
single attribute could be used to satisfy both.  I'm not yet convinced that
its either necessary for us to support such dynamic heirarchical filtering,
nor that it can be done with a single attribute.  The more I look at it,
the more I think that different attributes would be necessary.

Based on the input from the providers and on the amount of work in various
ietf WGs on VPN definition and usage, I do think that a 'vpn-id' attribute
(I'll no longer refer to it as scope) would be useful.  In this case,
splitting it out is simply syntactical, and its used as a way to map
overlapping addresses into a unique global address. 

Jonathan's scope is completely different.

- Matt
 





---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 11 21:48:31 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA17641
	for <iptel-archive@odin.ietf.org>; Thu, 11 Mar 1999 21:48:31 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D3CF57FEC; Thu, 11 Mar 1999 12:03:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2FDF97FF8; Thu, 11 Mar 1999 11:46:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <013b01be6a69$403925f0$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        "Matt Squire" <msquire@nortelnetworks.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 12:12:37 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

>> I think to handle this type of cases correctly, we need to introduce a
>> concept of scope for the advertisement. Gateways in CA will advertise
>> the 800 number for company X in CA while gateways in NY will
>> advertise the same 800 for company Y in NY. The advertisement from
>> the CA domain does not get propagated to the NY domain and vice
>> versa. This way the numbers that are locally significant and are being
>> administrated locally stays local. This however does not prevent the
>> same 800 number being used by upper level for a third customer.
>
>A sensible idea, but seems hard to do. How does one represent scope? How
>does one encode the notion that "this number is valid within the greater
>new york metropolitan area", or "northern california" into a protocol?

One approach is to introduce hierarchy among the GLP speaking Internet
telephony gateways. Instead of having a flat network of gateways, a
hierarchy is built among the gateways. In a way that is similar to PNNI,
the gateways becomes members of a peer group (PG). A peer group
leader (PGL) is elected in a group. PGLs in one level becomes peers
on the next level. And the process of PGL election process repeats
on the new level. The hierarchy building is finished when all gateways
are tied into one peer group at the highest global level.


Within a peer group, the relationships are peer to peer. The distributed
database is synchronized among all members in the peer group. At
the peer group leader, policies are installed to control the aggregation
and propagation of content of the database to the upper level. This way
a peer group becomes a natural scope. For numbers that are only
locally meaningful, they are not propagated by the peer group leader
to the next level.

With the establishment of peer groups and the hierarchy, updates are
limited to a peer group. They can be aggregated at peer group leader.
Updates and synchronization of a database are no longer global.
This will give the protocol a higher degree of scalability. The forming
of the peer groups and the hierarchy is dynamic. It allows the protocol
to react to the changes in topology or connectivity in the network.

- Greg


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 02:24:23 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA28849
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 02:24:23 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 95BE97FE8; Thu, 11 Mar 1999 11:42:40 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 309B37FE7; Thu, 11 Mar 1999 11:40:29 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <010901be6a62$ea0ea480$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>, <hsalama@cisco.com>
Cc: "John Schnizlein" <jschnizl@cisco.com>, "David R. Oran" <oran@cisco.com>,
        "Matt Squire" <msquire@nortelnetworks.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 11:27:15 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

>> Because of the above example, I think all types of translation (800
>> numbers or LNP) and mapping should occur prior to call routing. This is
>> particularly important for 800 numbers, because they have a regional
>> scope in many cases. The same 800 number may owned by company X in CA
>> and company Y in NY. In the above example, the call will be incorrectly
>> routed to company X instead of going to the originally intended
>> destination, company Y.
>
>This is another good example of why translation before gateway selection
>is a good idea. However, I'm not sure there is much we can do from a
>protocol perspective to prevent it. There may be valid reasons why a
>company would like to advertise an 800 number (only they own it, for
>example), and I'd rather not exclude this possibility. It seems like a
>sensible approach is to allow non-routing numbers in GLP, but mention
>that there are issues with them, and translation before routing is a
>good idea. Does this sound sensible?

I think to handle this type of cases correctly, we need to introduce a
concept of scope for the advertisement. Gateways in CA will advertise
the 800 number for company X in CA while gateways in NY will
advertise the same 800 for company Y in NY. The advertisement from
the CA domain does not get propagated to the NY domain and vice
versa. This way the numbers that are locally significant and are being
administrated locally stays local. This however does not prevent the
same 800 number being used by upper level for a third customer.

For example, while company X's 800 number is administrated by
PacBell and company Y's 800 number is administrated by NYNEX,
AT&T can assign the same 800 number to company Z located in
Houston, Texas.

The design of the protocol should support these scenarios. It is
up to the telephone network operators to sort out the issues of
conflict and confusion caused by the assignment of same numbers
to different entities in different part or level in the network. These
are business decisions. Ultimately, they are driven by the customers'
need and public's ability to handle the possible confusions. The
protocol should provide mechanisms to carry it out.

- Greg



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 02:35:08 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA28937
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 02:35:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 264E78195; Fri, 12 Mar 1999 02:15:44 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 19E098193; Fri, 12 Mar 1999 02:11:33 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BE6C15.84379220.umesh@seeyes.co.in>
From: Umesh Mallugari <umesh@seeyes.co.in>
Reply-To: "umesh@seeyes.co.in" <umesh@seeyes.co.in>
To: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Cc: "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>,
        "'rkumar@holontech.com'" <rkumar@holontech.com>
Subject: RE: Handling of Setup message by Gatekeeper.
Date: Thu, 11 Mar 1999 23:18:47 -0800
Organization: SeeYes Network Technologies Pvt Ltd
X-Mailer: Microsoft Internet E-mail/MAPI - 8.0.0.4211
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Ritu,

Won' t the fields "sourceAddress" and "sourceInfo" in the Setup UUIE be of any help to the Gatekeeper
to perform checks on the Setup message received?

-Umesh.

-----Original Message-----
From:	Hanlin_Fang@3com.com [SMTP:Hanlin_Fang@3com.com]
Sent:	Thursday, March 11, 1999 5:16 PM
To:	Deepak Kumar
Cc:	'rtgupta@hss.hns.com'; ronald h. davis; h323implementors@imtc.org; iptel@lists.research.bell-labs.com
Subject:	RE: Handling of Setup message by Gatekeeper.



How can Gatekeepers know that there are some endpoints in its zone not
registered? If GKs don't know their existence, then how can those endpoints be
managed?
Thanks for any information

Regards!

Hanlin




Deepak Kumar <deepak@microsoft.com> on 03/09/99 10:26:38 PM

To:   "'rtgupta @hss.hns.com'" <rtgupta@hss.hns.com>, "ronald h. davis"
      <ronaldd@LUCENT.COM>
cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com (Hanlin
      Fang/HQ/3Com)
Subject:  RE: Handling of Setup message by Gatekeeper.




Q.931 could originate at a rogue (ie. in the physical zone, but
unregistered, and doesn't send ARQ) end-point or an end-point/GK outside the
zone. Why is a distinction between these two scenarios meaningful?

An end-point that chooses (illegally) not to register with its Gatekeeper
could be treated exactly like an end-point outside the zone, and be denied
the benefits of Gatekeeper support - such as the ability to recv incoming
calls, QoS reservations, and the ability to place and recv calls using
network resources such as Gateways and proxies.

Deepak Kumar,

Program Manager, IP Telephony,
Windows Networking and Communications,
Microsoft Corp.

> -----Original Message-----
> From:   rtgupta@hss.hns.com [SMTP:rtgupta@hss.hns.com]
> Sent:   Tuesday, March 09, 1999 9:19 PM
> To:     ronald h. davis
> Cc:     h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject:     Re: Handling of Setup message by Gatekeeper.
>
> Hi Davis,
>     Thanks for your inputs.
>
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But
> the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
>
> In contrast to what you said, I think RAS signalling is must for all the
> endpoints in any GK's zone. RAS signaling is not required for endpoints,
> outside the zone. Now when GK receives a Setup message, it doesn't know
> whether it is from its own zone or from outside. Before doing any further
> checks, it needs to check that. And as I mentioned in my previous mails,
> to
> do that, it needs to maintain a static databse of its zone. But there is
> no
> mandatory field in Setup message,giving source info. So basic question is
> how to perform this check.
>
> regards,
> Ritu.
>
> _____________________________________________________________________
> Ritu Gupta                                   |       email:
> rtgupta@hss.hns.com
> Senior Software Engineer,       |                   rtgupta@hotmail.com
> Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
> Banglore, Karnataka                  |                    91-80-3374075(H)
> INDIA.                                            |       Fax    :
> 91-80-2286393
> _____________________________________________________________________
> Restlessness and discontent are the first necessities of progress.
>
>
>
>
>
> "ronald h. davis" <ronaldd@lucent.com> on 03/09/99 11:32:19 PM
>
> To:   rtgupta@hss.hns.com
> cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
> Subject:  Re: Handling of Setup message by Gatekeeper.
>
>
>
>
> rtgupta@hss.hns.com wrote:
> >
>                [scenarios deleted]
> > In all these cases, gatekeeper will receive Setup message without
> receiving
> > ARQ message. So how can gatekeeper check whether ARQ is expected from
> the
> > endpoint, which sent the Setup or not? No field in the Setup message
> seem
> > to give this information. If gatekeeper doesn't check for this, that
> means
> > even the endpoints in its own zone can bypass gatekeeper for RAS
> > signalling?
> >
> > Any help/information on this issue will be appreciated.
> >
> the gatekeeper is only responsible for arq/acf/arj with endpoints within
> its zone.  all endpoints within the zone establish a ras channel (over
> which these messages are sent) with their gatekeeper.
> endpoints in other zones don't establish a ras channel with this
> gatekeeper.  thus, when calls are received from outside of the zone
> the recipient (be through the direct or gatekeeper routed call model)
> has no idea how the calling endpoint came upon sending the setup
> message.
> this statement may be controversial but it is my opinion that h.323
> is lan-oriented, so if you think of gatekeepers as administering
> bandwidth allocation, &c for endpoints on different lans it would
> be difficult for a gatekeeper to perform this administration for
> a different lan.  and it would be even harder for a gatekeeper to
> perform such administration over a wide area network consisting of
> a number of zones.
> this fact leads me to believe that the ras channel should be *optional*
> for h.323 gatekeepers.  to abstract your scenarios further it is
> possible that gatekeepers could act as application layer routers, in
> which case they are terminating h.225 signalling and h.245 media control
> channels for endpoints with which they have no ras channel.
> --
>   __  ______  __  / __/ |       lucent technologies, naperville il, usa
> _/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
> "if you ain't the lead dog, the scenery never changes"  -- edmund wilson
>
>
>
>
> -------------------------------------------------------------------------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"

---------
This message came from the IETF IPTEL Working Group Mailing List.







---------
This message came from the IETF IPTEL Working Group Mailing List.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 02:51:49 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA28995
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 02:51:48 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 42E198186; Fri, 12 Mar 1999 02:46:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B70D18035; Fri, 12 Mar 1999 02:45:28 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BE6C11.9D45E540.umesh@seeyes.co.in>
From: Umesh Mallugari <umesh@seeyes.co.in>
Reply-To: "umesh@seeyes.co.in" <umesh@seeyes.co.in>
To: "'Hanlin_Fang@3com.com'" <Hanlin_Fang@3com.com>,
        Deepak Kumar <deepak@microsoft.com>,
        "'rkumar@holontech.com'" <rkumar@holontech.com>
Cc: "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>,
        "ronald h. davis" <ronaldd@LUCENT.COM>,
        "h323implementors@imtc.org" <h323implementors@imtc.org>,
        "iptel@lists.research.bell-labs.com" <iptel@lists.research.bell-labs.com>
Subject: RE: Handling of Setup message by Gatekeeper.
Date: Thu, 11 Mar 1999 22:50:51 -0800
Organization: SeeYes Network Technologies Pvt Ltd
X-Mailer: Microsoft Internet E-mail/MAPI - 8.0.0.4211
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hanlin,

There is no way a Gatekeeper can know that an endpoint in its zone has not registered except when
the administrator has statically configured all the endpoint IDs into it.
Simply, out of the zone of a Gatekeeper ( unregistered with Gatekeeper ) means out of management.

-Umesh.

-----Original Message-----
From:	Hanlin_Fang@3com.com [SMTP:Hanlin_Fang@3com.com]
Sent:	Thursday, March 11, 1999 5:16 PM
To:	Deepak Kumar
Cc:	'rtgupta@hss.hns.com'; ronald h. davis; h323implementors@imtc.org; iptel@lists.research.bell-labs.com
Subject:	RE: Handling of Setup message by Gatekeeper.



How can Gatekeepers know that there are some endpoints in its zone not
registered? If GKs don't know their existence, then how can those endpoints be
managed?
Thanks for any information

Regards!

Hanlin




Deepak Kumar <deepak@microsoft.com> on 03/09/99 10:26:38 PM

To:   "'rtgupta @hss.hns.com'" <rtgupta@hss.hns.com>, "ronald h. davis"
      <ronaldd@LUCENT.COM>
cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com (Hanlin
      Fang/HQ/3Com)
Subject:  RE: Handling of Setup message by Gatekeeper.




Q.931 could originate at a rogue (ie. in the physical zone, but
unregistered, and doesn't send ARQ) end-point or an end-point/GK outside the
zone. Why is a distinction between these two scenarios meaningful?

An end-point that chooses (illegally) not to register with its Gatekeeper
could be treated exactly like an end-point outside the zone, and be denied
the benefits of Gatekeeper support - such as the ability to recv incoming
calls, QoS reservations, and the ability to place and recv calls using
network resources such as Gateways and proxies.

Deepak Kumar,

Program Manager, IP Telephony,
Windows Networking and Communications,
Microsoft Corp.

> -----Original Message-----
> From:   rtgupta@hss.hns.com [SMTP:rtgupta@hss.hns.com]
> Sent:   Tuesday, March 09, 1999 9:19 PM
> To:     ronald h. davis
> Cc:     h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject:     Re: Handling of Setup message by Gatekeeper.
>
> Hi Davis,
>     Thanks for your inputs.
>
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But
> the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
>
> In contrast to what you said, I think RAS signalling is must for all the
> endpoints in any GK's zone. RAS signaling is not required for endpoints,
> outside the zone. Now when GK receives a Setup message, it doesn't know
> whether it is from its own zone or from outside. Before doing any further
> checks, it needs to check that. And as I mentioned in my previous mails,
> to
> do that, it needs to maintain a static databse of its zone. But there is
> no
> mandatory field in Setup message,giving source info. So basic question is
> how to perform this check.
>
> regards,
> Ritu.
>
> _____________________________________________________________________
> Ritu Gupta                                   |       email:
> rtgupta@hss.hns.com
> Senior Software Engineer,       |                   rtgupta@hotmail.com
> Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
> Banglore, Karnataka                  |                    91-80-3374075(H)
> INDIA.                                            |       Fax    :
> 91-80-2286393
> _____________________________________________________________________
> Restlessness and discontent are the first necessities of progress.
>
>
>
>
>
> "ronald h. davis" <ronaldd@lucent.com> on 03/09/99 11:32:19 PM
>
> To:   rtgupta@hss.hns.com
> cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
> Subject:  Re: Handling of Setup message by Gatekeeper.
>
>
>
>
> rtgupta@hss.hns.com wrote:
> >
>                [scenarios deleted]
> > In all these cases, gatekeeper will receive Setup message without
> receiving
> > ARQ message. So how can gatekeeper check whether ARQ is expected from
> the
> > endpoint, which sent the Setup or not? No field in the Setup message
> seem
> > to give this information. If gatekeeper doesn't check for this, that
> means
> > even the endpoints in its own zone can bypass gatekeeper for RAS
> > signalling?
> >
> > Any help/information on this issue will be appreciated.
> >
> the gatekeeper is only responsible for arq/acf/arj with endpoints within
> its zone.  all endpoints within the zone establish a ras channel (over
> which these messages are sent) with their gatekeeper.
> endpoints in other zones don't establish a ras channel with this
> gatekeeper.  thus, when calls are received from outside of the zone
> the recipient (be through the direct or gatekeeper routed call model)
> has no idea how the calling endpoint came upon sending the setup
> message.
> this statement may be controversial but it is my opinion that h.323
> is lan-oriented, so if you think of gatekeepers as administering
> bandwidth allocation, &c for endpoints on different lans it would
> be difficult for a gatekeeper to perform this administration for
> a different lan.  and it would be even harder for a gatekeeper to
> perform such administration over a wide area network consisting of
> a number of zones.
> this fact leads me to believe that the ras channel should be *optional*
> for h.323 gatekeepers.  to abstract your scenarios further it is
> possible that gatekeepers could act as application layer routers, in
> which case they are terminating h.225 signalling and h.245 media control
> channels for endpoints with which they have no ras channel.
> --
>   __  ______  __  / __/ |       lucent technologies, naperville il, usa
> _/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
> "if you ain't the lead dog, the scenery never changes"  -- edmund wilson
>
>
>
>
> -------------------------------------------------------------------------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"

---------
This message came from the IETF IPTEL Working Group Mailing List.







---------
This message came from the IETF IPTEL Working Group Mailing List.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 02:59:25 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA29013
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 02:59:25 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 4AE4B8092; Thu, 11 Mar 1999 11:39:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2E147804E; Thu, 11 Mar 1999 11:11:58 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990309140309.00a9dcf0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Tue, 09 Mar 1999 14:03:11 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>, hsalama@cisco.com,
        iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>> 
>> Because of the above example, I think all types of translation (800
>> numbers or LNP) and mapping should occur prior to call routing. This is
>> particularly important for 800 numbers, because they have a regional
>> scope in many cases. The same 800 number may owned by company X in CA
>> and company Y in NY. In the above example, the call will be incorrectly
>> routed to company X instead of going to the originally intended
>> destination, company Y.
>
>This is another good example of why translation before gateway selection
>is a good idea. However, I'm not sure there is much we can do from a
>protocol perspective to prevent it. There may be valid reasons why a
>company would like to advertise an 800 number (only they own it, for
>example), and I'd rather not exclude this possibility. It seems like a
>sensible approach is to allow non-routing numbers in GLP, but mention
>that there are issues with them, and translation before routing is a
>good idea. Does this sound sensible?
>

This sounds like the right thing to do.  I think the framework document
should mention where and how translation of such numbers is expected/allowed.

- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 03:17:42 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA29166
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 03:17:42 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 4239B810A; Fri, 12 Mar 1999 02:59:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0CE2081B7; Fri, 12 Mar 1999 02:55:43 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E8C6DB.52DF@hpsqf.sqf.hp.com>
Date: Fri, 12 Mar 1999 07:48:43 +0000
From: Eileen Christie <eileenc@hpsqf.sqf.hp.com>
Reply-To: South@sqf.hp.com, Queensferry@sqf.hp.com, West@sqf.hp.com,
        Lothian@sqf.hp.com, Scotland@sqf.hp.com, EH30@sqf.hp.com,
        9TG@sqf.hp.com
Organization: Hewlett-Packard
X-Mailer: Mozilla 3.03 (WinNT; I)
MIME-Version: 1.0
To: iptel@lists.research.bell-labs.com
Subject: [Fwd: (no subject)]
Content-Type: message/rfc822
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Message-ID: <36C0445C.119F@hpsqf.sqf.hp.com>
Date: Tue, 09 Feb 1999 14:21:16 +0000
From: Eileen Christie <eileenc@hpsqf.sqf.hp.com>
Reply-To: South, Queensferry, West, Lothian, Scotland, EH30, 9TG
Organization: Hewlett-Packard
X-Mailer: Mozilla 3.03 (WinNT; I)
MIME-Version: 1.0
To: extmajor@BayNetworks.COM
Subject: (no subject)
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

which

unsuscribe *


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 03:35:35 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA29312
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 03:35:34 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0D0098038; Thu, 11 Mar 1999 17:43:42 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7E8728039; Thu, 11 Mar 1999 17:33:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Wed, 10 Mar 1999 13:54:45 +0100
From: massin@col.bsf.alcatel.fr (Raphael Massin)
Message-Id: <199903101254.NAA06240@c5s144.clb>
To: iptel@lists.research.bell-labs.com
Subject: SAP
X-Sun-Charset: US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hello all,

I am looking for the draft about SAP 'Session Annoucement Protocol'.
Maybe it expired because i can't find it at ftp.ietf.org/internet-drafts

Can you tell me where i will find it, please ?

Regards,

	Raphael

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 03:40:30 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA29337
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 03:40:29 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 371C87FD3; Tue,  9 Mar 1999 12:13:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0D20F7FD4; Tue,  9 Mar 1999 12:13:20 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990309120938.00ad15c0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Tue, 09 Mar 1999 12:09:40 -0500
To: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>> This begs the question, then: does the translation of the 800 number get
>> done before the gateway selection process, or after? By doing it before,
>> a gateway can be chosen which is, in fact, close to the real number
>> being dialed. With 800 numbers, all gateways will be able to reach it.
>> It seems that this opens the possibility of very non-optimal routing. A
>> caller in NY that dials an 800 number, goes to a gateway in CA. The
>> phone network there determines that the 800 number translates to a
>> number in NY, and routes the call through the PSTN back to NY. However,
>> if the translation occurs before gateway selection, the caller can
>> choose a local NY gateway and do much better.
>
>Translation before the gateway selection is *obviously* the correct model. We
>should strive towards this goal.
>
>PatC
>
>

The next question is when before gateway selection.  Translation could be
done as a pre-cursor to call-routing by the ingress, could be done by
signalling itself via things like SIP redirections/proxies, or could be
done within the routing infrastructure itself.  

I'd prefer one of the first two alternatives.  

- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 03:45:12 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA29356
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 03:45:11 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E13AE8029; Thu, 11 Mar 1999 17:36:02 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 523F38015; Thu, 11 Mar 1999 17:27:01 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: Deepak Kumar <deepak@microsoft.com>
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <65256730.00270CE7.00@sampark.hss.hns.com>
Date: Wed, 10 Mar 1999 12:46:58 +0530
Subject: RE: Handling of Setup message by Gatekeeper.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi,
   Ok, Gatekeeper might decide that way. But how will a GK check whether
endpoint, sending Setup message belongs to his zone or not? If yes, then
whether it is registered or not? If registered whether it has sent an ARQ
or not? Again the basic fundamental question is how to get
index/information to make all these checks?

Once the results of above checks are avaialbale, GK can decide what to do
in different scenarios like Setup from unregistered endpoint in its own
zone but without sending an ARQ.

Regards,
Ritu.
_____________________________________________________________________
Ritu Gupta                                   |       email:
rtgupta@hss.hns.com
Senior Software Engineer,       |              rtgupta@hotmail.com
Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
Banglore, Karnataka                  |              91-80-3374075(H)
INDIA.                                            |       Fax  :
91-80-2286393
_____________________________________________________________________
Restlessness and discontent are the first necessities of progress.




Deepak Kumar <deepak@microsoft.com> on 03/10/99 11:56:38 AM

To:   "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>, "ronald h. davis"
      <ronaldd@LUCENT.COM>
cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject:  RE: Handling of Setup message by Gatekeeper.




Q.931 could originate at a rogue (ie. in the physical zone, but
unregistered, and doesn't send ARQ) end-point or an end-point/GK outside
the
zone. Why is a distinction between these two scenarios meaningful?
An end-point that chooses (illegally) not to register with its Gatekeeper
could be treated exactly like an end-point outside the zone, and be denied
the benefits of Gatekeeper support - such as the ability to recv incoming
calls, QoS reservations, and the ability to place and recv calls using
network resources such as Gateways and proxies.
Deepak Kumar,
Program Manager, IP Telephony,
Windows Networking and Communications,
Microsoft Corp.
> -----Original Message-----
> From:   rtgupta@hss.hns.com [SMTP:rtgupta@hss.hns.com]
> Sent:   Tuesday, March 09, 1999 9:19 PM
> To:     ronald h. davis
> Cc:     h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject:     Re: Handling of Setup message by Gatekeeper.
>
> Hi Davis,
>     Thanks for your inputs.
>
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But
> the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
>
> In contrast to what you said, I think RAS signalling is must for all the
> endpoints in any GK's zone. RAS signaling is not required for endpoints,
> outside the zone. Now when GK receives a Setup message, it doesn't know
> whether it is from its own zone or from outside. Before doing any further
> checks, it needs to check that. And as I mentioned in my previous mails,
> to
> do that, it needs to maintain a static databse of its zone. But there is
> no
> mandatory field in Setup message,giving source info. So basic question is
> how to perform this check.
>
> regards,
> Ritu.
>
> _____________________________________________________________________
> Ritu Gupta                                   |       email:
> rtgupta@hss.hns.com
> Senior Software Engineer,       |                   rtgupta@hotmail.com
> Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
> Banglore, Karnataka                  |
91-80-3374075(H)
> INDIA.                                            |       Fax    :
> 91-80-2286393
> _____________________________________________________________________
> Restlessness and discontent are the first necessities of progress.
>
>
>
>
>
> "ronald h. davis" <ronaldd@lucent.com> on 03/09/99 11:32:19 PM
>
> To:   rtgupta@hss.hns.com
> cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
> Subject:  Re: Handling of Setup message by Gatekeeper.
>
>
>
>
> rtgupta@hss.hns.com wrote:
> >
>                [scenarios deleted]
> > In all these cases, gatekeeper will receive Setup message without
> receiving
> > ARQ message. So how can gatekeeper check whether ARQ is expected from
> the
> > endpoint, which sent the Setup or not? No field in the Setup message
> seem
> > to give this information. If gatekeeper doesn't check for this, that
> means
> > even the endpoints in its own zone can bypass gatekeeper for RAS
> > signalling?
> >
> > Any help/information on this issue will be appreciated.
> >
> the gatekeeper is only responsible for arq/acf/arj with endpoints within
> its zone.  all endpoints within the zone establish a ras channel (over
> which these messages are sent) with their gatekeeper.
> endpoints in other zones don't establish a ras channel with this
> gatekeeper.  thus, when calls are received from outside of the zone
> the recipient (be through the direct or gatekeeper routed call model)
> has no idea how the calling endpoint came upon sending the setup
> message.
> this statement may be controversial but it is my opinion that h.323
> is lan-oriented, so if you think of gatekeepers as administering
> bandwidth allocation, &c for endpoints on different lans it would
> be difficult for a gatekeeper to perform this administration for
> a different lan.  and it would be even harder for a gatekeeper to
> perform such administration over a wide area network consisting of
> a number of zones.
> this fact leads me to believe that the ras channel should be *optional*
> for h.323 gatekeepers.  to abstract your scenarios further it is
> possible that gatekeepers could act as application layer routers, in
> which case they are terminating h.225 signalling and h.245 media control
> channels for endpoints with which they have no ras channel.
> --
>   __  ______  __  / __/ |       lucent technologies, naperville il, usa
> _/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
> "if you ain't the lead dog, the scenery never changes"  -- edmund wilson
>
>
>
>
> -------------------------------------------------------------------------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"





---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 03:51:18 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA29383
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 03:51:18 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C195E81B1; Fri, 12 Mar 1999 02:38:26 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9D1EA8030; Fri, 12 Mar 1999 02:38:13 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E8B293.FC214414@utstar.com>
Date: Thu, 11 Mar 1999 22:22:11 -0800
From: Greg Zhang <gzhang@utstar.com>
Organization: WACOS, Inc.
X-Mailer: Mozilla 4.5 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 2: lifetimes
References: <3.0.32.19990311141033.00ab25e0@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Matt Squire wrote:

> Given the potential size of the database, I would propose that we go
> with (a). Aging individual entries does not seem particularly
> attractive although it certainly would provide some neat benefits.

Agree quite strongly here :)

- Greg


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 04:01:03 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA29425
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 04:01:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 97598811B; Thu, 11 Mar 1999 18:48:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 14F4F811D; Thu, 11 Mar 1999 18:45:11 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E8543D.240CFE28@cisco.com>
Date: Thu, 11 Mar 1999 15:39:41 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Christian Huitema <huitema@research.bellcore.com>
Cc: Matt Squire <msquire@nortelnetworks.com>, "David R. Oran" <oran@cisco.com>,
        Greg Zhang <gzhang@utstar.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <3.0.32.19990309170219.00aa9dc0@bl-mail1.corpeast.baynetworks.com> <990311181245.ZM1353@seawind.research.bellcore.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Christian Huitema wrote:
> 
> Well, it seems that we are back again to the distinction between naming
> and routing. With GLP, the phone number in the E164 space is really an
> alias to a "local" phone number.  Basically, the translation is:
> 
> a) use the number to obtain the id of the local provider,
> 
> b) route to the local provider
> 

But there may be multiple local providers (multiple egress gateways
owned by different providers) servicing the same number. Therefore, I
think we should route to the destination number. Then local provider
selection will automatically occur as part of route selection, and it
can still be controlled by injecting appropriate policy at the various
GLP speakers.

Hussein

PS: Am I the only person receiving old, and out-of-order messages from
IPTEL?


> c) within the local provider, route to the user based on the number.
> 
> This means that we will have a two phase look-up:
> 
> a) maybe using DNS:
>         look up number, get provider ID.
> b) within GLP:
>         maintain routes based on provider ID, and maybe also prefix.
> 
> --
> Christian Huitema
> ------------------------------
> Please note my new address: huitema@research.telcordia.com
> http://www.telcordia.com/

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 04:08:16 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA29494
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 04:08:15 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 141047FF6; Thu, 11 Mar 1999 12:01:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E0672803C; Thu, 11 Mar 1999 11:48:29 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E58BE2.61439977@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 16:00:18 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: Greg Zhang <gzhang@utstar.com>, John Schnizlein <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <010901be6a62$ea0ea480$2821a8c0@greg_zhang.wacos.com> <36E57B5C.E357450@dnrc.bell-labs.com> <36E58845.BA234BEB@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> I prefer the second alternative. The scope attribute shouldn't be opaque
> to intermediate LSs. I think the scope will factor into aggregation and
> route selection. For example, an LS shouldn't aggregate two
> advertisements with different scopes. And if an LS receives multiple
> advertisements for the same 800 number, each with a different scope, it
> shouldn't apply route selection to these advertisements, but rather
> continue advertising them all. I think that's necessary for GLP to work
> in a topology where non-contiguous administrative domains can accept
> calls from the same scope?

I agree with your requirements, but don't see how they eliminate
approach 1 (the scope attribute). What you are saying is that:

* an LS can only aggregate two routes if their scope attributes match
* a route selection should not occur across routes for the same prefix
but with different scope attributes

I guess "opaque" is really the wrong word. What I mean is that we should
not try and code a representation of a scope beyond a numerical token.
The meaning of the token would be bound to geographic scopes by means of
registration with IANA (or even proprietary scopes between providers).
We would then be able to define aggregation and selection rules, like
those above, based on whether two scopes match or not, and nothing more.
LS routing policies would be based on specific values, and it would be
up to the administrator to know, for example, that value 10 means "New
York".

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 04:12:36 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA29524
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 04:12:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C1A0181AD; Fri, 12 Mar 1999 02:30:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A3DF781AE; Fri, 12 Mar 1999 02:30:08 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E8B1E8.50083661@utstar.com>
Date: Thu, 11 Mar 1999 22:19:20 -0800
From: Greg Zhang <gzhang@utstar.com>
Organization: WACOS, Inc.
X-Mailer: Mozilla 4.5 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 1: phone numbers
References: <3.0.32.19990311140813.00aca430@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Matt Squire wrote:

> a) Expressing a set of phone destinations as a prefix, range, regular
> expression, or other? The above methods have all been mentioned as
> possible methods of describing a colletion of phone numbers. I
> believe we've reached consensus on the use of prefixes as the best
> method, based on the use of heirarchical phone numbers and on the
> potential for aggregation.

How about allowing all three options including individual phone numbers
as a special case for regular expression? This will make the aggregation,
lookup and matching code somewhat complex. But it does provide
the maximum flexibility for real world use. With number portability, an
aggregrated entry will likely to have holes in it. If holes are to be avoided,
the entries in the database get very fragmented. They may end up to be
either ranges or descret numbers.

Of course, one can choose to make the holes hidden when propagating
the entries. Routing requests that reaches the wrong gateway will get
redirected. I am not sure if this approach can work. The wrong gateway
needs to know where to redirect the request to. In a case of multiple providers advertising the services for numbers in the same range, the holes will get aggregated/hidden. Gateways of some provider may not know which
gateway serves the numbers in the holes.

> c) What numbers are allowed in the database?

From a simplistic point of view, as long as the signalling server associated
with a LS is capable of handle signalling for calls to to a number, the LS
should be able to advertisement the number. I agree that local routable
numbers and 800 numbers should get fully qualified to E.164 numbers
before advertising. With corporate/private numbers, I would think with
proper scoping, they should be allowed to be advertised and stored in
the database. I am not sure whether you can or should convert corporate
private numbers to E.164 numbers. If not, we would need a type field
making the solution less elegant.

> For (b), I propose that we just go with a string of octet digits (ie
> 0x02 == '2'). If people feel the need to save space, we could encode
> two digits into a single octet, but in the interest of stating an
> opinion, I'll go with the 1 octet == 1 digit approach.

I vote for putting two digits in one octet. The code for encoding/
decoding this, and matching etc. is not complex or costly. Print it in hex,
it is perfectly readable to human too :)

> Scoping and portability will be discussed as separate issues.

Since number presentation and handling of corporate/private numbers
seems to be related to scoping and portability, we may not be able to
view them as separate issues. Having to handle them may make the
solution not as elegant as one would otherwise desire. But we may not
have a choice.

- Greg


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 04:23:29 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA29616
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 04:23:28 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B719D8018; Thu, 11 Mar 1999 17:40:55 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 02E0180D8; Thu, 11 Mar 1999 17:26:10 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: "ronald h. davis" <ronaldd@lucent.com>
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <65256730.001BC1A0.00@sampark.hss.hns.com>
Date: Wed, 10 Mar 1999 10:48:33 +0530
Subject: Re: Handling of Setup message by Gatekeeper.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi Davis,
    Thanks for your inputs.

I agree that GK is responsible only for the endpoints whixh exist in its
zone. And also that it can not administer endpoints in other zones. But the
scenarios mentioned in the query are all valid scenarios. GK CAN receive
Setup message without any RAS signalling taking place before that. So it
can not terminate h225 signalling channel because of absence of RAS
channel.

In contrast to what you said, I think RAS signalling is must for all the
endpoints in any GK's zone. RAS signaling is not required for endpoints,
outside the zone. Now when GK receives a Setup message, it doesn't know
whether it is from its own zone or from outside. Before doing any further
checks, it needs to check that. And as I mentioned in my previous mails, to
do that, it needs to maintain a static databse of its zone. But there is no
mandatory field in Setup message,giving source info. So basic question is
how to perform this check.

regards,
Ritu.

_____________________________________________________________________
Ritu Gupta                                   |       email:
rtgupta@hss.hns.com
Senior Software Engineer,       |                   rtgupta@hotmail.com
Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
Banglore, Karnataka                  |                    91-80-3374075(H)
INDIA.                                            |       Fax    :
91-80-2286393
_____________________________________________________________________
Restlessness and discontent are the first necessities of progress.





"ronald h. davis" <ronaldd@lucent.com> on 03/09/99 11:32:19 PM

To:   rtgupta@hss.hns.com
cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject:  Re: Handling of Setup message by Gatekeeper.




rtgupta@hss.hns.com wrote:
>
               [scenarios deleted]
> In all these cases, gatekeeper will receive Setup message without
receiving
> ARQ message. So how can gatekeeper check whether ARQ is expected from the
> endpoint, which sent the Setup or not? No field in the Setup message seem
> to give this information. If gatekeeper doesn't check for this, that
means
> even the endpoints in its own zone can bypass gatekeeper for RAS
> signalling?
>
> Any help/information on this issue will be appreciated.
>
the gatekeeper is only responsible for arq/acf/arj with endpoints within
its zone.  all endpoints within the zone establish a ras channel (over
which these messages are sent) with their gatekeeper.
endpoints in other zones don't establish a ras channel with this
gatekeeper.  thus, when calls are received from outside of the zone
the recipient (be through the direct or gatekeeper routed call model)
has no idea how the calling endpoint came upon sending the setup
message.
this statement may be controversial but it is my opinion that h.323
is lan-oriented, so if you think of gatekeepers as administering
bandwidth allocation, &c for endpoints on different lans it would
be difficult for a gatekeeper to perform this administration for
a different lan.  and it would be even harder for a gatekeeper to
perform such administration over a wide area network consisting of
a number of zones.
this fact leads me to believe that the ras channel should be *optional*
for h.323 gatekeepers.  to abstract your scenarios further it is
possible that gatekeepers could act as application layer routers, in
which case they are terminating h.225 signalling and h.245 media control
channels for endpoints with which they have no ras channel.
--
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
"if you ain't the lead dog, the scenery never changes"  -- edmund wilson





---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 04:49:28 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA29791
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 04:49:27 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3AE267FD8; Thu, 11 Mar 1999 14:35:44 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 810C28019; Thu, 11 Mar 1999 14:34:33 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990311141033.00ab25e0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 11 Mar 1999 14:10:34 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: data issue 2: lifetimes
Mime-Version: 1.0
Content-Type: text/enriched; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

<bigger>

2) Lifetime of routing entries.  

   

   There have been various requirements stated with respect to the

   period of time that a routing entry is valid.  Some alternatives 
that

   have been discussed include:

     a)Routing entries are valid indefinitely.  The protocol is

       responsible for explicitly removing/purging entries that are no 

       longer valid.

     b)Routing entries are given a lifetime at their origination.

       Unless its refreshed before the lifetime expires, the routing
entry

       is automatically deleted after a certain period of time.

     c)Routing entries can be scheduled.  Here, the entry is given both

       a start and end to its lifetime.  The times could be relative or

       absolute (requiring something like NTP). 


Given the potential size of the database, I would propose that we go

with (a).  Aging individual entries does not seem particularly

attractive although it certainly would provide some neat benefits.

</bigger>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 11:56:26 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA06359
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 11:56:25 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 76050805D; Thu, 11 Mar 1999 14:34:34 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2D2557FD8; Thu, 11 Mar 1999 14:34:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990311140813.00aca430@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 11 Mar 1999 14:08:14 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: data issue 1: phone numbers
Mime-Version: 1.0
Content-Type: text/enriched; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



<bigger>Phone numbers.  


Most of these issues have been discussed on the list.


   a) Expressing a set of phone destinations as a prefix, range, 
regular

   expression, or other?  The above methods have all been mentioned as

   possible methods of describing a colletion of phone numbers.  I

   believe we've reached consensus on the use of prefixes as the best

   method, based on the use of heirarchical phone numbers and on the

   potential for aggregation.


   b) Encoding number as string, binary string, hex string, or other?

   We've discussed multiple methods for encoding the phone numbers into

   the protocol.  I believe we've agreed on the use of some kind of

   string representation.  I don't know if we've reached agreement on

   the exact kind of string representation.  

   

   c) What numbers are allowed in the database?  There are (at least)

   the following categories of phone numbers which have been mentioned

   as in need of representation.

      * Global routable e164.  Standard international numbers like

        +19199929048.

      * Local routable (ie no country/area code).  An example would be

        9929048.  The number has local but not global significance,

        although it can be automatically mapped to a number with global

        significance.  

      * Private/corporate.  These are numbers which are routable within

        a particular area, but without global significance and without
an

        existing method for mapping the number to one with global

        significance.

      * "800" numbers.  These are numbers that are not routable in the

        PSTN.  In the PSTN, the "800" number is translated into a
routable

        number.  Such numbers have multiple purposes:

        - portability - the mapping to a routable number can be changed

        over time

        - scoping - the mapping can be different depending upon the

        source of the call, or the mapping is only visible to a

        limited set of servers

        - charging - since the caller is not exposed to the final

        destination, the callee is responsible for usage

        - a bunch of other reasons as well...


For (b), I propose that we just go with a string of octet digits (ie

0x02 == '2').  If people feel the need to save space, we could encode

two digits into a single octet, but in the interest of stating an

opinion, I'll go with the 1 octet == 1 digit approach.  


On the kinds of phone numbers supported, I assume we have consensus on

the necessity of routable e164 numbers and on the fact that local

numbers should not be permitted (ie they should be mapped into routable

e164s before going into the call routing infrastructure).  I believe

that we also have consensus that we can route to an "800" number, but

that THIS IS NOT THE SAME AS PROVIDING "800" SERVICES (ie we can treat

"800" numbers as a routable e164, but if this does not meet the needs 
of

the deployment, then such number should not be permitted into the

database - the protocol can route on any number that's in the database).  


The protocol does not place any restrictions on billing/charging.  If

the stated use of "800" numbers in the call routing infrastructure does

not satisfy the the billing constraints of the deployment, then "800"

numbers should not be permitted in the database.  The protocol also 
does

not require or prohibit a directory service that translates "800"

numbers to routable e164s.  


Scoping and portability will be discussed as separate issues.  Any other
'services' provided by "800" numbers should be dealt with as independent
issues and should be handled by attributes and not by handling a number
as 'special' because it has a certain form.





</bigger>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 12:01:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA06640
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 12:01:53 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 407E9809B; Thu, 11 Mar 1999 11:39:31 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1A06F8077; Thu, 11 Mar 1999 11:09:42 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: "ronald h. davis" <ronaldd@lucent.com>
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <65256731.001827B8.00@sampark.hss.hns.com>
Date: Thu, 11 Mar 1999 12:29:39 +0530
Subject: Re: Handling of Setup message by Gatekeeper.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Davis,
     You had mentioned in your previous mail

************
this fact leads me to believe that the ras channel should be *optional*
for h.323 gatekeepers.  to abstract your scenarios further it is
possible that gatekeepers could act as application layer routers, in
which case they are terminating h.225 signalling and h.245 media control
channels for endpoints with which they have no ras channel.
************

All I'm saying is how to perform this check that the endpoint is from
outside the zone and then terminate the h.225 connection? Just because RAS
signaling is not present, we can not terminate the h.225 conection. In this
case we can not assume that h.225 message is coming from outside the zone.
What if it is from an unregistered endpoint within the zone itself?

Regards,
Ritu.





"ronald h. davis" <ronaldd@lucent.com> on 03/10/99 10:14:21 PM

To:   rtgupta@hss.hns.com
cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject:  Re: Handling of Setup message by Gatekeeper.




rtgupta@hss.hns.com wrote:
>
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But
the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
>
i think that i don't understand what you're trying to say.  the setup
message is an h.225 call signalling message, not a ras message.
--
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
"if you ain't the lead dog, the scenery never changes"  -- edmund wilson





---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 12:07:31 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA06774
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 12:07:28 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0B327801B; Fri, 12 Mar 1999 04:12:10 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2527981D8; Fri, 12 Mar 1999 04:09:40 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E8D715.ADEED34D@alcatel.be>
Date: Fri, 12 Mar 1999 09:57:58 +0100
From: Sandrine Roth <sandrine.roth@alcatel.be>
Organization: Alcatel CRC Antwerp
X-Mailer: Mozilla 4.03 [en] (WinNT; I)
MIME-Version: 1.0
To: Raphael Massin <massin@col.bsf.alcatel.fr>
Cc: iptel@lists.research.bell-labs.com
Subject: Re: SAP
References: <199903101254.NAA06240@c5s144.clb>
Content-Type: multipart/mixed; boundary="------------D2660CC4B09DF93D2103A82A"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.
--------------D2660CC4B09DF93D2103A82A
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit



Raphael Massin wrote:

> Hello all,
>
> I am looking for the draft about SAP 'Session Annoucement Protocol'.
> Maybe it expired because i can't find it at ftp.ietf.org/internet-drafts
>
> Can you tell me where i will find it, please ?
>
> Regards,
>
>         Raphael
>
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.


See attachment SAP v2

Regards,

Sandrine

--------------D2660CC4B09DF93D2103A82A
Content-Type: text/plain; charset=us-ascii; name="draft-ietf-mmusic-sap-v2-00.txt"
Content-Disposition: inline; filename="draft-ietf-mmusic-sap-v2-00.txt"
Content-Transfer-Encoding: 7bit




Internet Engineering Task Force				       MMUSIC WG
Internet Draft						Maryann	P. Maher
draft-ietf-mmusic-sap-v2-00.txt					     ISI
16 November 1998					   Colin Perkins
Expires: May 30, 1998						     UCL



		Session	Announcement Protocol: Version 2


Status of this Memo


   This	document is an Internet	Draft. Internet	Drafts are working
   documents of	the Internet Engineering Task Force (IETF), its	Areas,
   and its Working Groups. Note	that other groups may also distribute
   working documents as	Internet Drafts.

   Internet Drafts are draft documents valid for a maximum of six
   months. Internet Drafts may be updated, replaced, or	obsoleted by
   other documents at any time.	It is not appropriate to use Internet
   Drafts as reference material	or to cite them	other than as a	"working
   draft" or "work in progress."

   To view the entire list of current Internet-Drafts, please check the
   "1id-abstracts.txt" listing contained in the	Internet-Drafts	Shadow
   Directories on ftp.is.co.za (Africa), ftp.nordu.net (Northern
   Europe), ftp.nic.it (Southern Europe), munnari.oz.au	(Pacific Rim),
   ftp.ietf.org	(US East Coast), or ftp.isi.edu	(US West Coast).

Abstract

   This	document describes modifications and enhancements to SAP, the
   session directory announcement protocol, for	support	of IPv6
   conferencing	environments. Along with support for IPv6, a couple new
   features are	also introduced.  This document	compliments [SAP], which
   fully describes SAP in the context of IPv4. Readers are assumed to be
   familiar with [SAP].

   Note: At this time, this document presents an initial set of	ideas
   aimed solely	at starting discussion within the working group. We
   await the RFC publication of	[SAP] before finalizing	any protocol
   specifications.

   This	document is a product of the Multiparty	Multimedia Session
   Control (MMUSIC) working group of the Internet Engineering Task
   Force. Comments are solicited and should be addressed to the	working



							        [Page 1]

Internet-Draft		     SAP Version 2			Nov 1998


   group's mailing list	at confctrl@isi.edu and/or the author.



   1.  Overview

   This	document describes a new version of SAP, the session directory
   announcement	protocol, for support of IPv6 conferencing environments.
   SAP is a protocol used for handling multicast and unicast session
   description packets and defines an encapsulating packet format to be
   used	by session directory clients. As currently defined, only IPv4
   session announcements are supported mainly due to the fact that an
   IPv4	address	field is contained in the SAP packet header.  This docu-
   ment	describes the modifications to SAP for supporting IPv6 session
   announcements and introduces	some additional	new features. The two
   new features	described in this document are:

      1) a MIME	Content-Type specifier,	and

      2) a URL scheme for referencing SAP announcements.

   Note: At this time, this document presents an initial set of	ideas
   aimed solely	at starting discussion within the working group. We
   await the RFC publication of	[SAP] before finalizing	any protocol
   specifications.


   2.  Modifications to	the SAP	Packet Format

   Current SAPv1 data packets are of the following format:

    0			1		    2			3
    0 1	2 3 4 5	6 7 8 9	0 1 2 3	4 5 6 7	8 9 0 1	2 3 4 5	6 7 8 9	0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   | V=1 | MT  |E|C|   auth len	   |	    msg	id hash		   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |			  originating source			   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |		    optional authentication header		   |
   |				 ....				   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |			   optional timeout			   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |			 optional random field			   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |			      text payload			   |
   |				 ....				   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+



							        [Page 2]

Internet-Draft		     SAP Version 2			Nov 1998


   In order to support session directory clients in IPv6 environments,
   the SAP packet format must be modified to contain a 128-bit IPv6
   source address instead of the 32-bit	IPv4 address.

   Therefore, SAP v2 data packets are of the following format:

    0			1		    2			3
    0 1	2 3 4 5	6 7 8 9	0 1 2 3	4 5 6 7	8 9 0 1	2 3 4 5	6 7 8 9	0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   | V=1 |A| MT|E|C|   auth len	   |	    msg	id hash		   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |								   |
   |			  originating source			   |
   |				 ....				   |
   |								   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |		    optional authentication header		   |
   |				 ....				   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |			   optional timeout			   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |		       optional	random field			   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |			    text payload			   |
   |				 ....				   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+


   Note: the version number in the packet header has NOT changed.

   A	   Address type: 0 = IPv4, 1 = IPv6

   If the A bit	is 0, the orig source contains a 32-bit	IPv4 address. If
   the A bit is	1, the orig source contains a 128-bit IPv6 address.


   3.  Scoping SAP Announcements

   A SAP client	that announces a conference session periodically multi-
   casts an announcement packet	to a well known	multicast address and
   port.  The well known IPv6 SAP address is FF0X:0:0:0:0:0:2:7FFE,
   where X is the 4-bit	scope value. The following scope values	are
   currently defined in	IPv6:








							        [Page 3]

Internet-Draft		     SAP Version 2			Nov 1998


	   Value  Scope			 Recommended Bandwidth Limit
	  ------ -----------		-----------------------------
	    0x1	  Node-local		      n/a
	    0x2	  Link-local		      20 Kbps
	    0x5	  Site-local		      10 Kbps
	    0x8	  Organization-local	      1	Kbps
	    0xE	  Global		      200 bps

   The announcement is multicast with the same scope as	the session it
   is announcing, ensuring that	the recipients of the announcement can
   also	be potential recipients	of the session being advertised.  Having
   a scope field within	the address itself removes the need to carve out
   scope ranges	within the multicast address space or scope addresses
   according to	a corresponding	TTL.  Therefore, unlike	in IPv4, TTL-
   scoped announcements	do not exist in	IPv6 environments.  The	scope
   value to be used in the well-know SAP address is simply the same used
   in the multicast address assigned for the session.

   For example,	an announcement	for a link-local session assigned the
   multicast address of	FF02:0:0:0:0:0:1234:5678, should be advertised
   on SAP address FF02:0:0:0:0:0:2:7FFE. (Note:	Issues related to IPv6
   multicast address allocation	are being addressed in the MALLOC work-
   ing group.)

   In the table	above, recommended bandwidth limits are	given for ses-
   sions within	the defined scopes. The	total bandwidth	to be used for
   SAP announcements should be one-fourth of the session bandwidth,
   though this may be inappropriate for	some uses.




   4.  SAP Payload

   In previous versions	of SAP,	the payload was	required to be a session
   description in the SDP format [RFC2327]. With the introduction of new
   session description formats,	such as	SMIL [SMIL], it	is believed that
   that	restriction is no longer appropriate. SAP v2 therefore allows
   for other content to	be carried. That content should	begin with a
   MIME	Content-Type specifier.

	   Content-Type: application/sdp

	   v=0
	   o=...

   If the content type is "application/sdp" the	MIME header may	be omit-
   ted,	for backwards compatibility with SAP v1	applications. If any



							        [Page 4]

Internet-Draft		     SAP Version 2			Nov 1998


   other content is carried the	MIME header MUST be present.


   5.  SAP URL scheme

   In certain cases, it	is desirable to	reference a SAP	announcement.
   For example,	if it is desired that a	new participant	join an	existing
   session yet it is not known if that participant is within the scope
   of the announcement then an explicit	reference to the announcement
   will	enable them to determine if it can be received.	Providing the
   session description contained within	that announcement merely allows
   them	to join	the session, when they then notice that	the media
   streams are not visible.  Moreover, the addresses contained in a ses-
   sion	description for	one scope may not be valid outside that	scope
   zone.

   We therefore	define a URL scheme for	SAP announcements.  The	combina-
   tion	of msg-id-hash and originating-source fields in	the SAP	header
   is sufficient to identify a particular announcement.

   sap:msg-id@orig-source

   TBD:	What is	the effect of 10.x.x.x assignments on this?


   6.  Compatibility

   SAPv2 announcement are backwards compatible with SAPv1 provided that
   IPv4	sessions are announced,	and that the payload is	SDP with the
   content-type	omitted.

   If IPv6 announcements are made, they	will be	discarded by SAPv1 tools
   since they represent	illegal	MT values in that protocol.

   If the Content-Type header is present in the	payload, the announce-
   ment	is invalid in SAPv1. Those tools require that SDP is used, hence
   the payload for a SAPv1 announcement	MUST start with	"v=".


   7.  Security	Considerations

   SAP contains	mechanisms for ensuring	integrity of session announce-
   ments, for authenticating the origin	of an announcement and for
   encrypting such announcements. The strengths	and limitations	of these
   mechanisms are described in the 'Security Considerations' section of
   [SAP]. Those	considerations apply to	this document as well.





							        [Page 5]

Internet-Draft		     SAP Version 2			Nov 1998


   8.  Acknowledgements

REFERENCES

   [SAP] Handley, M.J.,	Session	Announcement Protocol, draft-ietf-
       mmusic-sap-01.txt, work in progress.

   [ADDARCH] Hinden, R., and S.	Deering, "IP Version 6 Addressing Archi-
       tecture", RFC 2373, July	1998.

   [IPV6] Deering, S., and R. Hinden, Editors, "Internet Protocol, Ver-
       sion 6 (IPv6) Specification", RFC 1883, December	1995.

   [MASGN] Hinden, R., and S. Deering, "IPv6 Multicast Address Assign-
       ments", RFC 2375, July 1998.

   [RFC2119] Bradner, S., "Key words for use in	RFCs to	Indicate
       Requirement Levels", BCP	14, RFC	2119, March 1997.

   [SMIL]

   [RFC2327] Handley, M., and Jacobson,	V., "SDP: Session Description
       Protocol", RFC2327, April 1998.

Author's Address

   Maryann P. Maher <maher@isi.edu>
   USC/ISI
   4350	N. Fairfax Drive, Suite	620
   Arlington VA	22203

   Colin Perkins <c.perkins@cs.ucl.ac.uk>
   Department of Computer Science
   University College London
   Gower Street
   London WC1E 6BT















							        [Page 6]


--------------D2660CC4B09DF93D2103A82A--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 12:11:21 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA06884
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 12:11:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B17CD801C; Thu, 11 Mar 1999 17:45:34 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 581FD801E; Thu, 11 Mar 1999 17:36:00 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E688BE.5A134E19@dnrc.bell-labs.com>
Date: Wed, 10 Mar 1999 09:59:10 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: Scott Petrack <scott.petrack@metatel.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>, Alok Mittal <amittal@hss.hns.com>,
        hsalama@cisco.com, iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <3.0.32.19990310071252.00ab4390@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Matt Squire wrote:
> 
> >I agree with >> and less so with >  above. Translations
> >are just outside of Iptel scope.
> >It is a separate can of worms, and we have enough of them on our plate
> >already (I just love mixed metaphors).
> >
> 
> Right, so just mention that number translation is not within the scope of
> GLP, and that it is not an expected function of call routing.   Other
> things (front-end, intra-domain, etc) not within the scope of GLP are
> mentioned in order to provide an idea of where GLP fits in to a larger ip
> telephony picture.  I don't want details, just a mention that it ain't our
> problem.

I agree with Matt here. I think its appropriate to mention this in the
framework document.

-Jonathan R.


-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 12:15:20 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA06969
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 12:15:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E172481B3; Fri, 12 Mar 1999 02:44:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0C76E818D; Fri, 12 Mar 1999 02:11:28 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <000701be6c54$1fda3780$741f0e0a@Jlaf005.telkom.go.id>
From: "Henri Setiawan" <henry@risti.telkom.co.id>
To: <h323implementors@imtc.org>
Cc: <iptel@lists.research.bell-labs.com>
Subject: Parameter for interoperability test
Date: Fri, 12 Mar 1999 13:46:55 +0700
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_0004_01BE6C8E.CAF285C0"
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.1
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.

------=_NextPart_000_0004_01BE6C8E.CAF285C0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Dear folks,

We like to study about interoperability test for gateway and gateway =
keeper, do anyone know what kind of parameters that we should test for =
this testing ?

Thanks

Henri S
R&D of PT Telkom Indonesia.



------=_NextPart_000_0004_01BE6C8E.CAF285C0
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">
<HTML><HEAD>
<META content=3Dtext/html;charset=3Diso-8859-1 =
http-equiv=3DContent-Type>
<META content=3D'"MSHTML 5.00.0910.1309"' name=3DGENERATOR></HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT size=3D2>Dear folks,</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>We like to study about interoperability test for =
gateway and=20
gateway keeper, do anyone know what kind of parameters that we should =
test for=20
this testing ?</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>Thanks</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>Henri S</FONT></DIV>
<DIV><FONT size=3D2>R&amp;D of PT Telkom Indonesia.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV>&nbsp;</DIV></BODY></HTML>

------=_NextPart_000_0004_01BE6C8E.CAF285C0--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 12:31:35 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA07435
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 12:31:35 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 98AF38130; Thu, 11 Mar 1999 19:57:42 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7A3B18074; Thu, 11 Mar 1999 19:47:35 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D18C@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'Jonathan Rosenberg'" <jdrosen@dnrc.bell-labs.com>,
        Matt Squire <msquire@nortelnetworks.com>
Cc: Greg Zhang <gzhang@utstar.com>, hsalama <hsalama@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>, "David R. Oran" <oran@cisco.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Thu, 11 Mar 1999 17:52:47 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear Jonathan,

in AIN SS7 network, we call the first three digit of your local
phone number as prefix. 800, 212 in NY are called area code.
When a guy call 800 number, it is actually send a request to 
national SCP to get the real destination number across the country.
The billing will be paid by 800 number receiver. I hope it will help.
Thanks.

Henry W. Lu


-----Original Message-----
From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
Sent: Wednesday, March 10, 1999 7:16 AM
To: Matt Squire
Cc: Greg Zhang; hsalama; Scott Petrack; Alok Mittal; David R. Oran;
iptel
Subject: Re: GLP, number portability and wireless subscribers


Phew. Lots of mail flying around here. let me take a moment to summarize
what I think I've heard so far on scopes. Several definitions have been
explicitly or implicitly used. These include:


1. Scope is a way to disambiguate prefixes which are not globally
unique. For example, a particular 800 number might exist in NY, and
another in CA, and they are the same number, but who you get connected
to is different in NY and in CA. The scope would be used as an extra
identifier to make prefix+scope (or vpn-id as Matt calls it in this
email) globally unique. 

2. Scope is a way to disambiguate prefixes which are not globally
unique, but the scope further defines the places where the prefix is
valid. Ideally, this would mean "this prefix is only valid in NY" or
something like that. These scopes might be registered as tags with IANA,
or defines as prefixes themselves. For example, a prefix of Y with scope
X means that entities within scope X can call the prefix Y.

3. Scope is a way to define how far a prefix is propagated. If LS's were
organized hierarchically, the scope would indicate the level in the
hierarchy. Boundary LS's wouldn't propagate prefixes with a scope
corresponding to their level outside their level. Matt has also
mentioned a TTL for this purpose.


Functions (1) and (2) are similar, (3) is different. The question is:
(1) do we need all three functions, and (2) can a single identifier
serve both purposes? We might be able to get a grip on this by first
figuring the first one out - what is it we wish to accomplish with
scopes, and then we can decide how many attributes we are talking about.


-Jonathan R.

Matt Squire wrote:
> 
> At 06:56 PM 3/9/99 -0500, Jonathan Rosenberg wrote:
> >David R. Oran wrote:
> >>
> >> Then scopes are identical to a deeper hierarchy and splitting them out
> >> is simply a syntactical, not a semantic distinction, right?
> >
> >My understanding is that it is semantic. By "splitting them out", I
> >presume you mean having them be separate from the address itself. The
> >prefix represents information on "where you're calling to", and the
> >scope information on "where you are calling from". So, a prefix Y with
> >scope X means you can call Y from X. I think we need explicit scopes if
> >there are cases where route attributes for a prefix only apply when
> >calling from numbers outside of the prefix (and apply only when calling
> >from somewhere defined by the scope).
> 
> I don't see how we can guarantee any determination of 'where you are
> calling from.'  What does my pc call from?  Where is an LS?  A gateway
> might be considered to have a 'location', where the location is given by
> the area code of its PSTN connectivity, but other devices without PSTN
> connectivity aren't tied to a PSTN location.
> 
> It seems like any definition of scope along these lines requires alot of
> administrative help, ie where LSs are tied/located with certain geographic
> areas.  Would not a TTL-like attribute provide similar functionality
(again
> based on geographic ties/connectetions in LSs)?  A number could be
> advertised with a certain TTL that prevents it from being advertised after
> too many hops.  If your connections are geography-based, then this gives
> you geographic scoping.  We make an additional statement that given
> multiple potential routes to a destination, the one with the higher/lower
> TTL is preferred (because its 'closer').
> 
> >
> >>
> >> Do you think we would need hierarchical scopes? If so, do you see
> >> aggregation of hierarchical scopes? How would this differ from
> >> hierarchical aggregation of addresses? Are scopes always more global
> >> than addresses?
> >
> >We could potentially define hierarchical scopes (if they are also
> >represented as prefixes, they are hierarchical). The rules for
> >aggregation would presumably be the same. But, adding another
> >aggregatable entity in a route advertisement seems like it will make
> >route aggregation overall more complex, so I'd rather avoid this. Are
> >scopes global? I think they need to be. The whole idea is that a prefix
> >advertised in one place is only useful somewhere else.
> >
> 
> Unless someone says something to convice me otherwise, I'd vote against
> this concept of scoping.
> 
> - Matt

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 13:50:34 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA09560
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 13:50:34 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 520258034; Thu, 11 Mar 1999 17:49:51 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F350A80FB; Thu, 11 Mar 1999 17:34:42 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.5.32.19990310111830.00ab9960@hotair.hobl.lucent.com>
X-Sender: louise@hotair.hobl.lucent.com
X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.5 (32)
Date: Wed, 10 Mar 1999 11:18:30 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: "Louise F. A. Spergel" <spergel@lucent.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: Greg Zhang <gzhang@utstar.com>, "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
In-Reply-To: <36E53DE3.694C7D69@dnrc.bell-labs.com>
References: <010501be69aa$79336240$2821a8c0@greg_zhang.wacos.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

My comments below are specific to the US. I don't have any knowledge on
other countries.

With 800 numbers, the issue is different due to the current implementation
and regulatory constraints. In 800 service, the called party pays,
therefore the called party picks the carrier, which currently are all
PSTN-type carriers. The carrier for each number is contained in a database
which is accessed by SS7 TCAP. This database is maintained by a neutral
party. In the near-term, you will need to consult this particular database
and to find the right carrier and then hand the call to a gateway to that
carrier. 

At some point I think the SIGTRAN protocols will enable this.

Of course, in the long run, this may change somewhat.

Louise


At 10:27 AM 3/9/99 -0500, Jonathan Rosenberg wrote:
>Greg Zhang wrote:
>> 
>> Hello,
>> 
>> One issue needs to be considered is number portability. Number portability
>> (NP) is reality for wired POTS lines for some time now. Wireless Number
>> Portability will be a reality in a few years' time. The portable numbers
>> cannot
>> be routed by strictly following the hierarchy derived from the digits of
>> called
>> number. 
>
>I agree LNP makes the problem harder. I have two responses here:
>
>1. Simply "not aggregating" is always an option for a protocol which
>provides aggregation. I don't want to not support aggregation because of
>LNP. If LNP gets in the way, it can just be turned off if needed.
>
>2. I'm not yet convinced LNP actually prevents aggregation. Lets say
>gateway X can reach ten numbers, +1 607-1* (using glob matching), +1
>607-2*, etc. so these could be aggregated to +1 607*. Lets further
>assume all ten numbers are LNP, and actually terminate all over the
>country. These numbers can still all be reached through the gateway.
>They cannot be aggregated only if the attributes for each are different.
>The only attribute likely to differ is cost. So, the question is, does
>the cost (to the gateway provider) for terminating calls to each of the
>numbers differ? Even if they do, is the cost the gateway provider wants
>to charge to its users different for each number? If the answer to
>either is no, then they can still be aggregated. With long distance
>rates at 10 cents a minute *anywhere* in the U.S., this seems not so
>unlikely.
>
>
>There is a related issue of 800 numbers, and any of the other
>non-routing numbers. Should these be advertised by GLP? Or, should they
>be translated first, and then the translated numbers advertised? (Please
>correct me if I have misunderstood some aspect of how this works in the
>PSTN).
>
>-Jonathan R.
>-- 
>Jonathan D. Rosenberg                       Lucent Technologies
>Member of Technical Staff                   101 Crawfords Corner Rd.
>High Speed Networks Research                Holmdel, NJ 07733
>FAX:   (732) 834-5379                       Rm. 4C-526
>EMAIL: jdrosen@bell-labs.com
>URL: http://www.cs.columbia.edu/~jdrosen
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.
>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 13:59:42 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA09688
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 13:59:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 246D7816C; Thu, 11 Mar 1999 20:57:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 385468120; Thu, 11 Mar 1999 19:57:01 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D18D@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'David R. Oran'" <oran@cisco.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>, hsalama@cisco.com
Cc: iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Thu, 11 Mar 1999 18:16:05 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear David,

I think what we can not afford is place a call to wrong number.
Whether we will reveal the true number or not is not
very important from my perspective. You have to know 
that SS7 network is not like IP network, there is no
network topology database information in SCP like
OSPF or BGP in IP world. This create problem for reconstructing
call routing from IP world. SS7 is per call base and do 
a SCP query if there is 800 or ported number. No matter
SIP or H.323, will be per call based through gateway query. 
We are still lock the protocol in old ways per call based. We can
not update IP world PSTN topology like what we are doing
in OSPF or BGP for the whole network in one database sync.
I hope it will help. Thanks.

Henry W. Lu


-----Original Message-----
From: David R. Oran [mailto:oran@cisco.com]
Sent: Tuesday, March 09, 1999 1:12 PM
To: Jonathan Rosenberg; hsalama@cisco.com
Cc: iptel
Subject: RE: GLP, number portability and wireless subscribers


I'm reluctant to bring this up, because some would view it as a corner
case or just a "feeeeture" of the PSTN, but when the PSTN does number
translation (or call forwarding in the network as well), the resulting
number is not revealed to the calling party. The pure directory scheme
would reveal the translated number since that's what would be put in the
end-to-end signaling (e.g. SIP, H.323). Do we care about this?

> -----Original Message-----
> From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
> Sent: Tuesday, March 09, 1999 1:43 PM
> To: hsalama@cisco.com
> Cc: John Schnizlein; Greg Zhang; David R. Oran; Matt Squire; Scott
> Petrack; Alok Mittal; iptel
> Subject: Re: GLP, number portability and wireless subscribers
>
>
> Hussein F. Salama wrote:
> >
> > Jonathan Rosenberg wrote:
> >
> > >
> > > This begs the question, then: does the translation of the
> 800 number get
> > > done before the gateway selection process, or after? By
> doing it before,
> > > a gateway can be chosen which is, in fact, close to the
> real number
> > > being dialed. With 800 numbers, all gateways will be able
> to reach it.
> > > It seems that this opens the possibility of very
> non-optimal routing. A
> > > caller in NY that dials an 800 number, goes to a gateway
> in CA. The
> > > phone network there determines that the 800 number translates to a
> > > number in NY, and routes the call through the PSTN back
> to NY. However,
> > > if the translation occurs before gateway selection, the caller can
> > > choose a local NY gateway and do much better.
> >
> > Because of the above example, I think all types of translation (800
> > numbers or LNP) and mapping should occur prior to call
> routing. This is
> > particularly important for 800 numbers, because they have a regional
> > scope in many cases. The same 800 number may owned by
> company X in CA
> > and company Y in NY. In the above example, the call will be
> incorrectly
> > routed to company X instead of going to the originally intended
> > destination, company Y.
>
> This is another good example of why translation before
> gateway selection
> is a good idea. However, I'm not sure there is much we can do from a
> protocol perspective to prevent it. There may be valid reasons why a
> company would like to advertise an 800 number (only they own it, for
> example), and I'd rather not exclude this possibility. It seems like a
> sensible approach is to allow non-routing numbers in GLP, but mention
> that there are issues with them, and translation before routing is a
> good idea. Does this sound sensible?
>
> Scott P., of course, is right, that the translation service itself is
> outside the scope of iptel.
>
> -Jonathan R.
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen
>


---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 14:04:32 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09845
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 14:04:31 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2B31E80AE; Thu, 11 Mar 1999 15:55:55 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7269880AD; Thu, 11 Mar 1999 14:55:53 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990311140245.00ac3c60@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 11 Mar 1999 14:02:46 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: data representation issues
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


All - 

Over the next day or so, I'm going try to air on the list some of the data
representation issues that I'm planning to mention during our wg meeting
next week.  I know its a bit late, but I figure its better to start now
than to wait.  

I'll attempt to express the issue, mention some alternatives, and where I
think we have consensus and where we do not.   At the end, I'll probably
give an opinion on the issue so that everyone can tell me how wrong I am
and what we really need to do instead.  

Note that the two potential protocols differ in the set of attributes.  On
those issues that deal with representation of an attribute that is not
shared by both protocols, one of the issues is, of course, whether we need
the attribute at all.  

I will try to incorporate the received comments into what I present next
week when discussing the representation issues.  

Load your guns and prepare to fire.  

- Matt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 14:44:14 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA12151
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 14:44:13 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DB3CB807D; Thu, 11 Mar 1999 17:42:39 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6911880DC; Thu, 11 Mar 1999 17:31:33 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990310071252.00ab4390@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Wed, 10 Mar 1999 07:12:54 -0500
To: Scott Petrack <scott.petrack@metatel.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>, Alok Mittal <amittal@hss.hns.com>,
        hsalama@cisco.com, iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 07:01 PM 3/9/99 -0500, Scott Petrack wrote:
>
>
>> >This is another good example of why translation before gateway selection
>> >is a good idea. However, I'm not sure there is much we can do from a
>> >protocol perspective to prevent it. There may be valid reasons why a
>> >company would like to advertise an 800 number (only they own it, for
>> >example), and I'd rather not exclude this possibility. It seems like a
>> >sensible approach is to allow non-routing numbers in GLP, but mention
>> >that there are issues with them, and translation before routing is a
>> >good idea. Does this sound sensible?
>> >
>> 
>> This sounds like the right thing to do.  I think the framework document
>> should mention where and how translation of such numbers is
expected/allowed.
>> 
>
>I agree with >> and less so with >  above. Translations
>are just outside of Iptel scope.  
>It is a separate can of worms, and we have enough of them on our plate 
>already (I just love mixed metaphors). 
>

Right, so just mention that number translation is not within the scope of
GLP, and that it is not an expected function of call routing.   Other
things (front-end, intra-domain, etc) not within the scope of GLP are
mentioned in order to provide an idea of where GLP fits in to a larger ip
telephony picture.  I don't want details, just a mention that it ain't our
problem.  

Now you can have your worms and eat it too.

- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 14:45:10 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA12174
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 14:45:10 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DBB2D8194; Fri, 12 Mar 1999 02:08:47 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A322D8186; Fri, 12 Mar 1999 02:08:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: "umesh@seeyes.co.in" <umesh@seeyes.co.in>
Cc: iptel@lists.research.bell-labs.com
Message-ID: <65256732.0024B139.00@sampark.hss.hns.com>
Date: Fri, 12 Mar 1999 12:15:37 +0530
Subject: RE: Handling of Setup message by Gatekeeper.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Umesh,
      'source Address' is a sequence of alias addresses and is an optional
field. This ight not be present at all. even if it is there, we can not
bank on it. Because alias addresses are not globally unique. Endpoints in
different zones can have same alias address. And 'source Info' is just the
type of endpoint the source is. Values can be GK, GW, MSU or terminal etc.
with some vendr info. So we can not check on basis of thisfield also.

Thansk,
Ritu.

P.S. Check previous mails for a link to a proposal in Setup message.





Umesh Mallugari <umesh@seeyes.co.in> on 03/12/99 12:48:47 PM

Please respond to "umesh@seeyes.co.in" <umesh@seeyes.co.in>

To:   "'iptel@lists.research.bell-labs.com'"
      <iptel@lists.research.bell-labs.com>
cc:   "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>,
      "'rkumar@holontech.com'" <rkumar@holontech.com>
Subject:  RE: Handling of Setup message by Gatekeeper.




Ritu,
Won' t the fields "sourceAddress" and "sourceInfo" in the Setup UUIE be of
any help to the Gatekeeper
to perform checks on the Setup message received?
-Umesh.
-----Original Message-----
From:     Hanlin_Fang@3com.com [SMTP:Hanlin_Fang@3com.com]
Sent:     Thursday, March 11, 1999 5:16 PM
To:  Deepak Kumar
Cc:  'rtgupta@hss.hns.com'; ronald h. davis; h323implementors@imtc.org;
iptel@lists.research.bell-labs.com
Subject:  RE: Handling of Setup message by Gatekeeper.


How can Gatekeepers know that there are some endpoints in its zone not
registered? If GKs don't know their existence, then how can those endpoints
be
managed?
Thanks for any information
Regards!
Hanlin



Deepak Kumar <deepak@microsoft.com> on 03/09/99 10:26:38 PM
To:   "'rtgupta @hss.hns.com'" <rtgupta@hss.hns.com>, "ronald h. davis"
      <ronaldd@LUCENT.COM>
cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com (Hanlin
      Fang/HQ/3Com)
Subject:  RE: Handling of Setup message by Gatekeeper.



Q.931 could originate at a rogue (ie. in the physical zone, but
unregistered, and doesn't send ARQ) end-point or an end-point/GK outside
the
zone. Why is a distinction between these two scenarios meaningful?
An end-point that chooses (illegally) not to register with its Gatekeeper
could be treated exactly like an end-point outside the zone, and be denied
the benefits of Gatekeeper support - such as the ability to recv incoming
calls, QoS reservations, and the ability to place and recv calls using
network resources such as Gateways and proxies.
Deepak Kumar,
Program Manager, IP Telephony,
Windows Networking and Communications,
Microsoft Corp.
> -----Original Message-----
> From:   rtgupta@hss.hns.com [SMTP:rtgupta@hss.hns.com]
> Sent:   Tuesday, March 09, 1999 9:19 PM
> To:     ronald h. davis
> Cc:     h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject:     Re: Handling of Setup message by Gatekeeper.
>
> Hi Davis,
>     Thanks for your inputs.
>
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But
> the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
>
> In contrast to what you said, I think RAS signalling is must for all the
> endpoints in any GK's zone. RAS signaling is not required for endpoints,
> outside the zone. Now when GK receives a Setup message, it doesn't know
> whether it is from its own zone or from outside. Before doing any further
> checks, it needs to check that. And as I mentioned in my previous mails,
> to
> do that, it needs to maintain a static databse of its zone. But there is
> no
> mandatory field in Setup message,giving source info. So basic question is
> how to perform this check.
>
> regards,
> Ritu.
>
> _____________________________________________________________________
> Ritu Gupta                                   |       email:
> rtgupta@hss.hns.com
> Senior Software Engineer,       |                   rtgupta@hotmail.com
> Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
> Banglore, Karnataka                  |
91-80-3374075(H)
> INDIA.                                            |       Fax    :
> 91-80-2286393
> _____________________________________________________________________
> Restlessness and discontent are the first necessities of progress.
>
>
>
>
>
> "ronald h. davis" <ronaldd@lucent.com> on 03/09/99 11:32:19 PM
>
> To:   rtgupta@hss.hns.com
> cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
> Subject:  Re: Handling of Setup message by Gatekeeper.
>
>
>
>
> rtgupta@hss.hns.com wrote:
> >
>                [scenarios deleted]
> > In all these cases, gatekeeper will receive Setup message without
> receiving
> > ARQ message. So how can gatekeeper check whether ARQ is expected from
> the
> > endpoint, which sent the Setup or not? No field in the Setup message
> seem
> > to give this information. If gatekeeper doesn't check for this, that
> means
> > even the endpoints in its own zone can bypass gatekeeper for RAS
> > signalling?
> >
> > Any help/information on this issue will be appreciated.
> >
> the gatekeeper is only responsible for arq/acf/arj with endpoints within
> its zone.  all endpoints within the zone establish a ras channel (over
> which these messages are sent) with their gatekeeper.
> endpoints in other zones don't establish a ras channel with this
> gatekeeper.  thus, when calls are received from outside of the zone
> the recipient (be through the direct or gatekeeper routed call model)
> has no idea how the calling endpoint came upon sending the setup
> message.
> this statement may be controversial but it is my opinion that h.323
> is lan-oriented, so if you think of gatekeepers as administering
> bandwidth allocation, &c for endpoints on different lans it would
> be difficult for a gatekeeper to perform this administration for
> a different lan.  and it would be even harder for a gatekeeper to
> perform such administration over a wide area network consisting of
> a number of zones.
> this fact leads me to believe that the ras channel should be *optional*
> for h.323 gatekeepers.  to abstract your scenarios further it is
> possible that gatekeepers could act as application layer routers, in
> which case they are terminating h.225 signalling and h.245 media control
> channels for endpoints with which they have no ras channel.
> --
>   __  ______  __  / __/ |       lucent technologies, naperville il, usa
> _/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
> "if you ain't the lead dog, the scenery never changes"  -- edmund wilson
>
>
>
>
> -------------------------------------------------------------------------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"
---------
This message came from the IETF IPTEL Working Group Mailing List.






---------
This message came from the IETF IPTEL Working Group Mailing List.








---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 15:26:25 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA14350
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 15:26:25 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5DF2C8000; Thu, 11 Mar 1999 11:48:41 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 396347FE5; Thu, 11 Mar 1999 11:41:39 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E5A817.4D62790F@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 18:00:39 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: "David R. Oran" <oran@cisco.com>
Cc: hsalama@cisco.com, iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <NBBBIFCAKKOPNMMKHHEFEEJNEEAA.oran@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

David R. Oran wrote:
> 
> I'm reluctant to bring this up, because some would view it as a corner
> case or just a "feeeeture" of the PSTN, but when the PSTN does number
> translation (or call forwarding in the network as well), the resulting
> number is not revealed to the calling party. The pure directory scheme
> would reveal the translated number since that's what would be put in the
> end-to-end signaling (e.g. SIP, H.323). Do we care about this?

The pure directory scheme won't necessarily have this characteristic.
For SIP, if a proxy does the lookup, the called number is placed in the
Request-URI. The To and From fields remain unmodified. As such, the
caller will never see the translated number. The called party may see
this number, but if it does, its only because its the next hop. Having
the called party see this number seems a non-problem, since its the
number for that party anyway. The problem you describe only arises if
the caller itself performs the directory lookup. Whether service
providers allow end users to query their 800 number directories directly
is a good question, so this case may never happen. I'm not sure how this
would work in an H.323 context.

In any case, this is outside the scope of iptel. 

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 15:50:46 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA15438
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 15:50:45 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D49CA52BF; Fri, 12 Mar 1999 14:42:25 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4042A52D6; Fri, 12 Mar 1999 14:20:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D190@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'Henry Sinnreich'" <henry.sinnreich@mci.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        "David R. Oran" <oran@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: hsalama <hsalama@cisco.com>, John Schnizlein <jschnizl@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Fri, 12 Mar 1999 13:13:00 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear Henry,
Any carrier want to provide IP telephony service will need to
build their own "confederation" in IP world to register users and billing.
Since it is in IP world, the routing information will be able to exchange 
with each other via BGP-4 with attributes between  different "confederation 
serving domain". It won't be free so you don't have to worry. Thanks.
Henry W. Lu


-----Original Message-----
From: Henry Sinnreich [mailto:henry.sinnreich@mci.com]
Sent: Thursday, March 11, 1999 3:26 PM
To: Matt Squire; David R. Oran; Greg Zhang; Jonathan Rosenberg
Cc: hsalama; John Schnizlein; Scott Petrack; Alok Mittal; iptel
Subject: RE: GLP, number portability and wireless subscribers


Matt Squire writes:
>If there is a scope registration
> authority (maybe this wasn't clear before), then the combination of
> scope+number is globally unambiguous, with the default scope being global
> e164.

Is there such an authority?

Thanks, Henry

> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Matt
> Squire
> Sent: Tuesday, March 09, 1999 5:02 PM
> To: David R. Oran; Greg Zhang; Jonathan Rosenberg
> Cc: hsalama; John Schnizlein; Scott Petrack; Alok Mittal; iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> At 04:34 PM 3/9/99 -0500, David R. Oran wrote:
> >Or, alternatively, project all numbers into a global, unambiguous space.
> >This seems cleaner and much less error prone than topologically
> >established scope boundaries, and less administratively burdensome than
> >solving the problem of scope name collisions and information propagation
> >boundaries. It also avoids having to talk about translations between
> >scopes, which I'm sure will happen when the other shoe drops :-).
> >
>
> I'm a little confused by this comment. To me, scopes provide the
> unambiguous global number space.  If there is a scope registration
> authority (maybe this wasn't clear before), then the combination of
> scope+number is globally unambiguous, with the default scope being global
> e164.  If someone registers a scope and they wish it to have topological
> significance, then that's an administrative matter and they should be able
> to set up their peer association and policy accordingly.
>
> I'd probably be against scopes of the form "+1-405" because that
> would just
> lead to too many headaches in interpreting, aggregating, etc.
>
> An opaque scope using a equal/!equal test might provide what we desire.  I
> haven't convinced myself completely, but I think it has the potential.
>
> - Matt
>
>
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
>


---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 17:15:04 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17953
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 17:15:02 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 354C65304; Fri, 12 Mar 1999 15:44:08 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id ABA118253; Fri, 12 Mar 1999 11:51:34 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990312091046.00aefdd0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Fri, 12 Mar 1999 09:10:47 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: data issue 6: domain representation
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


6) Internet Telephony Administrative Domains

For many reasons, GLP requires a way to represent different internet
telphony domains such that every domain has a unique representation.  We
haven't explicitly discussed how to represent such a domain.  

a) What is the appropriate representation of a domain?  Should it be
numeric (like an AS number), character based (like a domain name), or a URL
(like a SIP address)?

b) If numeric, should it be taken from the same number space as AS numbers
or not?

Due to its potential uses, I think its relatively clear that a numeric
representation would be the most efficient.  

Which brings us to (b).  My opinion is that call-routing is an independent
problem from IP routing, as such, the name/number spaces should be
independent.  

If we go with integrating e164 numbers into BGP, then, of course, an IP
telephony domain is just an AS and the name/number spaces are shared.  If
TBGP is considered a separate protocol from BGP, then again the number
spaces should probably be independent.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 18:14:05 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA18945
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 18:14:05 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 933BA5312; Fri, 12 Mar 1999 18:07:57 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3FE6552FB; Fri, 12 Mar 1999 18:07:55 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <E32BD2062DAFD1119CFD0000F804F31702C0486F@cosexch003.mcit.com>
From: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: H.323 black phone???
Date: Fri, 12 Mar 1999 22:26:22 -0000
Importance: high
X-Priority: 1
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE6CD7.5BCF6BE2"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE6CD7.5BCF6BE2
Content-Type: text/plain

Hello,

I've been searching the web, and have not found an H.323 black phone.
Everything is a PC card or a videophone.  Some manufacturers claim they'll
have a unit in " a few months".
Does anyone know of an H.323 phone I can buy now?  FYI, Mediatrix makes a
SIP phone ... I have seen it.

Thanks,
Ralph

Ralph S. Hoefelmeyer
MCI WorldCom Engineer



------_=_NextPart_001_01BE6CD7.5BCF6BE2
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>H.323 black phone???</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Hello,</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">I've been searching the web, =
and have not found an H.323 black phone.&nbsp; Everything is a PC card =
or a videophone.&nbsp; Some manufacturers claim they'll have a unit in =
&quot; a few months&quot;.</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Does anyone know of an H.323 =
phone I can buy now?&nbsp; FYI, Mediatrix makes a SIP phone ... I have =
seen it.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Thanks,</FONT>
<BR><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Ralph</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Ralph S. Hoefelmeyer</FONT>
<BR><FONT COLOR=3D"#0000FF" FACE=3D"Arial">MCI WorldCom Engineer</FONT>
</P>
<BR>

</BODY>
</HTML>
------_=_NextPart_001_01BE6CD7.5BCF6BE2--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 18:19:10 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA19005
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 18:19:09 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6613B801A; Thu, 11 Mar 1999 14:54:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6F6DD8089; Thu, 11 Mar 1999 14:35:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990311141239.00ac5200@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 11 Mar 1999 14:12:39 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: data issue 3: signalling protocols
Mime-Version: 1.0
Content-Type: text/enriched; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

<bigger>


3) Expressing signalling protocols.  


   There seems to be consensus that we need an attribute for expressing

   the signalling protocols supported by the next-hop or gateway.  But

   there are still questions to address.  


     a) How to express a signalling protocol?   I don't believe there

        is any pre-existing standardized method for expressing

        signalling protocols, so we get to make one up. 

     b) What signalling protocols to express?  We certainly need to

        express both SIP and H323.  In the TBGP draft, H323 is

        parititioned into Q.931 and RAS.  Also, MGCP and SGCP are

        included.  

          - Is there consensus that H323 should be split into RAS and

          Q931?

          - Is there consensus MGCP and SGCP included?  

     c) Include port number?  Do we need to include the UDP/TCP on 
which

        signalling messages to the next-hop should be sent, and whether

        TCP or UDP should be used?  

     d) Include other/additional signalling parameters?

      

I would propose that a new registry be created for the purposes of

registering ip-telephony signalling protocols.  As usual, one registers

both a number, name, reference, contact, etc.  


On the protocols to express, I'm a little fuzzy on why MGCP and SGCP

need to have messages routed based on e164 numbers.  Not having read 
the

specs in a while, I thought that they were expressing a control 
protocol

between a call control agent and a media gateway.   I would assume that

they would know the gateways that they are managing and hence wouldn't

need to locate the gateways via routing on a phone number.  Can someone

enlighten me?


Since signalling messages can be sent/received on different ports and

using different protocols, I think that expressing whether a gateway

requires UDP or TCP, and on which port it expects to receive the

signalling protocol messages, is required.  


For (d), in draft-ietf-iptel-glp-00.txt, I included room for 
"Additional

Signalling Parameters."  The idea is to allow a place for other

signalling parameters that are required in order to establish 
signalling

connectivity to the next-hop.  Originally, I was going to place the IP

protocol and TCP/UDP port in here, but I separated them into a separate

question because they seemed required for any signalling protocol.  I

thought this field might prove useful as the list of signalling

protocols expands or as the capabilities of the protocols expand.  Note

that the purpose of this field is not to express everything about the

signalling protocol, only those parameters which are required for

connectivity.  If it can be negotiated after connecting, then it

shouldn't be here.



</bigger>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 19:09:39 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA19768
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 19:09:39 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E343B822B; Fri, 12 Mar 1999 12:50:15 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0463481A7; Fri, 12 Mar 1999 02:10:07 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990311225816.00ac35f0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 11 Mar 1999 22:58:19 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: data issue 5: price/cost
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


5) Price/cost.

We all agree that we need to somehow represent the cost of making a call to
a particular destination at a particular gateway.  The more difficult
question is how.

a) Can we define a set of cost primitives to sufficiently represent pricing
structures?  
b) If so, what are they?
c) If not, what then?

I believe that we had consensus at the last wg meeting that we would be
unable to determine a sufficient set of cost primitives to satisfy every
objective, and that satisfying a smaller subset of objectives was not
sufficient.  Which seemed to leave us under a few rocks and surrounded by
hard places.  

One option is to use an opaque (ie not necessarily understood by the
call-routers) attribute to express the cost of a call.  This would permit
arbitrary pricing structures to be expressed via something like a text
message or a URL.  This would be of great help to those that view GLP as
providing a service to the calling user to help them choose the 'cheapest'
gateway.  

In general, we all agreed that the cost of the IP segment of the call will
be negligible compared to the cost of the PSTN segment.   Since the effect
of the IP segment on price is deemed negligible, the fact the a call-router
cannot adjust the attribute is not too harmful.  Unfortunately, an opaque
attribute has significant impacts to aggregation and route selection.  If
the call-router does not understand the attribute, then it could not choose
the best route based on that attribute.  Also, routes with different prices
cannot be aggregated if the structure is not understood.    

Which brings us to a second option.  Even though we cannot agree on a set
of metrics that would be sufficient, a unitless metric can provide great
benefits.  Much like a unitless capacity, a unitless price metric would
permit comparisons of different routes based on a relative view of price.
The caller would not know the 'actual' price in terms of price/min etc, but
a unitless price would permit better aggregation and route selection.  As
with a unitless capacity, the set of providers within a ip telephony
confederation would have to coordinate the interpretation of the relative
metric between and within domains.  Since the protocol does not know how to
interpret the relative metric, it is difficult for the attribute to be
adjusted based on the cost of the IP path.  We could mandate that the
attribute (a) not be adjusted when propagated, or (b) make it a
configurable policy within GLP to increment this metric by x when
propagated.  

My personal belief is that we probably need both a pricing metric and a
pricing description.  For a pricing description, I suggest either a text
based message or a URL that can be used to describe a pricing structure.
To assist in aggretability, providers should use relatively few different
attributes (the attributes can be subtyped by provider and subtyped within
that) to describe their pricing structures, having each provide enough
information to be applicable at a number of gateways.  The intent of the
pricing description is that this attribute can be displayed to a user (or
interpreted by some process on behalf of an end-user) when/if the user
accesses the call routing database via some front-end protocol, and that
the information fairly accurately reflects what the user will actually pay
for the call.  A URL could be used to handle large amounts of data in a
language independent manner.  This information would not be modified by the
call-routers, only propagated so that the ingress can get an idea of how
much a call is going to cost.  The information reflects only the cost on
the PSTN side and nothing on the IP side.  
Additionally, there should be a pricing metric that can be used and
modified by the routers in route selection and aggregation.  

I think the combination of the two can provide both network and end-user
information for gateway selection.  

If we agree that a pricing description is needed, I would further suggest
that it be signed by the orginator of the information using a public-key
method.  Since people may be making decision based on this data, it should
be authenticated so that people have some reason to trust its validity.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 20:10:41 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA20402
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 20:10:40 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 688BE530D; Fri, 12 Mar 1999 15:47:55 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3790D5300; Fri, 12 Mar 1999 15:45:10 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
To: Sandrine Roth <sandrine.roth@alcatel.be>
Cc: Raphael Massin <massin@col.bsf.alcatel.fr>,
        iptel@lists.research.bell-labs.com
Subject: Re: SAP
In-reply-to: Your message of "Fri, 12 Mar 1999 09:57:58 +0100." <36E8D715.ADEED34D@alcatel.be>
Date: Fri, 12 Mar 1999 17:01:27 +0000
Message-ID: <6787.921258087@cs.ucl.ac.uk>
From: Colin Perkins <C.Perkins@cs.ucl.ac.uk>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--> Sandrine Roth writes:
>Raphael Massin wrote:
>> Hello all,
>> I am looking for the draft about SAP 'Session Annoucement Protocol'.
>> Maybe it expired because i can't find it at ftp.ietf.org/internet-drafts
>>
>> Can you tell me where i will find it, please ?
>>
>> Regards,
>>
>>         Raphael
>
>See attachment SAP v2
>
>Regards,
>Sandrine

The SAP v2 draft provides a few extensions to the base SAP spec, but is not
complete by itself. The main spec has expired, but I have a copy online at
http://www.cs.ucl.ac.uk/staff/c.perkins/internet-drafts/draft-ietf-mmusic-sap-00.txt

There will be a merger of the base spec, security and v2 drafts in the
not-too-distant future.

Cheers,
Colin

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 21:12:03 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA21017
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 21:12:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7E783530F; Fri, 12 Mar 1999 20:06:39 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 552355323; Fri, 12 Mar 1999 20:06:31 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Dean Willis" <Dean.Willis@MCI.COM>
To: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@MCI.COM>,
        "IPTEL IETF Mailing List" <iptel@lists.research.bell-labs.com>
Subject: RE: H.323 black phone???
Date: Fri, 12 Mar 1999 18:42:46 -0600
Message-ID: <000501be6cea$69bb9240$67e323a6@dwillispc3.mcit.com>
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_0006_01BE6CB8.1F212240"
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
In-reply-to: <E32BD2062DAFD1119CFD0000F804F31702C0486F@cosexch003.mcit.com>
Importance: Normal
X-MIMEOLE: Produced By Microsoft MimeOLE V4.72.3155.0
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.

------=_NextPart_000_0006_01BE6CB8.1F212240
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit

Symbol Technologies has a nifty cordless H.323 phone (called NetVision). I
think I have two of them in the lab here -- we showed them at the Feb 19th
STARtech show talking to MS NetMeeting 2.1

see http://www.symbol.com/

These phones use RF Ethernet, so you need a Symbol AccessPoint 24 repeater
somewhere nearby. Symbol also makes PCMCIA and ISA adapters for PC to this
RF ethernet. They seem to have about a 100' range indoors.

Cisco also makes a single-port H.323 adapter -- called the 2611 router with
one FXS and one Ethernet port. Of course, it is a little pricey at around
$5,000

--
Dean
    -----Original Message-----
    From: owner-iptel@lists.research.bell-labs.com
[mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Hoefelmeyer,
Ralph (MCI)
    Sent: Friday, March 12, 1999 4:26 PM
    To: IPTEL IETF Mailing List
    Subject: H.323 black phone???
    Importance: High


    Hello,

    I've been searching the web, and have not found an H.323 black phone.
Everything is a PC card or a videophone.  Some manufacturers claim they'll
have a unit in " a few months".

    Does anyone know of an H.323 phone I can buy now?  FYI, Mediatrix makes
a SIP phone ... I have seen it.

    Thanks,
    Ralph

    Ralph S. Hoefelmeyer
    MCI WorldCom Engineer




------=_NextPart_000_0006_01BE6CB8.1F212240
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">
<HTML>
<HEAD>

<META content=3D"text/html; charset=3Dus-ascii" =
http-equiv=3DContent-Type><TITLE>H.323 black phone???</TITLE>
<META content=3D'"MSHTML 4.72.3612.1700"' name=3DGENERATOR>
</HEAD>
<BODY>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =
size=3D2>Symbol=20
Technologies has a nifty cordless H.323 phone (called NetVision). I =
think I have=20
two of them in the lab here -- we showed them at the Feb 19th STARtech =
show=20
talking to MS NetMeeting 2.1</FONT></SPAN></DIV>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =

size=3D2></FONT></SPAN>&nbsp;</DIV>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =
size=3D2>see <A=20
href=3D"http://www.symbol.com/">http://www.symbol.com/</A></FONT></SPAN><=
/DIV>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =

size=3D2></FONT></SPAN>&nbsp;</DIV>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =
size=3D2>These=20
phones use RF Ethernet, so you need a Symbol AccessPoint 24 repeater =
somewhere=20
nearby. Symbol also makes PCMCIA and ISA adapters for PC to this RF =
ethernet.=20
They seem to have about a 100' range indoors.</FONT></SPAN></DIV>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =

size=3D2></FONT></SPAN>&nbsp;</DIV>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =
size=3D2>Cisco=20
also makes a single-port H.323 adapter -- called the 2611 router with =
one FXS=20
and one Ethernet port. Of course, it is a little pricey at around=20
$5,000</FONT></SPAN></DIV>
<DIV>&nbsp;</DIV>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =

size=3D2>--</FONT></SPAN></DIV>
<DIV><SPAN class=3D110022700-13031999><FONT color=3D#0000ff face=3DArial =

size=3D2></FONT></SPAN><SPAN class=3D110022700-13031999><FONT =
color=3D#0000ff=20
face=3DArial size=3D2>Dean</FONT></SPAN></DIV>
<BLOCKQUOTE=20
style=3D"BORDER-LEFT: #0000ff solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: =
5px">
    <DIV class=3DOutlookMessageHeader><FONT face=3D"Times New Roman"=20
    size=3D2>-----Original Message-----<BR><B>From:</B>=20
    owner-iptel@lists.research.bell-labs.com=20
    [mailto:owner-iptel@lists.research.bell-labs.com]<B>On Behalf Of</B> =

    Hoefelmeyer, Ralph (MCI)<BR><B>Sent:</B> Friday, March 12, 1999 4:26 =

    PM<BR><B>To:</B> IPTEL IETF Mailing List<BR><B>Subject:</B> H.323 =
black=20
    phone???<BR><B>Importance:</B> High<BR><BR></FONT></DIV>
    <P><FONT color=3D#0000ff face=3DArial>Hello,</FONT> </P>
    <P><FONT color=3D#0000ff face=3DArial>I've been searching the web, =
and have not=20
    found an H.323 black phone.&nbsp; Everything is a PC card or a=20
    videophone.&nbsp; Some manufacturers claim they'll have a unit in =
&quot; a=20
    few months&quot;.</FONT></P>
    <P><FONT color=3D#0000ff face=3DArial>Does anyone know of an H.323 =
phone I can=20
    buy now?&nbsp; FYI, Mediatrix makes a SIP phone ... I have seen =
it.</FONT>=20
    </P>
    <P><FONT color=3D#0000ff face=3DArial>Thanks,</FONT> <BR><FONT =
color=3D#0000ff=20
    face=3DArial>Ralph</FONT> </P>
    <P><FONT color=3D#0000ff face=3DArial>Ralph S. Hoefelmeyer</FONT> =
<BR><FONT=20
    color=3D#0000ff face=3DArial>MCI WorldCom Engineer</FONT>=20
</P><BR></BLOCKQUOTE></BODY></HTML>

------=_NextPart_000_0006_01BE6CB8.1F212240--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 12 21:15:55 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA21057
	for <iptel-archive@odin.ietf.org>; Fri, 12 Mar 1999 21:15:55 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 9363C5315; Fri, 12 Mar 1999 17:09:38 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2218C5318; Fri, 12 Mar 1999 17:09:07 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <4.1.19990312152230.00a20380@popd.ix.netcom.com>
X-Sender: rshockey@popd.ix.netcom.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1 
Date: Fri, 12 Mar 1999 15:28:20 -0600
To: Christian Huitema <huitema@research.bellcore.com>
From: Richard Shockey <rshockey@ix.netcom.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: Scott Petrack <scott.petrack@metatel.com>,
        iptel <iptel@lists.research.bell-labs.com>
In-Reply-To: <990311181245.ZM1353@seawind.research.bellcore.com>
References: <Matt Squire <msquire@nortelnetworks.com>
 <3.0.32.19990309170219.00aa9dc0@bl-mail1.corpeast.baynetworks.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>This means that we will have a two phase look-up:
>
>a) maybe using DNS:
>	look up number, get provider ID.
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^      

Was'nt this within the charter of the ENUM BOF?  Number translation sounds
a bit like number resolution to me.

>b) within GLP:
>	maintain routes based on provider ID, and maybe also prefix.
>
>-- 
>Christian Huitema


>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Richard Shockey
Shockey Consulting LLC                  
8045 Big Bend Blvd. Suite 110    
St. Louis, MO 63119
Voice 314.918.9020
Fax   314.918.9015
INTERNET Mail & IFAX : rshockey@ix.netcom.com
eFAX 815.333.1237  
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 00:13:52 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA24368
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 00:13:51 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A7CC1530C; Fri, 12 Mar 1999 15:48:42 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2D57B530A; Fri, 12 Mar 1999 15:45:50 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Lu, Henry" <hwlu@tri.sbc.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>, <hsalama@cisco.com>
Cc: "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Fri, 12 Mar 1999 12:02:11 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFGEBFEFAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Importance: Normal
In-Reply-To: <8391FE2580A1D11198CB00A0C99845D467D18D@tripacific.ttl.pactel.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

> -----Original Message-----
> From: Lu, Henry [mailto:hwlu@tri.sbc.com]
> Sent: Thursday, March 11, 1999 7:16 PM
> To: 'David R. Oran'; Jonathan Rosenberg; hsalama@cisco.com
> Cc: iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> Dear David,
>
> I think what we can not afford is place a call to wrong number.
Define "wrong number", please. If I have an 800 number and call to say
that I want it to ring 555-4545 instead of 555-7676 and then hang up, is
it "placing a call to a wrong number" if 555-7676 rings for the next 30
seconds while the tables are getting updated? This is no different from
the convergence properties of IP routing.

> Whether we will reveal the true number or not is not
> very important from my perspective.
Really? This is not what I hear from my service provider friends, but as
I said I'd rather not worry about this so let's drop it for now.
Jonathan rightly points out that as long as you limit translation result
visibility to parts of the network the owner of the translation trusts,
things will be OK. The porblem arises if you simultaneously want to have
optimal routing AND information hiding.

> You have to know
> that SS7 network is not like IP network, there is no
> network topology database information in SCP like
> OSPF or BGP in IP world.
Yes, I know this.

> This create problem for reconstructing
> call routing from IP world. SS7 is per call base and do
> a SCP query if there is 800 or ported number. No matter
> SIP or H.323, will be per call based through gateway query.
> We are still lock the protocol in old ways per call based. We can
> not update IP world PSTN topology like what we are doing
> in OSPF or BGP for the whole network in one database sync.
Ok, I'll bite. Why not? Remember, we're not doing destination selection
in GLP, we're only doing route selection once the final routing-number
is already known.

> I hope it will help. Thanks.
>
> Henry W. Lu
>
>
> -----Original Message-----
> From: David R. Oran [mailto:oran@cisco.com]
> Sent: Tuesday, March 09, 1999 1:12 PM
> To: Jonathan Rosenberg; hsalama@cisco.com
> Cc: iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> I'm reluctant to bring this up, because some would view it as a corner
> case or just a "feeeeture" of the PSTN, but when the PSTN does number
> translation (or call forwarding in the network as well), the resulting
> number is not revealed to the calling party. The pure directory scheme
> would reveal the translated number since that's what would be
> put in the
> end-to-end signaling (e.g. SIP, H.323). Do we care about this?
>
> > -----Original Message-----
> > From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
> > Sent: Tuesday, March 09, 1999 1:43 PM
> > To: hsalama@cisco.com
> > Cc: John Schnizlein; Greg Zhang; David R. Oran; Matt Squire; Scott
> > Petrack; Alok Mittal; iptel
> > Subject: Re: GLP, number portability and wireless subscribers
> >
> >
> > Hussein F. Salama wrote:
> > >
> > > Jonathan Rosenberg wrote:
> > >
> > > >
> > > > This begs the question, then: does the translation of the
> > 800 number get
> > > > done before the gateway selection process, or after? By
> > doing it before,
> > > > a gateway can be chosen which is, in fact, close to the
> > real number
> > > > being dialed. With 800 numbers, all gateways will be able
> > to reach it.
> > > > It seems that this opens the possibility of very
> > non-optimal routing. A
> > > > caller in NY that dials an 800 number, goes to a gateway
> > in CA. The
> > > > phone network there determines that the 800 number
> translates to a
> > > > number in NY, and routes the call through the PSTN back
> > to NY. However,
> > > > if the translation occurs before gateway selection, the
> caller can
> > > > choose a local NY gateway and do much better.
> > >
> > > Because of the above example, I think all types of
> translation (800
> > > numbers or LNP) and mapping should occur prior to call
> > routing. This is
> > > particularly important for 800 numbers, because they have
> a regional
> > > scope in many cases. The same 800 number may owned by
> > company X in CA
> > > and company Y in NY. In the above example, the call will be
> > incorrectly
> > > routed to company X instead of going to the originally intended
> > > destination, company Y.
> >
> > This is another good example of why translation before
> > gateway selection
> > is a good idea. However, I'm not sure there is much we can do from a
> > protocol perspective to prevent it. There may be valid reasons why a
> > company would like to advertise an 800 number (only they own it, for
> > example), and I'd rather not exclude this possibility. It
> seems like a
> > sensible approach is to allow non-routing numbers in GLP,
> but mention
> > that there are issues with them, and translation before routing is a
> > good idea. Does this sound sensible?
> >
> > Scott P., of course, is right, that the translation service
> itself is
> > outside the scope of iptel.
> >
> > -Jonathan R.
> > --
> > Jonathan D. Rosenberg                       Lucent Technologies
> > Member of Technical Staff                   101 Crawfords Corner Rd.
> > High Speed Networks Research                Holmdel, NJ 07733
> > FAX:   (732) 834-5379                       Rm. 4C-526
> > EMAIL: jdrosen@bell-labs.com
> > URL: http://www.cs.columbia.edu/~jdrosen
> >
>
>
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 00:33:06 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA24942
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 00:33:05 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A6ED052D1; Fri, 12 Mar 1999 14:09:56 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 805B28270; Fri, 12 Mar 1999 13:34:39 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D18F@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'Louise F. A. Spergel'" <spergel@lucent.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Greg Zhang <gzhang@utstar.com>, "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Fri, 12 Mar 1999 12:18:46 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear Louise,
Did SIGTRAN group promise to deliver that through SS7
gateway from IP world can get right PSTN call routing
in the case of 1-800 and number been ported? I think
the problem is in SS7 world and need to be fix there
in stead of doing patches in IP world. It is just like
shooting a moving duck. Thanks.
---Henry W. Lu

-----Original Message-----
From: Louise F. A. Spergel [mailto:spergel@lucent.com]
Sent: Wednesday, March 10, 1999 8:19 AM
To: Jonathan Rosenberg
Cc: Greg Zhang; David R. Oran; Matt Squire; Scott Petrack; Alok Mittal;
iptel
Subject: Re: GLP, number portability and wireless subscribers


My comments below are specific to the US. I don't have any knowledge on
other countries.

With 800 numbers, the issue is different due to the current implementation
and regulatory constraints. In 800 service, the called party pays,
therefore the called party picks the carrier, which currently are all
PSTN-type carriers. The carrier for each number is contained in a database
which is accessed by SS7 TCAP. This database is maintained by a neutral
party. In the near-term, you will need to consult this particular database
and to find the right carrier and then hand the call to a gateway to that
carrier. 

At some point I think the SIGTRAN protocols will enable this.

Of course, in the long run, this may change somewhat.

Louise


At 10:27 AM 3/9/99 -0500, Jonathan Rosenberg wrote:
>Greg Zhang wrote:
>> 
>> Hello,
>> 
>> One issue needs to be considered is number portability. Number
portability
>> (NP) is reality for wired POTS lines for some time now. Wireless Number
>> Portability will be a reality in a few years' time. The portable numbers
>> cannot
>> be routed by strictly following the hierarchy derived from the digits of
>> called
>> number. 
>
>I agree LNP makes the problem harder. I have two responses here:
>
>1. Simply "not aggregating" is always an option for a protocol which
>provides aggregation. I don't want to not support aggregation because of
>LNP. If LNP gets in the way, it can just be turned off if needed.
>
>2. I'm not yet convinced LNP actually prevents aggregation. Lets say
>gateway X can reach ten numbers, +1 607-1* (using glob matching), +1
>607-2*, etc. so these could be aggregated to +1 607*. Lets further
>assume all ten numbers are LNP, and actually terminate all over the
>country. These numbers can still all be reached through the gateway.
>They cannot be aggregated only if the attributes for each are different.
>The only attribute likely to differ is cost. So, the question is, does
>the cost (to the gateway provider) for terminating calls to each of the
>numbers differ? Even if they do, is the cost the gateway provider wants
>to charge to its users different for each number? If the answer to
>either is no, then they can still be aggregated. With long distance
>rates at 10 cents a minute *anywhere* in the U.S., this seems not so
>unlikely.
>
>
>There is a related issue of 800 numbers, and any of the other
>non-routing numbers. Should these be advertised by GLP? Or, should they
>be translated first, and then the translated numbers advertised? (Please
>correct me if I have misunderstood some aspect of how this works in the
>PSTN).
>
>-Jonathan R.
>-- 
>Jonathan D. Rosenberg                       Lucent Technologies
>Member of Technical Staff                   101 Crawfords Corner Rd.
>High Speed Networks Research                Holmdel, NJ 07733
>FAX:   (732) 834-5379                       Rm. 4C-526
>EMAIL: jdrosen@bell-labs.com
>URL: http://www.cs.columbia.edu/~jdrosen
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.
>

---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 00:40:21 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA24962
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 00:40:21 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C75588081; Thu, 11 Mar 1999 17:29:45 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1B80A8007; Thu, 11 Mar 1999 17:14:03 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: Santo Wiryaman <swiryaman@VIDEOSERVER.COM>
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <65256730.00190A40.00@sampark.hss.hns.com>
Date: Wed, 10 Mar 1999 10:26:09 +0530
Subject: RE: call clearing by GK
Mime-Version: 1.0
Content-type: multipart/mixed; 
	Boundary="0__=DksaaLS18ZEnsZmIZVi55jEew1uONru7FLQZIrduIcCns450Yfb6cIrk"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--0__=DksaaLS18ZEnsZmIZVi55jEew1uONru7FLQZIrduIcCns450Yfb6cIrk
Content-type: text/plain; charset=us-ascii

Hi All,
      I agree with Santo.
In case of normal scenario, when GK had received ARQ, then GK can release
the call using DRQ, irrespective of the call model. In this case, GK has
all the info about CRV and other parameters.

In case of preGranted Arq, two cases are direct model and GK-routed model.
In case of direct model, because neither ARQ was received not any call
signalling message, so there is no point of GK trying to disengage the call
(As Chris pointed out.) because GK is not even aware of the call. In other
GK-routed model, because GK does not know CRV and other essential
parameters, it can not send DRQ. As "whether RAS CRV and call siganlling
CRV have to be same or not", is still not clear, GK can not bank on Call
Signalling CRV to send DRQ. So Release Complete is the only option left to
release the call. Moreover there are other scenarios also where GK can send
Release Complete to any of the endpoint, involved in the call (like
mandatory elements missing in a message etc.), so from feasibility and
possiblity point of view this should not be an issue. Moreover according to
the scenario given by Santo, B can receive release Complete sent by A and
relayed by GK. In my opinion, endpoint B is transparent to whether Release
Complete was initiated by endpoint A or GK.

Thanks,
Ritu.





Santo Wiryaman <swiryaman@VIDEOSERVER.COM> on 03/10/99 01:16:49 AM

Please respond to Santo Wiryaman <swiryaman@VIDEOSERVER.COM>

To:   h323implementors@imtc.org
cc:
Subject:  RE: call clearing by GK




Hi Srini,
The key word in section 8.5.3 is "may".  Furthermore, the section refers to
Figure 34 which suggests a Direct-routed GK, in which case sending a DRQ is
the only thing the GK can do.  In a GK-routed case, the GK may also send
ReleaseComplete first, instead of a DRQ, in my opinion.
Consider A is in a call with B through GK (GK-routed).  When A initiates
the
disconnect, the following may take place:
1. A sends H.245 EndSession to GK
2. GK "relays" the EndSession to B
3. A sends ReleaseComplete to GK
4. GK "relays" the ReleaseComplete to B
5. A sends DRQ to GK
6. B sends DRQ to GK
The point is, endpoint B must handle a disconnect procedure that is
initiated by A. Therefore, the GK should be able to disconnect B using the
same sequence as if the the disconnect had been initiated by A.
Regards,
Santo Wiryaman
VideoServer Inc.
swiryaman@videoserver.com
> -----Original Message-----
> From:   srini@trillium.com [SMTP:srini@trillium.com]
> Sent:   Tuesday, March 09, 1999 2:15 PM
> To:     CPurvis@MADGE.COM; h323implementors@imtc.org;
> ITU-SG16@mailbag.cps.intel.com
> Subject:     Re: call clearing by GK
>
> Please find my comments inline.
>
> thanks,
> Srini
> Trillium Digital Systems, Inc.
>
>
> >
> > Oops!
> > Sorry about the blank response!  Let's try again!
> >
> > This isn't QUITE as simple a question as it looks.
> > I'd recommend looking at it from the two separate points of view:
> > >From the endpoint's point of view, I would say the answer is "yes" (as
> Ritu
> > says).  A gatekeeper is allowed to send you a DRQ to clear the call.
> > Depending on how the endpoint is written, though, this may not be
> trivial to
> > achieve, since without pregranted ARQs it is reasonable for an endpoint
> to
> > want to match up DRQs with individual ARQs.
> >
> > As a gatekeeper, though, I wonder why one would want to implement that.
> For
> > the gatekeeper to be aware of a call when using pregranted ARQs, it
must
> be
> > routing call signalling, so it can quite happily terminate the call by
> means
> > of ReleaseComplete - which will NOT break poorly written endpoints.
> >
> > The related question is whether an endpoint sends a DRQ on "normal"
call
> > termination in these cases.  If the gatekeeper is routing
> call-signalling it
> > is arguably unnecessary (ReleaseComplete clears the call); if not, then
> > potentially the first the gatekeeper knows about the call is the DRQ at
> its
> > end (not useful!)!  Note that H.323v2 section 8.5.2 assumes that ARQ
has
> > happened at the beginning of the call, so behaviour where this is
absent
> > (due to pregranted ARQ) is arguable.
>
>
> In normal case, where ARQ happend at the beginning of the call, H.323V2
> section 8.5.3 clearly says GK may terminate the call by sending DRQ and
> endpoints shall follow the steps 1-6 of section 8.5, where the step
> 6 is sending the Release Complete. Hence, IMO, GK shall not send Release
> Complete to clear the call in normal case, it shall use DRQ. In the
> preGrantedArq case, I agree with Chris that sending Release Complete is
> one possible way to clear the call by GK.
>
>
>
> >
> > This leaves a question mark over the use of the other call-specific RAS
> > message, BRQ.  Bandwidth issues with pregranted ARQs are discussed in
> both a
> > contribution and a TD from the recent Monterey meeting (I forget the
> > numbers).
> >
> > Regards,
> > Chris
> > --
> > Dr Chris Purvis - Senior Development Engineer, WAVE CC Software
> > Madge Networks Ltd, Wexham Springs, Framewood Road, Wexham, Berks.
> ENGLAND
> > Phone: +44 1753 661 359  email: cpurvis@madge.com
> -------------------------------------------------------------------------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"


--0__=DksaaLS18ZEnsZmIZVi55jEew1uONru7FLQZIrduIcCns450Yfb6cIrk
Content-type: application/octet-stream; 
	name="att1.unk"
Content-Transfer-Encoding: base64
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--0__=DksaaLS18ZEnsZmIZVi55jEew1uONru7FLQZIrduIcCns450Yfb6cIrk--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 00:46:33 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA25010
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 00:46:33 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 63515800A; Thu, 11 Mar 1999 11:08:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9B4C77FE0; Thu, 11 Mar 1999 11:07:47 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E56BB7.E1A0DB91@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 13:43:03 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <010501be69aa$79336240$2821a8c0@greg_zhang.wacos.com> <199903091625.IAA01401@mailman.cisco.com> <36E54E32.C8931AB2@dnrc.bell-labs.com> <36E55B77.B5105476@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> Jonathan Rosenberg wrote:
> 
> >
> > This begs the question, then: does the translation of the 800 number get
> > done before the gateway selection process, or after? By doing it before,
> > a gateway can be chosen which is, in fact, close to the real number
> > being dialed. With 800 numbers, all gateways will be able to reach it.
> > It seems that this opens the possibility of very non-optimal routing. A
> > caller in NY that dials an 800 number, goes to a gateway in CA. The
> > phone network there determines that the 800 number translates to a
> > number in NY, and routes the call through the PSTN back to NY. However,
> > if the translation occurs before gateway selection, the caller can
> > choose a local NY gateway and do much better.
> 
> Because of the above example, I think all types of translation (800
> numbers or LNP) and mapping should occur prior to call routing. This is
> particularly important for 800 numbers, because they have a regional
> scope in many cases. The same 800 number may owned by company X in CA
> and company Y in NY. In the above example, the call will be incorrectly
> routed to company X instead of going to the originally intended
> destination, company Y.

This is another good example of why translation before gateway selection
is a good idea. However, I'm not sure there is much we can do from a
protocol perspective to prevent it. There may be valid reasons why a
company would like to advertise an 800 number (only they own it, for
example), and I'd rather not exclude this possibility. It seems like a
sensible approach is to allow non-routing numbers in GLP, but mention
that there are issues with them, and translation before routing is a
good idea. Does this sound sensible?

Scott P., of course, is right, that the translation service itself is
outside the scope of iptel. 

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 00:53:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA25025
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 00:53:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 90DA58079; Thu, 11 Mar 1999 17:29:26 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F1D7780BE; Thu, 11 Mar 1999 17:14:02 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Tue, 9 Mar 1999 23:53:02 -0500 (EST)
Message-Id: <199903100453.XAA11049@conrail.cs.columbia.edu>
From: Jonathan Lennox <lennox@conrail.cs.columbia.edu>
To: List IPTel <iptel@lists.research.bell-labs.com>
Subject: Re: A question about CPL
In-Reply-To: <003001be677d$657cc040$9c4566ca@xmzhao>
References: <003001be677d$657cc040$9c4566ca@xmzhao>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

On Sat, March 6 1999, "Xiaoming Zhao" wrote to "iptel@lists.research.bell-labs.com" saying:

> Hi,
>   I'm a freshman on IP telephony, especially CPL. After I read the CPL
> draft, I have a question.
> 
> What's the relationship between several scripts for one user?
>     For example, Jones is an IPtel user. Of course a script for what
> actions should be taken when calls for Jones come is needed. However,
> another script used when Jones calls out for someone else is also
> needed. How can these two scripts work correctly together?

This is discussed somewhat in the CPL framework/requirements document; That
draft may have expired by now, but you can find it in several formats linked
from <http://www.cs.columbia.edu/~lennox/>
(draft-ietf-iptel-cpl-requirements-00).  I was hoping to do a revision of it
for the Minneapolis IETF, but didn't get it done in time.  This is probably
an issue which needs to be clarified further in the revision.

There are two questions here: one is can a user have several scripts, and
the second is how do several scripts interact.  In the first case, the
answer is yes; I envision a user typically having separate incoming and
outgoing scripts.  The first of these is invoked when a call reaches the
server with Jones in the To: field; the second one when Jones is in the From
field.  (How these scripts are distinguished to the server is an open issue
for draft-ietf-iptel-sip-reg-payload-00, and any other CPL transport
protocols.)  I'm not sure what other categories of scripts are necessary for
a CPL.  

The second question is theoretically simpler, though implementation of it
will probably be tricky.  This isn't so much a case of the interaction of
Jones's outgoing script and Jones's incoming script so much as it is the
interaction of Jones's outgoing script and Smith's incoming script, when
Jones calls Smith.  In this case, you don't want this to behave any
differently whether Jones and Smith are using the same CPL server or two
different ones -- the call will first be handled by Jones's script, and if
that script directs the call to be placed to Smith, Smith's script will be
invoked.  Whether signalling information was sent over the wire in between
these two invocations (two different CPL servers) or not (one CPL server
with two users' scripts) is irrelevant.

Things get a little trickier if you want to add administrative scripts to
the mix, because some administrative scripts have to run before user
scripts, and some afterward, particularly for things like administrative
incoming and outgoing call screening where the user is potentially trying to
evade the administrator's restrictions.  But the basic model holds: we can
take advantage of the distributed nature of IPTel call processing to model
multiple scripts as though they always ran on separate servers, even when
they actually don't.

-- 
Jonathan Lennox
lennox@cs.columbia.edu

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 01:12:32 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA25242
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 01:12:31 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 87A4B5316; Fri, 12 Mar 1999 17:10:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DBB9A5314; Fri, 12 Mar 1999 17:10:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <5CB32D2B7DECD111AEA200805F95D60F022B3707@NJ0117EXCH001U>
From: "Rollender, Douglas Harold (Douglas)" <rollender@lucent.com>
To: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>,
        Scott Petrack <scott.petrack@metatel.com>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Fri, 12 Mar 1999 16:24:09 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Translation before or after gateway selection seems to depend on the service
architecture (i.e., 800 is not the same as LNP) and the position of the
gateway service provider in the end-to-end routing of the call.  Is there
one "correct" model top apply in all cases?  Is one needed?...Doug.

	----------
	From:  Scott Petrack [SMTP:scott.petrack@metatel.com]
	Sent:  Tuesday, March 09, 1999 12:30 PM
	To:  pcalhoun@eng.sun.com
	Cc:  Jonathan Rosenberg; John Schnizlein; Greg Zhang; David R. Oran;
Matt Squire; Alok Mittal; iptel
	Subject:  Re: GLP, number portability and wireless subscribers



	On Tue, 9 Mar 1999, pcalhoun@eng.sun.com wrote:

	[Jonathan R:}
	> > This begs the question, then: does the translation of the 800
number get
	> > done before the gateway selection process, or after? By doing it
before,
	...
	> > if the translation occurs before gateway selection, the caller
can
	> > choose a local NY gateway and do much better.
	> 
	> Translation before the gateway selection is *obviously* the
correct model. We
	> should strive towards this goal.
	> 
	> PatC
	> 

	Well, yes, but there are lots of 800 numbers whose translations
change on
	time scales MUCH slower than the GLP advertisement changes. In fact,
most
	800 numbers are this way -- they are used by small-to-medium
	businesses, point to the 'real' telephone number, with no
	logic at all, and don't change at all.

	If a GW wants to advertise that it can service such a number, I
don't see
	why it should be forbidden from doing so. 

	[Of course there is a need for a separate look-up and translation
service,
	but this is outside the scope of Iptel - right Jonathan? ;-)]

	The general problem of for how long the advertisements are valid,
and how
	quickly change can be propogated, is just that general -- there is
nothing
	special about 800 numbers or other directory numbers. 

	Scott 


	---------
	This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 02:11:34 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA02194
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 02:11:33 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C84E98025; Thu, 11 Mar 1999 17:35:34 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C79FC8022; Thu, 11 Mar 1999 17:30:43 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990310082902.007e8860@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Wed, 10 Mar 1999 08:29:04 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: Greg Zhang <gzhang@utstar.com>, hsalama <hsalama@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>, "David R. Oran" <oran@cisco.com>,
        iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 06:56 PM 3/9/99 -0500, Jonathan Rosenberg wrote:
>David R. Oran wrote:
>> 
>> Then scopes are identical to a deeper hierarchy and splitting them out
>> is simply a syntactical, not a semantic distinction, right?
>
>My understanding is that it is semantic. By "splitting them out", I
>presume you mean having them be separate from the address itself. The
>prefix represents information on "where you're calling to", and the
>scope information on "where you are calling from". So, a prefix Y with
>scope X means you can call Y from X. I think we need explicit scopes if
>there are cases where route attributes for a prefix only apply when
>calling from numbers outside of the prefix (and apply only when calling
>from somewhere defined by the scope).


I don't see how we can guarantee any determination of 'where you are
calling from.'  What does my pc call from?  Where is an LS?  A gateway
might be considered to have a 'location', where the location is given by
the area code of its PSTN connectivity, but other devices without PSTN
connectivity aren't tied to a PSTN location.

It seems like any definition of scope along these lines requires alot of
administrative help, ie where LSs are tied/located with certain geographic
areas.  Would not a TTL-like attribute provide similar functionality (again
based on geographic ties/connectetions in LSs)?  A number could be
advertised with a certain TTL that prevents it from being advertised after
too many hops.  If your connections are geography-based, then this gives
you geographic scoping.  We make an additional statement that given
multiple potential routes to a destination, the one with the higher/lower
TTL is preferred (because its 'closer').  

>
>> 
>> Do you think we would need hierarchical scopes? If so, do you see
>> aggregation of hierarchical scopes? How would this differ from
>> hierarchical aggregation of addresses? Are scopes always more global
>> than addresses?
>
>We could potentially define hierarchical scopes (if they are also
>represented as prefixes, they are hierarchical). The rules for
>aggregation would presumably be the same. But, adding another
>aggregatable entity in a route advertisement seems like it will make
>route aggregation overall more complex, so I'd rather avoid this. Are
>scopes global? I think they need to be. The whole idea is that a prefix
>advertised in one place is only useful somewhere else. 
>

Unless someone says something to convice me otherwise, I'd vote against
this concept of scoping.  

- Matt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 06:12:46 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA03910
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 06:12:46 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8760E531D; Fri, 12 Mar 1999 17:14:40 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1AD4E531A; Fri, 12 Mar 1999 17:11:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D191@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'Matt Squire'" <msquire@nortelnetworks.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: RE: data issue 4: capacity
Date: Fri, 12 Mar 1999 15:08:45 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear Matt,

Great job you have done. You had asked right question and 
I do not have the answer for you on most of them. It look 
like you will need to solve them and I think you will be
able to get them done in the future.
Here is my two cents on capacity issues from my perspective.
I think the choice of BGP or SCSP should be based on what kind
of Gateway you are working with. If you are dealing with IP
per call based QoS, SCSP will be fine. But if you are dealing
with trunking, ATM gateway, you may need BGP-4 to have detail
QoS and mostly will follow BISDN or ATM Qos format as their 
attributes. It is a industry trend that ATM will handle trunking
QoS and beyond edge router/switch will be IP QoS. You will mostly
get a 56K channel in PSTN so there is no need to detail the QoS 
Request from IP side. BGP-4 in IP world is only for inter-AS routing
mostly. They may need to deal with trunking QoS. I hope it will
help. Wish you great success next week. Thanks.

Henry W. Lu

-----Original Message-----
From: Matt Squire [mailto:msquire@nortelnetworks.com]
Sent: Thursday, March 11, 1999 6:21 PM
To: IPTEL IETF Mailing List
Subject: data issue 4: capacity



4) Capacity.

I think we have consensus that a capacity attribute is needed.  What we
mean by capacity, unfortunately, is still up in the air.

  a) Should capacity be unitless and relative, or measurable and
     absolute?  
  b) If measurable/absolute, what are the appropriate metric(s).  Some
     possibilities are bandwidth, maximum circuits, load, (others?)...
  c) What are the timescales that capacity can be changed?  In other
     words, does it represent some maximum capacity (constant) or
     current capacity (depends on load).
  d) Is capacity an attribute of an egress gateway, or to the signalling
     path, both, or do we need two different capacity measurements (one
     for path, another for egress gw)?

The definition and use of capacity will most likely differ depending upon
your preference for the SCSP or BGP approach.  With the SCSP approach, I
suggested a unitless, relative capacity which represents the capacity of
the egress gateway.  I thought that we had come to such a consensus during
Jonathan's presentation at the last wg meeting.  

With the SCSP approach, this makes more sense because the its very
difficult to infer anything about the capacity of the IP path because of
the explicit separation of the call routing from the IP routing.  With the
SCSP approach, an LS that aggregates or changes the next hop information
could modify the capacity to include the capacity of the path between the
LS and its next-hop.  The value of such a modification is limited because
it does not reflect the capacity of the data path, which is much more
likely to be the bottleneck.

With the BGP approach, carrying e164 numbers in BGP integrates data and
call routing, and thus capacity information about the IP path can be
integrated into the capacity metric.  In the current TBGP draft, it is
considered an open issue whether TBGP is a simple extension to BGP (and
thus yields this integrated routing) or whether TBGP is a different
instance of the protocol.  I believe that if TBGP is a different instance
of the protocol (and is separated from the IP routing information), then it
will not be able yield an accurate reflection of path capacity.  

A unitless measurement has a disadvantage in that it is an administrative
manner to ensure that the weight of a gateway reflects is capacity relative
to other gateways.  In an inter-domain protocol, this requires some
administrative coordination across domains.  A measurable capacity would be
nice, but it seems that no single metric would suffice, and if we permit
multiple metrics, then we're faced with the problem of determining which
capacities are larger than others (ie is 20 circuits better than 100 Mbs or
not).  

In either case, I'd recommend that the capacity attribute NOT reflect
anything about the real-time load on a gateway or path.  Due to the
complications of multiple metrics, I'd also recommend that capacity remain
unitless and that it be an administrative exercise for the capacity to be
consistent across the confederation of iptel domains.  Hence, I'd propose a
simple 4-octet field that can be used do determine relative capacity.




---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 07:07:29 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA04237
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 07:07:29 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D54C8534B; Fri, 12 Mar 1999 23:06:45 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A97A55326; Fri, 12 Mar 1999 23:06:42 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Organization: NetSpeak Corporation
Message-Id: <4.1.19990312220713.0096bc30@crash.netspeak.com>
X-Sender: cstrickland@crash.netspeak.com (Unverified)
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1 
Date: Fri, 12 Mar 1999 22:34:12 -0500
To: Hanlin Fang <Hanlin_Fang@3com.com>
From: Craig Strickland <craig@netspeak.com>
Subject: RE: Handling of Setup message by Gatekeeper.
Cc: IETF IP Telephony <iptel@lists.research.bell-labs.com>,
        "H.323 Implementors Group" <h323implementors@imtc.org>,
        Ronald Davis <ronaldd@lucent.com>, Deepak Kumar <deepak@microsoft.com>,
        Ritu Gupta <rtgupta@hss.hns.com>
In-Reply-To: <88256732.000726A2.00@hqoutbound.ops.3com.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 17:15 03/11/99 -0800, Hanlin_Fang@3com.com wrote:
>How can Gatekeepers know that there are some endpoints in its zone
>not registered? If GKs don't know their existence, then how can those
>endpoints be managed?

Actually, this question is "self-disproving", based on the definition of a
"zone" and the relationship of endpoints to the zone.

A couple of references from H.323 v2:

3.49	zone: A Zone (see Figure 3) is the collection of all terminals (Tx),
Gateways (GW), and Multipoint Control Units (MCU) managed by a single
Gatekeeper (GK).  ...  A Zone has one and only one Gatekeeper.

7.2.2	Endpoint registration
Registration is the process by which an endpoint joins a Zone  ...  As part
of their configuration process, all endpoints shall register with the
Gatekeeper...  Registration shall occur before any calls are attempted ...

Essentially, by the very definition of "zone", it's not possible to have an
unregistered endpoint "within" the zone.  A "zone" is comprised of a
[single] Gatekeeper and the [registered] endpoints it manages.

All other endpoints are, by definition, "out of zone".  Thus, there is no
semantic meaning for the phrase "unregistered enpoint within the zone".

Although I see where you're trying to go, the following phrase touches on
the goal of attempting to equate an H.323 "zone" with an administrative
"zone of control" (such as a LAN or LAN segment):

3.49	zone:  ...  A Zone has one and only one Gatekeeper. A Zone may be
independent of network topology ...

The latter statement makes it clear that the meaning of a "zone" differs
from the network layout.  There's nothing you can rely on in the contents
of an H.323 message that will reliably inform you if a non-compliant
("rogue") endpoint is being truthful about where it's located.  Many source
identification fields are optional, and if the endpoint is refusing to
RRQ/ARQ, then it's likely to lie about its identity as well.

The best you could hope for if you desired to treat a "rogue" endpoint
differently from an "outsider" would be to examine the transport layers'
information made available about the network connection -- the IP address
and TCP port number.  Assuming the endpoint is not performing "IP spoofing"
(a more difficult technique), the Gatekeeper could be provided with a table
of IP addresses or address ranges which are to be considered "insiders", as
contrasted to "outsiders".  That's probably just about your only option.
--
Craig Strickland
NetSpeak Corporation (home of WebPhone)
http://www.netspeak.com/


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 08:07:25 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA04681
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 08:07:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1C9607FF0; Thu, 11 Mar 1999 11:59:47 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 41E3D7FF6; Thu, 11 Mar 1999 11:52:32 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D185@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'Jonathan Rosenberg'" <jdrosen@dnrc.bell-labs.com>,
        Greg Zhang <gzhang@utstar.com>
Cc: "'pcalhoun@eng.sun.com'" <Pat.Calhoun@Eng.Sun.COM>,
        Matt Squire <msquire@nortelnetworks.com>,
        John Schnizlein <jschnizl@cisco.com>, "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 15:35:39 -0600 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear Jonathan,
What if Lucent and Nortel solve this problem via SCP by including 
local number portability information, and the problem become much simpler. 
All we have to do is sync with SCP database. I understand 
that it is simpler to have local switch to keep local number
portability within one switch and do not have to consult
SCP, if it is local within a switch. The problem is that SCP
do not have these local information for ip world to sync with. 
I agree with you that IPTEL do not have to worry about this problem, 
but this database dependency on SCP and SSP may limit the calling 
scope from IP world. Do you know who we can refer this problem to?

Dear Greg,
I just want to provide information that I think will be helpful.
Based on my understanding, SS7 netwrok is very secured network
and their protocol is not like IP protocol at all. I do not think
that SS7 world will adapt the protocol we are discussing here, because 
the protocol nature of SS7. I hope that you will write the draft and
Lucent and Nortel work out their problem and we can accomplish the goal
to having wireless calling IP world or vis versa. Thanks.

Henry W. Lu
925-598-1245
SBC, Technology Resources, Inc.
4698 Willow Road
Pleasanton, CA 94588

-----Original Message-----
From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
Sent: Tuesday, March 09, 1999 11:31 AM
To: Lu, Henry
Cc: 'pcalhoun@eng.sun.com'; Matt Squire; John Schnizlein; Greg Zhang;
David R. Oran; Scott Petrack; Alok Mittal; iptel
Subject: Re: GLP, number portability and wireless subscribers


Lu, Henry wrote:
> 
> Hello,
> 
> I am in favor of Pat's point but I am still hoping that we can
> achieve a sound solution in order for us to handle number portability
> environment in the future, including wireless.
> 
> The issues I see is how to synchronize two database between GLP and SS7
> world. In AIN, 800 number information are kept in not only Signaling
> Control Point(SCP) but also Service Switching Point(SSP), normally in
> local DMS-100 or 5ESS switches. The 800 calling process had to go through
> a conversation session between SSP and SCP to figure out the final
> routing to the destination. There are two information databases been
> kept on both SCP and local SSP. Do not forget that there are many SSPs
> in the network, not like SCP, we have to deal with.
> The major problem occur when a new portable number been provisioned,
> into both SSP and SCP but there is no information on GLP side.
> GLP and SS7 database is out of sync basically.

As you mention, the somewhat unusual way in which the database is
distributed in the phone network makes this a hard problem. The best way
to keep these databases in sync is to simply not have a separate
database. When a proxy or gatekeeper is going to make a call, it first
queries the phone network (somehow) to translate the number into a
routing number. GLP only advertises routing numbers. As such, the
database of number translations is stored only in the phone network, as
it is today. 

If, however, one chooses to deploy a parallel database in the Internet,
accessible by LDAP or some other means (enum perhaps?), the
synchronization problem exists.

Fortunately, it is outside of our charter to worry about this problem.
We need only decide what numbers may or may not be propagated by GLP,
and how they are represented and aggregated.

-Jonathan R.



-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 08:13:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA04735
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 08:13:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B27AF7FFF; Thu, 11 Mar 1999 11:44:57 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3B576804E; Thu, 11 Mar 1999 11:42:14 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Matt Squire" <msquire@nortelnetworks.com>,
        "Greg Zhang" <gzhang@utstar.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: "hsalama" <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 18:14:50 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFGEKEEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Importance: Normal
In-Reply-To: <3.0.32.19990309170219.00aa9dc0@bl-mail1.corpeast.baynetworks.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Then scopes are identical to a deeper hierarchy and splitting them out
is simply a syntactical, not a semantic distinction, right?

Do you think we would need hierarchical scopes? If so, do you see
aggregation of hierarchical scopes? How would this differ from
hierarchical aggregation of addresses? Are scopes always more global
than addresses?

Are we in sync here or am I being obtuse?

> -----Original Message-----
> From: Matt Squire [mailto:msquire@nortelnetworks.com]
> Sent: Tuesday, March 09, 1999 5:02 PM
> To: David R. Oran; Greg Zhang; Jonathan Rosenberg
> Cc: hsalama; John Schnizlein; Scott Petrack; Alok Mittal; iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> At 04:34 PM 3/9/99 -0500, David R. Oran wrote:
> >Or, alternatively, project all numbers into a global,
> unambiguous space.
> >This seems cleaner and much less error prone than topologically
> >established scope boundaries, and less administratively
> burdensome than
> >solving the problem of scope name collisions and information
> propagation
> >boundaries. It also avoids having to talk about translations between
> >scopes, which I'm sure will happen when the other shoe drops :-).
> >
>
> I'm a little confused by this comment. To me, scopes provide the
> unambiguous global number space.  If there is a scope registration
> authority (maybe this wasn't clear before), then the combination of
> scope+number is globally unambiguous, with the default scope
> being global
> e164.  If someone registers a scope and they wish it to have
> topological
> significance, then that's an administrative matter and they
> should be able
> to set up their peer association and policy accordingly.
>
> I'd probably be against scopes of the form "+1-405" because
> that would just
> lead to too many headaches in interpreting, aggregating, etc.
>
> An opaque scope using a equal/!equal test might provide what
> we desire.  I
> haven't convinced myself completely, but I think it has the potential.
>
> - Matt
>
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 09:07:38 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA05284
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 09:07:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5E34B52E7; Fri, 12 Mar 1999 15:20:58 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3A89C52D4; Fri, 12 Mar 1999 15:20:38 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <009101be6cc3$a2907030$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "IPTEL IETF Mailing List" <iptel@lists.research.bell-labs.com>,
        "Matt Squire" <msquire@nortelnetworks.com>
Subject: Re: number portability
Date: Fri, 12 Mar 1999 12:04:36 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MIMEOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Matt Squire wrote:
>GLP as provides a service to the caller.  Things like 800 numbers provide
>services to the callee.  What motiviation do call-routing providers have to
>route 800 numbers efficiently?  Call-routing provider X could decide to use
>a GW on the west coast for all calls to my 800 number, which is actually
>translated to a number on the east coast.  I could bitch to my 800 number
>provider, but what could they do?  With 'normal' numbers, the caller gets
>billed, and poor routing means lost customers, but not here.  As an 800
>number owner, I do not want my 800 number handled by any other provider's
>call-routing unless I have assurance that it is routed efficiently.  And
>without regulation, I won't get that assurance.

I think what we are talking about is a case of multiple gateways in a
network
that advertise service or connectivity to a 800 number at the same time.
These gateways can be gateways attached to the network of one provider,
or gateways from serveral different providers. Since it is callee pay, the
callee or its provider will need to advertise a preference or priority or
even
cost as a attribute for its attached gateways. Other gateways also
advertising
the connectivity to this 800 number will have a lower preference or priority
value.
Only when the primary provider's gateways cannot be reached, a call is
routed
through the other gateways. Multiple gateways that attached to the primary
provider can also be differentiated through the preference/priority/cost or
distance metric. LS on the west coast will resolve the 800 number to the
primary provider's west coast gateway while a LS on the east coast will
resolve it to the provider's east coast gateway. Would this be a
satisfactory
protocol solution?

Best regards,

- Greg



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 10:13:43 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA05714
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 10:13:43 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id AE0357FF5; Thu, 11 Mar 1999 11:59:53 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id CA7C87FE9; Thu, 11 Mar 1999 11:44:32 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Tue, 9 Mar 1999 19:01:05 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: Matt Squire <msquire@nortelnetworks.com>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>, Alok Mittal <amittal@hss.hns.com>,
        hsalama@cisco.com, iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
In-Reply-To: <3.0.32.19990309140309.00a9dcf0@bl-mail1.corpeast.baynetworks.com>
Message-ID: <Pine.LNX.4.04.9903091854110.12539-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



> >This is another good example of why translation before gateway selection
> >is a good idea. However, I'm not sure there is much we can do from a
> >protocol perspective to prevent it. There may be valid reasons why a
> >company would like to advertise an 800 number (only they own it, for
> >example), and I'd rather not exclude this possibility. It seems like a
> >sensible approach is to allow non-routing numbers in GLP, but mention
> >that there are issues with them, and translation before routing is a
> >good idea. Does this sound sensible?
> >
> 
> This sounds like the right thing to do.  I think the framework document
> should mention where and how translation of such numbers is expected/allowed.
> 

I agree with >> and less so with >  above. Translations
are just outside of Iptel scope.  
It is a separate can of worms, and we have enough of them on our plate 
already (I just love mixed metaphors). 

I think these sorts of translations 
might end up being in ENUM's scope eventually, when we get around to 
that. But let's agree that
GLP doesn't know about non-routing numbers. 

On a separate note.
Is there anything from malloc stuff or scoping of multicast addresses that
we can learn? Probably not, since each address carries its own scope, but
maybe there is. I haven't been following it.

Scottt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 10:41:42 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA05891
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 10:41:42 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B9B138185; Fri, 12 Mar 1999 02:07:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 36E18817E; Thu, 11 Mar 1999 21:44:51 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990311212111.00ab7d80@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 11 Mar 1999 21:21:16 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: data issue 4: capacity
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


4) Capacity.

I think we have consensus that a capacity attribute is needed.  What we
mean by capacity, unfortunately, is still up in the air.

  a) Should capacity be unitless and relative, or measurable and
     absolute?  
  b) If measurable/absolute, what are the appropriate metric(s).  Some
     possibilities are bandwidth, maximum circuits, load, (others?)...
  c) What are the timescales that capacity can be changed?  In other
     words, does it represent some maximum capacity (constant) or
     current capacity (depends on load).
  d) Is capacity an attribute of an egress gateway, or to the signalling
     path, both, or do we need two different capacity measurements (one
     for path, another for egress gw)?

The definition and use of capacity will most likely differ depending upon
your preference for the SCSP or BGP approach.  With the SCSP approach, I
suggested a unitless, relative capacity which represents the capacity of
the egress gateway.  I thought that we had come to such a consensus during
Jonathan's presentation at the last wg meeting.  

With the SCSP approach, this makes more sense because the its very
difficult to infer anything about the capacity of the IP path because of
the explicit separation of the call routing from the IP routing.  With the
SCSP approach, an LS that aggregates or changes the next hop information
could modify the capacity to include the capacity of the path between the
LS and its next-hop.  The value of such a modification is limited because
it does not reflect the capacity of the data path, which is much more
likely to be the bottleneck.

With the BGP approach, carrying e164 numbers in BGP integrates data and
call routing, and thus capacity information about the IP path can be
integrated into the capacity metric.  In the current TBGP draft, it is
considered an open issue whether TBGP is a simple extension to BGP (and
thus yields this integrated routing) or whether TBGP is a different
instance of the protocol.  I believe that if TBGP is a different instance
of the protocol (and is separated from the IP routing information), then it
will not be able yield an accurate reflection of path capacity.  

A unitless measurement has a disadvantage in that it is an administrative
manner to ensure that the weight of a gateway reflects is capacity relative
to other gateways.  In an inter-domain protocol, this requires some
administrative coordination across domains.  A measurable capacity would be
nice, but it seems that no single metric would suffice, and if we permit
multiple metrics, then we're faced with the problem of determining which
capacities are larger than others (ie is 20 circuits better than 100 Mbs or
not).  

In either case, I'd recommend that the capacity attribute NOT reflect
anything about the real-time load on a gateway or path.  Due to the
complications of multiple metrics, I'd also recommend that capacity remain
unitless and that it be an administrative exercise for the capacity to be
consistent across the confederation of iptel domains.  Hence, I'd propose a
simple 4-octet field that can be used do determine relative capacity.




---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 15:57:21 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA08509
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 15:57:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6B0CE8060; Thu, 11 Mar 1999 17:50:06 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9B05880FD; Thu, 11 Mar 1999 17:35:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990310085301.007e7d10@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Wed, 10 Mar 1999 08:53:02 -0500
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: number portability
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


One other issue on number portability while we're discussing it... 

GLP as provides a service to the caller.  Things like 800 numbers provide
services to the callee.  What motiviation do call-routing providers have to
route 800 numbers efficiently?  Call-routing provider X could decide to use
a GW on the west coast for all calls to my 800 number, which is actually
translated to a number on the east coast.  I could bitch to my 800 number
provider, but what could they do?  With 'normal' numbers, the caller gets
billed, and poor routing means lost customers, but not here.  As an 800
number owner, I do not want my 800 number handled by any other provider's
call-routing unless I have assurance that it is routed efficiently.  And
without regulation, I won't get that assurance.  

So although I agree that we can include 800 numbers in GLP and route on
them just like any other number, we might be opening things up for
regulatory problems.  This is probably outside the scope of this wg, but
its another reason to perform translation before call-routing, and to be
silent on the 800 number issue in our specs.   

- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 16:04:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA08609
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 16:04:53 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BB19F8023; Thu, 11 Mar 1999 11:48:05 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2FD397FEA; Thu, 11 Mar 1999 11:46:30 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990309152825.00aae460@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Tue, 09 Mar 1999 15:28:27 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        Greg Zhang <gzhang@utstar.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: hsalama@cisco.com, John Schnizlein <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 02:49 PM 3/9/99 -0500, Jonathan Rosenberg wrote:
>Greg Zhang wrote:
>> 
>> >This is another good example of why translation before gateway selection
>> >is a good idea. However, I'm not sure there is much we can do from a
>> >protocol perspective to prevent it. There may be valid reasons why a
>> >company would like to advertise an 800 number (only they own it, for
>> >example), and I'd rather not exclude this possibility. It seems like a
>> >sensible approach is to allow non-routing numbers in GLP, but mention
>> >that there are issues with them, and translation before routing is a
>> >good idea. Does this sound sensible?
>> 
>> I think to handle this type of cases correctly, we need to introduce a
>> concept of scope for the advertisement. Gateways in CA will advertise
>> the 800 number for company X in CA while gateways in NY will
>> advertise the same 800 for company Y in NY. The advertisement from
>> the CA domain does not get propagated to the NY domain and vice
>> versa. This way the numbers that are locally significant and are being
>> administrated locally stays local. This however does not prevent the
>> same 800 number being used by upper level for a third customer.
>
>A sensible idea, but seems hard to do. How does one represent scope? How
>does one encode the notion that "this number is valid within the greater
>new york metropolitan area", or "northern california" into a protocol? 
>
>Perhaps one solution is to allow scopes to be represented as opaque
>tokens. These tokens can be registered, perhaps with IANA, along with a
>desription of the what the scope means. This would allow peers to
>program their LS's to recognize certain scope tokens, and do something
>appropriate with them. Prefixes with unrecognized scope tokens probably
>shouldn't be propagated.
>
>As an alternative, scope could be represented by a separate phone number
>prefix. This second prefix indicates the set of numbers you need to call
>from for this scope to be defined. For example:
>
>phone numbers: +1-800-55*
>scope: +1-415*
>
>means that 800-55* is only defined within parts of California.
>
>Any other ideas? Or do people think that this isn't needed?
>

I've received similar comments from several providers.  This is one of
things I was going to bring up when discussing general data representation
issues.  Scope has applicability to 1-800 numbers, but it also has
applicability to private/corporate numbers.  I've received several comments
from providers saying that they need the ability to route on a
private/corporate number (in addition to 1-800 comments), where the same
address may have different destinations depending on its scope and the
number itself has no global significance.  In one sense, scope may be
indicate a vpn routing object, and in another it may indicate 1-800-type
limitation.  

Note that these concepts of scope are not exactly the same, but I think a
single attribute might be able to cover them both.  I'm trying to flush
this out a little better before the meeting.  

What do folks think about the necessity of a 'scope' attribute for
 (a) 1-800
 (b) corporate/private addresses
(ie would I be wasting my time trying to address this concern?)

- Matt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 16:20:01 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA08664
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 16:20:00 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A0C93820E; Fri, 12 Mar 1999 12:50:13 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 70D0581A3; Fri, 12 Mar 1999 02:10:01 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <E32BD2062DAFD1119CFD0000F804F31702C04866@cosexch003.mcit.com>
From: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>
To: "'Scott Petrack'" <scott.petrack@metatel.com>,
        "David R. Oran" <oran@cisco.com>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>, hsalama@cisco.com,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Fri, 12 Mar 1999 03:00:53 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE6C34.8B386C24"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE6C34.8B386C24
Content-Type: text/plain

This has interesting implications for call tracing and interception, which
is mandated by some governments.  One could, by proxy, make untraceable
calls.

Ralph

> -----Original Message-----
> From:	Scott Petrack [SMTP:scott.petrack@metatel.com]
> Sent:	Thursday, March 11, 1999 5:22 PM
> To:	David R. Oran
> Cc:	Jonathan Rosenberg; hsalama@cisco.com; iptel
> Subject:	RE: GLP, number portability and wireless subscribers
> 
> On Tue, 9 Mar 1999, David R. Oran wrote:
> 
> > I'm reluctant to bring this up, because some would view it as a corner
> > case or just a "feeeeture" of the PSTN, but when the PSTN does number
> > translation (or call forwarding in the network as well), the resulting
> > number is not revealed to the calling party. The pure directory scheme
> > would reveal the translated number since that's what would be put in the
> > end-to-end signaling (e.g. SIP, H.323). Do we care about this?
> 
> This question has nothing to do with PSTN gateways or GLP, right? It has
> to do with a system based on end-to-end signalling, like SIP and H.323.
> 
> If a called party doesn't want to reveal his/her "real" address, then s/he
> must set up a SIP proxy server which will get the invitation to the "fake"
> address, and the proxy will pass the invitation to the real address. 
> 
> In other words, if you need this feeeeeture within a SIP/H.323 system,
> then you need to use a SIP proxy server or H.323 GK routed signalling, and
> then the "translation" service will return the "fake" routing number.
> 
> Scott
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

------_=_NextPart_001_01BE6C34.8B386C24
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>RE: GLP, number portability and wireless subscribers</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">This has interesting =
implications for call tracing and interception, which is mandated by =
some governments.&nbsp; One could, by proxy, make untraceable =
calls.</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Ralph</FONT>
</P>
<UL>
<P><FONT SIZE=3D1 FACE=3D"Arial">-----Original Message-----</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">From:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Scott Petrack =
[SMTP:scott.petrack@metatel.com]</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">Sent:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Thursday, March 11, 1999 5:22 PM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">David R. Oran</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Cc:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">Jonathan Rosenberg; hsalama@cisco.com; iptel</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT>=
</B> <FONT SIZE=3D1 FACE=3D"Arial">RE: GLP, number portability and =
wireless subscribers</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">On Tue, 9 Mar =
1999, David R. Oran wrote:</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt; I'm =
reluctant to bring this up, because some would view it as a =
corner</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt; case or =
just a &quot;feeeeture&quot; of the PSTN, but when the PSTN does =
number</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt; =
translation (or call forwarding in the network as well), the =
resulting</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt; number =
is not revealed to the calling party. The pure directory scheme</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt; would =
reveal the translated number since that's what would be put in =
the</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">&gt; =
end-to-end signaling (e.g. SIP, H.323). Do we care about this?</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">This question =
has nothing to do with PSTN gateways or GLP, right? It has</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">to do with a =
system based on end-to-end signalling, like SIP and H.323.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">If a called =
party doesn't want to reveal his/her &quot;real&quot; address, then =
s/he</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">must set up a =
SIP proxy server which will get the invitation to the =
&quot;fake&quot;</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">address, and =
the proxy will pass the invitation to the real address. </FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">In other =
words, if you need this feeeeeture within a SIP/H.323 system,</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">then you need =
to use a SIP proxy server or H.323 GK routed signalling, and</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">then the =
&quot;translation&quot; service will return the &quot;fake&quot; =
routing number.</FONT>
</P>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">Scott</FONT>
</P>
<BR>

<P><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier =
New">---------</FONT>
<BR><FONT COLOR=3D"#000000" SIZE=3D2 FACE=3D"Courier New">This message =
came from the IETF IPTEL Working Group Mailing List.</FONT>
</P>
</UL>
</BODY>
</HTML>
------_=_NextPart_001_01BE6C34.8B386C24--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 17:03:52 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA08948
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 17:03:51 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 4D17ABF; Sat, 13 Mar 1999 17:03:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 55FDE808C; Thu, 11 Mar 1999 11:27:32 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6803B808B; Thu, 11 Mar 1999 11:12:49 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: bsramesh@netlab.hcltech.com
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <65256731.001E8D14.00@sampark.hss.hns.com>
Date: Thu, 11 Mar 1999 11:19:59 +0530
Subject: Re: Handling of Setup message by Gatekeeper.
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi,
    Thanks for the link. I think adding EndpointId as a mandatory field
will solve the problem of making checks on sender of the message. After
checking the authenticity of the sender, decisions can be made by the
gatekeeper for various scenarios like setup message without RAS signalling
, both in valid and invalid cases. But for version 1 and 2 endpoints, I
think gatekeeper needs to drop this check, if source info is not available.
Again because standards do not say anything so solution will be
proprietary.

I know, registration is mandatory for all the endpoints, in a zone but I
pointed out the scenario from security point of view. There can be
unregitered endpoints, trying the access the network for making calls.

Thanks,
Ritu.






bsramesh@netlab.hcltech.com on 03/10/99 07:44:05 PM

To:   rtgupta@hss.hns.com
cc:   janar@netlab.hcltech.com, bose@netlab.hcltech.com
Subject:  Re: Handling of Setup message by Gatekeeper.




>In contrast to what you said, I think RAS signalling is must for all the
>endpoints in any GK's zone. RAS signaling is not required for endpoints,
>outside the zone. Now when GK receives a Setup message, it doesn't know
>whether it is from its own zone or from outside. Before doing any further
>checks, it needs to check that. And as I mentioned in my previous mails,
to
>do that, it needs to maintain a static databse of its zone. But there is
no
>mandatory field in Setup message,giving source info. So basic question is
There is a proposal in SG-16 to include the EndPointIdentifier as a
mandatory field in the Setup message for GK routed calls.
You can get the document from the PictureTel's SG16 site, on
ftp://standard.pictel.com/avc-site/9902_Mon/TD-46.zip
Also note that registration is mandatory before making a call
(in an earlier email you had mentioned a scenario where an
endpoint tries to make a call after powering up. Starting up
after a power down is one of the situations when an endpoint
will send an RRQ to the GK.)
<EOM>

>how to perform this check.
_____________________________________________________________
>Ritu Gupta                                   |       email:
>rtgupta@hss.hns.com
>Senior Software Engineer,       |                   rtgupta@hotmail.com
>Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
>Banglore, Karnataka                  |                    91-80-3374075(H)
>INDIA.                                            |       Fax    :
>91-80-2286393
>_____________________________________________________________________
>Restlessness and discontent are the first necessities of progress.
]




---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 17:11:13 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA08979
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 17:11:13 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 2C74A118; Sat, 13 Mar 1999 17:09:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D191B7FEA; Thu, 11 Mar 1999 11:50:47 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1C8087FEE; Thu, 11 Mar 1999 11:48:30 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E5B795.910EA47C@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 19:06:45 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: "David R. Oran" <oran@cisco.com>
Cc: Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
References: <NBBBIFCAKKOPNMMKHHEFCEGLEEAA.oran@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

David R. Oran wrote:
> 
> I'd make one exception to this principle. If there's a set of cases
> where the aggregator can be told a purported "hole" in a set of
> otherwise aggregatable prefixes isn't actually present somewhere, then
> the aggregator can do a much better job of coalescing parts of the
> address space. For example, if there's block of numbers like "+1-600"
> which are known not to exist then telling the aggregator this will allow
> for "+1-***" to be aggregated when otherwise it would be missed due to
> the hole.

I agree that semantic blind aggregation is a better idea. Otherwise,
each poor LS needs to know the structure of the numbers everywhere in
the world. 

This hole idea is interesting. Would you also send the hole along with
the aggregated prefix? If not, an LS is right back to needing to know
about the structure of number allocations in order to aggregate. If
there is an inconsistency in the views of two LS's, you could have some
interesting mis-routing problems.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 17:12:12 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA08990
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 17:12:12 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id A1FDA140; Sat, 13 Mar 1999 17:10:58 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6440180FF; Thu, 11 Mar 1999 18:40:58 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BB8968105; Thu, 11 Mar 1999 18:39:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Thu, 11 Mar 1999 18:12:47 -0500 (EST)
From: Christian Huitema <huitema@research.bellcore.com>
Message-Id: <990311181245.ZM1353@seawind.research.bellcore.com>
In-Reply-To: Matt Squire <msquire@nortelnetworks.com>
        "RE: GLP, number portability and wireless subscribers" (Mar  9,  5:02pm)
References: <3.0.32.19990309170219.00aa9dc0@bl-mail1.corpeast.baynetworks.com>
X-Mailer: Z-Mail (5.0.0 30July97)
To: Matt Squire <msquire@nortelnetworks.com>, "David R. Oran" <oran@cisco.com>,
        Greg Zhang <gzhang@utstar.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Cc: hsalama <hsalama@cisco.com>, John Schnizlein <jschnizl@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Well, it seems that we are back again to the distinction between naming
and routing. With GLP, the phone number in the E164 space is really an
alias to a "local" phone number.  Basically, the translation is:

a) use the number to obtain the id of the local provider,

b) route to the local provider

c) within the local provider, route to the user based on the number.

This means that we will have a two phase look-up:

a) maybe using DNS:
	look up number, get provider ID.
b) within GLP:
	maintain routes based on provider ID, and maybe also prefix.

-- 
Christian Huitema
------------------------------
Please note my new address: huitema@research.telcordia.com
http://www.telcordia.com/

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 17:18:03 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA09020
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 17:18:03 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 6B0731F9; Sat, 13 Mar 1999 17:16:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id AF14852F8; Fri, 12 Mar 1999 15:39:58 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DD79452E5; Fri, 12 Mar 1999 15:16:32 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <008801be6cc1$7715adf0$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "IPTEL IETF Mailing List" <iptel@lists.research.bell-labs.com>,
        "Matt Squire" <msquire@nortelnetworks.com>
Subject: Re: number portability
Date: Fri, 12 Mar 1999 11:49:07 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MIMEOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

There are some significant differences between 900, 800, 700 numbers
and portable numbers. The former can be viewed as virtual numbers.
They are translated and/or routed to a real number call by call routing
process. In 800 case, callee pays. In 900 and 700 cases, caller pays.
In all cases, the real number is not revealed to the caller. The network
can routes the call to different real numbers based on caller's
geographical location and time of day etc. Even in the case of 900
and 700 numbers, caller has no choice of how to route the call. True?

The portable numbers as in Local Number Portability are real numbers.
They are used by caller in dialing. Caller pays for the call. Caller can
choose the carrier to deliver the call. The call is routed based on a
static mapping of the physical phone.

There are some other cases we also need to think about and cover
in the protocol too. With CT-2 and PCS, some phone companies
has been trying to offer a new type of service that will let a subscriber
to have one phone number and ring that subscriber either at his/her
office phone, home phone, the mobile, or his/her voice mail box.
Note this service is different from regular call forwarding. With
minimal notification or logic, your call will follow you and try to
track you down whereever you are.

Another class of problem will be wireless subscribers, to support
mobility in one operator's network and allow perhaps seamless
roaming across the network, you will need to keep track of where
the mobile is currently at and advertisement that location information.

All these cases ideally should be handled by the GLP nicely. With the
flexibility and robustness of the TCP/IP based Internet, the
benefit of hind sight and starting from scratch , a Internet based
telephony solution should not only match the capabilities of current
state-of-the-art practise in the PSTN world, but also do it better and
be able to provide many value added services that is not possible
or just too costly to provision in the PSTN network. I think with some
proper design, we can achieve this goal.

- Greg


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 17:31:02 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA09082
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 17:31:01 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 75926204; Sat, 13 Mar 1999 17:17:01 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2509380D8; Thu, 11 Mar 1999 17:52:30 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C3E7480FF; Thu, 11 Mar 1999 17:36:44 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E68C99.3ABADA78@dnrc.bell-labs.com>
Date: Wed, 10 Mar 1999 10:15:37 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: Greg Zhang <gzhang@utstar.com>, hsalama <hsalama@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>, "David R. Oran" <oran@cisco.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <3.0.32.19990310082902.007e8860@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Phew. Lots of mail flying around here. let me take a moment to summarize
what I think I've heard so far on scopes. Several definitions have been
explicitly or implicitly used. These include:


1. Scope is a way to disambiguate prefixes which are not globally
unique. For example, a particular 800 number might exist in NY, and
another in CA, and they are the same number, but who you get connected
to is different in NY and in CA. The scope would be used as an extra
identifier to make prefix+scope (or vpn-id as Matt calls it in this
email) globally unique. 

2. Scope is a way to disambiguate prefixes which are not globally
unique, but the scope further defines the places where the prefix is
valid. Ideally, this would mean "this prefix is only valid in NY" or
something like that. These scopes might be registered as tags with IANA,
or defines as prefixes themselves. For example, a prefix of Y with scope
X means that entities within scope X can call the prefix Y.

3. Scope is a way to define how far a prefix is propagated. If LS's were
organized hierarchically, the scope would indicate the level in the
hierarchy. Boundary LS's wouldn't propagate prefixes with a scope
corresponding to their level outside their level. Matt has also
mentioned a TTL for this purpose.


Functions (1) and (2) are similar, (3) is different. The question is:
(1) do we need all three functions, and (2) can a single identifier
serve both purposes? We might be able to get a grip on this by first
figuring the first one out - what is it we wish to accomplish with
scopes, and then we can decide how many attributes we are talking about.


-Jonathan R.

Matt Squire wrote:
> 
> At 06:56 PM 3/9/99 -0500, Jonathan Rosenberg wrote:
> >David R. Oran wrote:
> >>
> >> Then scopes are identical to a deeper hierarchy and splitting them out
> >> is simply a syntactical, not a semantic distinction, right?
> >
> >My understanding is that it is semantic. By "splitting them out", I
> >presume you mean having them be separate from the address itself. The
> >prefix represents information on "where you're calling to", and the
> >scope information on "where you are calling from". So, a prefix Y with
> >scope X means you can call Y from X. I think we need explicit scopes if
> >there are cases where route attributes for a prefix only apply when
> >calling from numbers outside of the prefix (and apply only when calling
> >from somewhere defined by the scope).
> 
> I don't see how we can guarantee any determination of 'where you are
> calling from.'  What does my pc call from?  Where is an LS?  A gateway
> might be considered to have a 'location', where the location is given by
> the area code of its PSTN connectivity, but other devices without PSTN
> connectivity aren't tied to a PSTN location.
> 
> It seems like any definition of scope along these lines requires alot of
> administrative help, ie where LSs are tied/located with certain geographic
> areas.  Would not a TTL-like attribute provide similar functionality (again
> based on geographic ties/connectetions in LSs)?  A number could be
> advertised with a certain TTL that prevents it from being advertised after
> too many hops.  If your connections are geography-based, then this gives
> you geographic scoping.  We make an additional statement that given
> multiple potential routes to a destination, the one with the higher/lower
> TTL is preferred (because its 'closer').
> 
> >
> >>
> >> Do you think we would need hierarchical scopes? If so, do you see
> >> aggregation of hierarchical scopes? How would this differ from
> >> hierarchical aggregation of addresses? Are scopes always more global
> >> than addresses?
> >
> >We could potentially define hierarchical scopes (if they are also
> >represented as prefixes, they are hierarchical). The rules for
> >aggregation would presumably be the same. But, adding another
> >aggregatable entity in a route advertisement seems like it will make
> >route aggregation overall more complex, so I'd rather avoid this. Are
> >scopes global? I think they need to be. The whole idea is that a prefix
> >advertised in one place is only useful somewhere else.
> >
> 
> Unless someone says something to convice me otherwise, I'd vote against
> this concept of scoping.
> 
> - Matt

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 17:44:06 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA09166
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 17:44:06 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id A3CF11E6; Sat, 13 Mar 1999 17:16:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5CBA581EA; Fri, 12 Mar 1999 04:44:09 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F2E21813F; Thu, 11 Mar 1999 19:56:37 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Thu, 11 Mar 1999 19:09:26 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: Christian Huitema <huitema@research.bellcore.com>
Cc: Matt Squire <msquire@nortelnetworks.com>, "David R. Oran" <oran@cisco.com>,
        Greg Zhang <gzhang@utstar.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        hsalama <hsalama@cisco.com>, John Schnizlein <jschnizl@cisco.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
In-Reply-To: <990311181245.ZM1353@seawind.research.bellcore.com>
Message-ID: <Pine.LNX.4.04.9903111856190.6794-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

The fact that this situation is a mess in the PSTN does not mean that it
must be a mess in GLP. And the fact that the "scope" of routing
numbers in the PSTN is the "provider ID" does not
mean that this is the only scope that we should allow within GLP.
Certainly, 1-700 numbers will absolutely require a scope which is a
"provider id", but 800 numbers might not. For example, my company owns a
certain 800 number, and it will resolve to our office telephone number no
matter what provider we happen to choose to give us service. Nothing in
GLP will require that the scope will be "provider id."

Routing is reasonably well defined in the Internet, and it is totally
separate from name-to-anything translations. 

Although there will be a need to have some sort of name-to-something
translation within the Internet, I do not think that this is within the
scope of Iptel. The use of DNS, LDAP, or anything else to effect the translations really
is outside the scope of iptel work, no? 

One of the reasons that I did not fire up ENUM in recent weeks is that we
needed to see better the structure required of the routing
numbers that GLP will use. These routing numbers will be (at least one of)
the outputs of some future translation protocol. I hope that this ENUM
service will be used by gateways, as well as by other applications.

I think that iptel should concentrate on the GLP problem, and deal only
with routing numbers. Once we decide what those numbers look like, it will
be much simpler to translate from "names" to them. 

Scott



On Thu, 11 Mar 1999, Christian Huitema wrote:

> Well, it seems that we are back again to the distinction between naming
> and routing. With GLP, the phone number in the E164 space is really an
> alias to a "local" phone number.  Basically, the translation is:
> 
> a) use the number to obtain the id of the local provider,
> 
> b) route to the local provider
> 
> c) within the local provider, route to the user based on the number.
> 
> This means that we will have a two phase look-up:
> 
> a) maybe using DNS:
> 	look up number, get provider ID.
> b) within GLP:
> 	maintain routes based on provider ID, and maybe also prefix.
> 
> -- 
> Christian Huitema
> ------------------------------
> Please note my new address: huitema@research.telcordia.com
> http://www.telcordia.com/
> 


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 17:58:33 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA09202
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 17:58:32 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 3FDCD1FD; Sat, 13 Mar 1999 17:16:56 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D20AB7FB7; Thu, 11 Mar 1999 17:40:39 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6AB8480BF; Thu, 11 Mar 1999 17:26:47 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <2F2DC5CE035DD1118C8E00805FFE354C0C90F17B@RED-MSG-56>
From: Deepak Kumar <deepak@microsoft.com>
To: "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>,
        "ronald h. davis" <ronaldd@LUCENT.COM>
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject: RE: Handling of Setup message by Gatekeeper.
Date: Tue, 9 Mar 1999 22:26:38 -0800 
X-Mailer: Internet Mail Service (5.5.2524.0)
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Q.931 could originate at a rogue (ie. in the physical zone, but
unregistered, and doesn't send ARQ) end-point or an end-point/GK outside the
zone. Why is a distinction between these two scenarios meaningful?

An end-point that chooses (illegally) not to register with its Gatekeeper
could be treated exactly like an end-point outside the zone, and be denied
the benefits of Gatekeeper support - such as the ability to recv incoming
calls, QoS reservations, and the ability to place and recv calls using
network resources such as Gateways and proxies.

Deepak Kumar,

Program Manager, IP Telephony,
Windows Networking and Communications,
Microsoft Corp.

> -----Original Message-----
> From:	rtgupta@hss.hns.com [SMTP:rtgupta@hss.hns.com]
> Sent:	Tuesday, March 09, 1999 9:19 PM
> To:	ronald h. davis
> Cc:	h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject:	Re: Handling of Setup message by Gatekeeper.
> 
> Hi Davis,
>     Thanks for your inputs.
> 
> I agree that GK is responsible only for the endpoints whixh exist in its
> zone. And also that it can not administer endpoints in other zones. But
> the
> scenarios mentioned in the query are all valid scenarios. GK CAN receive
> Setup message without any RAS signalling taking place before that. So it
> can not terminate h225 signalling channel because of absence of RAS
> channel.
> 
> In contrast to what you said, I think RAS signalling is must for all the
> endpoints in any GK's zone. RAS signaling is not required for endpoints,
> outside the zone. Now when GK receives a Setup message, it doesn't know
> whether it is from its own zone or from outside. Before doing any further
> checks, it needs to check that. And as I mentioned in my previous mails,
> to
> do that, it needs to maintain a static databse of its zone. But there is
> no
> mandatory field in Setup message,giving source info. So basic question is
> how to perform this check.
> 
> regards,
> Ritu.
> 
> _____________________________________________________________________
> Ritu Gupta                                   |       email:
> rtgupta@hss.hns.com
> Senior Software Engineer,       |                   rtgupta@hotmail.com
> Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
> Banglore, Karnataka                  |                    91-80-3374075(H)
> INDIA.                                            |       Fax    :
> 91-80-2286393
> _____________________________________________________________________
> Restlessness and discontent are the first necessities of progress.
> 
> 
> 
> 
> 
> "ronald h. davis" <ronaldd@lucent.com> on 03/09/99 11:32:19 PM
> 
> To:   rtgupta@hss.hns.com
> cc:   h323implementors@imtc.org, iptel@lists.research.bell-labs.com
> Subject:  Re: Handling of Setup message by Gatekeeper.
> 
> 
> 
> 
> rtgupta@hss.hns.com wrote:
> >
>                [scenarios deleted]
> > In all these cases, gatekeeper will receive Setup message without
> receiving
> > ARQ message. So how can gatekeeper check whether ARQ is expected from
> the
> > endpoint, which sent the Setup or not? No field in the Setup message
> seem
> > to give this information. If gatekeeper doesn't check for this, that
> means
> > even the endpoints in its own zone can bypass gatekeeper for RAS
> > signalling?
> >
> > Any help/information on this issue will be appreciated.
> >
> the gatekeeper is only responsible for arq/acf/arj with endpoints within
> its zone.  all endpoints within the zone establish a ras channel (over
> which these messages are sent) with their gatekeeper.
> endpoints in other zones don't establish a ras channel with this
> gatekeeper.  thus, when calls are received from outside of the zone
> the recipient (be through the direct or gatekeeper routed call model)
> has no idea how the calling endpoint came upon sending the setup
> message.
> this statement may be controversial but it is my opinion that h.323
> is lan-oriented, so if you think of gatekeepers as administering
> bandwidth allocation, &c for endpoints on different lans it would
> be difficult for a gatekeeper to perform this administration for
> a different lan.  and it would be even harder for a gatekeeper to
> perform such administration over a wide area network consisting of
> a number of zones.
> this fact leads me to believe that the ras channel should be *optional*
> for h.323 gatekeepers.  to abstract your scenarios further it is
> possible that gatekeepers could act as application layer routers, in
> which case they are terminating h.225 signalling and h.245 media control
> channels for endpoints with which they have no ras channel.
> --
>   __  ______  __  / __/ |       lucent technologies, naperville il, usa
> _/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
> "if you ain't the lead dog, the scenery never changes"  -- edmund wilson
> 
> 
> 
> 
> -------------------------------------------------------------------------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:09:06 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09363
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:09:06 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id D6195F0; Sat, 13 Mar 1999 17:08:01 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3F317823C; Fri, 12 Mar 1999 11:50:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 545118204; Fri, 12 Mar 1999 11:49:57 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990312103751.00ab4990@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Fri, 12 Mar 1999 10:37:52 -0500
To: Mike Gettles <Mike.Gettles@mail.sprint.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: RE: number portability
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


My SLA as a 800 number owner is with provider X.  Caller A may have a SLA
as an iptel user with provider Y.  Unless provider X and Y have an
agreement, caller A could get routed to any gateway when connecting to my
800 number, and since the call is free to caller A, he won't care, leaving
me to pay the PSTN charges to provider X.  To me, fixing this problem
implies mandated SLAs between providers, and mandated inter-provider SLAs
equates to regulation.  

I just realized that several other folks expressed similar concerns in
recent postings, pointing out the make the 800 number translations visible
on the IP side.  What a lousy time for the mail exploder to get screwy.  

Anyway, I guess my whole point was that unless we try to solve the reverse
charging problem, we cannot claim that GLP provides "800" services.  We may
be able to provide some features of 800 numbers, for example, portability,
but we need to be careful and separate the individual features that we can
provide from what people expect from an 800 number service.  The service
may be built using GLP and other stuff (directory lookups, integrated
databases with SCPs, etc), but GLP itself does not provide the service.  

I tried to express this when addressing 'data issue 1: phone numbers' in a
recent post.  This will probably show up on the list in a day or two:(


At 09:00 AM 3/12/99 -0600, Mike Gettles wrote:
>It's called SLA (Service Level Agreement).  You specify as part of the
>service contract what level of performance you require.  Look to the
>carriers that are willing to put their money where their mouth is, i.e.
>financial penalties for failure to perform. It is incumbent on you to
>measure that performance but you can get the carrier to supply you with
>the information you need as part of the management reports package that
>goes with the service.
>
>-----Original Message-----
>From: msquire [mailto:msquire@nortelnetworks.com]
>Sent: Wednesday, March 10, 1999 7:53 AM
>To: iptel
>Cc: msquire
>Subject: number portability
>
>
>
>One other issue on number portability while we're discussing it...  
>
>GLP as provides a service to the caller.  Things like 800 numbers
>provide
>services to the callee.  What motiviation do call-routing providers have
>to
>route 800 numbers efficiently?  Call-routing provider X could decide to
>use
>a GW on the west coast for all calls to my 800 number, which is actually
>translated to a number on the east coast.  I could bitch to my 800
>number
>provider, but what could they do?  With 'normal' numbers, the caller
>gets
>billed, and poor routing means lost customers, but not here.  As an 800
>number owner, I do not want my 800 number handled by any other
>provider's
>call-routing unless I have assurance that it is routed efficiently.  And
>without regulation, I won't get that assurance.   
>
>So although I agree that we can include 800 numbers in GLP and route on
>them just like any other number, we might be opening things up for
>regulatory problems.  This is probably outside the scope of this wg, but
>its another reason to perform translation before call-routing, and to be
>silent on the 800 number issue in our specs.    
>
>- Matt
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.
>
>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:17:13 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09415
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:17:13 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 3C832134; Sat, 13 Mar 1999 17:10:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 4BC4F7FFB; Thu, 11 Mar 1999 11:56:27 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9C08C7FFE; Thu, 11 Mar 1999 11:44:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <013201be6a65$493a9610$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "Lu, Henry" <hwlu@tri.sbc.com>,
        "'pcalhoun@eng.sun.com'" <Pat.Calhoun@Eng.Sun.COM>,
        "Matt Squire" <msquire@nortelnetworks.com>
Cc: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "John Schnizlein" <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 11:44:13 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

>I am in favor of Pat's point but I am still hoping that we can
>achieve a sound solution in order for us to handle number portability
>environment in the future, including wireless.

I agree whole heartedly that this should be one of the goals that we try to
achieve.

>The issues I see is how to synchronize two database between GLP and SS7
>world. In AIN, 800 number information are kept in not only Signaling
>Control Point(SCP) but also Service Switching Point(SSP), normally in
>local DMS-100 or 5ESS switches. The 800 calling process had to go through
>a conversation session between SSP and SCP to figure out the final
>routing to the destination. There are two information databases been
>kept on both SCP and local SSP. Do not forget that there are many SSPs
>in the network, not like SCP, we have to deal with.
>The major problem occur when a new portable number been provisioned,
>into both SSP and SCP but there is no information on GLP side.
>GLP and SS7 database is out of sync basically.
>
>How to solve this database synchronization problem, I do not have
>a solution for you but I hope that we can see the nature of the
>problem first.

When you have a Internet telephony gateway that is attached to a
PSTN network, the gateway needs to advertisement the reachability
of the numbers in the PSTN network. When provisioning of a new
portable number occurs, the portable number and its routing information
("office code") needs to be sent to the gateway. The gateway in turn
propagate this information to other gateways on the Internet through
GLP. The mechanism with which the information in SCP and SSPs
are kept in sync with the attached Internet telephony gateway could
be the same mechanism used to update this information from the
subscriber management system to SCP and SSPs. It is outside the
scope of GLP though.

- Greg



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:23:49 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09448
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:23:48 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 5373D1DB; Sat, 13 Mar 1999 17:16:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E3E8653C8; Sat, 13 Mar 1999 16:43:07 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3CDEA53C7; Sat, 13 Mar 1999 16:43:02 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Internal-ID: 36EA31020000244B
Message-ID: <36EA31020000244B@mailnew.fibertel.com.ar> (added by mailnew.fibertel.com.ar)
Date: Sat, 13 Mar 1999 15:32:37 -0300
From: IIBA <iiba@bitsmart.com>
Subject: Cursos Intensivos de InformÂtica de SÂbados y Domingos.
Reply-To: iiba@bitsmart.com
To: iptel@lists.research.bell-labs.com
X-Mailer: dbMail v1.3b1 by Mach5 Software. Please report spamming to dbspam@mach5.com.
Content-Transfer-Encoding: Quoted-Printable
MIME-Version: 1.0
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: Quoted-Printable

AT.: iptel@lists.research.bell-labs.com

ATENCION: Cursos Intensivos de Inform=E1tica de fin de semana.

NO DEJE PASAR ESTA OPORTUNIDAD DE CAPACITARSE YA!

OBJETO: Los cursos apuntan a resolver la problem=E1tica de las personas q=
ue
no disponen del tiempo necesario para capacitarse durante la semana.=20

Se entregar=E1 diploma de realizaci=F3n del curso.

Pr=E1ctica intensiva personalizada sobre computadoras (1 PC por alumno, 1=
00%
de pr=E1ctica).

CURSOS DISPONIBLES:

Windows 95 (2 d=EDas - 16 hs.).  $ 150,--
Windows 98 - Operaci=F3n (2 d=EDas 16 hs.). $ 150,--
Windows 98 - Comunicaciones e Internet (1 d=EDa 8 hs.). $ 90,--
Internet (1 d=EDa 8 hs.). $ 90,--
Word 97 - B=E1sico (2 d=EDas 16 hs.). $ 150,--
Word 97 - Avanzado (1 d=EDa 8 hs.). $ 90,--
Excel 97 - B=E1sico (2 d=EDas 16 hs.). $ 150,--
Excel 97 - Avanzado (2 d=EDas 16 hs.). $ 150,--
AutoCAD 14 - Dibujo en 2 dimensiones (2 d=EDas 16 hs.). $ 150,--
AutoCAD 14 - Dibujo en 3D y Modelado de S=F3lidos (2 d=EDas 16 hs.). $ 15=
0,--

OTROS CURSOS:

Access 95
Access 97
Corel Draw 6.0
Corel Draw 8.0
Excel 95
Introducci=F3n al PC
Office 95
Office 97
PowerPoint 95
PowerPoint 97
Visual Basic 4.0
Windows NT 4 - Workstation
Word 95

Puede consultar los contenidos de los cursos, sus duraciones y costos
enviandonos un e-mail a:
iiba@bitsmart.com

FECHAS Y HORARIOS DISPONIBLES (Marzo de 1999):

S=E1bado 	13/03
Domingo 	14/03
S=E1bado 	20/03
Domingo 	21/03
S=E1bado 	27/03
Domingo 	28/03

En horarios de 9 a 13 hs. y  15 a 19 hs (consultar por horarios
especiales).

LUGAR:

Los cursos de realizar=E1n en Capital Federal (tambi=E9n se realizan curs=
os a
distancia).

RESERVAR VACANTE (Cupos limitados).

A los interesados rogamos enviar los datos siguientes y nos pondremos en
contacto.

NOMBRE Y APELLIDO:
EMPRESA y CARGO (Si corresponde):
TELEFONO:
CURSO(S) DE INTERES:
FECHAS DE INTERES:


Cordialmente,

Depto. de Capacitaci=F3n
Instituto Inform=E1tico Buenos Aires
iiba@bitsmart.com

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:30:22 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09496
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:30:22 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 610B5153; Sat, 13 Mar 1999 17:11:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 764C77FDB; Tue,  9 Mar 1999 12:35:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C72897FDE; Tue,  9 Mar 1999 12:35:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E55B77.B5105476@cisco.com>
Date: Tue, 09 Mar 1999 09:33:43 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <010501be69aa$79336240$2821a8c0@greg_zhang.wacos.com> <199903091625.IAA01401@mailman.cisco.com> <36E54E32.C8931AB2@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Jonathan Rosenberg wrote:

> 
> This begs the question, then: does the translation of the 800 number get
> done before the gateway selection process, or after? By doing it before,
> a gateway can be chosen which is, in fact, close to the real number
> being dialed. With 800 numbers, all gateways will be able to reach it.
> It seems that this opens the possibility of very non-optimal routing. A
> caller in NY that dials an 800 number, goes to a gateway in CA. The
> phone network there determines that the 800 number translates to a
> number in NY, and routes the call through the PSTN back to NY. However,
> if the translation occurs before gateway selection, the caller can
> choose a local NY gateway and do much better.

Because of the above example, I think all types of translation (800
numbers or LNP) and mapping should occur prior to call routing. This is
particularly important for 800 numbers, because they have a regional
scope in many cases. The same 800 number may owned by company X in CA
and company Y in NY. In the above example, the call will be incorrectly
routed to company X instead of going to the originally intended
destination, company Y.

Hussein



> 
> -Jonathan R.
> 
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-2499

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:36:10 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09573
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:36:10 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 0B1D516B; Sat, 13 Mar 1999 17:12:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D67EF8203; Fri, 12 Mar 1999 11:43:57 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DB5318207; Fri, 12 Mar 1999 11:43:48 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <4110BC526A58D211B56C0000F6B9636DA851E4@ws-exch01.dev.madge.com>
From: Chris Purvis <CPurvis@Madge.com>
To: "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>, bsramesh@netlab.hcltech.com
Cc: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Subject: RE: Handling of Setup message by Gatekeeper.
Date: Fri, 12 Mar 1999 11:14:25 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain;
	charset="windows-1252"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

I fail to see what is gained by adding EndpointId as a "mandatory" field
(which will not be filled in by v1 & v2 endpoints anyway) - quite apart from
being impossible (it's gatekeeper-allocated, so in an environment without
gatekeepers you can't fill it in).  It is optional in TD-46, and for good
reasons.

After all, if I'm in the business of wanting to break things I can easily
send an EndpointID I make up in any setup message.  To try to return to the
original problem, if I as an endpoint want to send a setup message without
having sent an ARQ, I can soon make something up to put in any field I
choose.  If I want to obey the spec and there are any gatekeepers around at
all, I will not send a setup without first sending an ARQ to my gatekeeper.

Regards,
Chris

> -----Original Message-----
> From: rtgupta@hss.hns.com [mailto:rtgupta@hss.hns.com]
> Sent: 11 March 1999 5:50
> To: bsramesh@netlab.hcltech.com
> Cc: h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject: Re: Handling of Setup message by Gatekeeper.
> 
> 
> Hi,
>     Thanks for the link. I think adding EndpointId as a 
> mandatory field
> will solve the problem of making checks on sender of the 
> message. After
> checking the authenticity of the sender, decisions can be made by the
> gatekeeper for various scenarios like setup message without 
> RAS signalling
> , both in valid and invalid cases. But for version 1 and 2 
> endpoints, I
> think gatekeeper needs to drop this check, if source info is 
> not available.
> Again because standards do not say anything so solution will be
> proprietary.
> 
> I know, registration is mandatory for all the endpoints, in a 
> zone but I
> pointed out the scenario from security point of view. There can be
> unregitered endpoints, trying the access the network for making calls.
> 
> Thanks,
> Ritu.
> 
> 
> 
> 
> 
> 
> bsramesh@netlab.hcltech.com on 03/10/99 07:44:05 PM
> 
> To:   rtgupta@hss.hns.com
> cc:   janar@netlab.hcltech.com, bose@netlab.hcltech.com
> Subject:  Re: Handling of Setup message by Gatekeeper.
> 
> 
> 
> 
> >In contrast to what you said, I think RAS signalling is must 
> for all the
> >endpoints in any GK's zone. RAS signaling is not required 
> for endpoints,
> >outside the zone. Now when GK receives a Setup message, it 
> doesn't know
> >whether it is from its own zone or from outside. Before 
> doing any further
> >checks, it needs to check that. And as I mentioned in my 
> previous mails,
> to
> >do that, it needs to maintain a static databse of its zone. 
> But there is
> no
> >mandatory field in Setup message,giving source info. So 
> basic question is
> There is a proposal in SG-16 to include the EndPointIdentifier as a
> mandatory field in the Setup message for GK routed calls.
> You can get the document from the PictureTel's SG16 site, on
> ftp://standard.pictel.com/avc-site/9902_Mon/TD-46.zip
> Also note that registration is mandatory before making a call
> (in an earlier email you had mentioned a scenario where an
> endpoint tries to make a call after powering up. Starting up
> after a power down is one of the situations when an endpoint
> will send an RRQ to the GK.)
> <EOM>
> 
> >how to perform this check.
> _____________________________________________________________
> >Ritu Gupta                                   |       email:
> >rtgupta@hss.hns.com
> >Senior Software Engineer,       |                   
> rtgupta@hotmail.com
> >Hughes Software Systems,     |       phone: 91-80-2286390/91/92(W)
> >Banglore, Karnataka                  |                    
> 91-80-3374075(H)
> >INDIA.                                            |       Fax    :
> >91-80-2286393
> >_____________________________________________________________________
> >Restlessness and discontent are the first necessities of progress.
> ]
> 
> 
> 
> --------------------------------------------------------------
> -----------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"
> 
 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:40:46 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09596
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:40:46 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id E651F1DE; Sat, 13 Mar 1999 17:16:05 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id AB2EF8015; Thu, 11 Mar 1999 17:39:05 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E52138018; Thu, 11 Mar 1999 17:26:51 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: "Matt Squire" <msquire@nortelnetworks.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: Comments on draft-ietf-iptel-glp-00.txt
Date: Wed, 10 Mar 1999 02:01:56 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFEEKMEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Importance: Normal
In-Reply-To: <36E5B795.910EA47C@dnrc.bell-labs.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

> -----Original Message-----
> From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
> Sent: Tuesday, March 09, 1999 7:07 PM
> To: David R. Oran
> Cc: Matt Squire; Scott Petrack; Alok Mittal; iptel
> Subject: Re: Comments on draft-ietf-iptel-glp-00.txt
>
>
> David R. Oran wrote:
> >
> > I'd make one exception to this principle. If there's a set of cases
> > where the aggregator can be told a purported "hole" in a set of
> > otherwise aggregatable prefixes isn't actually present
> somewhere, then
> > the aggregator can do a much better job of coalescing parts of the
> > address space. For example, if there's block of numbers
> like "+1-600"
> > which are known not to exist then telling the aggregator
> this will allow
> > for "+1-***" to be aggregated when otherwise it would be
> missed due to
> > the hole.
>
> I agree that semantic blind aggregation is a better idea. Otherwise,
> each poor LS needs to know the structure of the numbers everywhere in
> the world.
>
> This hole idea is interesting. Would you also send the hole along with
> the aggregated prefix? If not, an LS is right back to needing to know
> about the structure of number allocations in order to aggregate. If
> there is an inconsistency in the views of two LS's, you could
> have some
> interesting mis-routing problems.
>
> -Jonathan R.
>
I'm beginning to feel like I'm living inside a Beatles song - "...fixing
a hole where the rain gets in, to stop my mind from wandering"... now
where was I...

...oh yes. Well I hadn't actually thought of the idea of propagating
"hole advertisements". I was thinking of simply making an assertion to
the aggregator via some local configuration. I don't see how this
necessarily leads to the LS (remind me what the acronym stands for - I'm
typing this at 2AM...) having to know the structure of allocations? Just
like you have to know whether you can reach prefix "+44 365" in order to
inject a route to "+44 365" you likewise have to know that all routes to
"+777" lead to a black hole. If this is what you mean by knowing the
structure of allocations, then yes, but this isn't global knowledge
needed by every LS, only those LSs serving gateways advertising routes
to prefixes "surrounding" the hole.

It may be feasible to actually inject the route to the hole, but this
begs the question of which LS(s) actually do it and whether it looks
like a route to a black hole or is something special. In any event, I
threw out the idea without a whole lot of thought behind it, and now
that you've called my bluff I'll have to go off and see exactly what the
further implications are.

Dave.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:44:25 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09611
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:44:24 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 46DDB1C8; Sat, 13 Mar 1999 17:15:31 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5932A81EE; Fri, 12 Mar 1999 04:44:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4E2378144; Thu, 11 Mar 1999 19:57:36 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Henry Sinnreich" <henry.sinnreich@mci.com>
To: "Matt Squire" <msquire@nortelnetworks.com>,
        "David R. Oran" <oran@cisco.com>, "Greg Zhang" <gzhang@utstar.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: "hsalama" <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Thu, 11 Mar 1999 18:25:52 -0500
Message-ID: <NBBBIIJFOKPMFOOILMBKMECPDMAA.henry.sinnreich@mci.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
In-reply-to: <3.0.32.19990309170219.00aa9dc0@bl-mail1.corpeast.baynetworks.com>
X-MIMEOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Importance: Normal
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Matt Squire writes:
>If there is a scope registration
> authority (maybe this wasn't clear before), then the combination of
> scope+number is globally unambiguous, with the default scope being global
> e164.

Is there such an authority?

Thanks, Henry

> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Matt
> Squire
> Sent: Tuesday, March 09, 1999 5:02 PM
> To: David R. Oran; Greg Zhang; Jonathan Rosenberg
> Cc: hsalama; John Schnizlein; Scott Petrack; Alok Mittal; iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> At 04:34 PM 3/9/99 -0500, David R. Oran wrote:
> >Or, alternatively, project all numbers into a global, unambiguous space.
> >This seems cleaner and much less error prone than topologically
> >established scope boundaries, and less administratively burdensome than
> >solving the problem of scope name collisions and information propagation
> >boundaries. It also avoids having to talk about translations between
> >scopes, which I'm sure will happen when the other shoe drops :-).
> >
>
> I'm a little confused by this comment. To me, scopes provide the
> unambiguous global number space.  If there is a scope registration
> authority (maybe this wasn't clear before), then the combination of
> scope+number is globally unambiguous, with the default scope being global
> e164.  If someone registers a scope and they wish it to have topological
> significance, then that's an administrative matter and they should be able
> to set up their peer association and policy accordingly.
>
> I'd probably be against scopes of the form "+1-405" because that
> would just
> lead to too many headaches in interpreting, aggregating, etc.
>
> An opaque scope using a equal/!equal test might provide what we desire.  I
> haven't convinced myself completely, but I think it has the potential.
>
> - Matt
>
>
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:47:35 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09627
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:47:35 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 48CC1142; Sat, 13 Mar 1999 17:11:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B24708275; Fri, 12 Mar 1999 13:43:44 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 639A881A7; Fri, 12 Mar 1999 13:06:00 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D18E@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'David R. Oran'" <oran@cisco.com>, "Lu, Henry" <hwlu@ttl.pactel.com>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>, hsalama@cisco.com
Cc: iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Fri, 12 Mar 1999 11:47:43 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear David,

My scope is from allowing IP telephony world can call
customer on PSTN and vise versa. The current architecture
is through SS7 gateway from IP world. There are works ongoing
in other groups but I have not been convinced that the
gateway will provide the correct information to IP world
due to SS7 architecture and the way they are going. 1-800
and number portability are two examples. I heard that we
need to spend hundreds of million dollar just to upgrade
the switches to do number portability. I think there should
be money spent to make sure that SS7 world will be able to
do it on SCP level so that a query from IP world can get the
right route in PSTN. Only this way, we can truly open up local
phone competition. Otherwise, I do not see that will happen.
I hope it will help. Thanks.

Henry W. Lu
925-598-1245
SBC, Technology Resources, Inc.
4698 willow road,
Pleasanton, CA 94588


-----Original Message-----
From: David R. Oran [mailto:oran@cisco.com]
Sent: Friday, March 12, 1999 9:02 AM
To: Lu, Henry; Jonathan Rosenberg; hsalama@cisco.com
Cc: iptel
Subject: RE: GLP, number portability and wireless subscribers


> -----Original Message-----
> From: Lu, Henry [mailto:hwlu@tri.sbc.com]
> Sent: Thursday, March 11, 1999 7:16 PM
> To: 'David R. Oran'; Jonathan Rosenberg; hsalama@cisco.com
> Cc: iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> Dear David,
>
> I think what we can not afford is place a call to wrong number.
Define "wrong number", please. If I have an 800 number and call to say
that I want it to ring 555-4545 instead of 555-7676 and then hang up, is
it "placing a call to a wrong number" if 555-7676 rings for the next 30
seconds while the tables are getting updated? This is no different from
the convergence properties of IP routing.

> Whether we will reveal the true number or not is not
> very important from my perspective.
Really? This is not what I hear from my service provider friends, but as
I said I'd rather not worry about this so let's drop it for now.
Jonathan rightly points out that as long as you limit translation result
visibility to parts of the network the owner of the translation trusts,
things will be OK. The porblem arises if you simultaneously want to have
optimal routing AND information hiding.

> You have to know
> that SS7 network is not like IP network, there is no
> network topology database information in SCP like
> OSPF or BGP in IP world.
Yes, I know this.

> This create problem for reconstructing
> call routing from IP world. SS7 is per call base and do
> a SCP query if there is 800 or ported number. No matter
> SIP or H.323, will be per call based through gateway query.
> We are still lock the protocol in old ways per call based. We can
> not update IP world PSTN topology like what we are doing
> in OSPF or BGP for the whole network in one database sync.
Ok, I'll bite. Why not? Remember, we're not doing destination selection
in GLP, we're only doing route selection once the final routing-number
is already known.

> I hope it will help. Thanks.
>
> Henry W. Lu
>
>
> -----Original Message-----
> From: David R. Oran [mailto:oran@cisco.com]
> Sent: Tuesday, March 09, 1999 1:12 PM
> To: Jonathan Rosenberg; hsalama@cisco.com
> Cc: iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> I'm reluctant to bring this up, because some would view it as a corner
> case or just a "feeeeture" of the PSTN, but when the PSTN does number
> translation (or call forwarding in the network as well), the resulting
> number is not revealed to the calling party. The pure directory scheme
> would reveal the translated number since that's what would be
> put in the
> end-to-end signaling (e.g. SIP, H.323). Do we care about this?
>
> > -----Original Message-----
> > From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
> > Sent: Tuesday, March 09, 1999 1:43 PM
> > To: hsalama@cisco.com
> > Cc: John Schnizlein; Greg Zhang; David R. Oran; Matt Squire; Scott
> > Petrack; Alok Mittal; iptel
> > Subject: Re: GLP, number portability and wireless subscribers
> >
> >
> > Hussein F. Salama wrote:
> > >
> > > Jonathan Rosenberg wrote:
> > >
> > > >
> > > > This begs the question, then: does the translation of the
> > 800 number get
> > > > done before the gateway selection process, or after? By
> > doing it before,
> > > > a gateway can be chosen which is, in fact, close to the
> > real number
> > > > being dialed. With 800 numbers, all gateways will be able
> > to reach it.
> > > > It seems that this opens the possibility of very
> > non-optimal routing. A
> > > > caller in NY that dials an 800 number, goes to a gateway
> > in CA. The
> > > > phone network there determines that the 800 number
> translates to a
> > > > number in NY, and routes the call through the PSTN back
> > to NY. However,
> > > > if the translation occurs before gateway selection, the
> caller can
> > > > choose a local NY gateway and do much better.
> > >
> > > Because of the above example, I think all types of
> translation (800
> > > numbers or LNP) and mapping should occur prior to call
> > routing. This is
> > > particularly important for 800 numbers, because they have
> a regional
> > > scope in many cases. The same 800 number may owned by
> > company X in CA
> > > and company Y in NY. In the above example, the call will be
> > incorrectly
> > > routed to company X instead of going to the originally intended
> > > destination, company Y.
> >
> > This is another good example of why translation before
> > gateway selection
> > is a good idea. However, I'm not sure there is much we can do from a
> > protocol perspective to prevent it. There may be valid reasons why a
> > company would like to advertise an 800 number (only they own it, for
> > example), and I'd rather not exclude this possibility. It
> seems like a
> > sensible approach is to allow non-routing numbers in GLP,
> but mention
> > that there are issues with them, and translation before routing is a
> > good idea. Does this sound sensible?
> >
> > Scott P., of course, is right, that the translation service
> itself is
> > outside the scope of iptel.
> >
> > -Jonathan R.
> > --
> > Jonathan D. Rosenberg                       Lucent Technologies
> > Member of Technical Staff                   101 Crawfords Corner Rd.
> > High Speed Networks Research                Holmdel, NJ 07733
> > FAX:   (732) 834-5379                       Rm. 4C-526
> > EMAIL: jdrosen@bell-labs.com
> > URL: http://www.cs.columbia.edu/~jdrosen
> >
>
>
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:49:11 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09651
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:49:11 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 4F1B61AB; Sat, 13 Mar 1999 17:14:31 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 053F17FD7; Tue,  9 Mar 1999 12:31:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4D5677FDB; Tue,  9 Mar 1999 12:31:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Tue, 9 Mar 1999 12:30:23 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
In-Reply-To: <Roam.SIMCSD.2.0.4.920998470.9397.pcalhoun@hsmpka>
Message-ID: <Pine.LNX.4.04.9903091223400.3873-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



On Tue, 9 Mar 1999, pcalhoun@eng.sun.com wrote:

[Jonathan R:}
> > This begs the question, then: does the translation of the 800 number get
> > done before the gateway selection process, or after? By doing it before,
...
> > if the translation occurs before gateway selection, the caller can
> > choose a local NY gateway and do much better.
> 
> Translation before the gateway selection is *obviously* the correct model. We
> should strive towards this goal.
> 
> PatC
> 

Well, yes, but there are lots of 800 numbers whose translations change on
time scales MUCH slower than the GLP advertisement changes. In fact, most
800 numbers are this way -- they are used by small-to-medium
businesses, point to the 'real' telephone number, with no
logic at all, and don't change at all.

If a GW wants to advertise that it can service such a number, I don't see
why it should be forbidden from doing so. 

[Of course there is a need for a separate look-up and translation service,
but this is outside the scope of Iptel - right Jonathan? ;-)]

The general problem of for how long the advertisements are valid, and how
quickly change can be propogated, is just that general -- there is nothing
special about 800 numbers or other directory numbers. 

Scott 


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:51:19 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09673
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:51:19 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 1DCBD1A3; Sat, 13 Mar 1999 17:14:13 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BDBFC7FFE; Thu, 11 Mar 1999 11:58:08 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BB0857FF5; Thu, 11 Mar 1999 11:51:42 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Greg Zhang" <gzhang@utstar.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Matt Squire" <msquire@nortelnetworks.com>
Cc: <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 16:34:50 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFGEJPEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Importance: Normal
In-Reply-To: <01a301be6a6e$2ff450c0$2821a8c0@greg_zhang.wacos.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Or, alternatively, project all numbers into a global, unambiguous space.
This seems cleaner and much less error prone than topologically
established scope boundaries, and less administratively burdensome than
solving the problem of scope name collisions and information propagation
boundaries. It also avoids having to talk about translations between
scopes, which I'm sure will happen when the other shoe drops :-).

> -----Original Message-----
> From: Greg Zhang [mailto:gzhang@utstar.com]
> Sent: Tuesday, March 09, 1999 3:46 PM
> To: Jonathan Rosenberg; Matt Squire
> Cc: hsalama@cisco.com; John Schnizlein; David R. Oran; Scott Petrack;
> Alok Mittal; iptel
> Subject: Re: GLP, number portability and wireless subscribers
>
>
> Matt Squire wrote:
> >I've received similar comments from several providers.  This
> is one of
> >things I was going to bring up when discussing general data
> representation
> >issues.  Scope has applicability to 1-800 numbers, but it also has
> >applicability to private/corporate numbers.
>
> < some comments cut>
>
> >What do folks think about the necessity of a 'scope' attribute for
> > (a) 1-800
> > (b) corporate/private addresses
> >(ie would I be wasting my time trying to address this concern?)
>
>
> I think as long as there is a real world requirement for it,
> the protocol
> would need to support it.
>
> With the peer group and a hierarchy scheme I outlined earlier, a scope
> becomes a proper definition/configuration of a peer group. A gateway
> can belong to more than one peer group on one level. The members
> in a peer group does not need to be geographically/topologically
> adjacent. Corporate/private numbers can be supported by defining
> a peer group for the corporate VPN and having the gateways in the
> peer group exchanging their number advertisement privately within
> the peer group.
>
> - Greg
>
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:52:35 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09684
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:52:35 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 70EE5192; Sat, 13 Mar 1999 17:13:51 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D3887819F; Fri, 12 Mar 1999 13:34:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E40F780EA; Fri, 12 Mar 1999 13:34:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Lu, Henry" <hwlu@tri.sbc.com>, "Lu, Henry" <hwlu@ttl.pactel.com>,
        "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>, <hsalama@cisco.com>
Cc: "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Fri, 12 Mar 1999 13:17:19 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFOEBJEFAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
Importance: Normal
In-Reply-To: <8391FE2580A1D11198CB00A0C99845D467D18E@tripacific.ttl.pactel.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

This is excellent sentiment, Henry, and I agree entirely. Luckily, it
isn't in the IPTEL charter to solve this problem. Maybe Enum will take
it on (poor Scott, I can see him cringing already), or maybe we can
define a whole new architecture to replace today's SCP-based number
translation.

It's a big problem space. We have to attack it incrmentally.

> -----Original Message-----
> From: Lu, Henry [mailto:hwlu@tri.sbc.com]
> Sent: Friday, March 12, 1999 12:48 PM
> To: 'David R. Oran'; Lu, Henry; Jonathan Rosenberg; hsalama@cisco.com
> Cc: iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> Dear David,
>
> My scope is from allowing IP telephony world can call
> customer on PSTN and vise versa. The current architecture
> is through SS7 gateway from IP world. There are works ongoing
> in other groups but I have not been convinced that the
> gateway will provide the correct information to IP world
> due to SS7 architecture and the way they are going. 1-800
> and number portability are two examples. I heard that we
> need to spend hundreds of million dollar just to upgrade
> the switches to do number portability. I think there should
> be money spent to make sure that SS7 world will be able to
> do it on SCP level so that a query from IP world can get the
> right route in PSTN. Only this way, we can truly open up local
> phone competition. Otherwise, I do not see that will happen.
> I hope it will help. Thanks.
>
> Henry W. Lu
> 925-598-1245
> SBC, Technology Resources, Inc.
> 4698 willow road,
> Pleasanton, CA 94588
>
>
> -----Original Message-----
> From: David R. Oran [mailto:oran@cisco.com]
> Sent: Friday, March 12, 1999 9:02 AM
> To: Lu, Henry; Jonathan Rosenberg; hsalama@cisco.com
> Cc: iptel
> Subject: RE: GLP, number portability and wireless subscribers
>
>
> > -----Original Message-----
> > From: Lu, Henry [mailto:hwlu@tri.sbc.com]
> > Sent: Thursday, March 11, 1999 7:16 PM
> > To: 'David R. Oran'; Jonathan Rosenberg; hsalama@cisco.com
> > Cc: iptel
> > Subject: RE: GLP, number portability and wireless subscribers
> >
> >
> > Dear David,
> >
> > I think what we can not afford is place a call to wrong number.
> Define "wrong number", please. If I have an 800 number and call to say
> that I want it to ring 555-4545 instead of 555-7676 and then
> hang up, is
> it "placing a call to a wrong number" if 555-7676 rings for
> the next 30
> seconds while the tables are getting updated? This is no
> different from
> the convergence properties of IP routing.
>
> > Whether we will reveal the true number or not is not
> > very important from my perspective.
> Really? This is not what I hear from my service provider
> friends, but as
> I said I'd rather not worry about this so let's drop it for now.
> Jonathan rightly points out that as long as you limit
> translation result
> visibility to parts of the network the owner of the
> translation trusts,
> things will be OK. The porblem arises if you simultaneously
> want to have
> optimal routing AND information hiding.
>
> > You have to know
> > that SS7 network is not like IP network, there is no
> > network topology database information in SCP like
> > OSPF or BGP in IP world.
> Yes, I know this.
>
> > This create problem for reconstructing
> > call routing from IP world. SS7 is per call base and do
> > a SCP query if there is 800 or ported number. No matter
> > SIP or H.323, will be per call based through gateway query.
> > We are still lock the protocol in old ways per call based. We can
> > not update IP world PSTN topology like what we are doing
> > in OSPF or BGP for the whole network in one database sync.
> Ok, I'll bite. Why not? Remember, we're not doing destination
> selection
> in GLP, we're only doing route selection once the final routing-number
> is already known.
>
> > I hope it will help. Thanks.
> >
> > Henry W. Lu
> >
> >
> > -----Original Message-----
> > From: David R. Oran [mailto:oran@cisco.com]
> > Sent: Tuesday, March 09, 1999 1:12 PM
> > To: Jonathan Rosenberg; hsalama@cisco.com
> > Cc: iptel
> > Subject: RE: GLP, number portability and wireless subscribers
> >
> >
> > I'm reluctant to bring this up, because some would view it
> as a corner
> > case or just a "feeeeture" of the PSTN, but when the PSTN
> does number
> > translation (or call forwarding in the network as well),
> the resulting
> > number is not revealed to the calling party. The pure
> directory scheme
> > would reveal the translated number since that's what would be
> > put in the
> > end-to-end signaling (e.g. SIP, H.323). Do we care about this?
> >
> > > -----Original Message-----
> > > From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
> > > Sent: Tuesday, March 09, 1999 1:43 PM
> > > To: hsalama@cisco.com
> > > Cc: John Schnizlein; Greg Zhang; David R. Oran; Matt Squire; Scott
> > > Petrack; Alok Mittal; iptel
> > > Subject: Re: GLP, number portability and wireless subscribers
> > >
> > >
> > > Hussein F. Salama wrote:
> > > >
> > > > Jonathan Rosenberg wrote:
> > > >
> > > > >
> > > > > This begs the question, then: does the translation of the
> > > 800 number get
> > > > > done before the gateway selection process, or after? By
> > > doing it before,
> > > > > a gateway can be chosen which is, in fact, close to the
> > > real number
> > > > > being dialed. With 800 numbers, all gateways will be able
> > > to reach it.
> > > > > It seems that this opens the possibility of very
> > > non-optimal routing. A
> > > > > caller in NY that dials an 800 number, goes to a gateway
> > > in CA. The
> > > > > phone network there determines that the 800 number
> > translates to a
> > > > > number in NY, and routes the call through the PSTN back
> > > to NY. However,
> > > > > if the translation occurs before gateway selection, the
> > caller can
> > > > > choose a local NY gateway and do much better.
> > > >
> > > > Because of the above example, I think all types of
> > translation (800
> > > > numbers or LNP) and mapping should occur prior to call
> > > routing. This is
> > > > particularly important for 800 numbers, because they have
> > a regional
> > > > scope in many cases. The same 800 number may owned by
> > > company X in CA
> > > > and company Y in NY. In the above example, the call will be
> > > incorrectly
> > > > routed to company X instead of going to the originally intended
> > > > destination, company Y.
> > >
> > > This is another good example of why translation before
> > > gateway selection
> > > is a good idea. However, I'm not sure there is much we
> can do from a
> > > protocol perspective to prevent it. There may be valid
> reasons why a
> > > company would like to advertise an 800 number (only they
> own it, for
> > > example), and I'd rather not exclude this possibility. It
> > seems like a
> > > sensible approach is to allow non-routing numbers in GLP,
> > but mention
> > > that there are issues with them, and translation before
> routing is a
> > > good idea. Does this sound sensible?
> > >
> > > Scott P., of course, is right, that the translation service
> > itself is
> > > outside the scope of iptel.
> > >
> > > -Jonathan R.
> > > --
> > > Jonathan D. Rosenberg                       Lucent Technologies
> > > Member of Technical Staff                   101 Crawfords
> Corner Rd.
> > > High Speed Networks Research                Holmdel, NJ 07733
> > > FAX:   (732) 834-5379                       Rm. 4C-526
> > > EMAIL: jdrosen@bell-labs.com
> > > URL: http://www.cs.columbia.edu/~jdrosen
> > >
> >
> >
> > ---------
> > This message came from the IETF IPTEL Working Group Mailing List.
> >
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:53:35 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09711
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:53:35 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id D5C59182; Sat, 13 Mar 1999 17:13:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 373A67FEF; Thu, 11 Mar 1999 11:55:03 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 433487FFB; Thu, 11 Mar 1999 11:43:29 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E5B511.8D8EE034@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 18:56:01 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: "David R. Oran" <oran@cisco.com>
Cc: Matt Squire <msquire@nortelnetworks.com>, Greg Zhang <gzhang@utstar.com>,
        hsalama <hsalama@cisco.com>, John Schnizlein <jschnizl@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <NBBBIFCAKKOPNMMKHHEFGEKEEEAA.oran@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

David R. Oran wrote:
> 
> Then scopes are identical to a deeper hierarchy and splitting them out
> is simply a syntactical, not a semantic distinction, right?

My understanding is that it is semantic. By "splitting them out", I
presume you mean having them be separate from the address itself. The
prefix represents information on "where you're calling to", and the
scope information on "where you are calling from". So, a prefix Y with
scope X means you can call Y from X. I think we need explicit scopes if
there are cases where route attributes for a prefix only apply when
calling from numbers outside of the prefix (and apply only when calling
from somewhere defined by the scope).

For example, for corporate addresses, I would say scopes are not needed.
For some local address, don't propagate it beyond the boundaries of the
enterprise. No scope is needed since if you've received the prefix,
you're allowed to call it. Same idea as enterprise IP networks with
internal IP addressing that never goes outside. No special routing
protocols needed,or extra attributes. Just a policy at the periphery
which causes nothing to be propagated further.

> 
> Do you think we would need hierarchical scopes? If so, do you see
> aggregation of hierarchical scopes? How would this differ from
> hierarchical aggregation of addresses? Are scopes always more global
> than addresses?

We could potentially define hierarchical scopes (if they are also
represented as prefixes, they are hierarchical). The rules for
aggregation would presumably be the same. But, adding another
aggregatable entity in a route advertisement seems like it will make
route aggregation overall more complex, so I'd rather avoid this. Are
scopes global? I think they need to be. The whole idea is that a prefix
advertised in one place is only useful somewhere else. 

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:54:53 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09723
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:54:52 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 5A0DB15D; Sat, 13 Mar 1999 17:11:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0F8BD7FFC; Thu, 11 Mar 1999 11:42:09 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D304A80A9; Thu, 11 Mar 1999 11:40:45 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "David R. Oran" <oran@cisco.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Greg Zhang" <gzhang@utstar.com>
Cc: <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "Matt Squire" <msquire@nortelnetworks.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 16:24:31 -0500
Message-ID: <NBBBIFCAKKOPNMMKHHEFAEJOEEAA.oran@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Importance: Normal
In-Reply-To: <36E57B5C.E357450@dnrc.bell-labs.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

I don't think it's either needed or desirable to have locally-scoped
numbers, at least at the scale we're talking about for GLP. It would be
like asking IP routing to handle scoping for Net 10 addresses rather
than banning them from crossing any administrative boundaries.

I'm of course willing to be convinced otherwise, but this seems a recipe
for wierd failure modes and horrendous complexity.

Dave.

> -----Original Message-----
> From: Jonathan Rosenberg [mailto:jdrosen@dnrc.bell-labs.com]
> Sent: Tuesday, March 09, 1999 2:50 PM
> To: Greg Zhang
> Cc: hsalama@cisco.com; John Schnizlein; David R. Oran; Matt Squire;
> Scott Petrack; Alok Mittal; iptel
> Subject: Re: GLP, number portability and wireless subscribers
>
>
> Greg Zhang wrote:
> >
> > >This is another good example of why translation before
> gateway selection
> > >is a good idea. However, I'm not sure there is much we can
> do from a
> > >protocol perspective to prevent it. There may be valid
> reasons why a
> > >company would like to advertise an 800 number (only they
> own it, for
> > >example), and I'd rather not exclude this possibility. It
> seems like a
> > >sensible approach is to allow non-routing numbers in GLP,
> but mention
> > >that there are issues with them, and translation before
> routing is a
> > >good idea. Does this sound sensible?
> >
> > I think to handle this type of cases correctly, we need to
> introduce a
> > concept of scope for the advertisement. Gateways in CA will
> advertise
> > the 800 number for company X in CA while gateways in NY will
> > advertise the same 800 for company Y in NY. The advertisement from
> > the CA domain does not get propagated to the NY domain and vice
> > versa. This way the numbers that are locally significant
> and are being
> > administrated locally stays local. This however does not prevent the
> > same 800 number being used by upper level for a third customer.
>
> A sensible idea, but seems hard to do. How does one represent
> scope? How
> does one encode the notion that "this number is valid within
> the greater
> new york metropolitan area", or "northern california" into a
> protocol?
>
> Perhaps one solution is to allow scopes to be represented as opaque
> tokens. These tokens can be registered, perhaps with IANA,
> along with a
> desription of the what the scope means. This would allow peers to
> program their LS's to recognize certain scope tokens, and do something
> appropriate with them. Prefixes with unrecognized scope
> tokens probably
> shouldn't be propagated.
>
> As an alternative, scope could be represented by a separate
> phone number
> prefix. This second prefix indicates the set of numbers you
> need to call
> from for this scope to be defined. For example:
>
> phone numbers: +1-800-55*
> scope: +1-415*
>
> means that 800-55* is only defined within parts of California.
>
> Any other ideas? Or do people think that this isn't needed?
>
> -Jonathan R.
>
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX:   (732) 834-5379                       Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:56:15 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09738
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:56:15 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 5BC8E162; Sat, 13 Mar 1999 17:11:58 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 9C57B804E; Thu, 11 Mar 1999 11:54:43 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 31CB68003; Thu, 11 Mar 1999 11:50:36 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <003001be6a93$b11283f0$2821a8c0@greg_zhang.wacos.com>
From: "Greg Zhang" <gzhang@utstar.com>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>
Cc: <hsalama@cisco.com>, "John Schnizlein" <jschnizl@cisco.com>,
        "David R. Oran" <oran@cisco.com>,
        "Matt Squire" <msquire@nortelnetworks.com>,
        "Scott Petrack" <scott.petrack@metatel.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
Date: Tue, 9 Mar 1999 17:16:24 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Jonathan Rosenberg wrote:
>In the current model, synchronization of the databases occurs from a
>point outward until an aggregation point, at which a route change is no
>longer propagated. This effectively creates a hierarchy along the
>boundaries of the addressing hierarchy. Of course, if the LS topology
>doesn't follow the topology of the phone number allocations in the
>gateways, aggregation will be more limited, and the system appears more
>flat. However, a system with explicitly defined boundaries, as you
>propose, will do no better in this scenario either. As such, I don't see
>much advantage to explicitly defining hierarchies.


I guess the difference really comes down to whether there is a high
degree of aggregatability among the phone numbers in the network.
If aggregation can be achieved for a large majority of the telephone
numbers in the network, the two approach's effect are similar.
Building a hierarchy may seems to be an unnecessary exercise.
However, if the aggregatability of phone number is low in the network,
the hierarchical approach used with some techniques outlined below
and some of my earlier messages will cut down the amount of update
needed in the network while still providing full reachability of all
numbers in the whole network.

In our target network, we have 1 million plus mobile subscribers and
about 400 to 500 telephony gateways. The numbering plan within the
network is flat. Any subscriber can be at any gateway at any given
moment in time. Within the network, the phone numbers are not
aggregatable by any means. Each one of them needs to be treated
individually.

A GLP or location update protocol as we have been calling it with
a hierarchy will allow us to limit the scope of each location
update as it happens when a handoff occurs to a small group of
gateways. This update will get propagated up through the hierarchy.
But it will not be propagated to other peer groups on the same level.
On each level, you only needs to propagate it in one peer group. When
a location look up request or a SETUP message gets generated in one
peer group, if the location look up fails, it will get passed up the
hierarchy by the peer group leader. Eventually, the location information
for the subscriber is resolved and SETUP message gets forwarded
to the gateway where the subscriber is currently at.

With this design, the location information is resolvable from any point
in the network at any given time. However, the updates does not
need to be propagated to every gateway in the network even though
the phone numbers cannot aggregated. The updates are only
propagated within small groups on each level. When used with
the concept of originator and a hello mechanism to check the status
of the connectivity to the originator, the location advertisement does
not need to be refreshed as long as no change has occurred, i.e. the
said user has not moved to a new gateway. This cuts down the
amount of update messages considerably thus providing the
scalability that is needed for a carrier class mobile telephone
network.

- Greg



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:56:49 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09749
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:56:49 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 9CF2918A; Sat, 13 Mar 1999 17:13:36 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3B9FF7FFD; Thu, 11 Mar 1999 11:53:33 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3B2FC7FEF; Thu, 11 Mar 1999 11:49:43 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990310123647.010299a8@pop.ozemail.com.au>
X-Sender: dclowes@pop.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Wed, 10 Mar 1999 12:36:47 +1100
To: Scott Petrack <scott.petrack@metatel.com>
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: Re: 800 numbers, portability, and aggregation
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
In-Reply-To: <Pine.LNX.4.04.9903091159110.3156-100000@petrack.metatel.co
 m>
References: <36E54E32.C8931AB2@dnrc.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Scott,

I agree with you - almost.

At 12:07 1999-03-09 -0500, Scott Petrack wrote:
>
>The translation of "directory numbers" to "network numbers" can have a lot
>of dependencies that are unrelated to GLP -- the calling party id, the
>time of day, some ACD info. 

"Directory" numbers, and speed-dial numbers are examples of dialed numbers.
They do need to be translated, in the caller's context. Some of them,
particularly the "directory" ones will translate into obvious network
numbers, but the translation should not be overlooked or ignored. For
example, 555-1212 may translate into +1-415-555-1212 in one location, but
+1-212-555-1212 in a different location. I believe that 1-800 numbers
observe a similar characteristic, they can translate into different network
addresses which are dependent on some characteristic of the caller's
context - their location.

Down under, we have 13 (one-three) numbers. They're like 1-800 numbers,
except the caller pays the cost of a local call while the callee also pays,
based on distance. Take-out pizza chains, banks, insurance companies, and
government departments all love them. You dial the number and get routed to
the local entity - if there is one.

I believe that 1-800 numbers are much the same.

Portable numbers are just another example of a dialed number. They need to
be extended with default local context, before being translated into a
routing number. When I dial 555-1212 in Kalamazoo (MI), I really mean
+1-616-555-1212. In the morning, this translates into routing number
+1-212-555-1212. After noon, the routing number becomes +1-415-555-1212,
and late at night +1-313-555-1212.

>I suppose that GLP advertisements are meant to be valid for some time
>which is related to an estimate of how long the particular gateway will be
>in service, rather than (say) that it can service a number until close
>of business at 5pm EST.

This sort of thing bothers me. Are you talking about the dialed number, or
the network/routing number?

If it's the dialed number, then there may be a different translation,
depending on location, or time of day. At some times, there may be no valid
translation, or a translation to a network number that is busy, or unmanned.

If it's a network number, then it should translate into something routable,
almost physical. It shouldn't suddenly cease to exist (outside the quantum
theory of Internet Telephony, the ID for which I've yet to see), but may be
busy or unanswered.
 
>That being said, of course Dave is right that there are a lot of 800
>numbers which are perfectly static, and it would be a very useful
>optimization to be able to advertise those numbers. 

Because the callee pays, probably distance related charges, the callee
would like to control which gateways advertise their numbers. If their only
office is in NYC, they would rather restrict advertising to NYC, than have
some gateway in SFO advertising "their" number.

Every gateway connected to the PSTN has the same "costs" to the caller for
every 800 number. If the callee has their own gateway (no cost to them)
then they should be permitted to ban other gateways from advertising their
number at no cost.

In addition, callees who have devices that can handle calls, should be able
to stake an exclusive claim to their particular network number, as well as
their directory number. In which case, their directory number may translate
into a "network number", which translates into a transport address, but
would probably translate directly into a transport address.

>Bottom line: I don't see why we should forbid a gateway from advertising
>an 800 number if it wishes to, but we still need a separate "directory
>number to network number" lookup protocol. 

Isn't that "Alias Address to Transport Address" lookup protocol. It may be
the same protocol. It may be used once, or several times. It may be used
differently, or the same.

Aggregation:
============

Portable numbers, and 800 numbers, are issued by somebody - you don't just
invent your own. These numbers are allocated from a range under the control
of somebody. These ranges are aggregated before they are allocated.

Routing numbers are also allocated by somebody. They are also aggregated
before allocation. Same deal. Frequently, or at least historically, the
allocating entities have been the same, and the allocation algorithms have
been linked.

Now we are talking about them being separated. And they are separated in
the mobile telephony world. Now, people with expertise in mobile telephony
will probably tell me this isn't the way it works, but it is a way it could
work.

I buy a mobile telephone, and select a service provider. I get a SIM card
and am told my phone number is X. Transparent to me, my SIM card has its
own number Y. Both X and Y are from a range allocated to my country, and my
service provider, and possibly my locality. My subscriber enters (X, Y)
into their database.

I switch off my mobile phone, and get on a plane (or the other way round).
When I get off, I turn on my phone. My phone tells the local carrier that
my SIM card number is X. The carrier can tell by the aggregation rules of
X, who is the responsible service provider, and establishes that they will
be paid for calls made. They also announce to the service provider how X
can be called.

When someone dials Y, the rules of aggregation, which are different from X,
direct the caller to my service provider. My service provider translates X
into Y, and redirects the call to the mobile carrier described above.

Presumably, as I roam about, the mobile carrier is updating the
relationship. It could either be in my service provider, or in the mobile
carrier, as I roam from tower to tower. It will be in my service provider,
as I roam from carrier to carrier, and country to country.

That is at least two distinct sets of aggregation. Perhaps the mobile
carrier address is also aggregated as well.

My phone number is countryCode.carrierCode.directoryNumber,
 while my SIM number is countryCode.carrierCode.subscriberNumber,
 and my temporary roaming id is countryCode.carrierCode.temporaryNumber!

In DNS, it might be 555-1212.616.1.dialled-number returning
555-1212.415.1.network-number, then translating that returns
555-1212@gw7.sfo.megatel.com etc....

Here the allocating authority for +1-616 is hit for the first translation,
while the gateway provider for +1-415 is hit for the second, and
recursively translated by ns.megatel.com.

>Scott 

Douglas


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 13 18:57:16 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA09760
	for <iptel-archive@odin.ietf.org>; Sat, 13 Mar 1999 18:57:15 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 7318C171; Sat, 13 Mar 1999 17:12:31 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 68AEE7FE7; Thu, 11 Mar 1999 11:45:32 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F3EE87FF1; Thu, 11 Mar 1999 11:41:31 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E5770B.647D7D36@dnrc.bell-labs.com>
Date: Tue, 09 Mar 1999 14:31:23 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: "Lu, Henry" <hwlu@tri.sbc.com>
Cc: "'pcalhoun@eng.sun.com'" <Pat.Calhoun@Eng.Sun.COM>,
        Matt Squire <msquire@nortelnetworks.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <8391FE2580A1D11198CB00A0C99845D467D183@tripacific.ttl.pactel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Lu, Henry wrote:
> 
> Hello,
> 
> I am in favor of Pat's point but I am still hoping that we can
> achieve a sound solution in order for us to handle number portability
> environment in the future, including wireless.
> 
> The issues I see is how to synchronize two database between GLP and SS7
> world. In AIN, 800 number information are kept in not only Signaling
> Control Point(SCP) but also Service Switching Point(SSP), normally in
> local DMS-100 or 5ESS switches. The 800 calling process had to go through
> a conversation session between SSP and SCP to figure out the final
> routing to the destination. There are two information databases been
> kept on both SCP and local SSP. Do not forget that there are many SSPs
> in the network, not like SCP, we have to deal with.
> The major problem occur when a new portable number been provisioned,
> into both SSP and SCP but there is no information on GLP side.
> GLP and SS7 database is out of sync basically.

As you mention, the somewhat unusual way in which the database is
distributed in the phone network makes this a hard problem. The best way
to keep these databases in sync is to simply not have a separate
database. When a proxy or gatekeeper is going to make a call, it first
queries the phone network (somehow) to translate the number into a
routing number. GLP only advertises routing numbers. As such, the
database of number translations is stored only in the phone network, as
it is today. 

If, however, one chooses to deploy a parallel database in the Internet,
accessible by LDAP or some other means (enum perhaps?), the
synchronization problem exists.

Fortunately, it is outside of our charter to worry about this problem.
We need only decide what numbers may or may not be propagated by GLP,
and how they are represented and aggregated.

-Jonathan R.



-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 14 03:23:20 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA00865
	for <iptel-archive@odin.ietf.org>; Sun, 14 Mar 1999 03:23:20 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id CD64EFB; Sun, 14 Mar 1999 03:22:39 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id F08D252AA; Sun, 14 Mar 1999 03:21:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3C25B52BB; Sun, 14 Mar 1999 03:21:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Sun, 14 Mar 1999 03:20:02 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>
Cc: iptel <iptel@lists.research.bell-labs.com>
Subject: RE: untraceable calls
In-Reply-To: <E32BD2062DAFD1119CFD0000F804F31702C04866@cosexch003.mcit.com>
Message-ID: <Pine.LNX.4.04.9903140317190.1750-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



On Fri, 12 Mar 1999, Hoefelmeyer, Ralph (MCI) wrote:

> This has interesting implications for call tracing and interception, which
> is mandated by some governments.  One could, by proxy, make untraceable
> calls.
> 

It is quite possible today to make "untraceable calls". I can call into a
PBX, and then from this PBX to another PBX in the same organization but in
a different country, and then dial out that PBX to the public network.
If I do it right, the police will only see the last PBX, and won't be able
to see my own CLI ("caller id") at all. 

If an administration wants to prevent "untraceable" calls, it can always
forbid calls from proxies (and PBXs ;-)).

Scott


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 14 15:11:10 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA05140
	for <iptel-archive@odin.ietf.org>; Sun, 14 Mar 1999 15:11:10 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id DEA47107; Sun, 14 Mar 1999 15:10:39 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7277E52AA; Sun, 14 Mar 1999 15:09:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C665F52BB; Sun, 14 Mar 1999 15:09:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Henry Sinnreich" <henry.sinnreich@mci.com>
To: "Scott Petrack" <scott.petrack@metatel.com>,
        "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>
Cc: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "John Schnizlein" <jschnizl@cisco.com>,
        "Greg Zhang" <gzhang@utstar.com>, "David R. Oran" <oran@cisco.com>,
        "Matt Squire" <msquire@nortelnetworks.com>,
        "Alok Mittal" <amittal@hss.hns.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Sun, 14 Mar 1999 14:06:56 -0600
Message-ID: <NBBBIIJFOKPMFOOILMBKAEFODMAA.henry.sinnreich@mci.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2212 (4.71.2419.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
In-Reply-To: <Pine.LNX.4.04.9903091223400.3873-100000@petrack.metatel.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Scott writes:
>Well, yes, but there are lots of 800 numbers whose translations change on
>time scales MUCH slower than the GLP advertisement changes. In fact, most
>800 numbers are this way -- they are used by small-to-medium
>businesses, point to the 'real' telephone number, with no
>logic at all, and don't change at all.

Still there is time of day routing for 800 numbers.

Henry

-----Original Message-----
From: owner-iptel@lists.research.bell-labs.com
[mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Scott
Petrack
Sent: Tuesday, March 09, 1999 11:30 AM
To: pcalhoun@eng.sun.com
Cc: Jonathan Rosenberg; John Schnizlein; Greg Zhang; David R. Oran; Matt
Squire; Alok Mittal; iptel
Subject: Re: GLP, number portability and wireless subscribers




On Tue, 9 Mar 1999, pcalhoun@eng.sun.com wrote:

[Jonathan R:}
> > This begs the question, then: does the translation of the 800 number get
> > done before the gateway selection process, or after? By doing it before,
...
> > if the translation occurs before gateway selection, the caller can
> > choose a local NY gateway and do much better.
>
> Translation before the gateway selection is *obviously* the correct model.
We
> should strive towards this goal.
>
> PatC
>

Well, yes, but there are lots of 800 numbers whose translations change on
time scales MUCH slower than the GLP advertisement changes. In fact, most
800 numbers are this way -- they are used by small-to-medium
businesses, point to the 'real' telephone number, with no
logic at all, and don't change at all.

If a GW wants to advertise that it can service such a number, I don't see
why it should be forbidden from doing so.

[Of course there is a need for a separate look-up and translation service,
but this is outside the scope of Iptel - right Jonathan? ;-)]

The general problem of for how long the advertisements are valid, and how
quickly change can be propogated, is just that general -- there is nothing
special about 800 numbers or other directory numbers.

Scott


---------
This message came from the IETF IPTEL Working Group Mailing List.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 14 16:57:08 1999
Received: from plts.org (plts.research.bell-labs.com [204.178.16.48] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA05936
	for <iptel-archive@odin.ietf.org>; Sun, 14 Mar 1999 16:57:08 -0500 (EST)
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by plts.org (Postfix) with ESMTP
	id 83191122; Sun, 14 Mar 1999 16:56:48 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2170652AA; Sun, 14 Mar 1999 16:55:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 69AFF52BB; Sun, 14 Mar 1999 16:55:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Sun, 14 Mar 1999 16:54:49 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: Henry Sinnreich <henry.sinnreich@mci.com>
Cc: "pcalhoun@eng.sun.com" <Pat.Calhoun@Eng.Sun.COM>,
        Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        John Schnizlein <jschnizl@cisco.com>, Greg Zhang <gzhang@utstar.com>,
        "David R. Oran" <oran@cisco.com>,
        Matt Squire <msquire@nortelnetworks.com>,
        Alok Mittal <amittal@hss.hns.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
In-Reply-To: <NBBBIIJFOKPMFOOILMBKAEFODMAA.henry.sinnreich@mci.com>
Message-ID: <Pine.LNX.4.04.9903141652030.4130-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

yes, of course, Henry, you are right. My point was that there are 800
numbers without time-of-day routing, and such numbers could clearly be
routed via GLP without any problem at all. 

My guess is that 800-numbers-with-time-of-day-routing are ones that we
don't want to advertise via GLP. Of course there is a slippery slope here
-- maybe we could add a time of day "scope" to the numbers -- but I think
that we don't want to go down that path (too much soap). 

Scott

On Sun, 14 Mar 1999, Henry Sinnreich wrote:

> Scott writes:
> >Well, yes, but there are lots of 800 numbers whose translations change on
> >time scales MUCH slower than the GLP advertisement changes. In fact, most
> >800 numbers are this way -- they are used by small-to-medium
> >businesses, point to the 'real' telephone number, with no
> >logic at all, and don't change at all.
> 
> Still there is time of day routing for 800 numbers.
> 
> Henry
> 
> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Scott
> Petrack
> Sent: Tuesday, March 09, 1999 11:30 AM
> To: pcalhoun@eng.sun.com
> Cc: Jonathan Rosenberg; John Schnizlein; Greg Zhang; David R. Oran; Matt
> Squire; Alok Mittal; iptel
> Subject: Re: GLP, number portability and wireless subscribers
> 
> 
> 
> 
> On Tue, 9 Mar 1999, pcalhoun@eng.sun.com wrote:
> 
> [Jonathan R:}
> > > This begs the question, then: does the translation of the 800 number get
> > > done before the gateway selection process, or after? By doing it before,
> ...
> > > if the translation occurs before gateway selection, the caller can
> > > choose a local NY gateway and do much better.
> >
> > Translation before the gateway selection is *obviously* the correct model.
> We
> > should strive towards this goal.
> >
> > PatC
> >
> 
> Well, yes, but there are lots of 800 numbers whose translations change on
> time scales MUCH slower than the GLP advertisement changes. In fact, most
> 800 numbers are this way -- they are used by small-to-medium
> businesses, point to the 'real' telephone number, with no
> logic at all, and don't change at all.
> 
> If a GW wants to advertise that it can service such a number, I don't see
> why it should be forbidden from doing so.
> 
> [Of course there is a need for a separate look-up and translation service,
> but this is outside the scope of Iptel - right Jonathan? ;-)]
> 
> The general problem of for how long the advertisements are valid, and how
> quickly change can be propogated, is just that general -- there is nothing
> special about 800 numbers or other directory numbers.
> 
> Scott
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 14 20:57:33 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA07377
	for <iptel-archive@odin.ietf.org>; Sun, 14 Mar 1999 20:57:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B412052AA; Sun, 14 Mar 1999 20:53:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1B80D52B7; Sun, 14 Mar 1999 20:53:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36E91472.65B6BC71@telcel.net.ve>
Date: Fri, 12 Mar 1999 09:19:46 -0400
From: Omar Ferrer <oferrer@telcel.net.ve>
Organization: TELCEL Celular, C.A.
X-Mailer: Mozilla 4.5 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Henri Setiawan <henry@risti.telkom.co.id>,
        H323 Implementors <h323implementors@imtc.org>,
        IPTEL <iptel@lists.research.bell-labs.com>
Subject: [Fwd: Parameter for interoperability test]
Content-Type: multipart/alternative;
 boundary="------------D39C298268573A8F6EFE1EE5"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


--------------D39C298268573A8F6EFE1EE5
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Dear Mr. Setiawan:

I've recently subscribe to this newsgroup asking for some help on
testing VoIP products. I'm beginning a project in which we want to test
Gateways, Gatekeepers and Network Managers, but we do not have
experience with that because this field is completely new for the Latin
America market.

Do you now what I need to measure or test on each equipment and how?

We are beginning a project for VoIP. In the first phase we want to test
each equipment on a point to point link (TDM) with a higher bandwith and
then decreasing as the traffic increase. In the second phase, we want to
test the same products but with direct connection to the Internet. Here
we need to prove some QoS tools and bandwith management techniques, and
also, some testing with a Global Network for call terminations
worldwide.

Could you give some approach for this?

Thanks!!!
Omar Ferrer
Caracas - Venezuela

-------- Original Message --------
Subject: Parameter for interoperability test
   Date: Fri, 12 Mar 1999 13:46:55 +0700
   From: "Henri Setiawan" <henry@risti.telkom.co.id>
     To: <h323implementors@imtc.org>
     CC: <iptel@lists.research.bell-labs.com>

 Dear folks, We like to study about interoperability test for gateway
and gateway keeper, do anyone know what kind of parameters that we
should test for this testing ? Thanks Henri SR&D of PT Telkom
Indonesia.

--------------D39C298268573A8F6EFE1EE5
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<body bgcolor="#FFFFFF">
Dear Mr. Setiawan:
<p>I've recently subscribe to this newsgroup asking for some help on testing
VoIP products. I'm beginning a project in which we want to test Gateways,
Gatekeepers and Network Managers, but we do not have experience with that
because this field is completely new for the Latin America market.
<p>Do you now what I need to measure or test on each equipment and how?
<p>We are beginning a project for VoIP. In the first phase we want to test
each equipment on a point to point link (TDM) with a higher bandwith and
then decreasing as the traffic increase. In the second phase, we want to
test the same products but with direct connection to the Internet. Here
we need to prove some QoS tools and bandwith management techniques, and
also, some testing with a Global Network for call terminations worldwide.
<p>Could you give some approach for this?
<p>Thanks!!!
<br>Omar Ferrer
<br>Caracas - Venezuela
<p>-------- Original Message --------
<table BORDER=0 CELLSPACING=0 CELLPADDING=0 >
<tr>
<th ALIGN=RIGHT VALIGN=BASELINE NOWRAP>Subject:&nbsp;</th>

<td>Parameter for interoperability test</td>
</tr>

<tr>
<th ALIGN=RIGHT VALIGN=BASELINE NOWRAP>Date:&nbsp;</th>

<td>Fri, 12 Mar 1999 13:46:55 +0700</td>
</tr>

<tr>
<th ALIGN=RIGHT VALIGN=BASELINE NOWRAP>From:&nbsp;</th>

<td>"Henri Setiawan" &lt;henry@risti.telkom.co.id></td>
</tr>

<tr>
<th ALIGN=RIGHT VALIGN=BASELINE NOWRAP>To:&nbsp;</th>

<td>&lt;h323implementors@imtc.org></td>
</tr>

<tr>
<th ALIGN=RIGHT VALIGN=BASELINE NOWRAP>CC:&nbsp;</th>

<td>&lt;iptel@lists.research.bell-labs.com></td>
</tr>
</table>

<br>&nbsp;<font size=-1>Dear folks,</font>&nbsp;<font size=-1>We like to
study about interoperability test for gateway and gateway keeper, do anyone
know what kind of parameters that we should test for this testing ?</font>&nbsp;<font size=-1>Thanks</font>&nbsp;<font size=-1>Henri
S</font><font size=-1>R&amp;D of PT Telkom Indonesia.</font>&nbsp;&nbsp;
</body>
</html>

--------------D39C298268573A8F6EFE1EE5--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 14 22:03:13 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA07736
	for <iptel-archive@odin.ietf.org>; Sun, 14 Mar 1999 22:03:12 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B5DEB52BB; Sun, 14 Mar 1999 21:59:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0ACF152BF; Sun, 14 Mar 1999 21:59:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <E32BD2062DAFD1119CFD0000F804F31702C0487F@cosexch003.mcit.com>
From: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>
To: "'Omar Ferrer'" <oferrer@telcel.net.ve>,
        Henri Setiawan <henry@risti.telkom.co.id>,
        H323 Implementors <h323implementors@imtc.org>,
        IPTEL <iptel@lists.research.bell-labs.com>
Subject: RE: [Fwd: Parameter for interoperability test]
Date: Mon, 15 Mar 1999 02:56:57 -0000
Importance: high
X-Priority: 1
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE6E8F.7D7874DC"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE6E8F.7D7874DC
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Omar (& Henri),

The first thing you need to test gatekeeper/gateway interoperability is =
a
LAN protocol analyzer, preferably one that will decode H.323.
Hewlett-Packard has one in Beta, it works well.
 =20
Cisco has a product called Netflow for it's products, which allows one =
to
measure various attributes of the routers' performance (note that =
Cisco's
gatekeeper and gateway products use a version of their IOS).  If you =
use
Cisco, this is a possibility.

Various Measures of Merit (MOM) could be:

- time of call initiation to call setup
- interoperability of H.323 products; verify claims of H.323 v2
compatibility between products from different firms.  This has been an
important issue.
- the ease of configuration of the gatekeepers for number translation
(provisioning), if you are not accessing an existing number translation
mechanism
- the ability to hook into the gatekeeper for your outside number
translation setup (API access to the gatekeeper)

I would suggest you use an Ethernet (or Fast Ethernet)  LAN between the
gateways and gatekeeper.  Plug all of the devices into one hub, and it =
keeps
things simple, for your first try.   This is a simpler scope than what =
you
propose below, but interoperability is a key issue before moving to =
bigger
ideas.

QoS is a complex question, dependent on the target network(s).  Router
configurations, network topologies and firewalls all play a role here.

I would be interested in what you think, and what results you come up =
with,
if you can share this with us.

Good Luck!

Ralph

Ralph S. Hoefelmeyer
MCI WorldCom Engineer


> -----Original Message-----
> From:	Omar Ferrer [SMTP:oferrer@telcel.net.ve]
> Sent:	Friday, March 12, 1999 6:20 AM
> To:	Henri Setiawan; H323 Implementors; IPTEL
> Subject:	[Fwd: Parameter for interoperability test]
>=20
> Dear Mr. Setiawan:=20
>=20
> I've recently subscribe to this newsgroup asking for some help on =
testing
> VoIP products. I'm beginning a project in which we want to test =
Gateways,
> Gatekeepers and Network Managers, but we do not have experience with =
that
> because this field is completely new for the Latin America market.=20
>=20
> Do you now what I need to measure or test on each equipment and how?=20
>=20
> We are beginning a project for VoIP. In the first phase we want to =
test
> each equipment on a point to point link (TDM) with a higher bandwith =
and
> then decreasing as the traffic increase. In the second phase, we want =
to
> test the same products but with direct connection to the Internet. =
Here we
> need to prove some QoS tools and bandwith management techniques, and =
also,
> some testing with a Global Network for call terminations worldwide.=20
>=20
> Could you give some approach for this?=20
>=20
> Thanks!!!=20
> Omar Ferrer=20
> Caracas - Venezuela=20
>=20
> -------- Original Message --------=20
> Subject:=A0	 Parameter for interoperability test	=20
> Date:=A0	 Fri, 12 Mar 1999 13:46:55 +0700	=20
> From:=A0	 "Henri Setiawan" <henry@risti.telkom.co.id>	=20
> To:=A0	 <h323implementors@imtc.org>	=20
> CC:=A0	 <iptel@lists.research.bell-labs.com>	=20
>=20
> =A0Dear folks,=A0We like to study about interoperability test for =
gateway and
> gateway keeper, do anyone know what kind of parameters that we should =
test
> for this testing ?=A0Thanks=A0Henri SR&D of PT Telkom =
Indonesia.=A0=A0=20

------_=_NextPart_001_01BE6E8F.7D7874DC
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Diso-8859-1">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2232.0">
<TITLE>RE: [Fwd: Parameter for interoperability test]</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Omar (&amp; Henri),</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">The first thing you need to =
test gatekeeper/gateway interoperability is a LAN protocol analyzer, =
preferably one that will decode H.323.&nbsp; Hewlett-Packard has one in =
Beta, it works well.</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">&nbsp; </FONT>
<BR><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Cisco has a product called =
Netflow for it's products, which allows one to measure various =
attributes of the routers' performance (note that Cisco's gatekeeper =
and gateway products use a version of their IOS).&nbsp; If you use =
Cisco, this is a possibility.</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Various Measures of Merit =
(MOM) could be:</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">- time of call initiation to =
call setup</FONT>
<BR><FONT COLOR=3D"#0000FF" FACE=3D"Arial">- interoperability of H.323 =
products; verify claims of H.323 v2 compatibility between products from =
different firms.&nbsp; This has been an important issue.</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">- the ease of configuration =
of the gatekeepers for number translation (provisioning), if you are =
not accessing an existing number translation mechanism</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">- the ability to hook into =
the gatekeeper for your outside number translation setup (API access to =
the gatekeeper)</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">I would suggest you use an =
Ethernet (or Fast Ethernet)&nbsp; LAN between the gateways and =
gatekeeper.&nbsp; Plug all of the devices into one hub, and it keeps =
things simple, for your first try.&nbsp;&nbsp; This is a simpler scope =
than what you propose below, but interoperability is a key issue before =
moving to bigger ideas.</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">QoS is a complex question, =
dependent on the target network(s).&nbsp; Router configurations, =
network topologies and firewalls all play a role here.</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">I would be interested in what =
you think, and what results you come up with, if you can share this =
with us.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Good Luck!</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Ralph</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Ralph S. Hoefelmeyer</FONT>
<BR><FONT COLOR=3D"#0000FF" FACE=3D"Arial">MCI WorldCom Engineer</FONT>
</P>
<BR>
<UL>
<P><FONT SIZE=3D1 FACE=3D"Arial">-----Original Message-----</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">From:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Omar Ferrer [SMTP:oferrer@telcel.net.ve]</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">Sent:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Friday, March 12, 1999 6:20 AM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">Henri Setiawan; H323 Implementors; IPTEL</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT>=
</B> <FONT SIZE=3D1 FACE=3D"Arial">[Fwd: Parameter for interoperability =
test]</FONT>
</P>

<P><FONT COLOR=3D"#000000" FACE=3D"Arial">Dear Mr. Setiawan: </FONT>
</P>

<P><FONT COLOR=3D"#000000" FACE=3D"Arial">I've recently subscribe to =
this newsgroup asking for some help on testing VoIP products. I'm =
beginning a project in which we want to test Gateways, Gatekeepers and =
Network Managers, but we do not have experience with that because this =
field is completely new for the Latin America market. </FONT></P>

<P><FONT COLOR=3D"#000000" FACE=3D"Arial">Do you now what I need to =
measure or test on each equipment and how? </FONT>
</P>

<P><FONT COLOR=3D"#000000" FACE=3D"Arial">We are beginning a project =
for VoIP. In the first phase we want to test each equipment on a point =
to point link (TDM) with a higher bandwith and then decreasing as the =
traffic increase. In the second phase, we want to test the same =
products but with direct connection to the Internet. Here we need to =
prove some QoS tools and bandwith management techniques, and also, some =
testing with a Global Network for call terminations worldwide. =
</FONT></P>

<P><FONT COLOR=3D"#000000" FACE=3D"Arial">Could you give some approach =
for this? </FONT>
</P>

<P><FONT COLOR=3D"#000000" FACE=3D"Arial">Thanks!!!<BR>
Omar Ferrer<BR>
Caracas - Venezuela </FONT>
</P>

<P><FONT COLOR=3D"#000000" FACE=3D"Arial">-------- Original Message =
-------- </FONT>
<BR><B><FONT =
FACE=3D"Arial">Subject:=A0</FONT></B>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
;&nbsp;<FONT FACE=3D"Arial"> Parameter for interoperability =
test&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </FONT>
<BR><B><FONT FACE=3D"Arial">Date:=A0</FONT></B>&nbsp;&nbsp;<FONT =
FACE=3D"Arial"> Fri, 12 Mar 1999 13:46:55 =
+0700&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </FONT>
<BR><B><FONT FACE=3D"Arial">From:=A0</FONT></B>&nbsp;&nbsp;<FONT =
FACE=3D"Arial"> &quot;Henri Setiawan&quot; =
&lt;henry@risti.telkom.co.id&gt;&nbsp;&nbsp;&nbsp;&nbsp; </FONT>
<BR><B><FONT =
FACE=3D"Arial">To:=A0</FONT></B>&nbsp;&nbsp;&nbsp;&nbsp;<FONT =
FACE=3D"Arial"> =
&lt;h323implementors@imtc.org&gt;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =
</FONT>
<BR><B><FONT =
FACE=3D"Arial">CC:=A0</FONT></B>&nbsp;&nbsp;&nbsp;&nbsp;<FONT =
FACE=3D"Arial"> =
&lt;iptel@lists.research.bell-labs.com&gt;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=
 </FONT>
</P>

<P><FONT COLOR=3D"#000000" FACE=3D"Arial">=A0</FONT><FONT =
COLOR=3D"#000000" SIZE=3D2 FACE=3D"Arial">Dear folks,</FONT><FONT =
COLOR=3D"#000000" FACE=3D"Arial">=A0</FONT><FONT COLOR=3D"#000000" =
SIZE=3D2 FACE=3D"Arial">We like to study about interoperability test =
for gateway and gateway keeper, do anyone know what kind of parameters =
that we should test for this testing ?</FONT><FONT COLOR=3D"#000000" =
FACE=3D"Arial">=A0</FONT><FONT COLOR=3D"#000000" SIZE=3D2 =
FACE=3D"Arial">Thanks</FONT><FONT COLOR=3D"#000000" =
FACE=3D"Arial">=A0</FONT><FONT COLOR=3D"#000000" SIZE=3D2 =
FACE=3D"Arial">Henri SR&amp;D of PT Telkom Indonesia.</FONT><FONT =
COLOR=3D"#000000" FACE=3D"Arial">=A0=A0 </FONT></P>
</UL>
</BODY>
</HTML>
------_=_NextPart_001_01BE6E8F.7D7874DC--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 18:08:32 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA06491
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 18:08:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D2E81533F; Mon, 15 Mar 1999 13:53:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 39130534B; Mon, 15 Mar 1999 13:53:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36ED579F.D487D82A@cisco.com>
Date: Mon, 15 Mar 1999 10:55:27 -0800
From: Hussein F Salama <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: "ronald h. davis" <ronaldd@lucent.com>
Cc: Alok Mittal <amittal@hss.hns.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: Re: glp draft based on BGP
References: <01BE663A.B6B0A9C0@hss056.hss.hns.com> <36DEE85A.25189194@cisco.com> <36ED4C70.BB6CF1B@lucent.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



"ronald h. davis" wrote:
> 
> Hussein F. Salama wrote:
> >
> > An administrative domain, AD1, may wish that all incoming H.323 calls
> > from neighboring domains be routing by its gatekeeper first. The
> > gatekeeper then directs the incoming call to the appropriate gateway or
> > proxy. In this case, AD1, specifies the address of a gatekeeper as the
> > next hop and RAS as the signaling protocol. Alternatively, AD1, may
> > specify an H.323 gateway or proxy as the next hop and Q.931 as the call
> > signaling protocol. In the first case the gatekeeper is involved in both
> > call routing and admission control. In the second case, the gatekeeper
> > does admission control only.
> >
> 
> h.323 signalling does not occur over the ras channel, but anyway, my

Above, I suggested using RAS for call routing and admission control.
Didn't mention anything about call signaling. 

> question has to do with the overhead of querying ts1 first and then
> sending the setup message (i am referencing figure 1 of your draft).
> is it valid for ua to send the setup message (this assumes that it
> knows the protocol that it is using) to ts1 straight on, with ts1
> using its routing information to determine the next hop for the
> message?
> 
> the thrust of my suggestion is to view the tbgp speaker as being an
> appliation layer router.  depending upon network policies, the tbgp
> speaker could send the equivalent of an icmp redirect message to ua
> that could be cached for future calls.

I think the proposed query/reply mechanism is simpler. Redirecting the
call involves more state setup/teardown/modification overhead.

Hussein

> 
> --
>   __  ______  __  / __/ |       lucent technologies, naperville il, usa
> _/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
> "if you ain't the lead dog, the scenery never changes"  -- edmund wilson

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC-6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: (408) 527 7147, Fax: (408) 527 1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 18:47:16 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA06804
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 18:47:16 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D341752C2; Wed, 17 Mar 1999 18:31:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 16FFA534C; Wed, 17 Mar 1999 18:13:07 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Reply-To: <pmutooni@vipcalling.com>
From: "Philip K. Mutooni" <pmutooni@vipcalling.com>
To: "'Scott Petrack'" <scott.petrack@metatel.com>,
        "'Hoefelmeyer, Ralph (MCI)'" <Ralph.Hoefelmeyer@mci.com>
Cc: "'iptel'" <iptel@lists.research.bell-labs.com>
Subject: RE: untraceable calls
Date: Mon, 15 Mar 1999 17:56:18 -0600
Message-ID: <000101be6f3f$6bcecb90$3acbc0d1@pmutooni.vipcalling.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Importance: Normal
In-Reply-To: <Pine.LNX.4.04.9903140317190.1750-100000@petrack.metatel.com>
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

This would work well If I had an additional PBX abroad -
how could one do this if the organization did not have
this ability of routing calls to another country? Is
there a way to encrypt or strip the CLI - I came across
a company called microlink (in Israel) that develops a
device for doing this sort of thing. Any thoughts on
alternative solutions?

Phil

-----Original Message-----
From: owner-iptel@lists.research.bell-labs.com
[mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Scott
Petrack
Sent: Sunday, March 14, 1999 2:20 AM
To: Hoefelmeyer, Ralph (MCI)
Cc: iptel
Subject: RE: untraceable calls




On Fri, 12 Mar 1999, Hoefelmeyer, Ralph (MCI) wrote:

> This has interesting implications for call tracing and interception, which
> is mandated by some governments.  One could, by proxy, make untraceable
> calls.
>

It is quite possible today to make "untraceable calls". I can call into a
PBX, and then from this PBX to another PBX in the same organization but in
a different country, and then dial out that PBX to the public network.
If I do it right, the police will only see the last PBX, and won't be able
to see my own CLI ("caller id") at all.

If an administration wants to prevent "untraceable" calls, it can always
forbid calls from proxies (and PBXs ;-)).

Scott


---------
This message came from the IETF IPTEL Working Group Mailing List.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 19:10:57 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA07067
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 19:10:56 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 607135367; Wed, 17 Mar 1999 18:22:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E7F365357; Wed, 17 Mar 1999 18:20:57 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <002301be701c$4ca92020$46711fd1@sathomas>
From: "Stephen Thomas" <stephen.thomas@transnexus.com>
To: "IPTEL IETF Mailing List" <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 5: price/cost
Date: Tue, 16 Mar 1999 20:43:35 -0500
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3155.0
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Sorry for getting this out after the meeting, but I'm just now clearing
my inbox. But then again, we didn't get to discuss this in the meeting
anyway. Comments below.

-----Original Message-----
From: Matt Squire <msquire@nortelnetworks.com>
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Date: Friday, 12 March 1999 7:09 PM
Subject: data issue 5: price/cost


|5) Price/cost.
|
|We all agree that we need to somehow represent the cost of making a call to
|a particular destination at a particular gateway.  The more difficult
|question is how.
|
|a) Can we define a set of cost primitives to sufficiently represent pricing
|structures?
|b) If so, what are they?
|c) If not, what then?
|
|I believe that we had consensus at the last wg meeting that we would be
|unable to determine a sufficient set of cost primitives to satisfy every
|objective, and that satisfying a smaller subset of objectives was not
|sufficient.  Which seemed to leave us under a few rocks and surrounded by
|hard places.
|
|One option is to use an opaque (ie not necessarily understood by the
|call-routers) attribute to express the cost of a call.  This would permit
|arbitrary pricing structures to be expressed via something like a text
|message or a URL.  This would be of great help to those that view GLP as
|providing a service to the calling user to help them choose the 'cheapest'
|gateway.

Assuming that the model allows for gateway-to-gateway calls, it's worth
noting that in GW-to-GW calls the price charged to the calling user (by
that user's service provider) may well be completely independent of the
price charged to the originating service provider (by the terminating
service provider). Obviously, the terminating provider wants to express
its price and the originating service provider wants calls routed on
the basis of its cost. In that case GLP is not providing a service to
the calling user but rather to the call-router. As such, the price
attribute MUST be understandable by the call-router.

|In general, we all agreed that the cost of the IP segment of the call will
|be negligible compared to the cost of the PSTN segment.   Since the effect
|of the IP segment on price is deemed negligible, the fact the a call-router
|cannot adjust the attribute is not too harmful.  Unfortunately, an opaque
|attribute has significant impacts to aggregation and route selection.  If
|the call-router does not understand the attribute, then it could not choose
|the best route based on that attribute.  Also, routes with different prices
|cannot be aggregated if the structure is not understood.

I can see how price (with or without units) might aid route comparison,
but how does it help aggregation? (Possibly I'm being dense here and
someone will surely rescue me soon.) Isn't it true that "neighbor"
trees can only be aggregated if the price is identical. If for, example,
the prices for various prefixes are:

 +330*   $0.10      +331*   $0.10      +332*   $0.10      +333*   $0.10
 +334*   $0.10      +335*   $0.10      +336*   $0.10      +337*   $0.10
 +338*   $0.10      +339*   $0.10

then it may be possible (everything else being equal) to aggregate
those into a single prefix +33* at $0.10. But if even one of the original
prefixes has a price of, say, $0.11, then no aggregation is possible.
(Assuming prefixes rather than ranges for simplicity.)

Since we're talking about inter-domain gateway location here, I have
to ask what is the probability that two domains will charge the same
amount for "neighbor" trees? I suspect that probability is very low.

Is it maybe a question of my misunderstanding of the term "aggregation"
and what we're actually talking about here is the subsuming of a subtree
by a parent when the parent is of equal or better cost? (E.g. a
price of $0.05 for +3* would take priority over all of the above prefixes.)

|Which brings us to a second option.  Even though we cannot agree on a set
|of metrics that would be sufficient, a unitless metric can provide great
|benefits.  Much like a unitless capacity, a unitless price metric would
|permit comparisons of different routes based on a relative view of price.
|The caller would not know the 'actual' price in terms of price/min etc, but
|a unitless price would permit better aggregation and route selection. As
|with a unitless capacity, the set of providers within a ip telephony
|confederation would have to coordinate the interpretation of the relative
|metric between and within domains.  Since the protocol does not know how to
|interpret the relative metric, it is difficult for the attribute to be
|adjusted based on the cost of the IP path.  We could mandate that the
|attribute (a) not be adjusted when propagated, or (b) make it a
|configurable policy within GLP to increment this metric by x when
|propagated.

Admittedly it may be difficult to achieve politically, but I think
interoperability would be better served if we could agree at least
on a default, absolute interpretation for the price attribute and
perhaps allow confederations to choose to override this meaning
with their own interpetation. It might be helpful to have a flag
that signaled private interpretation. If, in the worst case, no
one ended up liking the default interpretation, the situation
would simply degenerate into the relative metric case.

To procede along these lines, the default interpretation would have
to include the following components in its definition.

  currency
  unit of measure: minutes, bits, packets, pages (for fax), etc.
  rounding increment: minute, second, etc.
  precise definition of start/stop events

Clearly specifying a single currency will be difficult politically,
but it is essential for comparison. You can't just say "convert
using standard currency exchange rates" because, (a) there is no
such thing as a standard exchange rate, and (b) most world
currencies are not (legally) freely traded. FWIW, GSM addressed
this problem by creating an artificial currency called "special
drawing rights" and the GSM MoU Association maintains exchange
rates for SDRs. Absent an equivalent organization to set exchange
rates, this approach seems impractical in our case.

As for the other components, I suggest that most service providers
would prefer to base their prices as closely as possible to their
costs, which means minutes rather than bits or packets. That does
require extra work in defining the rounding increment and start/stop
times (since time is continuous rather than discrete), but other
pricing schemes are likely to encounter more resistance.

Of course, this approach doesn't give providers nearly the flexibility
they have today, at least at the retail level. But, on the other hand,
there are some attributes of existing SCN retail pricing that make
predicting the cost in advance of the call (i.e. for route selection
or comparison) impossible. One example is MCI's $1 for the first 20
minutes and $0.10/minute thereafter. (If I'm remembering the ads
correctly.) I think we have to accept the fact that, at least for
the initial version of the protocol, we simply won't be able to deal
with these scenarios.


|My personal belief is that we probably need both a pricing metric and a
|pricing description.  For a pricing description, I suggest either a text
|based message or a URL that can be used to describe a pricing structure.
|To assist in aggretability, providers should use relatively few different
|attributes (the attributes can be subtyped by provider and subtyped within
|that) to describe their pricing structures, having each provide enough
|information to be applicable at a number of gateways. The intent of the
|pricing description is that this attribute can be displayed to a user (or
|interpreted by some process on behalf of an end-user) when/if the user
|accesses the call routing database via some front-end protocol, and that
|the information fairly accurately reflects what the user will actually pay
|for the call. A URL could be used to handle large amounts of data in a
|language independent manner.

I guess I'm still worried about the distinction between cost to the
end user and cost to the originating service provider. If the originating
service provider is making the choice, then the call-router is going
to have to interpret the attribute, not the user.

|This information would not be modified by the
|call-routers, only propagated so that the ingress can get an idea of how
|much a call is going to cost.  The information reflects only the cost on
|the PSTN side and nothing on the IP side.
|Additionally, there should be a pricing metric that can be used and
|modified by the routers in route selection and aggregation.
|
|I think the combination of the two can provide both network and end-user
|information for gateway selection.

Maybe now it's finally starting to sink into my feeble brain. (Or
maybe not?) By differentiating between network and end-user do you
mean to differentiate between calling parties that are using a
gateway (network) and calling parties that are using intelligent
devices such as SIP clients in their PC (end-user)? If that's
the case, then I might see the value in two different attributes.
One that's interpretable by call-routers and another that can
be presented to the end user. Of course, I wonder how many
terminating gateway operators are really going to price the
two differently. And, if they do, how long it will be before
some end-user pretends to be a gateway (or vice versa) in order
to get the most favorable pricing.

|If we agree that a pricing description is needed, I would further suggest
|that it be signed by the orginator of the information using a public-key
|method.  Since people may be making decision based on this data, it should
|be authenticated so that people have some reason to trust its validity.


Authentication is essential, but I think there is more to this than
simply signing the price attribute. Also important is what we might
call payment method. That also needs to be an (authenticated)
component of the attribute. For example, suppose Provider X has
bilateral peering agreements with Providers Y and Z, and relationships
with brokers (sometimes called "clearinghouses") A and B. Then
it really wants to advertise a price of N, but only if the originator
is provider Y or Z or if the originator is authorized by broker A
or B. (In reality, the price N will likely vary for each option.)
A LS belonging to provider W could take that information, determine,
for example, that although it wasn't Y or Z it did belong to a
common broker, say B, and should thus interpret the call price
accordingly. If we want to get into this in more detail, I would,
of course, be happy to elaborate.

Stephen


____________________________________________________________________

Stephen Thomas                                       +1 404 872 4745
TransNexus, Chief Technical Officer    stephen.thomas@transnexus.com
____________________________________________________________________



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 20:10:26 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA07485
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 20:10:26 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C616A532A; Mon, 15 Mar 1999 12:53:10 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 30F925340; Mon, 15 Mar 1999 12:53:10 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Cc: "'hampton@cisco.com'" <hampton@cisco.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Message-ID: <36ED479D.23FD4FFF@lucent.com>
Date: Mon, 15 Mar 1999 11:47:09 -0600
From: "ronald h. davis" <ronaldd@lucent.com>
Organization: Lucent Technologies
X-Mailer: Mozilla 4.05 [en]C-EMS-1.4  (WinNT; U)
MIME-Version: 1.0
To: Alok Mittal <amittal@hss.hns.com>
Original-CC: "'hampton@cisco.com'" <hampton@cisco.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: Re: glp draft based on BGP
References: <01BE663A.B6B0A9C0@hss056.hss.hns.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Alok Mittal wrote:
> 
> 1. IT_SIG_PROTO should specify H.323 instead of Q.931/RAS -- a version
>     number should also be supported here.
>

i think that q.931 should stay in the document.  however, ras should
not as the 's' in 'ras' stands for status and not signalling.  instead
h.323 (or h.225) should be specified.


> 2. Some fields present in the SCSP based draft, such as lifetime and support
>     for range of numbers (for E.164) are useful -- Should we include those?
>

as i recall, the bgp-4 model does not use information refresh as is
done in intra-domain protocols like ospf.  instead, information is
only sent when there is an update.  as far as e.164 numbers goes
wouldn't that go into nlri?


> 3. One drawback that I find in this approach is that the policy for forwarding
>     route information is present at every TBGP speaker (section 7.3) and hence
>     when the information arrives at a particular speaker, it is already subjected
>     to a cascade of policies, which it doesn't even know. Support for exchange
>     of full database (as in SCSP based draft) seems to be useful, if the policy
>     leading to decision, should be applied independently in each domain. Of
>     course, changes are exchanged incrementally so the network wouldn't be flooded.
>

i'm not sure how much comparison should be made between bgp and scsp.
bgp is an inter-domain routing protocol, where scsp (which is a
derivative of ospf) is more suitable in an intra-domain routing context.

-- 
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
"if you ain't the lead dog, the scenery never changes"  -- edmund wilson

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 20:15:05 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA07545
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 20:15:05 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 57F285374; Wed, 17 Mar 1999 18:28:32 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3F29952C2; Wed, 17 Mar 1999 18:17:41 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8391FE2580A1D11198CB00A0C99845D467D193@tripacific.ttl.pactel.com>
From: "Lu, Henry" <hwlu@tri.sbc.com>
To: "'Hoefelmeyer, Ralph (MCI)'" <Ralph.Hoefelmeyer@mci.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: RE: H.323 black phone???
Date: Tue, 16 Mar 1999 10:56:09 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dear Ralph, 
Check with Net2phone, They are start to offer IP telephony services and
claim that 
you can dial from IP to any PSTN phone.

http://www.net2phone.com/english/

They are NJ based IDT sub company and they are promoting it.
Thanks.

Henry W. Lu
-----Original Message-----
From: Hoefelmeyer, Ralph (MCI) [mailto:Ralph.Hoefelmeyer@mci.com]
Sent: Friday, March 12, 1999 2:26 PM
To: IPTEL IETF Mailing List
Subject: H.323 black phone???
Importance: High

Hello, 
I've been searching the web, and have not found an H.323 black phone.
Everything is a PC card or a videophone. Some manufacturers claim they'll
have a unit in " a few months".
Does anyone know of an H.323 phone I can buy now? FYI, Mediatrix makes a SIP
phone ... I have seen it. 
Thanks, 
Ralph 
Ralph S. Hoefelmeyer 
MCI WorldCom Engineer 




---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 21:10:58 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA07977
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 21:10:58 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B4F4B5378; Wed, 17 Mar 1999 18:31:43 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DFD86534E; Wed, 17 Mar 1999 18:15:38 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: rtgupta@hss.hns.com
To: h323implementors@imtc.org, iptel@lists.research.bell-labs.com
Message-ID: <65256737.0029C486.00@sampark.hss.hns.com>
Date: Wed, 17 Mar 1999 15:15:17 +0530
Subject: h.245 channel establishment
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi All,
    This is regarding h.245 control channel establishment.

Specifications (see sections 6.2.8 and 8.2 of h.323 ) say that either
endpoint, involved in the call, will initiate setup of TCP connection for
h.245 signalling. But there has not been any explicit mention of type of
'call model' supported. So should I assume that in its absence,
irrespective of call model, TCP conection will be initiated by endpoint
ONLY and not by gatekeeper?

In case GK decides to route h.245 signalling, then two h.245 channels will
be required, one between calling endpoint and GK and other between GK and
called party. So can I assume safely that in this case, both calling party
and called party will  initiate a h.245 channel towards gatekeeper? But in
this case how will called endpoint get GK's h.245 control channel transport
address?(For calling end, that is conveyed in call proceeding or Alert or
Connect or Progress message)

Also while switching from tunneling to a seperate connection for h.245,
both endpoints will have to initiate TCP connection towards GK to have two
channels.

In any of above two cases,what if one endpoint sets up h.245 connection
towards GK but other endpoint does not initiate? Should GK initiate a
channel setup or just send a Facility message with a reason of 'starth245'?

Any experiences or ideas to be shared?

Thanks,
Ritu.

standards discuss this setup.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 21:20:14 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA08035
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 21:20:14 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5B05952B7; Mon, 15 Mar 1999 01:11:44 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A6C8152D5; Mon, 15 Mar 1999 01:11:43 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36ECA53F.7F8A06F9@research.bellcore.com>
Date: Mon, 15 Mar 1999 01:14:23 -0500
From: Christian Huitema <huitema@research.bellcore.com>
X-Mailer: Mozilla 4.04 [en] (Win95; U)
MIME-Version: 1.0
To: Scott Petrack <scott.petrack@metatel.com>
Cc: iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <Pine.LNX.4.04.9903111856190.6794-100000@petrack.metatel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Scott Petrack wrote:
> 
> The fact that this situation is a mess in the PSTN does not mean that it
> must be a mess in GLP. And the fact that the "scope" of routing
> numbers in the PSTN is the "provider ID" does not
> mean that this is the only scope that we should allow within GLP.
> Certainly, 1-700 numbers will absolutely require a scope which is a
> "provider id", but 800 numbers might not. For example, my company owns a
> certain 800 number, and it will resolve to our office telephone number no
> matter what provider we happen to choose to give us service. Nothing in
> GLP will require that the scope will be "provider id."

I think that this all go back to a two phases approach for telephony
routing:

1) take a dip in a mapping directory to understand the "provider" for
the 
   phone number, which will result in a <provider, number> pair -- maybe
   with a specific provider that says "e164"

2) route based on the pair.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 21:25:20 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA08065
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 21:25:19 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 50E1B5380; Wed, 17 Mar 1999 18:12:25 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F3459534F; Wed, 17 Mar 1999 17:47:08 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Cc: Alok Mittal <amittal@hss.hns.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Message-ID: <36ED752D.E64DE513@lucent.com>
Date: Mon, 15 Mar 1999 15:01:33 -0600
From: "ronald h. davis" <ronaldd@lucent.com>
Organization: Lucent Technologies
X-Mailer: Mozilla 4.05 [en]C-EMS-1.4  (WinNT; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Original-CC: Alok Mittal <amittal@hss.hns.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: Re: glp draft based on BGP
References: <01BE663A.B6B0A9C0@hss056.hss.hns.com> <36DEE85A.25189194@cisco.com> <36ED4C70.BB6CF1B@lucent.com> <36ED579F.D487D82A@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F Salama wrote:
> 
> > the thrust of my suggestion is to view the tbgp speaker as being an
> > appliation layer router.  depending upon network policies, the tbgp
> > speaker could send the equivalent of an icmp redirect message to ua
> > that could be cached for future calls.
> 
> I think the proposed query/reply mechanism is simpler. Redirecting the
> call involves more state setup/teardown/modification overhead.
> 

this is certainly a reasonable model if you view the tbgp speaker as
providing an address resolution function such as in the atmarp server
in classical ip over atm.  

pgrp was coming from the standpoint of endpoints providing trunk
connections to class 5 switches.  in this case, the query/reply 
mechanism requires full mesh connectivity among the endpoints in the 
packet network.

	[i am interested in hearing other views on the following
	 proposition.  i'm going out on a limb here; hope it
	 doesn't break! :-0 ]

full mesh connectivity would gas out in a large network.  using an
application layer routing scheme would support hierarchy in networks
and, thereby, make them more scaleable.  consequently, the view of
the tbgp speaker as providing a forwarding function at the application
layer.  this is also consistent with the gatekeeper routed call model
from h.323 (i think).

a side benefit is that you would eliminate the delays associated with 
executing a query/reply cycle before initiating a call.

-- 
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
"if you ain't the lead dog, the scenery never changes"  -- edmund wilson

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 22:34:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA10252
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 22:34:50 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 407825356; Wed, 17 Mar 1999 18:22:33 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 995BC535A; Wed, 17 Mar 1999 18:18:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "villuzzi" <vincent.illuzzi@mci.com>
To: "IPTEL IETF Mailing List" <iptel@lists.research.bell-labs.com>
Subject: Looking for large scale VoIP Equipment
Date: Tue, 16 Mar 1999 13:51:50 -0500
Message-ID: <000401be6fde$0cfa36a0$70182ca6@vai.mci201.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

I am looking to either migrate then replace or upgrade large scale PBX
equipment with VoIP capable gear. Any one have vendor information for large
scale VoIP capable equipment with feature rich PBX like capabilities?

Thanks,

Vince


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 22:39:52 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA10283
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 22:39:49 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D0135536C; Wed, 17 Mar 1999 18:13:42 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A8B765352; Wed, 17 Mar 1999 17:45:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <4110BC526A58D211B56C0000F6B9636DA85207@ws-exch01.dev.madge.com>
From: Chris Purvis <CPurvis@Madge.com>
To: "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>, h323implementors@imtc.org,
        iptel@lists.research.bell-labs.com
Subject: RE: h.245 channel establishment
Date: Wed, 17 Mar 1999 11:07:17 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain;
	charset="windows-1252"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Ritu,

1. I see no reason why a gatekeeper can't bring up a TCP connection for
H.245 on the basis of an address an endpoint has given it.  This would seem
to be perfectly normal behaviour in any model where the gatekeeper is
routing H.245.

2. It's perfectly possible for one "leg" of a call to be tunnelling H.245,
and the other to have a separate TCP connection.

3. Called endpoing may get GK's H.245 control channel transport address from
the (optional) h245address field in the setup message (if the GK chooses to
fill it in).

Does that give you enough options?

I actually think this is one of the few places where direct-routed vs
GK-routed is not a major issue.

As an aside, we shouldn't assume that TCP is the only right answer for
anything in the H.323 world, because the H.323 standards do not require that
they're being implemented over IP.

Regards,
Chris

> -----Original Message-----
> From: rtgupta@hss.hns.com [mailto:rtgupta@hss.hns.com]
> Sent: 17 March 1999 9:45
> To: h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> Subject: h.245 channel establishment
> 
> 
> Hi All,
>     This is regarding h.245 control channel establishment.
> 
> Specifications (see sections 6.2.8 and 8.2 of h.323 ) say that either
> endpoint, involved in the call, will initiate setup of TCP 
> connection for
> h.245 signalling. But there has not been any explicit mention 
> of type of
> 'call model' supported. So should I assume that in its absence,
> irrespective of call model, TCP conection will be initiated 
> by endpoint
> ONLY and not by gatekeeper?
> 
> In case GK decides to route h.245 signalling, then two h.245 
> channels will
> be required, one between calling endpoint and GK and other 
> between GK and
> called party. So can I assume safely that in this case, both 
> calling party
> and called party will  initiate a h.245 channel towards 
> gatekeeper? But in
> this case how will called endpoint get GK's h.245 control 
> channel transport
> address?(For calling end, that is conveyed in call proceeding 
> or Alert or
> Connect or Progress message)
> 
> Also while switching from tunneling to a seperate connection 
> for h.245,
> both endpoints will have to initiate TCP connection towards 
> GK to have two
> channels.
> 
> In any of above two cases,what if one endpoint sets up h.245 
> connection
> towards GK but other endpoint does not initiate? Should GK initiate a
> channel setup or just send a Facility message with a reason 
> of 'starth245'?
> 
> Any experiences or ideas to be shared?
> 
> Thanks,
> Ritu.
> 
> standards discuss this setup.
> 
> 
> --------------------------------------------------------------
> -----------
> for list subscription instructions,
> send email to h323implementors-request@imtc.org with text "info"
> 
 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 22:48:27 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA10340
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 22:48:26 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id ADF4252F8; Mon, 15 Mar 1999 08:15:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0C35E530C; Mon, 15 Mar 1999 08:15:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36ED0769.E688B426@telcel.net.ve>
Date: Mon, 15 Mar 1999 09:13:13 -0400
From: Omar Ferrer <oferrer@telcel.net.ve>
Organization: TELCEL Celular, C.A.
X-Mailer: Mozilla 4.5 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
To: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>
Cc: Henri Setiawan <henry@risti.telkom.co.id>,
        H323 Implementors <h323implementors@imtc.org>,
        IPTEL <iptel@lists.research.bell-labs.com>
Subject: Re: [Fwd: Parameter for interoperability test]
References: <E32BD2062DAFD1119CFD0000F804F31702C0487F@cosexch003.mcit.com>
Content-Type: multipart/alternative;
 boundary="------------491F3C4C96653791CD758EA2"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


--------------491F3C4C96653791CD758EA2
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Thank you for your answer Mr. Hoefelmeyer:

As can be seen, we do not have knowledge of this testing. Even having an
ISP, VoIP serrvices are completely new because of the regulatory
enviroment that we are facing now.

Most of our equipments (Access Servers/5200 - 5300) are from Cisco, but
Lucent is offering to us a complete solution to replace our entire
platform. So, we need to test products that can operate with those.
Based on that, we are planning to call Cisco, Lucent, Ericsson, Siemens
and 3Com for Gateways and Gatekeepers.

I will very appreciated is you help me on build a testing project that
include H323 interoperability, QoS, Router configurations, billing
software, link with a Global Network, etc.

Thanks again Mr. Hoefelmeyer!!!
OF.

"Hoefelmeyer, Ralph (MCI)" wrote:

>
>
> Omar (& Henri),
>
> The first thing you need to test gatekeeper/gateway interoperability
> is a LAN protocol analyzer, preferably one that will decode H.323.
> Hewlett-Packard has one in Beta, it works well.
>
>
> Cisco has a product called Netflow for it's products, which allows one
> to measure various attributes of the routers' performance (note that
> Cisco's gatekeeper and gateway products use a version of their IOS).
> If you use Cisco, this is a possibility.
>
> Various Measures of Merit (MOM) could be:
>
> - time of call initiation to call setup
> - interoperability of H.323 products; verify claims of H.323 v2
> compatibility between products from different firms.  This has been an
> important issue.
>
> - the ease of configuration of the gatekeepers for number translation
> (provisioning), if you are not accessing an existing number
> translation mechanism
>
> - the ability to hook into the gatekeeper for your outside number
> translation setup (API access to the gatekeeper)
>
> I would suggest you use an Ethernet (or Fast Ethernet)  LAN between
> the gateways and gatekeeper.  Plug all of the devices into one hub,
> and it keeps things simple, for your first try.   This is a simpler
> scope than what you propose below, but interoperability is a key issue
> before moving to bigger ideas.
>
> QoS is a complex question, dependent on the target network(s).  Router
> configurations, network topologies and firewalls all play a role here.
>
> I would be interested in what you think, and what results you come up
> with, if you can share this with us.
>
> Good Luck!
>
> Ralph
>
> Ralph S. Hoefelmeyer
> MCI WorldCom Engineer
>
>      -----Original Message-----
>      From:   Omar Ferrer [SMTP:oferrer@telcel.net.ve]
>      Sent:   Friday, March 12, 1999 6:20 AM
>      To:     Henri Setiawan; H323 Implementors; IPTEL
>      Subject:        [Fwd: Parameter for interoperability test]
>
>      Dear Mr. Setiawan:
>
>      I've recently subscribe to this newsgroup asking for some help on
>      testing VoIP products. I'm beginning a project in which we want
>      to test Gateways, Gatekeepers and Network Managers, but we do not
>      have experience with that because this field is completely new
>      for the Latin America market.
>
>      Do you now what I need to measure or test on each equipment and
>      how?
>
>      We are beginning a project for VoIP. In the first phase we want
>      to test each equipment on a point to point link (TDM) with a
>      higher bandwith and then decreasing as the traffic increase. In
>      the second phase, we want to test the same products but with
>      direct connection to the Internet. Here we need to prove some QoS
>      tools and bandwith management techniques, and also, some testing
>      with a Global Network for call terminations worldwide.
>
>      Could you give some approach for this?
>
>      Thanks!!!
>      Omar Ferrer
>      Caracas - Venezuela
>
>      -------- Original Message --------
>      Subject:         Parameter for interoperability test
>      Date:    Fri, 12 Mar 1999 13:46:55 +0700
>      From:    "Henri Setiawan" <henry@risti.telkom.co.id>
>      To:      <h323implementors@imtc.org>
>      CC:      <iptel@lists.research.bell-labs.com>
>
>       Dear folks, We like to study about interoperability test for
>      gateway and gateway keeper, do anyone know what kind of
>      parameters that we should test for this testing ? Thanks Henri
>      SR&D of PT Telkom Indonesia.
>

--------------491F3C4C96653791CD758EA2
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Thank you for your answer Mr. Hoefelmeyer:
<p>As can be seen, we do not have knowledge of this testing. Even having
an ISP, VoIP serrvices are completely new because of the regulatory enviroment
that we are facing now.
<p>Most of our equipments (Access Servers/5200 - 5300) are from Cisco,
but Lucent is offering to us a complete solution to replace our entire
platform. So, we need to test products that can operate with those. Based
on that, we are planning to call Cisco, Lucent, Ericsson, Siemens and 3Com
for Gateways and Gatekeepers.
<p>I will very appreciated is you help me on build a testing project that
include H323 interoperability, QoS, Router configurations, billing software,
link with a Global Network, etc.
<p>Thanks again Mr. Hoefelmeyer!!!
<br>OF.
<p>"Hoefelmeyer, Ralph (MCI)" wrote:
<blockquote TYPE=CITE>&nbsp;
<p><font face="Arial"><font color="#0000FF">Omar (&amp; Henri),</font></font>
<p><font face="Arial"><font color="#0000FF">The first thing you need to
test gatekeeper/gateway interoperability is a LAN protocol analyzer, preferably
one that will decode H.323.&nbsp; Hewlett-Packard has one in Beta, it works
well.</font></font>
<br>&nbsp;
<p><font face="Arial"><font color="#0000FF">Cisco has a product called
Netflow for it's products, which allows one to measure various attributes
of the routers' performance (note that Cisco's gatekeeper and gateway products
use a version of their IOS).&nbsp; If you use Cisco, this is a possibility.</font></font>
<p><font face="Arial"><font color="#0000FF">Various Measures of Merit (MOM)
could be:</font></font>
<p><font face="Arial"><font color="#0000FF">- time of call initiation to
call setup</font></font>
<br><font face="Arial"><font color="#0000FF">- interoperability of H.323
products; verify claims of H.323 v2 compatibility between products from
different firms.&nbsp; This has been an important issue.</font></font>
<p><font face="Arial"><font color="#0000FF">- the ease of configuration
of the gatekeepers for number translation (provisioning), if you are not
accessing an existing number translation mechanism</font></font>
<p><font face="Arial"><font color="#0000FF">- the ability to hook into
the gatekeeper for your outside number translation setup (API access to
the gatekeeper)</font></font>
<p><font face="Arial"><font color="#0000FF">I would suggest you use an
Ethernet (or Fast Ethernet)&nbsp; LAN between the gateways and gatekeeper.&nbsp;
Plug all of the devices into one hub, and it keeps things simple, for your
first try.&nbsp;&nbsp; This is a simpler scope than what you propose below,
but interoperability is a key issue before moving to bigger ideas.</font></font>
<p><font face="Arial"><font color="#0000FF">QoS is a complex question,
dependent on the target network(s).&nbsp; Router configurations, network
topologies and firewalls all play a role here.</font></font>
<p><font face="Arial"><font color="#0000FF">I would be interested in what
you think, and what results you come up with, if you can share this with
us.</font></font>
<p><font face="Arial"><font color="#0000FF">Good Luck!</font></font>
<p><font face="Arial"><font color="#0000FF">Ralph</font></font>
<p><font face="Arial"><font color="#0000FF">Ralph S. Hoefelmeyer</font></font>
<br><font face="Arial"><font color="#0000FF">MCI WorldCom Engineer</font></font>
<ul><font face="Arial"><font size=-2>-----Original Message-----</font></font>
<br><b><font face="Arial"><font size=-2>From:&nbsp;&nbsp;</font></font></b>
<font face="Arial"><font size=-2>Omar Ferrer [SMTP:oferrer@telcel.net.ve]</font></font>
<br><b><font face="Arial"><font size=-2>Sent:&nbsp;&nbsp;</font></font></b>
<font face="Arial"><font size=-2>Friday, March 12, 1999 6:20 AM</font></font>
<br><b><font face="Arial"><font size=-2>To:&nbsp;&nbsp;&nbsp;&nbsp;</font></font></b>
<font face="Arial"><font size=-2>Henri Setiawan; H323 Implementors; IPTEL</font></font>
<br><b><font face="Arial"><font size=-2>Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</font></font></b>
<font face="Arial"><font size=-2>[Fwd: Parameter for interoperability test]</font></font>
<p><font face="Arial"><font color="#000000">Dear Mr. Setiawan:</font></font>
<p><font face="Arial"><font color="#000000">I've recently subscribe to
this newsgroup asking for some help on testing VoIP products. I'm beginning
a project in which we want to test Gateways, Gatekeepers and Network Managers,
but we do not have experience with that because this field is completely
new for the Latin America market.</font></font>
<p><font face="Arial"><font color="#000000">Do you now what I need to measure
or test on each equipment and how?</font></font>
<p><font face="Arial"><font color="#000000">We are beginning a project
for VoIP. In the first phase we want to test each equipment on a point
to point link (TDM) with a higher bandwith and then decreasing as the traffic
increase. In the second phase, we want to test the same products but with
direct connection to the Internet. Here we need to prove some QoS tools
and bandwith management techniques, and also, some testing with a Global
Network for call terminations worldwide.</font></font>
<p><font face="Arial"><font color="#000000">Could you give some approach
for this?</font></font>
<p><font face="Arial"><font color="#000000">Thanks!!!</font></font>
<br><font face="Arial"><font color="#000000">Omar Ferrer</font></font>
<br><font face="Arial"><font color="#000000">Caracas - Venezuela</font></font>
<p><font face="Arial"><font color="#000000">-------- Original Message --------</font></font>
<br><b><font face="Arial">Subject:&nbsp;</font></b>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<font face="Arial">
Parameter for interoperability test</font>
<br><b><font face="Arial">Date:&nbsp;</font></b>&nbsp;&nbsp;<font face="Arial">
Fri, 12 Mar 1999 13:46:55 +0700</font>
<br><b><font face="Arial">From:&nbsp;</font></b>&nbsp;&nbsp;<font face="Arial">
"Henri Setiawan" &lt;henry@risti.telkom.co.id></font>
<br><b><font face="Arial">To:&nbsp;</font></b>&nbsp;&nbsp;&nbsp;&nbsp;<font face="Arial">
&lt;h323implementors@imtc.org></font>
<br><b><font face="Arial">CC:&nbsp;</font></b>&nbsp;&nbsp;&nbsp;&nbsp;<font face="Arial">
&lt;iptel@lists.research.bell-labs.com></font>
<p><font face="Arial"><font color="#000000">&nbsp;<font size=-1>Dear folks,</font>
<font size=-1>We like to study about interoperability test for gateway
and gateway keeper, do anyone know what kind of parameters that we should
test for this testing ?</font> <font size=-1>Thanks</font> <font size=-1>Henri
SR&amp;D of PT Telkom Indonesia.</font></font></font></ul>
</blockquote>
</html>

--------------491F3C4C96653791CD758EA2--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 23:32:39 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA10932
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 23:32:37 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 4BC3953A3; Wed, 17 Mar 1999 22:30:26 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A570053A4; Wed, 17 Mar 1999 22:30:23 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <BDF8B3934BA6D211920400A0C94B11B60E853A@zuk15exch1.corp.isg.mot.com>
From: Grandclaude Sylvain-LSG012 <LSG012@email.mot.com>
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>,
        "'villuzzi'" <vincent.illuzzi@mci.com>
Subject: RE: Looking for large scale VoIP Equipment
Date: Thu, 18 Mar 1999 02:44:35 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2417.0)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi


I am a bit sad that you will, probably, not find something satisfying your
needs AT THE PRESENT TIME.

NOT ANY vendor has got large VoIP gear ready. Many do claim, but ???
You can still try with 3COM or NORTEL, but I am dubious they really have
something ready and working - well - NOW.
Start-Up companies likewise SONUS Networks and the likes have great
potential - but no working products yet as far as I am aware.
You will have to be patient !

Then, PABX-like "intelligence" is something else. 

This is NOT something any H.323-compliant GATEWAY is capable of providing on
its own.
This is NOT something any H.323-compliant GATEKEEPER is capable of providing
on its own.
This is something a GATEKEEPER might help to, if providing an information
feed to a 3rd-Party CTI SERVER S/W application.

Telephone sets + gateway + gatekeeper + complementary CTI server S/W
altogether may look somewhat equivalent to a PABX.
You need all the pieces in that puzzle.

And you'd better choose for FULLY standards-compliant pieces !



Hope this helps,


		Sylvain Grandclaude.

		Application Engineer Motorola ING EMEA
		tel 	+33 1 49 69 85 80
		fax 	+33 1 46 64 43 00
		mobile +33 6 07 25 78 70
		e-mail	LSG012@email.mot.com <mailto:LSG012@email.mot.com> 

		The opinions expressed above represent me only as an
individual and don't represent in any extent Motorola Company strategy or
commitments.


	----------
	From:  villuzzi [SMTP:vincent.illuzzi@mci.com]
	Sent:  mardi 16 mars 1999 19:52
	To:  IPTEL IETF Mailing List
	Subject:  Looking for large scale VoIP Equipment

	I am looking to either migrate then replace or upgrade large scale
PBX
	equipment with VoIP capable gear. Any one have vendor information
for large
	scale VoIP capable equipment with feature rich PBX like
capabilities?

	Thanks,

	Vince


	---------
	This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 23:46:29 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA11071
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 23:46:28 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 815AD52CF; Wed, 17 Mar 1999 22:49:46 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7B59852D1; Wed, 17 Mar 1999 22:49:44 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <BDF8B3934BA6D211920400A0C94B11B60E853E@zuk15exch1.corp.isg.mot.com>
From: Grandclaude Sylvain-LSG012 <LSG012@email.mot.com>
To: "'Hoefelmeyer, Ralph (MCI)'" <Ralph.Hoefelmeyer@mci.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>,
        "'Lu, Henry'" <hwlu@tri.sbc.com>
Subject: RE: H.323 black phone???
Date: Thu, 18 Mar 1999 03:48:22 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2417.0)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hello

Have you tried CISCO / SELSIUS ?

SELSIUS has an IP-based PABX and H.323 IP / Ethernet phones with NO
videoconferencing capability (no costly TFT color screen).
I am NOT, still, quite sure than it can work on its own. (It is a DHCP
client, and I think it is a LDAP client as well for phone directory - this
is quite a clever implementation but this is also MORE than a simple,
standard H.323 EP !).

Also, it is supposed to be fed by 48V DC power from unused wires in the
10BaseT RJ45 outlet.



http://www.selsius.com/Products/selsius_phone.htm
<http://www.selsius.com/Products/selsius_phone.htm> 


Hope this helps.


		Sylvain Grandclaude.

		Application Engineer Motorola ING EMEA
		tel 	+33 1 49 69 85 80
		fax 	+33 1 46 64 43 00
		mobile +33 6 07 25 78 70
		e-mail	LSG012@email.mot.com


	----------
	From:  Lu, Henry [SMTP:hwlu@tri.sbc.com]
	Sent:  mardi 16 mars 1999 17:56
	To:  'Hoefelmeyer, Ralph (MCI)'; IPTEL IETF Mailing List
	Subject:  RE: H.323 black phone???

	Dear Ralph, 
	Check with Net2phone, They are start to offer IP telephony services
and
	claim that 
	you can dial from IP to any PSTN phone.
	http://www.net2phone.com/english/

	They are NJ based IDT sub company and they are promoting it.
	Thanks.

	Henry W. Lu
	-----Original Message-----
	From: Hoefelmeyer, Ralph (MCI) [mailto:Ralph.Hoefelmeyer@mci.com]
	Sent: Friday, March 12, 1999 2:26 PM
	To: IPTEL IETF Mailing List
	Subject: H.323 black phone???
	Importance: High

	Hello, 
	I've been searching the web, and have not found an H.323 black
phone.
	Everything is a PC card or a videophone. Some manufacturers claim
they'll
	have a unit in " a few months".
	Does anyone know of an H.323 phone I can buy now? FYI, Mediatrix
makes a SIP
	phone ... I have seen it. 
	Thanks, 
	Ralph 
	Ralph S. Hoefelmeyer 
	MCI WorldCom Engineer 




	---------
	This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 23:49:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA11108
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 23:49:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B71E95341; Mon, 15 Mar 1999 13:13:34 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1C2D75343; Mon, 15 Mar 1999 13:13:34 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Cc: Alok Mittal <amittal@hss.hns.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Message-ID: <36ED4C70.BB6CF1B@lucent.com>
Date: Mon, 15 Mar 1999 12:07:44 -0600
From: "ronald h. davis" <ronaldd@lucent.com>
Organization: Lucent Technologies
X-Mailer: Mozilla 4.05 [en]C-EMS-1.4  (WinNT; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Original-CC: Alok Mittal <amittal@hss.hns.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: Re: glp draft based on BGP
References: <01BE663A.B6B0A9C0@hss056.hss.hns.com> <36DEE85A.25189194@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> An administrative domain, AD1, may wish that all incoming H.323 calls
> from neighboring domains be routing by its gatekeeper first. The
> gatekeeper then directs the incoming call to the appropriate gateway or
> proxy. In this case, AD1, specifies the address of a gatekeeper as the
> next hop and RAS as the signaling protocol. Alternatively, AD1, may
> specify an H.323 gateway or proxy as the next hop and Q.931 as the call
> signaling protocol. In the first case the gatekeeper is involved in both
> call routing and admission control. In the second case, the gatekeeper
> does admission control only.
> 

h.323 signalling does not occur over the ras channel, but anyway, my
question has to do with the overhead of querying ts1 first and then
sending the setup message (i am referencing figure 1 of your draft).
is it valid for ua to send the setup message (this assumes that it
knows the protocol that it is using) to ts1 straight on, with ts1
using its routing information to determine the next hop for the
message?

the thrust of my suggestion is to view the tbgp speaker as being an
appliation layer router.  depending upon network policies, the tbgp
speaker could send the equivalent of an icmp redirect message to ua
that could be cached for future calls.

-- 
  __  ______  __  / __/ |       lucent technologies, naperville il, usa
_/ (_(_) / (_(_/_/_(_/  .       ronald.h.davis@lucent.com
"if you ain't the lead dog, the scenery never changes"  -- edmund wilson

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 17 23:51:39 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA11131
	for <iptel-archive@odin.ietf.org>; Wed, 17 Mar 1999 23:51:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D7C685345; Mon, 15 Mar 1999 13:41:33 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 44E395348; Mon, 15 Mar 1999 13:41:33 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <1371C1A3868BD21190BE0008C71EA5F3235E7F@MCHH9EKA>
From: Klaghofer Karl ICN IB NL IP 7 <Karl.Klaghofer@icn.siemens.de>
To: "'Hoefelmeyer, Ralph (MCI)'" <Ralph.Hoefelmeyer@mci.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: AW: H.323 black phone???
Date: Mon, 15 Mar 1999 19:40:23 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by ietf.org id XAA11131

Ralph,

You can see an H.323 IP Phone from Siemens at CeBIT ´99 in Hannover
(Germany) in two weeks.

Regards,
Karl Klaghofer
Siemens AG


> -----Ursprüngliche Nachricht-----
> Von:	Hoefelmeyer, Ralph (MCI) [SMTP:Ralph.Hoefelmeyer@mci.com]
> Gesendet am:	Freitag, 12. März 1999 23:26
> An:	IPTEL IETF Mailing List
> Betreff:	H.323 black phone???
> Wichtigkeit:	Hoch
> 
> Hello, 
> 
> I've been searching the web, and have not found an H.323 black phone.
> Everything is a PC card or a videophone.  Some manufacturers claim they'll
> have a unit in " a few months".
> 
> Does anyone know of an H.323 phone I can buy now?  FYI, Mediatrix makes a
> SIP phone ... I have seen it. 
> 
> Thanks, 
> Ralph 
> 
> Ralph S. Hoefelmeyer 
> MCI WorldCom Engineer 
> 
> 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 00:14:48 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA11280
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 00:14:48 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0266652C1; Sun, 14 Mar 1999 22:05:36 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3C03452BF; Sun, 14 Mar 1999 22:05:34 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36EC7808.B74EAC13@cisco.com>
Date: Sun, 14 Mar 1999 19:01:28 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990311212111.00ab7d80@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Matt Squire wrote:
> 
> 4) Capacity.
> 
> I think we have consensus that a capacity attribute is needed.  What we
> mean by capacity, unfortunately, is still up in the air.
> 
>   a) Should capacity be unitless and relative, or measurable and
>      absolute?
>   b) If measurable/absolute, what are the appropriate metric(s).  Some
>      possibilities are bandwidth, maximum circuits, load, (others?)...
>   c) What are the timescales that capacity can be changed?  In other
>      words, does it represent some maximum capacity (constant) or
>      current capacity (depends on load).
>   d) Is capacity an attribute of an egress gateway, or to the signalling
>      path, both, or do we need two different capacity measurements (one
>      for path, another for egress gw)?
> 
> The definition and use of capacity will most likely differ depending upon
> your preference for the SCSP or BGP approach.  With the SCSP approach, I
> suggested a unitless, relative capacity which represents the capacity of
> the egress gateway.  I thought that we had come to such a consensus during
> Jonathan's presentation at the last wg meeting.
> 
> With the SCSP approach, this makes more sense because the its very
> difficult to infer anything about the capacity of the IP path because of
> the explicit separation of the call routing from the IP routing.  With the
> SCSP approach, an LS that aggregates or changes the next hop information
> could modify the capacity to include the capacity of the path between the
> LS and its next-hop.  The value of such a modification is limited because
> it does not reflect the capacity of the data path, which is much more
> likely to be the bottleneck.
> 
> With the BGP approach, carrying e164 numbers in BGP integrates data and
> call routing, and thus capacity information about the IP path can be
> integrated into the capacity metric.  In the current TBGP draft, it is
> considered an open issue whether TBGP is a simple extension to BGP (and
> thus yields this integrated routing) or whether TBGP is a different
> instance of the protocol.  I believe that if TBGP is a different instance
> of the protocol (and is separated from the IP routing information), then it
> will not be able yield an accurate reflection of path capacity.

Depends on how similar the IP ASes topology and TBGP Administrative
Domains (AD) topology are.

The capacity metric could also be used to enforce policy, in addition to
its use for advertising physical resources (bandwidth, ports, ... etc).
For example, it can be used for constraining the number of calls from
one AD to another. For example, AD1 may wish not to route more than x
calls from AD2 towards +1408*. This can be achieved by setting capacity
= x in advertisement of 1408* frowarded from AD1 to Ad2. 

It may be worthwhile to define two metrics for capacity, one for
physical resources and another one for policy.

Hussein


> 
> A unitless measurement has a disadvantage in that it is an administrative
> manner to ensure that the weight of a gateway reflects is capacity relative
> to other gateways.  In an inter-domain protocol, this requires some
> administrative coordination across domains.  A measurable capacity would be
> nice, but it seems that no single metric would suffice, and if we permit
> multiple metrics, then we're faced with the problem of determining which
> capacities are larger than others (ie is 20 circuits better than 100 Mbs or
> not).
> 
> In either case, I'd recommend that the capacity attribute NOT reflect
> anything about the real-time load on a gateway or path.  Due to the
> complications of multiple metrics, I'd also recommend that capacity remain
> unitless and that it be an administrative exercise for the capacity to be
> consistent across the confederation of iptel domains.  Hence, I'd propose a
> simple 4-octet field that can be used do determine relative capacity.
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 00:17:08 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA11295
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 00:17:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C520A535A; Wed, 17 Mar 1999 18:28:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3D3CF5355; Wed, 17 Mar 1999 18:16:24 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <004c01be6f47$23b3ce20$741f0e0a@Jlaf005.telkom.go.id>
From: "Henri Setiawan" <henry@risti.telkom.co.id>
To: "Omar Ferrer" <oferrer@telcel.net.ve>,
        "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>
Cc: "H323 Implementors" <h323implementors@imtc.org>,
        "IPTEL" <iptel@lists.research.bell-labs.com>
Subject: Re: [Fwd: Parameter for interoperability test]
Date: Tue, 16 Mar 1999 07:51:28 +0700
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_0047_01BE6F81.CC822C60"
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.1
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.

------=_NextPart_000_0047_01BE6F81.CC822C60
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hi Folks,

Thanks for the comments. Regarding what Omar said, we also have such =
kind of difficulty, but not only cause the test equipments it's more =
about the parameters for testing.

What we plan to use for evaluating each products is:
1. We measure the QoS of the Internet/intranet network to get the value =
of throughtput, delay, jitter, performance etc.
2. After we get those information, we compare with our general =
requirement document. If fulfill with our provision, we'll install the =
equipments directly to internet/intranet backbone to limit the node =
(cause node make more delay).
3. When the equipments already install we will test it as an acceptance =
test prosedure with the parameters that already signed together from our =
side and vendor.
4. The last thing is we test it ourself using H.323 parameter from ITU =
standard body.

From those testing, we make comparation from each products and make =
evaluation for the business case, cause we think that not only the =
technology that we should understand but the business factor we should =
make to accommodate techno-economic analysis.

If anyone have already doing the testing, please share with us ? Did =
anyone have business case analysis for IPTel already ? Would you kindly =
share with me ?

Thanks

Henri Setiawan
Access Network Engineer
RisTI - Telkom Indonesia


  -----Original Message-----
  From: Omar Ferrer <oferrer@telcel.net.ve>
  To: Hoefelmeyer, Ralph (MCI) <Ralph.Hoefelmeyer@mci.com>
  Cc: Henri Setiawan <henry@risti.telkom.co.id>; H323 Implementors =
<h323implementors@imtc.org>; IPTEL <iptel@lists.research.bell-labs.com>
  Date: Tuesday, March 16, 1999 1009
  Subject: Re: [Fwd: Parameter for interoperability test]


  Thank you for your answer Mr. Hoefelmeyer:=20
  As can be seen, we do not have knowledge of this testing. Even having =
an ISP, VoIP serrvices are completely new because of the regulatory =
enviroment that we are facing now.=20

  Most of our equipments (Access Servers/5200 - 5300) are from Cisco, =
but Lucent is offering to us a complete solution to replace our entire =
platform. So, we need to test products that can operate with those. =
Based on that, we are planning to call Cisco, Lucent, Ericsson, Siemens =
and 3Com for Gateways and Gatekeepers.=20

  I will very appreciated is you help me on build a testing project that =
include H323 interoperability, QoS, Router configurations, billing =
software, link with a Global Network, etc.=20

  Thanks again Mr. Hoefelmeyer!!!=20
  OF.=20

  "Hoefelmeyer, Ralph (MCI)" wrote:=20

     =20
    Omar (& Henri),=20

    The first thing you need to test gatekeeper/gateway interoperability =
is a LAN protocol analyzer, preferably one that will decode H.323.  =
Hewlett-Packard has one in Beta, it works well.=20
     =20

    Cisco has a product called Netflow for it's products, which allows =
one to measure various attributes of the routers' performance (note that =
Cisco's gatekeeper and gateway products use a version of their IOS).  If =
you use Cisco, this is a possibility.=20

    Various Measures of Merit (MOM) could be:=20

    - time of call initiation to call setup=20
    - interoperability of H.323 products; verify claims of H.323 v2 =
compatibility between products from different firms.  This has been an =
important issue.=20

    - the ease of configuration of the gatekeepers for number =
translation (provisioning), if you are not accessing an existing number =
translation mechanism=20

    - the ability to hook into the gatekeeper for your outside number =
translation setup (API access to the gatekeeper)=20

    I would suggest you use an Ethernet (or Fast Ethernet)  LAN between =
the gateways and gatekeeper.  Plug all of the devices into one hub, and =
it keeps things simple, for your first try.   This is a simpler scope =
than what you propose below, but interoperability is a key issue before =
moving to bigger ideas.=20

    QoS is a complex question, dependent on the target network(s).  =
Router configurations, network topologies and firewalls all play a role =
here.=20

    I would be interested in what you think, and what results you come =
up with, if you can share this with us.=20

    Good Luck!=20

    Ralph=20

    Ralph S. Hoefelmeyer=20
    MCI WorldCom Engineer=20

      -----Original Message-----=20
      From:   Omar Ferrer [SMTP:oferrer@telcel.net.ve]=20
      Sent:   Friday, March 12, 1999 6:20 AM=20
      To:     Henri Setiawan; H323 Implementors; IPTEL=20
      Subject:        [Fwd: Parameter for interoperability test]=20
      Dear Mr. Setiawan:=20

      I've recently subscribe to this newsgroup asking for some help on =
testing VoIP products. I'm beginning a project in which we want to test =
Gateways, Gatekeepers and Network Managers, but we do not have =
experience with that because this field is completely new for the Latin =
America market.=20

      Do you now what I need to measure or test on each equipment and =
how?=20

      We are beginning a project for VoIP. In the first phase we want to =
test each equipment on a point to point link (TDM) with a higher =
bandwith and then decreasing as the traffic increase. In the second =
phase, we want to test the same products but with direct connection to =
the Internet. Here we need to prove some QoS tools and bandwith =
management techniques, and also, some testing with a Global Network for =
call terminations worldwide.=20

      Could you give some approach for this?=20

      Thanks!!!=20
      Omar Ferrer=20
      Caracas - Venezuela=20

      -------- Original Message --------=20
      Subject:         Parameter for interoperability test=20
      Date:    Fri, 12 Mar 1999 13:46:55 +0700=20
      From:    "Henri Setiawan" <henry@risti.telkom.co.id>=20
      To:      <h323implementors@imtc.org>=20
      CC:      <iptel@lists.research.bell-labs.com>=20

       Dear folks, We like to study about interoperability test for =
gateway and gateway keeper, do anyone know what kind of parameters that =
we should test for this testing ? Thanks Henri SR&D of PT Telkom =
Indonesia.


------=_NextPart_000_0047_01BE6F81.CC822C60
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">
<HTML><HEAD>
<META content=3Dtext/html;charset=3Diso-8859-1 =
http-equiv=3DContent-Type><!doctype html public "-//w3c//dtd html 4.0 =
transitional//en">
<META content=3D'"MSHTML 5.00.0910.1309"' name=3DGENERATOR></HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT size=3D2>Hi Folks,</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>Thanks for the comments. Regarding what Omar said, =
we also=20
have such kind of difficulty, but not only cause the test equipments =
it's more=20
about the parameters for testing.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>What we plan to use for evaluating each products=20
is:</FONT></DIV>
<DIV><FONT size=3D2>1. We measure the QoS of the Internet/intranet =
network to get=20
the value of throughtput, delay, jitter, performance etc.</FONT></DIV>
<DIV><FONT size=3D2>2. After we get those information, we compare with =
our general=20
requirement document. If fulfill with our provision, we'll install the=20
equipments directly to internet/intranet backbone to limit the node =
(cause node=20
make more delay).</FONT></DIV>
<DIV><FONT size=3D2>3. When the equipments already install we will test =
it as an=20
acceptance test prosedure with the parameters that already signed =
together from=20
our side and vendor.</FONT></DIV>
<DIV><FONT size=3D2>4. The last thing is we test it ourself using H.323 =
parameter=20
from ITU standard body.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>From those testing, we make comparation from each =
products and=20
make evaluation for the business case, cause we think that not only the=20
technology that we should understand but the business factor we should =
make to=20
accommodate techno-economic analysis.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>If anyone have already doing the testing, please =
share with us=20
? Did anyone have business case analysis for IPTel already ? Would you =
kindly=20
share with me ?</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>Thanks</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>Henri Setiawan</FONT></DIV>
<DIV><FONT size=3D2>Access Network Engineer</FONT></DIV>
<DIV><FONT size=3D2>RisTI - Telkom Indonesia</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV>&nbsp;</DIV>
<BLOCKQUOTE=20
style=3D"BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: =
0px; PADDING-LEFT: 5px">
  <DIV><FONT face=3DArial size=3D2><B>-----Original =
Message-----</B><BR><B>From:=20
  </B>Omar Ferrer &lt;<A=20
  =
href=3D"mailto:oferrer@telcel.net.ve">oferrer@telcel.net.ve</A>&gt;<BR><B=
>To:=20
  </B>Hoefelmeyer, Ralph (MCI) &lt;<A=20
  =
href=3D"mailto:Ralph.Hoefelmeyer@mci.com">Ralph.Hoefelmeyer@mci.com</A>&g=
t;<BR><B>Cc:=20
  </B>Henri Setiawan &lt;<A=20
  =
href=3D"mailto:henry@risti.telkom.co.id">henry@risti.telkom.co.id</A>&gt;=
; H323=20
  Implementors &lt;<A=20
  =
href=3D"mailto:h323implementors@imtc.org">h323implementors@imtc.org</A>&g=
t;;=20
  IPTEL &lt;<A=20
  =
href=3D"mailto:iptel@lists.research.bell-labs.com">iptel@lists.research.b=
ell-labs.com</A>&gt;<BR><B>Date:=20
  </B>Tuesday, March 16, 1999 1009<BR><B>Subject: </B>Re: [Fwd: =
Parameter for=20
  interoperability test]<BR><BR></DIV></FONT>Thank you for your answer =
Mr.=20
  Hoefelmeyer:=20
  <P>As can be seen, we do not have knowledge of this testing. Even =
having an=20
  ISP, VoIP serrvices are completely new because of the regulatory =
enviroment=20
  that we are facing now.=20
  <P>Most of our equipments (Access Servers/5200 - 5300) are from Cisco, =
but=20
  Lucent is offering to us a complete solution to replace our entire =
platform.=20
  So, we need to test products that can operate with those. Based on =
that, we=20
  are planning to call Cisco, Lucent, Ericsson, Siemens and 3Com for =
Gateways=20
  and Gatekeepers.=20
  <P>I will very appreciated is you help me on build a testing project =
that=20
  include H323 interoperability, QoS, Router configurations, billing =
software,=20
  link with a Global Network, etc.=20
  <P>Thanks again Mr. Hoefelmeyer!!! <BR>OF.=20
  <P>&quot;Hoefelmeyer, Ralph (MCI)&quot; wrote:=20
  <BLOCKQUOTE TYPE =3D CITE>&nbsp;=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>Omar (&amp; =
Henri),</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>The first thing you need =
to test=20
    gatekeeper/gateway interoperability is a LAN protocol analyzer, =
preferably=20
    one that will decode H.323.&nbsp; Hewlett-Packard has one in Beta, =
it works=20
    well.</FONT></FONT> <BR>&nbsp;=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>Cisco has a product =
called Netflow=20
    for it's products, which allows one to measure various attributes of =
the=20
    routers' performance (note that Cisco's gatekeeper and gateway =
products use=20
    a version of their IOS).&nbsp; If you use Cisco, this is a=20
    possibility.</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>Various Measures of =
Merit (MOM)=20
    could be:</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>- time of call =
initiation to call=20
    setup</FONT></FONT> <BR><FONT face=3DArial><FONT color=3D#0000ff>-=20
    interoperability of H.323 products; verify claims of H.323 v2 =
compatibility=20
    between products from different firms.&nbsp; This has been an =
important=20
    issue.</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>- the ease of =
configuration of the=20
    gatekeepers for number translation (provisioning), if you are not =
accessing=20
    an existing number translation mechanism</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>- the ability to hook =
into the=20
    gatekeeper for your outside number translation setup (API access to =
the=20
    gatekeeper)</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>I would suggest you use =
an Ethernet=20
    (or Fast Ethernet)&nbsp; LAN between the gateways and =
gatekeeper.&nbsp; Plug=20
    all of the devices into one hub, and it keeps things simple, for =
your first=20
    try.&nbsp;&nbsp; This is a simpler scope than what you propose =
below, but=20
    interoperability is a key issue before moving to bigger =
ideas.</FONT></FONT>=20

    <P><FONT face=3DArial><FONT color=3D#0000ff>QoS is a complex =
question, dependent=20
    on the target network(s).&nbsp; Router configurations, network =
topologies=20
    and firewalls all play a role here.</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>I would be interested in =
what you=20
    think, and what results you come up with, if you can share this with =

    us.</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>Good Luck!</FONT></FONT> =

    <P><FONT face=3DArial><FONT color=3D#0000ff>Ralph</FONT></FONT>=20
    <P><FONT face=3DArial><FONT color=3D#0000ff>Ralph S. =
Hoefelmeyer</FONT></FONT>=20
    <BR><FONT face=3DArial><FONT color=3D#0000ff>MCI WorldCom =
Engineer</FONT></FONT>=20

    <UL><FONT face=3DArial><FONT size=3D-2>-----Original=20
      Message-----</FONT></FONT> <BR><B><FONT face=3DArial><FONT=20
      size=3D-2>From:&nbsp;&nbsp;</FONT></FONT></B> <FONT =
face=3DArial><FONT=20
      size=3D-2>Omar Ferrer [SMTP:oferrer@telcel.net.ve]</FONT></FONT>=20
      <BR><B><FONT face=3DArial><FONT =
size=3D-2>Sent:&nbsp;&nbsp;</FONT></FONT></B>=20
      <FONT face=3DArial><FONT size=3D-2>Friday, March 12, 1999 6:20=20
      AM</FONT></FONT> <BR><B><FONT face=3DArial><FONT=20
      size=3D-2>To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></FONT></B> <FONT=20
      face=3DArial><FONT size=3D-2>Henri Setiawan; H323 Implementors;=20
      IPTEL</FONT></FONT> <BR><B><FONT face=3DArial><FONT=20
      =
size=3D-2>Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT></FONT=
></B>=20
      <FONT face=3DArial><FONT size=3D-2>[Fwd: Parameter for =
interoperability=20
      test]</FONT></FONT>=20
      <P><FONT face=3DArial><FONT color=3D#000000>Dear Mr. =
Setiawan:</FONT></FONT>=20
      <P><FONT face=3DArial><FONT color=3D#000000>I've recently =
subscribe to this=20
      newsgroup asking for some help on testing VoIP products. I'm =
beginning a=20
      project in which we want to test Gateways, Gatekeepers and Network =

      Managers, but we do not have experience with that because this =
field is=20
      completely new for the Latin America market.</FONT></FONT>=20
      <P><FONT face=3DArial><FONT color=3D#000000>Do you now what I need =
to measure=20
      or test on each equipment and how?</FONT></FONT>=20
      <P><FONT face=3DArial><FONT color=3D#000000>We are beginning a =
project for=20
      VoIP. In the first phase we want to test each equipment on a point =
to=20
      point link (TDM) with a higher bandwith and then decreasing as the =
traffic=20
      increase. In the second phase, we want to test the same products =
but with=20
      direct connection to the Internet. Here we need to prove some QoS =
tools=20
      and bandwith management techniques, and also, some testing with a =
Global=20
      Network for call terminations worldwide.</FONT></FONT>=20
      <P><FONT face=3DArial><FONT color=3D#000000>Could you give some =
approach for=20
      this?</FONT></FONT>=20
      <P><FONT face=3DArial><FONT =
color=3D#000000>Thanks!!!</FONT></FONT> <BR><FONT=20
      face=3DArial><FONT color=3D#000000>Omar Ferrer</FONT></FONT> =
<BR><FONT=20
      face=3DArial><FONT color=3D#000000>Caracas - =
Venezuela</FONT></FONT>=20
      <P><FONT face=3DArial><FONT color=3D#000000>-------- Original =
Message=20
      --------</FONT></FONT> <BR><B><FONT=20
      =
face=3DArial>Subject:&nbsp;</FONT></B>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
;&nbsp;<FONT=20
      face=3DArial> Parameter for interoperability test</FONT> =
<BR><B><FONT=20
      face=3DArial>Date:&nbsp;</FONT></B>&nbsp;&nbsp;<FONT face=3DArial> =
Fri, 12 Mar=20
      1999 13:46:55 +0700</FONT> <BR><B><FONT=20
      face=3DArial>From:&nbsp;</FONT></B>&nbsp;&nbsp;<FONT face=3DArial> =
&quot;Henri=20
      Setiawan&quot; &lt;henry@risti.telkom.co.id&gt;</FONT> =
<BR><B><FONT=20
      face=3DArial>To:&nbsp;</FONT></B>&nbsp;&nbsp;&nbsp;&nbsp;<FONT =
face=3DArial>=20
      &lt;h323implementors@imtc.org&gt;</FONT> <BR><B><FONT=20
      face=3DArial>CC:&nbsp;</FONT></B>&nbsp;&nbsp;&nbsp;&nbsp;<FONT =
face=3DArial>=20
      &lt;iptel@lists.research.bell-labs.com&gt;</FONT>=20
      <P><FONT face=3DArial><FONT color=3D#000000>&nbsp;<FONT =
size=3D-1>Dear=20
      folks,</FONT> <FONT size=3D-1>We like to study about =
interoperability test=20
      for gateway and gateway keeper, do anyone know what kind of =
parameters=20
      that we should test for this testing ?</FONT> <FONT =
size=3D-1>Thanks</FONT>=20
      <FONT size=3D-1>Henri SR&amp;D of PT Telkom=20
      =
Indonesia.</FONT></FONT></FONT></P></UL></BLOCKQUOTE></BLOCKQUOTE></BODY>=
</HTML>

------=_NextPart_000_0047_01BE6F81.CC822C60--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 10:41:30 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA21380
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 10:41:30 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7387D52C5; Thu, 18 Mar 1999 09:39:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C839952C7; Thu, 18 Mar 1999 09:39:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <011801be714c$8b0b9d80$6531a38d@see.plym.ac.uk>
From: "Bogdan Ghita" <b.ghita@jack.see.plym.ac.uk>
To: <iptel@lists.research.bell-labs.com>
Subject: data content identification
Date: Thu, 18 Mar 1999 14:35:10 -0000
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.0810.800
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hello everybody

I don't know if there is an answer to my problem, but it worth to try at
least. I am looking for a method to identify voice packets on a data stream.
Considering the level of standardisation for the IP telephony, I thought to
start, for the beginning, with standard protocol, like HTTP, and then, if
the results are encouraging, to go on with my study.

I looked through the internet and the RFCs and I haven't found any
suggestion for determining the content of the packet by any method (looking
at its header, watching the data flows going through a point, and so on). It
is not supposed to be only a port-identifying at the TCP level, because
there is no standard, yet, for using specific ports for IP telephony, so
this could be consider a dead end.

Do you have any idea how could I solve this problem?

Thank you
Bogdan Ghita




---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 10:49:21 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA21461
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 10:49:21 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6B17152D5; Thu, 18 Mar 1999 10:17:32 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8D04852CD; Thu, 18 Mar 1999 10:17:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: storassa@hotmail.com
Message-ID: <36EEE96D.759ADDD2@hotmail.com>
Date: Tue, 16 Mar 1999 17:29:49 -0600
Reply-To: storassa@hotmail.com
X-Mailer: Mozilla 4.51 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
To: pint@lists.research.bell-labs.com, ietf-pin@lists.research.bell-labs.com,
        sigtran@baynetworks.com, confctrl@isi.edu,
        iptel@lists.research.bell-labs.com, megaco@baynetworks.com
Cc: abrusilovsky@lucent.com
Subject: FYI: PIN mail exploder
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

At the PIN BOF a number of people asked me to re-post an announcement
regarding PIN mailing list. We aim to please:

The PIN mailing list is located at
ietf-pin@lists.research.bell-labs.com.

To subscribe or unsubscribe yourself to the mailing list, send E-mail to
ietf-pin-request@lists.research.bell-labs.com with the single word
"subscribe" or "unsubscribe" (without the quotes) in the body of the
message. Please do not send subscription requests to
ietf-pin@lists.research.bell-labs.com.

Regards,
Alec Brusilovsky

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 11:25:00 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA21717
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 11:24:59 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1DD5352C8; Thu, 18 Mar 1999 11:19:42 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6A41F52C9; Thu, 18 Mar 1999 11:19:41 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F12715.A02FE83E@dnrc.bell-labs.com>
Date: Thu, 18 Mar 1999 11:17:25 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Chris Purvis <CPurvis@madge.com>
Cc: "'rtgupta@hss.hns.com'" <rtgupta@hss.hns.com>, h323implementors@imtc.org,
        iptel@lists.research.bell-labs.com
Subject: Re: h.245 channel establishment
References: <4110BC526A58D211B56C0000F6B9636DA85207@ws-exch01.dev.madge.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

This thread somehow got added to the iptel mailing list at some point.
It really seems like it belongs entirely on the h323 implementors list.
I would ask that people responding to this thread remove iptel from the
distribution list.

Thanks,
Jonathan R.

Chris Purvis wrote:
> 
> Ritu,
> 
> 1. I see no reason why a gatekeeper can't bring up a TCP connection for
> H.245 on the basis of an address an endpoint has given it.  This would seem
> to be perfectly normal behaviour in any model where the gatekeeper is
> routing H.245.
> 
> 2. It's perfectly possible for one "leg" of a call to be tunnelling H.245,
> and the other to have a separate TCP connection.
> 
> 3. Called endpoing may get GK's H.245 control channel transport address from
> the (optional) h245address field in the setup message (if the GK chooses to
> fill it in).
> 
> Does that give you enough options?
> 
> I actually think this is one of the few places where direct-routed vs
> GK-routed is not a major issue.
> 
> As an aside, we shouldn't assume that TCP is the only right answer for
> anything in the H.323 world, because the H.323 standards do not require that
> they're being implemented over IP.
> 
> Regards,
> Chris
> 
> > -----Original Message-----
> > From: rtgupta@hss.hns.com [mailto:rtgupta@hss.hns.com]
> > Sent: 17 March 1999 9:45
> > To: h323implementors@imtc.org; iptel@lists.research.bell-labs.com
> > Subject: h.245 channel establishment
> >
> >
> > Hi All,
> >     This is regarding h.245 control channel establishment.
> >
> > Specifications (see sections 6.2.8 and 8.2 of h.323 ) say that either
> > endpoint, involved in the call, will initiate setup of TCP
> > connection for
> > h.245 signalling. But there has not been any explicit mention
> > of type of
> > 'call model' supported. So should I assume that in its absence,
> > irrespective of call model, TCP conection will be initiated
> > by endpoint
> > ONLY and not by gatekeeper?
> >
> > In case GK decides to route h.245 signalling, then two h.245
> > channels will
> > be required, one between calling endpoint and GK and other
> > between GK and
> > called party. So can I assume safely that in this case, both
> > calling party
> > and called party will  initiate a h.245 channel towards
> > gatekeeper? But in
> > this case how will called endpoint get GK's h.245 control
> > channel transport
> > address?(For calling end, that is conveyed in call proceeding
> > or Alert or
> > Connect or Progress message)
> >
> > Also while switching from tunneling to a seperate connection
> > for h.245,
> > both endpoints will have to initiate TCP connection towards
> > GK to have two
> > channels.
> >
> > In any of above two cases,what if one endpoint sets up h.245
> > connection
> > towards GK but other endpoint does not initiate? Should GK initiate a
> > channel setup or just send a Facility message with a reason
> > of 'starth245'?
> >
> > Any experiences or ideas to be shared?
> >
> > Thanks,
> > Ritu.
> >
> > standards discuss this setup.
> >
> >
> > --------------------------------------------------------------
> > -----------
> > for list subscription instructions,
> > send email to h323implementors-request@imtc.org with text "info"
> >
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 12:12:52 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA22272
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 12:12:51 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 540F952CC; Thu, 18 Mar 1999 11:26:56 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5421F52CD; Thu, 18 Mar 1999 11:26:55 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F128C0.AE0D634A@dnrc.bell-labs.com>
Date: Thu, 18 Mar 1999 11:24:32 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Bogdan Ghita <b.ghita@jack.see.plym.ac.uk>
Cc: iptel@lists.research.bell-labs.com
Subject: Re: data content identification
References: <011801be714c$8b0b9d80$6531a38d@see.plym.ac.uk>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

The media (voice) for Internet telephony is transported using RTP, which
usually runs on UDP. Unfortunately, RTP does not have its own port
number or protocol number. The protocol is that for UDP, and the port is
established dynamically via the signaling protocols (SIP/SDP or H.323).
Therefore, there is no way a router can statelessly determine that a
packet is RTP. The determination is possible by stateful means, as is
done with RTP header compression (rfc2508). You can also snoop the
signaling packets and extract the port numbers.

I had written a proposal way back when for making RTP its own protocol
for this reason
(http://www.cs.columbia.edu/~jdrosen/papers/draft-rosenberg-rtpproto-00.txt),
but there was solid consensus not to do this, since it introduces all
kinds of backwards compatibility issues.

-Jonathan R.

Bogdan Ghita wrote:
> 
> Hello everybody
> 
> I don't know if there is an answer to my problem, but it worth to try at
> least. I am looking for a method to identify voice packets on a data stream.
> Considering the level of standardisation for the IP telephony, I thought to
> start, for the beginning, with standard protocol, like HTTP, and then, if
> the results are encouraging, to go on with my study.
> 
> I looked through the internet and the RFCs and I haven't found any
> suggestion for determining the content of the packet by any method (looking
> at its header, watching the data flows going through a point, and so on). It
> is not supposed to be only a port-identifying at the TCP level, because
> there is no standard, yet, for using specific ports for IP telephony, so
> this could be consider a dead end.
> 
> Do you have any idea how could I solve this problem?
> 
> Thank you
> Bogdan Ghita
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 16:25:54 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA26155
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 16:25:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id ED30352C3; Thu, 18 Mar 1999 14:01:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 51F0752C7; Thu, 18 Mar 1999 14:01:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <19B014596613D111AA08006097937C10014B9595@stca902a.eng.sc.rolm.com>
From: "Karapetkov, Stefan H" <Stefan.H.Karapetkov@icn.siemens.com>
To: "'Grandclaude Sylvain-LSG012'" <LSG012@email.mot.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>,
        "'villuzzi'" <vincent.illuzzi@mci.com>
Subject: RE: Looking for large scale VoIP Equipment
Date: Thu, 18 Mar 1999 10:55:18 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Looking for distributed PBX-like intelligence in the IP network and full
H.323 compliance (clients, IP phones, gateway, gatekeeper)? Check out
Siemens' HiNet RC 3000 at the Spring'99 VON show in Las Vegas, April 13-16. 
You will also see CTI servers and cool applications.
See you there, 
Stefan 

s 
Information & Communication Networks
Product Manager
IP Data Networks
voice: +1 (408) 492 6044
fax:    +1 (408) 492 4666
Stefan.Karapetkov@icn.siemens.com


> -----Original Message-----
> From:	Grandclaude Sylvain-LSG012 [SMTP:LSG012@email.mot.com]
> Sent:	Wednesday, March 17, 1999 6:45 PM
> To:	IPTEL IETF Mailing List; 'villuzzi'
> Subject:	RE: Looking for large scale VoIP Equipment
> 
> Hi
> 
> 
> I am a bit sad that you will, probably, not find something satisfying your
> needs AT THE PRESENT TIME.
> 
> NOT ANY vendor has got large VoIP gear ready. Many do claim, but ???
> You can still try with 3COM or NORTEL, but I am dubious they really have
> something ready and working - well - NOW.
> Start-Up companies likewise SONUS Networks and the likes have great
> potential - but no working products yet as far as I am aware.
> You will have to be patient !
> 
> Then, PABX-like "intelligence" is something else. 
> 
> This is NOT something any H.323-compliant GATEWAY is capable of providing
> on
> its own.
> This is NOT something any H.323-compliant GATEKEEPER is capable of
> providing
> on its own.
> This is something a GATEKEEPER might help to, if providing an information
> feed to a 3rd-Party CTI SERVER S/W application.
> 
> Telephone sets + gateway + gatekeeper + complementary CTI server S/W
> altogether may look somewhat equivalent to a PABX.
> You need all the pieces in that puzzle.
> 
> And you'd better choose for FULLY standards-compliant pieces !
> 
> 
> 
> Hope this helps,
> 
> 
> 		Sylvain Grandclaude.
> 
> 		Application Engineer Motorola ING EMEA
> 		tel 	+33 1 49 69 85 80
> 		fax 	+33 1 46 64 43 00
> 		mobile +33 6 07 25 78 70
> 		e-mail	LSG012@email.mot.com <mailto:LSG012@email.mot.com> 
> 
> 		The opinions expressed above represent me only as an
> individual and don't represent in any extent Motorola Company strategy or
> commitments.
> 
> 
> 	----------
> 	From:  villuzzi [SMTP:vincent.illuzzi@mci.com]
> 	Sent:  mardi 16 mars 1999 19:52
> 	To:  IPTEL IETF Mailing List
> 	Subject:  Looking for large scale VoIP Equipment
> 
> 	I am looking to either migrate then replace or upgrade large scale
> PBX
> 	equipment with VoIP capable gear. Any one have vendor information
> for large
> 	scale VoIP capable equipment with feature rich PBX like
> capabilities?
> 
> 	Thanks,
> 
> 	Vince
> 
> 
> 	---------
> 	This message came from the IETF IPTEL Working Group Mailing List.
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 22:12:49 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA00109
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 22:12:49 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A583B52CB; Thu, 18 Mar 1999 22:08:09 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id EA3A352CD; Thu, 18 Mar 1999 22:08:08 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F1BF7A.60C05106@dnrc.bell-labs.com>
Date: Thu, 18 Mar 1999 22:07:38 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: "ronald h. davis" <ronaldd@lucent.com>
Cc: hsalama@cisco.com, Alok Mittal <amittal@hss.hns.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: Re: glp draft based on BGP
References: <01BE663A.B6B0A9C0@hss056.hss.hns.com> <36DEE85A.25189194@cisco.com> <36ED4C70.BB6CF1B@lucent.com> <36ED579F.D487D82A@cisco.com> <36ED752D.E64DE513@lucent.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

ronald h. davis wrote:
> 
> Hussein F Salama wrote:
> >
> > > the thrust of my suggestion is to view the tbgp speaker as being an
> > > appliation layer router.  depending upon network policies, the tbgp
> > > speaker could send the equivalent of an icmp redirect message to ua
> > > that could be cached for future calls.
> >
> > I think the proposed query/reply mechanism is simpler. Redirecting the
> > call involves more state setup/teardown/modification overhead.
> >
> 
> this is certainly a reasonable model if you view the tbgp speaker as
> providing an address resolution function such as in the atmarp server
> in classical ip over atm.

How the database built up by a location server (LS) is used is out of
scope for GLP itself. In the framework document, this interaction is
called the "front end". Several possibilities are identified:

1. query/response, called "direct mode", perhaps with LDAP or RAS LRQ or
SLP
2. "proxy mode" - send a signaling message to the LS, and have it do the
resolution (I believe this is what Ron is suggesting), then forwarding
it towards the next hop

Redirects are also possible, easily done with SIP 300 responses. This is
sort of a hybrid between proxy mode and direct mode. I'll add something
about it to the framework document. 

However, we don't decide how this is done. We only define the protocol
for populating the databases.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 22:22:38 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA00162
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 22:22:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C9F9752D0; Thu, 18 Mar 1999 22:13:54 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5E1E552D2; Thu, 18 Mar 1999 22:13:52 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F1C0CC.80073FA6@dnrc.bell-labs.com>
Date: Thu, 18 Mar 1999 22:13:16 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Christian Huitema <huitema@research.bellcore.com>
Cc: Scott Petrack <scott.petrack@metatel.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <Pine.LNX.4.04.9903111856190.6794-100000@petrack.metatel.com> <36ECA53F.7F8A06F9@research.bellcore.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Christian Huitema wrote:
> 
> Scott Petrack wrote:
> >
> > The fact that this situation is a mess in the PSTN does not mean that it
> > must be a mess in GLP. And the fact that the "scope" of routing
> > numbers in the PSTN is the "provider ID" does not
> > mean that this is the only scope that we should allow within GLP.
> > Certainly, 1-700 numbers will absolutely require a scope which is a
> > "provider id", but 800 numbers might not. For example, my company owns a
> > certain 800 number, and it will resolve to our office telephone number no
> > matter what provider we happen to choose to give us service. Nothing in
> > GLP will require that the scope will be "provider id."
> 
> I think that this all go back to a two phases approach for telephony
> routing:
> 
> 1) take a dip in a mapping directory to understand the "provider" for
> the
>    phone number, which will result in a <provider, number> pair -- maybe
>    with a specific provider that says "e164"
> 
> 2) route based on the pair.

I think there is an important point here. To date, we have discussed two
models for handling number translations:

1. we do it in the IP side (perhaps via enum), before GLP, so that GLP
advertises routable numbers only
2. we do it in the PSTN side, so that GLP needs to advertise "names"
like 800 numbers

What Christian is really saying is that there is a third possibility:

1. a partial translation is done on the IP side, yielding a provider ID
or something like that. GLP also propagates provider ID's, so the call
is routed there. The final translation is done inside the PSTN.

I think its a valid requirement to support this. Comments?

-Jonathan R.


-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 23:27:41 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA02269
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 23:27:40 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 659AB52D2; Thu, 18 Mar 1999 22:24:12 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9B9E752D1; Thu, 18 Mar 1999 22:24:11 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F1C338.F5C59185@dnrc.bell-labs.com>
Date: Thu, 18 Mar 1999 22:23:36 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: Matt Squire <msquire@nortelnetworks.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990311212111.00ab7d80@bl-mail1.corpeast.baynetworks.com> <36EC7808.B74EAC13@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> Depends on how similar the IP ASes topology and TBGP Administrative
> Domains (AD) topology are.

I believe we agreed that independent of whether SCSP or TBGP is used,
the protocol runs independent and outside of the current routing
infrastructure.

> 
> The capacity metric could also be used to enforce policy, in addition to
> its use for advertising physical resources (bandwidth, ports, ... etc).
> For example, it can be used for constraining the number of calls from
> one AD to another. For example, AD1 may wish not to route more than x
> calls from AD2 towards +1408*. This can be achieved by setting capacity
> = x in advertisement of 1408* frowarded from AD1 to Ad2.

This introduces the following problem:

           -- AD1--
          /        \
        /           \
      AD2             AD3
       \              /
        \            / 
         \          /
          \        /
           \      /
              AD4

AD1 propagates a prefix to AD2, and also to AD3. The prefix is to a
gateway with capacity X. So, AD1 propagates capacity X to AD2, and also
capacity X to AD3. Both AD2 and AD3 advertise this to AD4, which now
thinks there is a total of 2X capacity to the prefix.

For your approach to work, the advertisements originated by an AD for a
prefix in a gateway  must have the sum of the capacities of the metrics
less than the actual gateway capacity. This results in a very
inefficient, static division of resources.

So, I'd rather not associate any "guarantees" on capacity with this
metric. I think the metric needs to have a unit, else providers won't
know how to select values that are comparable across providers (I think
there was consensus on this too during the meeting, but speak up if I'm
confused). But, we should treat this as nothing more than a "hint",
useful for making (1) policy decisions (as any attribute is useful for
that purpose), and (2) load sharing decisions between gateways.

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 18 23:51:22 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA03026
	for <iptel-archive@odin.ietf.org>; Thu, 18 Mar 1999 23:51:22 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D361352C7; Thu, 18 Mar 1999 23:46:49 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4C21252CA; Thu, 18 Mar 1999 23:46:49 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F1D69A.3D27C79F@dnrc.bell-labs.com>
Date: Thu, 18 Mar 1999 23:46:18 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: list iptel <iptel@lists.research.bell-labs.com>
Subject: Slides available
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

The slides from the meeting this past week are now available at:

http://www.bell-labs.com/mailing-lists/iptel/

Thanks,
Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 02:24:22 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA12185
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 02:24:21 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A316852C9; Fri, 19 Mar 1999 02:19:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E385E52CA; Fri, 19 Mar 1999 02:19:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36CD1041.1CDBA0F9@cisco.com>
Date: Thu, 18 Feb 1999 23:18:25 -0800
From: Hussein Salama <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.03 [en] (Win95; U)
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Matt Squire <msquire@nortelnetworks.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990311212111.00ab7d80@bl-mail1.corpeast.baynetworks.com> <36EC7808.B74EAC13@cisco.com> <36F1C338.F5C59185@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Jonathan Rosenberg wrote:
> 
> Hussein F. Salama wrote:
> >
> > Depends on how similar the IP ASes topology and TBGP Administrative
> > Domains (AD) topology are.
> 
> I believe we agreed that independent of whether SCSP or TBGP is used,
> the protocol runs independent and outside of the current routing
> infrastructure.
> 
> >
> > The capacity metric could also be used to enforce policy, in addition to
> > its use for advertising physical resources (bandwidth, ports, ... etc).
> > For example, it can be used for constraining the number of calls from
> > one AD to another. For example, AD1 may wish not to route more than x
> > calls from AD2 towards +1408*. This can be achieved by setting capacity
> > = x in advertisement of 1408* frowarded from AD1 to Ad2.
> 
> This introduces the following problem:
> 
>            -- AD1--
>           /        \
>         /           \
>       AD2             AD3
>        \              /
>         \            /
>          \          /
>           \        /
>            \      /
>               AD4
> 
> AD1 propagates a prefix to AD2, and also to AD3. The prefix is to a
> gateway with capacity X. So, AD1 propagates capacity X to AD2, and also
> capacity X to AD3. Both AD2 and AD3 advertise this to AD4, which now
> thinks there is a total of 2X capacity to the prefix.
> 

AD4 won't merge two routes to the same prefix. It will select either
vthe route coming from AD2 or the route coming from AD3. In both cases
the capacity of calls towards the prefix served by AD1 as seen by AD4
will still be X.

If you use such a metric, we have to make sure that it's value can only
decrease, but not increase, as it is propogated further from the egress
gateway.

> For your approach to work, the advertisements originated by an AD for a
> prefix in a gateway  must have the sum of the capacities of the metrics
> less than the actual gateway capacity. This results in a very
> inefficient, static division of resources.

In your example, assume the gateway at AD1 has a total capacity of X. It
advertises a capacity of Y to AD2 and a capacity of Z to Ad3. 

I think the only constraints on Y and Z should be:
0 <= Y <= X and 0 <= Z <= X. 

However Y + Z shouldn't necessarily be less than or equal to X. I think
it should be possible to achieve efficient utilization of resources by
selecting Y and Z such that Y + Z > X based on some statistical analysis
of call arrivals and call durations.

> 
> So, I'd rather not associate any "guarantees" on capacity with this
> metric. I think the metric needs to have a unit, else providers won't

From the above, you only guarantee that a minimum number of calls ( = X
- Z) will be accepted from  the AD2 route, and a minimum number of calls
(= X - Y) will be accepted from the AD3 route. No guarantees beyond
that.

> know how to select values that are comparable across providers (I think
> there was consensus on this too during the meeting, but speak up if I'm
> confused). But, we should treat this as nothing more than a "hint",
> useful for making (1) policy decisions (as any attribute is useful for
> that purpose), and (2) load sharing decisions between gateways.
> 

It's different from the originally proposed capacity metric, because it
can be modified hop by hop based on local policy. It could lead to a
fair utilizations of resources among calls from different origins. I
don't think it's much more difficult to define that the originally
proposed metric. The original capacity metric was proposed to be set
equal to the total capacity at the egress gateway and does not change
hop by hop. This is a special case of what I am proposing.

Hussein


> -Jonathan R.
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX: (732) 834-5379                         Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 04:02:16 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA12834
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 04:02:16 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 312EF52CA; Fri, 19 Mar 1999 03:57:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8DD6952D3; Fri, 19 Mar 1999 03:57:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F2148F.FC8F43BB@seeyes.co.in>
Date: Fri, 19 Mar 1999 14:40:39 +0530
From: Venkata Narasimha Murthy Nanduri <nvnmurthy@seeyes.co.in>
Reply-To: nvnmurthy@seeyes.co.in
Organization: SeeYes
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: iptel@lists.research.bell-labs.com
Subject: Reg. RTP Header compression/decompression
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hai,

I have to work on RTP header compession/decompression.
For this I am reading RFC1144 - TCP/IP header comression  and
RFC 2508 - RTP header compression.

Other than these, R there any other information sources which will be
useful?
Is there any sample implementation available on the net?

Any information reg. this is appreciated.

Thanks in advance
Murthy NVN


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 09:26:39 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA14317
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 09:26:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C4FC452D7; Fri, 19 Mar 1999 08:07:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1EB7652D6; Fri, 19 Mar 1999 08:07:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990319080424.00860100@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Fri, 19 Mar 1999 08:04:25 -0500
To: hsalama@cisco.com, Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: data issue 4: capacity
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>> > The capacity metric could also be used to enforce policy, in addition to
>> > its use for advertising physical resources (bandwidth, ports, ... etc).
>> > For example, it can be used for constraining the number of calls from
>> > one AD to another. For example, AD1 may wish not to route more than x
>> > calls from AD2 towards +1408*. This can be achieved by setting capacity
>> > = x in advertisement of 1408* frowarded from AD1 to Ad2.
>> 
>> This introduces the following problem:
>> 
>>            -- AD1--
>>           /        \
>>         /           \
>>       AD2             AD3
>>        \              /
>>         \            /
>>          \          /
>>           \        /
>>            \      /
>>               AD4
>> 
>> AD1 propagates a prefix to AD2, and also to AD3. The prefix is to a
>> gateway with capacity X. So, AD1 propagates capacity X to AD2, and also
>> capacity X to AD3. Both AD2 and AD3 advertise this to AD4, which now
>> thinks there is a total of 2X capacity to the prefix.
>> 
>
>AD4 won't merge two routes to the same prefix. It will select either
>vthe route coming from AD2 or the route coming from AD3. In both cases
>the capacity of calls towards the prefix served by AD1 as seen by AD4
>will still be X.

I'm guessing Jonathan was thinking about a load-balancing scenario, where
AD4 wants to use multiple paths.  To perform load-balancing, an LS would
have to have multiple routes.  Whether you implement this as cycling
through a selection of individual routes, or actually selecting multiple
routes isn't important.  Using a capacity attribute to load-balance, which
is one of the reasons people wanted this attribute (as opposed to just
route selection), interferes with your scenario of partitioning capacity. 

>
>If you use such a metric, we have to make sure that it's value can only
>decrease, but not increase, as it is propogated further from the egress
>gateway.

I'm not sure that this is any different then the capacity attribute as
discussed at the wg.  We decided (I think) that we'd really like to have a
metric on the capacity attribute.  Say we make that metric "max # calls".  

We've always been a little fuzzy on what happens with this attribute at
intermediate LSs.  If an LS is just passing the (prefix,gw) address along,
then I think its clear that it shouldn't modify the capacity attribute.
But if the LS is inserting itself as the next hop, then it of course has
the right to modify the attributes, capacity included.  

But aggregation and deaggregation certainly screws things up.  In the above
example, AD2 could have aggregated +1408 into +140.  Most likely the
capacity attribute changed.  If you sum the capacity attribute (a likely
possibility), then you lose information.  For example, AD2 could be
advertising +140 with a capacity of 5X.  AD4 could then send 5X calls for
+1408 even though, at AD1, there are only X calls for +1408.  Deaggregation
causes similar problems.  

This is one of the many problems with using any true metric in the face of
aggregation and deaggregation.  Its relationship to reality is lost as
information is aggregated and deaggregated.  

Ignoring the problems of aggregation, in order to get any kind of
statistical significance as the information is propagated, I think you need
to mandate similar behavior at each LS.  Although any LS can make a
decision to break up a capacity of X into Y+Z>X.  Say you use a rule to do
Y+Z=1.2X, and another LS breaks up Y into A+B=1.9Y, and another LS breaks
up Z into C+D=1.1Z, etc.  Although you could make statistical decisions,
you would have to make assumptions about Y and Z, but these assumptions
would most likely not be valid unless you know what decisions other LSs are
using.  So in order to get some overall statistical significance to the
system, I think you'd need to have a defined method for handling capacity
in such manner at all LSs.  

Of course, the last time I did any real mathematics was in grad school, and
I had hair then, so maybe you know of a method for getting some statistical
significance out of the system without mandating intermediate behavior - I
just don't see it.  

I also don't see any capacity attribute as providing "guarantees" in the
face of aggreg/deaggreg.  Guidelines, maybe.  Heuristics, sure.  But not
guarantees.

- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 11:28:03 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA16256
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 11:28:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 98CF252DA; Fri, 19 Mar 1999 10:27:25 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A4C2752DD; Fri, 19 Mar 1999 10:27:24 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Fri, 19 Mar 1999 10:25:19 -0500 (EST)
From: Scott Bradner <sob@harvard.edu>
Message-Id: <199903191525.KAA02518@newdev.harvard.edu>
To: b.ghita@jack.see.plym.ac.uk, iptel@lists.research.bell-labs.com
Subject: Re: data content identification
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

> I am looking for a method to identify voice packets on a data stream

why do you want to do this?

note that if its so someone can bill differently for the packets
I would expect that many of the users will adopt simple ways
to hide this from you

Scott

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 11:55:55 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA16576
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 11:55:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0E4D352CD; Fri, 19 Mar 1999 11:51:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6021052D6; Fri, 19 Mar 1999 11:51:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Fri, 19 Mar 1999 08:48:35 -0800 (PST)
From: Edward Vielmetti <evielmet@cisco.com>
To: Scott Bradner <sob@harvard.edu>
Cc: b.ghita@jack.see.plym.ac.uk, iptel@lists.research.bell-labs.com
Subject: Re: data content identification
In-Reply-To: <199903191525.KAA02518@newdev.harvard.edu>
Message-ID: <Pine.HPX.4.02.9903190835260.18239-100000@lint.cisco.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

> > I am looking for a method to identify voice packets on a data stream

It would seem to be straightforward to look for long-lived flows at
roughly voice-sized constant data rates.  No matter what protocol is
being used to carry these flows, they should stick out pretty clearly.
(As long as no one is doing crazy "spread spectrum" stuff across the IP
port space).  Constant bit rate encoders will be easier to find than
variable bit rate encoders, and if someone is doing streaming media like
realaudio or video you might be confused by those.

It'll be harder to pick out voice the less it is real time - a "push to
talk" scheme running over any of the internet "chat" protocols will be so
much more bursty than the canonical H.323 application than it might as
well be a totally different beast for content identification purposes. And
you know that the people hacking voice over chat don't go to international
standards committee meetings very often....

-- 
Edward M. Vielmetti                                 ||        ||        
Consulting Engineer, Office of the CTO,             ||        ||        
Cisco Systems                                      ||||      ||||       
Ann Arbor, MI                                  ..:||||||:...:||||||:..   
mailto:evielmet@cisco.com                      C i s c o S y s t e m s 



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 12:11:55 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA16746
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 12:11:55 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2176D52D1; Fri, 19 Mar 1999 12:07:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7459352D8; Fri, 19 Mar 1999 12:07:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <013b01be722a$cd5c8c60$6531a38d@see.plym.ac.uk>
From: "Bogdan Ghita" <b.ghita@jack.see.plym.ac.uk>
To: <iptel@lists.research.bell-labs.com>
Subject: Re: data content identification
Date: Fri, 19 Mar 1999 17:06:12 -0000
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.0810.800
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

>> I am looking for a method to identify voice packets on a data stream
>
>why do you want to do this?
>
>note that if its so someone can bill differently for the packets
>I would expect that many of the users will adopt simple ways
>to hide this from you
>
>Scott

I am not planning to do it for billing purposes.
I am doing a PhD project about monitoring the parameters of Voice over IP
traffic. It is not going to be specifically for one type of network of
application, that is why I hoped there is a specific header, or something
like this, inside the voice packets, no matter which is the application that
builds it.
I haven't found, at least until now, a good method to separe the voice
packets from the other types.

Regards
Bogdan Ghita



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 12:25:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA16835
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 12:25:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7CF1F52D9; Fri, 19 Mar 1999 12:21:13 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E579952DB; Fri, 19 Mar 1999 12:21:12 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <014401be722b$d9e29d20$6531a38d@see.plym.ac.uk>
From: "Bogdan Ghita" <b.ghita@jack.see.plym.ac.uk>
To: <iptel@lists.research.bell-labs.com>
Subject: Re: data content identification
Date: Fri, 19 Mar 1999 17:13:42 -0000
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.0810.800
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.0810.800
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

>> > I am looking for a method to identify voice packets on a data stream
>
>It would seem to be straightforward to look for long-lived flows at
>roughly voice-sized constant data rates.  No matter what protocol is
>being used to carry these flows, they should stick out pretty clearly.
>(As long as no one is doing crazy "spread spectrum" stuff across the IP
>port space).  Constant bit rate encoders will be easier to find than
>variable bit rate encoders, and if someone is doing streaming media like
>realaudio or video you might be confused by those.
I thought about looking at the encoders, but, again, there is no standard,
and, as you say, I cannot separate a voice conference from a live video
transmission, for example. Still, there should be something specific about
the voice (hopefully)...
>
>It'll be harder to pick out voice the less it is real time - a "push to
>talk" scheme running over any of the internet "chat" protocols will be so
>much more bursty than the canonical H.323 application than it might as
>well be a totally different beast for content identification purposes. And
>you know that the people hacking voice over chat don't go to international
>standards committee meetings very often....
Yes, that is true; the voice traffic doesn't have to be constant, or at a
constant speed.
 And I have to admit that a 'hacking voice conference' would be very useful
to my project, because it could clarify some of the aspects I'm looking for.

Regards
Bogdan Ghita

>
>--
>Edward M. Vielmetti                                 ||        ||
>Consulting Engineer, Office of the CTO,             ||        ||
>Cisco Systems                                      ||||      ||||
>Ann Arbor, MI                                  ..:||||||:...:||||||:..
>mailto:evielmet@cisco.com                      C i s c o S y s t e m s
>

>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 12:42:15 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA17008
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 12:42:14 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C69A652DF; Fri, 19 Mar 1999 12:37:13 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 553F352DD; Fri, 19 Mar 1999 12:37:13 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <4.1.19990319103125.009ad100@orinoco.cisco.com>
X-Sender: rtbell@orinoco.cisco.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1 
Date: Fri, 19 Mar 1999 10:35:06 -0700
To: iptel@lists.research.bell-labs.com
From: Bob Bell <rtbell@cisco.com>
Subject: Re: data content identification
Cc: b.ghita@jack.see.plym.ac.uk
In-Reply-To: <014401be722b$d9e29d20$6531a38d@see.plym.ac.uk>
Mime-Version: 1.0
Content-Type: multipart/alternative;
	boundary="=====================_10018693==_.ALT"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--=====================_10018693==_.ALT
Content-Type: text/plain; charset="us-ascii"


Maybe the simplest method would be to look to the RTP headers. This should
identify the packet types for you so that you can scan them at your leisure.
This only applies to those forms of VoIP which encapsulate in RTP but that
should be a significant number of them.

At 10:13 AM  3/19/99 , you wrote:
>>> > I am looking for a method to identify voice packets on a data stream
>>
>>It would seem to be straightforward to look for long-lived flows at
>>roughly voice-sized constant data rates.  No matter what protocol is
>>being used to carry these flows, they should stick out pretty clearly.
>>(As long as no one is doing crazy "spread spectrum" stuff across the IP
>>port space).  Constant bit rate encoders will be easier to find than
>>variable bit rate encoders, and if someone is doing streaming media like
>>realaudio or video you might be confused by those.
>I thought about looking at the encoders, but, again, there is no standard,
>and, as you say, I cannot separate a voice conference from a live video
>transmission, for example. Still, there should be something specific about
>the voice (hopefully)...
>>
>>It'll be harder to pick out voice the less it is real time - a "push to
>>talk" scheme running over any of the internet "chat" protocols will be so
>>much more bursty than the canonical H.323 application than it might as
>>well be a totally different beast for content identification purposes. And
>>you know that the people hacking voice over chat don't go to international
>>standards committee meetings very often....
>Yes, that is true; the voice traffic doesn't have to be constant, or at a
>constant speed.
> And I have to admit that a 'hacking voice conference' would be very useful
>to my project, because it could clarify some of the aspects I'm looking for.
>
>Regards
>Bogdan Ghita
>
>>
>>--
>>Edward M. Vielmetti                                 ||        ||
>>Consulting Engineer, Office of the CTO,             ||        ||
>>Cisco Systems                                      ||||      ||||
>>Ann Arbor, MI                                  ..:||||||:...:||||||:..
>>mailto:evielmet@cisco.com                      C i s c o S y s t e m s
>>
>
>>
>
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.

Bob Bell
Distinguished Member of the Technical Staff
Cisco Systems Inc.
801-294-3034(v)
801-294-3023(f)
--=====================_10018693==_.ALT
Content-Type: text/html; charset="us-ascii"

<html><div>Maybe the simplest method would be to look to the RTP headers.
This should identify the packet types for you so that you can scan them
at your leisure. This only applies to those forms of VoIP which
encapsulate in RTP but that should be a significant number of
them.</div>
<br>
<div>At 10:13 AM&nbsp; 3/19/99 , you wrote:</div>
<div>&gt;&gt;&gt; &gt; I am looking for a method to identify voice
packets on a data stream</div>
<div>&gt;&gt;</div>
<div>&gt;&gt;It would seem to be straightforward to look for long-lived
flows at</div>
<div>&gt;&gt;roughly voice-sized constant data rates.&nbsp; No matter
what protocol is</div>
<div>&gt;&gt;being used to carry these flows, they should stick out
pretty clearly.</div>
<div>&gt;&gt;(As long as no one is doing crazy &quot;spread
spectrum&quot; stuff across the IP</div>
<div>&gt;&gt;port space).&nbsp; Constant bit rate encoders will be easier
to find than</div>
<div>&gt;&gt;variable bit rate encoders, and if someone is doing
streaming media like</div>
<div>&gt;&gt;realaudio or video you might be confused by those.</div>
<div>&gt;I thought about looking at the encoders, but, again, there is no
standard,</div>
<div>&gt;and, as you say, I cannot separate a voice conference from a
live video</div>
<div>&gt;transmission, for example. Still, there should be something
specific about</div>
<div>&gt;the voice (hopefully)...</div>
<div>&gt;&gt;</div>
<div>&gt;&gt;It'll be harder to pick out voice the less it is real time -
a &quot;push to</div>
<div>&gt;&gt;talk&quot; scheme running over any of the internet
&quot;chat&quot; protocols will be so</div>
<div>&gt;&gt;much more bursty than the canonical H.323 application than
it might as</div>
<div>&gt;&gt;well be a totally different beast for content identification
purposes. And</div>
<div>&gt;&gt;you know that the people hacking voice over chat don't go to
international</div>
<div>&gt;&gt;standards committee meetings very often....</div>
<div>&gt;Yes, that is true; the voice traffic doesn't have to be
constant, or at a</div>
<div>&gt;constant speed.</div>
<div>&gt; And I have to admit that a 'hacking voice conference' would be
very useful</div>
<div>&gt;to my project, because it could clarify some of the aspects I'm
looking for.</div>
<div>&gt;</div>
<div>&gt;Regards</div>
<div>&gt;Bogdan Ghita</div>
<div>&gt;</div>
<div>&gt;&gt;</div>
<div>&gt;&gt;--</div>
<div>&gt;&gt;Edward M.
Vielmetti&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
||&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ||</div>
<div>&gt;&gt;Consulting Engineer, Office of the
CTO,&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
||&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ||</div>
<div>&gt;&gt;Cisco
Systems&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
||||&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ||||</div>
<div>&gt;&gt;Ann Arbor,
MI&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
..:||||||:...:||||||:..</div>
<div>&gt;&gt;<a href="mailto:evielmet@cisco.com" EUDORA=AUTOURL>mailto:evielmet@cisco.com</a>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
C i s c o S y s t e m s</div>
<div>&gt;&gt;</div>
<div>&gt;</div>
<div>&gt;&gt;</div>
<div>&gt;</div>
<div>&gt;</div>
<div>&gt;---------</div>
<div>&gt;This message came from the IETF IPTEL Working Group Mailing
List.</div>
<br>

<font size=5 color="#0000FF"><b>Bob Bell<br>
</font></b>Distinguished Member of the Technical Staff<br>
Cisco Systems Inc.<br>
801-294-3034(v)<br>
801-294-3023(f)</html>

--=====================_10018693==_.ALT--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 12:51:12 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA17077
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 12:51:12 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6461552DE; Fri, 19 Mar 1999 12:41:25 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 660F252E2; Fri, 19 Mar 1999 12:41:24 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <AF62C6C0EC3DD211929900805FEA97CF60808E@wltg-msg01.mccaw-stg.com>
From: "Hutton, Charles" <charles.hutton@attws.com>
To: iptel@lists.research.bell-labs.com,
        "'nvnmurthy@seeyes.co.in'" <nvnmurthy@seeyes.co.in>
Subject: RE: Reg. RTP Header compression/decompression
Date: Fri, 19 Mar 1999 09:39:58 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

SInce RFC2508 is IP-UDP-RTP compression in one step, it's not clear to me
whether you are interested in that or just RTP compression as you wrote.

Assuming you are interested in IP-UDP-RTP all together compression, you
might want to look at:

ATM Forum STR-SAA-RMOA-01.00
ATM Forum SAA-98-0139

But RFC2508 seems to be the answer right now. I don't know what other
vendors are doing, but Cisco supports it in all their products that I have
looked at. 

Charles Hutton
Systems Engineering
Wireless Local Technology Group
AT&T Wireless Services
PO Box 97059
Redmond, WA 98073
14520 NE 87th Street
Redmond, WA 98052 
425-702-2938 Voice
425-702-2568 FAX
charles.hutton@attws.com

> ----------
> From: 	Venkata Narasimha Murthy
> Nanduri[SMTP:nvnmurthy@seeyes.co.in]
> Sent: 	Friday, March 19, 1999 1:10 AM
> To: 	iptel@lists.research.bell-labs.com
> Subject: 	Reg. RTP Header compression/decompression
> 
> Hai,
> 
> I have to work on RTP header compession/decompression.
> For this I am reading RFC1144 - TCP/IP header comression  and
> RFC 2508 - RTP header compression.
> 
> Other than these, R there any other information sources which will be
> useful?
> Is there any sample implementation available on the net?
> 
> Any information reg. this is appreciated.
> 
> Thanks in advance
> Murthy NVN
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 13:49:13 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA17630
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 13:49:08 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5FA7452DC; Fri, 19 Mar 1999 13:42:31 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 078C852D8; Fri, 19 Mar 1999 13:42:27 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F296BC.626F@mitel.com>
Date: Fri, 19 Mar 1999 13:26:04 -0500
From: Robert Wood <robert_wood@Mitel.COM>
Organization: Mitel Corporation
X-Mailer: Mozilla 3.01 (X11; I; HP-UX B.10.20 9000/715)
MIME-Version: 1.0
To: iptel@lists.research.bell-labs.com
Subject: Re: data content identification
References: <4.1.19990319103125.009ad100@orinoco.cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Bob Bell wrote:
> 
> Maybe the simplest method would be to look to the RTP headers. This
> should identify the packet types for you so that you can scan them at
> your leisure. This only applies to those forms of VoIP which
> encapsulate in RTP but that should be a significant number of them.
> 
	And  how do you do this as RTP  doesn't have a protocol ID?

[\] Robert Wood   

The St. Lawrence river - fresh, warm, visible diving.

mailto:robert_wood@mitel.com http://www.magma.ca/~rgwood/

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 14:29:44 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18092
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 14:29:44 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id F10FD52E0; Fri, 19 Mar 1999 13:42:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9C74952DB; Fri, 19 Mar 1999 13:42:29 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: TERADYNE
From: "Kevin Davis" <Kevin_Davis@notes.teradyne.com>
To: b.ghita@jack.see.plym.ac.uk
Cc: iptel@lists.research.bell-labs.com
Message-ID: <85256739.0065C763.00@notes.teradyne.com>
Date: Fri, 19 Mar 1999 13:31:59 -0500
Subject: Re: data content identification
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


If you do not monitor the call setup which specifies UDP port numbers
(and also the call disconnect), it is still very feasible to examine
UDP packets for what appears to be RTP audio content. However, when
you find something that looks like RTP audio, you should validate
it and make sure that the header recurs as expected (not a mimic).
Hammer Technologies already offers a statistics monitor for RTP audio;
it is included as a feature of Hammer's VoIP Test Suite.
- Kevin

>Maybe the simplest method would be to look to the RTP headers. This should
>identify the packet types for you so that you can scan them at your leisure.
>This only applies to those forms of VoIP which encapsulate in RTP but that
>should be a significant number of them.
>
> >I am not planning to do it for billing purposes.
> >I am doing a PhD project about monitoring the parameters of Voice over IP
> >traffic. It is not going to be specifically for one type of network of
> >application, that is why I hoped there is a specific header, or something
> >like this, inside the voice packets, no matter which is the application that
> >builds it.
> >I haven't found, at least until now, a good method to separe the voice
> >packets from the other types.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 15:33:41 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA18803
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 15:33:40 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0DF1F52E2; Fri, 19 Mar 1999 13:53:15 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5691252E4; Fri, 19 Mar 1999 13:53:14 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Hanlin_Fang@3com.com
X-Lotus-FromDomain: 3COM
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: list iptel <iptel@lists.research.bell-labs.com>
Message-ID: <88256739.006783E3.00@hqoutbound.ops.3com.com>
Date: Fri, 19 Mar 1999 10:50:16 -0800
Subject: question about supplementary services for SIP
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



Hello, all:

Does SIP support the following services:
   Message Waiting Indication
   Conference out of Consultation
   Call Completion on Busy Subscriber
If it does, how?  which header field(s) and what kind of messages should be used
for them?

Thanks for any information?

Hanlin



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 19 18:48:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA20874
	for <iptel-archive@odin.ietf.org>; Fri, 19 Mar 1999 18:48:53 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8154452C5; Fri, 19 Mar 1999 18:42:49 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D272D52CE; Fri, 19 Mar 1999 18:42:48 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F2E0D9.4F9A580C@dnrc.bell-labs.com>
Date: Fri, 19 Mar 1999 18:42:17 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Hanlin_Fang@3com.com
Cc: list iptel <iptel@lists.research.bell-labs.com>
Subject: Re: question about supplementary services for SIP
References: <88256739.006783E3.00@hqoutbound.ops.3com.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hanlin_Fang@3com.com wrote:
> 
> Hello, all:
> 
> Does SIP support the following services:
>    Message Waiting Indication

No, and its something I think is out of scope for SIP. Rather, something
like IMAP would probably be the right thing for this. I might have lots
of messages, in fact - voice, email, faxes. Using a voice specific
protocol for messaging seems a bad idea. In fact, I think there was a
BoF on VPIM that might be relevant (not sure).

>    Conference out of Consultation

Not in the base spec, but something that is defined in the call control
extensions. We're in the process of reworking this, as I mentioned
during the meeting. Expect the draft soon.

>    Call Completion on Busy Subscriber

There are several ways to implement this, some of which are supported in
the base spec.

Approach 1: Callee sends 181 Queued responses back to the caller until
they're not busy, and then sends a 200 OK when free.

Approach 2: Callee notes the call attempt, sends a 4XX response. When
the callee is free, it calls back the caller.

An alternate approach not available in the base spec is to use a
SUBSCRIBE/NOTIFY mechanism to have the caller ask for call state
updates. Once the callee is free, a NOTIFY gets sent to the caller, and
it can retry the callee.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 21 12:37:57 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA03064
	for <iptel-archive@odin.ietf.org>; Sun, 21 Mar 1999 12:37:56 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E0BFA52AA; Sun, 21 Mar 1999 12:33:15 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 412C752BB; Sun, 21 Mar 1999 12:33:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F52CB5.199214D2@cisco.com>
Date: Sun, 21 Mar 1999 09:30:29 -0800
From: Hussein Salama <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.03 [en] (Win95; U)
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990319080424.00860100@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Matt Squire wrote:
> 
> >> > The capacity metric could also be used to enforce policy, in addition to
> >> > its use for advertising physical resources (bandwidth, ports, ... etc).
> >> > For example, it can be used for constraining the number of calls from
> >> > one AD to another. For example, AD1 may wish not to route more than x
> >> > calls from AD2 towards +1408*. This can be achieved by setting capacity
> >> > = x in advertisement of 1408* frowarded from AD1 to Ad2.
> >>
> >> This introduces the following problem:
> >>
> >>            -- AD1--
> >>           /        \
> >>         /           \
> >>       AD2             AD3
> >>        \              /
> >>         \            /
> >>          \          /
> >>           \        /
> >>            \      /
> >>               AD4
> >>
> >> AD1 propagates a prefix to AD2, and also to AD3. The prefix is to a
> >> gateway with capacity X. So, AD1 propagates capacity X to AD2, and also
> >> capacity X to AD3. Both AD2 and AD3 advertise this to AD4, which now
> >> thinks there is a total of 2X capacity to the prefix.
> >>
> >
> >AD4 won't merge two routes to the same prefix. It will select either
> >vthe route coming from AD2 or the route coming from AD3. In both cases
> >the capacity of calls towards the prefix served by AD1 as seen by AD4
> >will still be X.
> 
> I'm guessing Jonathan was thinking about a load-balancing scenario, where
> AD4 wants to use multiple paths.  To perform load-balancing, an LS would
> have to have multiple routes.  Whether you implement this as cycling
> through a selection of individual routes, or actually selecting multiple
> routes isn't important.  Using a capacity attribute to load-balance, which
> is one of the reasons people wanted this attribute (as opposed to just
> route selection), interferes with your scenario of partitioning capacity.

I prefer calling it capacity multiplexing, not capacity partitioning.

I don't think load balancing conflicts with capacity multiplexing as
long as you don't do load balancing between two path leading to the same
egress gateway. This is easy to avoid by comparing the egress
administrative domains in the AS_PATH attribute of TBGP (or its
equivalent) for the various alternate routes. If they are identical for
two routes A and B, no load balancing should be done among A and B.

> 
> >
> >If you use such a metric, we have to make sure that it's value can only
> >decrease, but not increase, as it is propogated further from the egress
> >gateway.
> 
> I'm not sure that this is any different then the capacity attribute as
> discussed at the wg.  We decided (I think) that we'd really like to have a
> metric on the capacity attribute.  Say we make that metric "max # calls".
> 
> We've always been a little fuzzy on what happens with this attribute at
> intermediate LSs.  If an LS is just passing the (prefix,gw) address along,
> then I think its clear that it shouldn't modify the capacity attribute.
> But if the LS is inserting itself as the next hop, then it of course has
> the right to modify the attributes, capacity included.
> 
> But aggregation and deaggregation certainly screws things up.  In the above
> example, AD2 could have aggregated +1408 into +140.  Most likely the
> capacity attribute changed.  If you sum the capacity attribute (a likely
> possibility), then you lose information.  For example, AD2 could be
> advertising +140 with a capacity of 5X.  AD4 could then send 5X calls for
> +1408 even though, at AD1, there are only X calls for +1408.  Deaggregation
> causes similar problems.
> 
> This is one of the many problems with using any true metric in the face of
> aggregation and deaggregation.  Its relationship to reality is lost as
> information is aggregated and deaggregated.

I agree, that's why we have to be very careful when defining aggregation
rules.

> 
> Ignoring the problems of aggregation, in order to get any kind of
> statistical significance as the information is propagated, I think you need
> to mandate similar behavior at each LS.  Although any LS can make a
> decision to break up a capacity of X into Y+Z>X.  Say you use a rule to do
> Y+Z=1.2X, and another LS breaks up Y into A+B=1.9Y, and another LS breaks
> up Z into C+D=1.1Z, etc.  Although you could make statistical decisions,
> you would have to make assumptions about Y and Z, but these assumptions
> would most likely not be valid unless you know what decisions other LSs are
> using.  So in order to get some overall statistical significance to the
> system, I think you'd need to have a defined method for handling capacity
> in such manner at all LSs.
> 

The problem you're describing above disappears if we adopt the
semi-dynamic route advertising/withdrawal mechanism I proposed at the
IPTEL meeting. Just a refresher. We don't want to advertise the
available resources on a route, because it is a dynamic metric. However,
if the utilized resources on a given route as a fraction of the total
capacity exceed a given high water mark, we withdraw that route, and
send an explicit Update message removing it from all call routing
tables. Then when the utilized resources go below a certain low water
mark, we start re-advertising the route once again. No dynamic metrics
involved. 

I still haven't thought about the rules for computing the paremeter "a"
at an LS, where Y+Z=a*X, but I don't think coordinating the value of "a"
at the various LSs is important, because it will not cause routing calls
to egress gateways which have run out of resources, except may be in
some transient states of the network.

Let me briefly summarize the benefits of my proposal for using capacity
multiplexing and semi-dynamic route advertising/withdrawal:
- It permits policy-based sharing of resources among calls from various
origins towards the same egress point.
- The existance of a route in a call routing table means that route
actually has sufficient resources available to accept new calls.
- We can still do load balancing among routes to different egress
points.
- Avoids static partitioning of capacity.
- Avoids dynamic metrics.

The only function a fixed capacity metric (representing the total
resources of an egress point) can achieve is load balancing among routes
to different egress points, and even that is done in a random fashion,
because LSs have no idea whether or not an egress point currently has
resources available.

Hussein


> Of course, the last time I did any real mathematics was in grad school, and
> I had hair then, so maybe you know of a method for getting some statistical
> significance out of the system without mandating intermediate behavior - I
> just don't see it.
> 
> I also don't see any capacity attribute as providing "guarantees" in the
> face of aggreg/deaggreg.  Guidelines, maybe.  Heuristics, sure.  But not
> guarantees.
> 
> - Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 21 16:51:00 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA04209
	for <iptel-archive@odin.ietf.org>; Sun, 21 Mar 1999 16:51:00 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 21BCC52BB; Sun, 21 Mar 1999 16:45:27 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7BEF352C2; Sun, 21 Mar 1999 16:45:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Sun, 21 Mar 1999 16:21:17 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Christian Huitema <huitema@research.bellcore.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
In-Reply-To: <36F1C0CC.80073FA6@dnrc.bell-labs.com>
Message-ID: <Pine.LNX.4.10.9903211558010.23536-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



On Thu, 18 Mar 1999, Jonathan Rosenberg wrote:

> I think there is an important point here. To date, we have discussed two
> models for handling number translations:
> 
> 1. we do it in the IP side (perhaps via enum), before GLP, so that GLP
> advertises routable numbers only
> 2. we do it in the PSTN side, so that GLP needs to advertise "names"
> like 800 numbers
> 
> What Christian is really saying is that there is a third possibility:
> 
> 1. a partial translation is done on the IP side, yielding a provider ID
> or something like that. GLP also propagates provider ID's, so the call
> is routed there. The final translation is done inside the PSTN.
> 
> I think its a valid requirement to support this. Comments?
> 

I want to make sure that I understand what you think Christian has in
mind. Are you suggesting the following scheme:

1. An endpoint wishes to route a call to 1-800-7387225. 
2. The endpoint does some IP lookup to discover that this number has a
   provider-ID of (say) "Qwest".
3. As a result of this lookup, the endpoint knows to route the call to 
   the provider-ID "Qwest." 
4. If there is a gateway which has advertised a route to "Qwest," the 
   endpoint will send it the call?

Is your point to use "Provider IDs" as a way to aggregate routes? 

In this way, rather arbitrary aggregations of phone numbers can be hidden
behind arbitrary "provider-IDs". And then it's up to ENUM servers 
to handle such globs. 

If this is what you meant, I'd just like to add the requirement that a
phone number might belong to several "provider IDs". This can't happen 
in the PSTN, of course, but there is no business
reason that I can't allow BOTH Qwest and Sprint to route 800 numbers for
me, putting them in competition for my business. (Of course the PSTN can't
handle this, but that is not GLP's problem. this whole issue
becomes an ENUM issue, not a GLP issue.)

Sorry if I misunderstood.

Scott






---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 21 17:25:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA04449
	for <iptel-archive@odin.ietf.org>; Sun, 21 Mar 1999 17:25:53 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6C19152C3; Sun, 21 Mar 1999 16:45:50 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C5E4C52BF; Sun, 21 Mar 1999 16:45:42 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Authentication-Warning: petrack.metatel.com: scott.petrack owned process doing -bs
Date: Sun, 21 Mar 1999 15:25:31 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
To: Christian Huitema <huitema@research.bellcore.com>
Cc: iptel <iptel@lists.research.bell-labs.com>, e164-to-ip@lserv.vocaltec.com
Subject: Re: GLP, number portability and wireless subscribers
In-Reply-To: <36ECA53F.7F8A06F9@research.bellcore.com>
Message-ID: <Pine.LNX.4.10.9903211518080.21679-100000@petrack.metatel.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



On Mon, 15 Mar 1999, Christian Huitema wrote:

> I think that this all go back to a two phases approach for telephony
> routing:
> 
> 1) take a dip in a mapping directory to understand the "provider" for
> the 
>    phone number, which will result in a <provider, number> pair -- maybe
>    with a specific provider that says "e164"
> 
> 2) route based on the pair.

Well, if "e164" is a provider, then I can see why there was some
confusion.

I am getting more and more partial to the basic requirement for the 
ENUM service being a translation from "Telephone Numbers" to "URLs". 

That is admittedly still a bit fuzzy, but I think it covers what everyone
wants: it will cover things like:

input: 1800xxxyyyy  to 
output: phone:1800xxxyyyy@someprovider.com

as well as

input: 1800xxxyyyy to 
output phone:1800xxxyyyy@provider1.com,sip:1800xxxyyyy@provider2.com 
			
etc. etc. 

It would even allow me to list two separate phone URLs, in case I want to 
allow to competitors to give me 800 number service.The URLs that are 
returned could indicate "provider" in all the senses of
the word. 

But this will be part of the upcoming ENUM work.

Scott


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 05:28:33 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA16545
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 05:28:33 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2D2FB52C1; Mon, 22 Mar 1999 04:09:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8725652C2; Mon, 22 Mar 1999 04:09:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <29F16C82368FD111914F00805F858E4F02096EF9@exchvs01.bcs.cs.philips.com>
From: "Derks, Frank " <F.Derks@pbc.be.philips.com>
To: "Hutton, Charles" <charles.hutton@attws.com>,
        iptel@lists.research.bell-labs.com,
        "'nvnmurthy@seeyes.co.in'" <nvnmurthy@seeyes.co.in>
Subject: RE: Reg. RTP Header compression/decompression - Support for it in
	 CISCO equipment
Date: Mon, 22 Mar 1999 10:08:23 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.0.1460.8)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi Charles,

to be correct, you should say that RTP is supported by IOS version 11.3. The
thing that we found, was that it is only effective for the port range
16384-65535! Be aware of this.

Frank

-----------------------------------------------------
Frank Derks                    |Tel  +31 35 6893238 |
Advanced Development           |Fax  +31 35 6891030 |
Philip Business Communications |P.O. Box 32         |
                               |1200 JD  Hilversum  |
                               |The Netherlands     |
----------------------------------------------------|
E-mail: mailto:f.derks@pbc.be.philips.com           |
WWW: http://www.business-comms.be.philips.com       |
-----------------------------------------------------



-----Original Message-----
From: Hutton, Charles [mailto:charles.hutton@attws.com]
Sent: 19 March 1999 18:40
To: iptel@lists.research.bell-labs.com; 'nvnmurthy@seeyes.co.in'
Subject: RE: Reg. RTP Header compression/decompression


SInce RFC2508 is IP-UDP-RTP compression in one step, it's not clear to me
whether you are interested in that or just RTP compression as you wrote.

Assuming you are interested in IP-UDP-RTP all together compression, you
might want to look at:

ATM Forum STR-SAA-RMOA-01.00
ATM Forum SAA-98-0139

But RFC2508 seems to be the answer right now. I don't know what other
vendors are doing, but Cisco supports it in all their products that I have
looked at. 

Charles Hutton
Systems Engineering
Wireless Local Technology Group
AT&T Wireless Services
PO Box 97059
Redmond, WA 98073
14520 NE 87th Street
Redmond, WA 98052 
425-702-2938 Voice
425-702-2568 FAX
charles.hutton@attws.com

> ----------
> From: 	Venkata Narasimha Murthy
> Nanduri[SMTP:nvnmurthy@seeyes.co.in]
> Sent: 	Friday, March 19, 1999 1:10 AM
> To: 	iptel@lists.research.bell-labs.com
> Subject: 	Reg. RTP Header compression/decompression
> 
> Hai,
> 
> I have to work on RTP header compession/decompression.
> For this I am reading RFC1144 - TCP/IP header comression  and
> RFC 2508 - RTP header compression.
> 
> Other than these, R there any other information sources which will be
> useful?
> Is there any sample implementation available on the net?
> 
> Any information reg. this is appreciated.
> 
> Thanks in advance
> Murthy NVN
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 

---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 08:28:21 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA17355
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 08:28:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 98E8752B7; Mon, 22 Mar 1999 07:49:15 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0578052C2; Mon, 22 Mar 1999 07:49:14 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: amittal@hss.hns.com
To: iptel@lists.research.bell-labs.com
Message-ID: <6525673C.0046B00A.00@sampark.hss.hns.com>
Date: Mon, 22 Mar 1999 18:22:49 +0530
Subject: ILS
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi

 Is Internet Location Server (ILS) a standard? Would appreciate any
pointers as to where I could find more details. How does it compare to a
directory service such as LDAP?

Thanks!
Alok


---
Alok Mittal
Technical Leader, Hughes Software Systems
Tel: (91) 80 228 6390, 91, 92    Fax: (91) 80 228 6393
E-mail: amittal@hss.hns.com



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 09:48:28 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA18013
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 09:48:27 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E898152C2; Mon, 22 Mar 1999 09:43:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5295752C7; Mon, 22 Mar 1999 09:43:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F65A26.3545575B@seeyes.co.in>
Date: Mon, 22 Mar 1999 20:26:39 +0530
From: Venkata Narasimha Murthy Nanduri <nvnmurthy@seeyes.co.in>
Reply-To: nvnmurthy@seeyes.co.in
Organization: SeeYes
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: "Hutton, Charles" <charles.hutton@attws.com>
Cc: iptel@lists.research.bell-labs.com
Subject: Re: Reg. RTP Header compression/decompression
References: <AF62C6C0EC3DD211929900805FEA97CF60808E@wltg-msg01.mccaw-stg.com>
Content-Type: multipart/mixed;
 boundary="------------69B1FEDD8AF8CA2333D8282D"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.
--------------69B1FEDD8AF8CA2333D8282D
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hai Charles,
Thank U very much for the information. I could not find the below
listed documents. Could U please send me a copy of them if U have?

Thanks in advance
Murthy NVN

"Hutton, Charles" wrote:

> SInce RFC2508 is IP-UDP-RTP compression in one step, it's not clear to me
> whether you are interested in that or just RTP compression as you wrote.
>
> Assuming you are interested in IP-UDP-RTP all together compression, you
> might want to look at:
>
> ATM Forum STR-SAA-RMOA-01.00
> ATM Forum SAA-98-0139
>
> But RFC2508 seems to be the answer right now. I don't know what other
> vendors are doing, but Cisco supports it in all their products that I have
> looked at.
>
> Charles Hutton
> Systems Engineering
> Wireless Local Technology Group
> AT&T Wireless Services
> PO Box 97059
> Redmond, WA 98073
> 14520 NE 87th Street
> Redmond, WA 98052
> 425-702-2938 Voice
> 425-702-2568 FAX
> charles.hutton@attws.com
>
> > ----------
> > From:         Venkata Narasimha Murthy
> > Nanduri[SMTP:nvnmurthy@seeyes.co.in]
> > Sent:         Friday, March 19, 1999 1:10 AM
> > To:   iptel@lists.research.bell-labs.com
> > Subject:      Reg. RTP Header compression/decompression
> >
> > Hai,
> >
> > I have to work on RTP header compession/decompression.
> > For this I am reading RFC1144 - TCP/IP header comression  and
> > RFC 2508 - RTP header compression.
> >
> > Other than these, R there any other information sources which will be
> > useful?
> > Is there any sample implementation available on the net?
> >
> > Any information reg. this is appreciated.
> >
> > Thanks in advance
> > Murthy NVN
> >
> >
> > ---------
> > This message came from the IETF IPTEL Working Group Mailing List.
> >

--
-------------------------------------
Venkata Narasimha Murthy Nanduri,
Software Engineer,
SeeYes Network Technologies Pvt. Ltd
H-4, DD Colony, Bagh Amberpet,
Hyderabad, Andhra Pradesh,
India - 500 013

Office : 91-40-760 7842
Home   : 91-40-712 3437
Fax    : 91-40-761 2860
-------------------------------------


--------------69B1FEDD8AF8CA2333D8282D
Content-Type: text/x-vcard; charset=us-ascii;
 name="nvnmurthy.vcf"
Content-Description: Card for Venkata Narasimha Murthy Nanduri
Content-Disposition: attachment;
 filename="nvnmurthy.vcf"
Content-Transfer-Encoding: 7bit

begin:vcard 
n:Nanduri;Venkata Narasimha Murthy
tel;fax:91-40-761 2860
tel;home:91-40-712 3437
tel;work:91-40-760 7842
x-mozilla-html:FALSE
adr:;;;;;;
version:2.1
email;internet:nvnmurthy@seeyes.co.in
x-mozilla-cpt:;-8368
fn:Murthy NVN
end:vcard

--------------69B1FEDD8AF8CA2333D8282D--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 11:12:13 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA20206
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 11:12:13 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C596052CA; Mon, 22 Mar 1999 10:21:11 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1AD5D52CC; Mon, 22 Mar 1999 10:21:10 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F65F4D.FFB933FF@dnrc.bell-labs.com>
Date: Mon, 22 Mar 1999 10:18:37 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Scott Petrack <scott.petrack@metatel.com>
Cc: Christian Huitema <huitema@research.bellcore.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <Pine.LNX.4.10.9903211558010.23536-100000@petrack.metatel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Scott Petrack wrote:
> 
> On Thu, 18 Mar 1999, Jonathan Rosenberg wrote:
> 
> > I think there is an important point here. To date, we have discussed two
> > models for handling number translations:
> >
> > 1. we do it in the IP side (perhaps via enum), before GLP, so that GLP
> > advertises routable numbers only
> > 2. we do it in the PSTN side, so that GLP needs to advertise "names"
> > like 800 numbers
> >
> > What Christian is really saying is that there is a third possibility:
> >
> > 1. a partial translation is done on the IP side, yielding a provider ID
> > or something like that. GLP also propagates provider ID's, so the call
> > is routed there. The final translation is done inside the PSTN.
> >
> > I think its a valid requirement to support this. Comments?
> >
> 
> I want to make sure that I understand what you think Christian has in
> mind. Are you suggesting the following scheme:
> 
> 1. An endpoint wishes to route a call to 1-800-7387225.
> 2. The endpoint does some IP lookup to discover that this number has a
>    provider-ID of (say) "Qwest".
> 3. As a result of this lookup, the endpoint knows to route the call to
>    the provider-ID "Qwest."
> 4. If there is a gateway which has advertised a route to "Qwest," the
>    endpoint will send it the call?

That was basically the idea.

> 
> Is your point to use "Provider IDs" as a way to aggregate routes?
> 
> In this way, rather arbitrary aggregations of phone numbers can be hidden
> behind arbitrary "provider-IDs". And then it's up to ENUM servers
> to handle such globs.

Yup. Its not clear to me yet how to represent such provider ID's. In one
way, they can (like the "scope" attribute we've been discussing), be
consider part of the address itself. So, if some gateway can route 800
numbers belonging to Bell Atlantic (who is provider 3, for example),
they might advertise the route:

3 1 800 *

On the other hand, I'm not so certain this is part of the address; its
more like an ID which represents a group of addresses (as you mention).
Perhaps what we need is two ways to represent "phone numbers" in GLP -
one is the phone-prefix type we've been discussing to date. Another
would be provider ID. That, of course, leads me to wonder about other
types of "addresses" in the PSTN one could advertise - point codes, for
example. Its a slippery slope....

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 11:20:05 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA20409
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 11:20:05 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1F6B752CD; Mon, 22 Mar 1999 10:09:43 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5592652CA; Mon, 22 Mar 1999 10:09:37 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <19990322150736.25759.rocketmail@send103.yahoomail.com>
Date: Mon, 22 Mar 1999 07:07:36 -0800 (PST)
From: Tom Gray <tom_gray_intp@yahoo.com>
Subject: Re: GLP, number portability and wireless subscribers
To: iptel@lists.research.bell-labs.com, tom_gray_intp@yahoo.com
Cc: Christian Huitema <huitema@research.bellcore.com>,
        iptel <iptel@lists.research.bell-labs.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



Hi

Has the issue of a private network been considered here along with the
issues of tenanting? I'm new so please excuse me if this has been
considered. A server may be used to service several enterprises say in
a shared office building. Each may well want their help desk to be
HELP or 4357, the president to be 1111, etc. They will certainly not
agree to have special tenant suffixes applied. Additionally it will
probably be a requirement to extend this  over multiple servers with
overlapping sets of tenants at each server.

Tom Gray




_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 11:27:25 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA20645
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 11:27:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 03FD852CB; Mon, 22 Mar 1999 10:09:26 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 601AC52CC; Mon, 22 Mar 1999 10:09:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <19990322150736.25759.rocketmail@send103.yahoomail.com>
Date: Mon, 22 Mar 1999 07:07:36 -0800 (PST)
From: Tom Gray <tom_gray_intp@yahoo.com>
Subject: Re: GLP, number portability and wireless subscribers
To: iptel@lists.research.bell-labs.com, tom_gray_intp@yahoo.com
Cc: Christian Huitema <huitema@research.bellcore.com>,
        iptel <iptel@lists.research.bell-labs.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



Hi

Has the issue of a private network been considered here along with the
issues of tenanting? I'm new so please excuse me if this has been
considered. A server may be used to service several enterprises say in
a shared office building. Each may well want their help desk to be
HELP or 4357, the president to be 1111, etc. They will certainly not
agree to have special tenant suffixes applied. Additionally it will
probably be a requirement to extend this  over multiple servers with
overlapping sets of tenants at each server.

Tom Gray




_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 11:44:52 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA21268
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 11:44:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A6BA652C7; Mon, 22 Mar 1999 09:55:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0C8EB52CA; Mon, 22 Mar 1999 09:55:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <19990322145430.7904.rocketmail@send102.yahoomail.com>
Date: Mon, 22 Mar 1999 06:54:30 -0800 (PST)
From: Tom Gray <tom_gray_intp@yahoo.com>
Subject: CPL -- Two questions about is future
To: iptel@lists.research.bell-labs.com, tom_gray_intp@yahoo.com
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


Hi

(1)
Reading the CPL documents, I get the impression is that only the first
part of a larger envisaged cpapbility is being described.

(1a)
 What are the currrent features that being planned for specificaiton
in CPL?

(1b)
How do these fit into the larger planned capability?


(2)
In the requirements document, mention is made of the feature
interaction problem of siganlling ambiguity. The example given is the
classic 'call waiting' and 'three way calling' interaction in which
the user may be confused by since the hook switch signal given may
result in two incompatible actions - answering the call waiting party
or going to conference dial tone to add another party to a three way
call. The answer given  in the requirement document is to refer to the
explicit nature of signalling in internet telephony protocols as
resolving this ambiguity.

This seems to miss the  point of the interaction. The interation
problem does not arise here internal to the protocols. There is no
indeterminacy which results in unpredictable system operation. Rather
the interaction occurs because of race condition between the user
becoming aware of call waiting and his actions. The interaction
manifiests itself at the user in his confusion.

We must recall that some people will be using ordinary telephones to
access IPTel servcie and the hook switch ambiguity will still cause
confusion. Secondly even if someone is using a special set which has
more explicit signals, the problem still arises since:

a) keys or soft keys must be provided and controlled which indicate
the possibiliites at any point in the  call. Thus the software which
controls the set must handle the interaction and worse yet must be
aware  of the state of the call in the server. A whole host of new
interaction bugs lie await here.

b) since it is user confusion which is the manisfistation of the
undesired interaction. The user must be trained on the use of the
multiple possibilities presented by such special set so that one cause
of confusion is not subsitituted for another.

In brief, I do no think that the  explicit signalling in IPTEL
protocols has much to add to the solution to this interaction. 


Tom Gray
_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 17:18:13 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA01475
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 17:18:13 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1683552C8; Mon, 22 Mar 1999 16:07:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 64D1152C9; Mon, 22 Mar 1999 16:07:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Dean Willis" <Dean.Willis@MCI.COM>
To: "Jonathan Rosenberg" <jdrosen@dnrc.bell-labs.com>,
        "Scott Petrack" <scott.petrack@metatel.com>
Cc: "Christian Huitema" <huitema@research.bellcore.com>,
        "iptel" <iptel@lists.research.bell-labs.com>
Subject: RE: GLP, number portability and wireless subscribers
Date: Mon, 22 Mar 1999 15:04:35 -0600
Message-ID: <000001be74a7$974fc360$aaa22ca6@dwillispc3.mcit.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Importance: Normal
In-Reply-To: <36F65F4D.FFB933FF@dnrc.bell-labs.com>
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Jonathan said:
> Scott Petrack wrote:
> > Is your point to use "Provider IDs" as a way to aggregate routes?
> >
> > In this way, rather arbitrary aggregations of phone numbers
> can be hidden
> > behind arbitrary "provider-IDs". And then it's up to ENUM servers
> > to handle such globs.
>
> Yup. Its not clear to me yet how to represent such provider
> ID's. In one
> way, they can (like the "scope" attribute we've been discussing), be
> consider part of the address itself. So, if some gateway can route 800
> numbers belonging to Bell Atlantic (who is provider 3, for example),
> they might advertise the route:
>

Some phrases to consider before using the provider-ID for aggregation:

1) Dual Homed Network: i.e., This number is serviced by GTE and MCI --
can't do this today because of provider aggregation.

2) Number Portability (local AND 800): A number once serviced by carrier
A that is now serviced exclusively by carrier B. LNP as we know and love
it as a relic of provider aggregation.

--
Dean


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 17:29:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA01979
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 17:29:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D1B0452C5; Mon, 22 Mar 1999 16:37:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 263D052D5; Mon, 22 Mar 1999 16:37:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Mon, 22 Mar 1999 16:35:28 -0500 (EST)
Message-Id: <199903222135.QAA48125@conrail.cs.columbia.edu>
From: Jonathan Lennox <lennox@conrail.cs.columbia.edu>
To: John Hearty <John.H.Hearty@mci.com>
Cc: Confctrl <confctrl@ISI.EDU>, IPTel <iptel@lists.research.bell-labs.com>
Subject: Caller preference draft and CPL
In-Reply-To: <19990319161453.CSVK20027@localHost>
References: <19990319161453.CSVK20027@localHost>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

On Fri, March 19 1999, "John Hearty" wrote to "Confctrl" saying:

>  I think the header/parameter approach to express the type of termination
> the caller wants can work well.  I agree with Jonathan that interacting
> CPL scripts between the caller and callee can get extremely complex.
> With a finite set of preference parameters to consider, it should be
> possible to update the CPL draft to allow these to be considered in
> callee CPL scripts.  That seems to be the best place to direct the
> interoperability efforts, since CPL can have more flexibility to deal
> with it.

Yes, I'm planning on considering this for the next draft of the CPL, though
I'm not quite sure yet how the parameters would work -- as Henning points
out, there's a complicated issue about what happens if these parameters
aren't specified.

One possibility (which has the virtue of simplicity) is to have a "location
filter" which simply filters the currently chosen location set according to
the caller's preferences.  This might even, potentially, be
parameterizable -- the filter could enable or disable certain preference
types.  The syntax for this is unclear, though.

Discussion about this issue is welcome, though it should probably be
restricted to the iptel mailing list
<mailto:iptel@lists.research.bell-labs.com>.  This message has been Cc'd
there.

-- 
Jonathan Lennox
lennox@cs.columbia.edu

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 22 19:23:46 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA05393
	for <iptel-archive@odin.ietf.org>; Mon, 22 Mar 1999 19:23:45 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5265352CC; Mon, 22 Mar 1999 17:27:48 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 240AA52BF; Mon, 22 Mar 1999 17:27:47 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Mon, 22 Mar 1999 17:26:36 -0500 (EST)
Message-Id: <199903222226.RAA48174@conrail.cs.columbia.edu>
From: Jonathan Lennox <lennox@conrail.cs.columbia.edu>
To: Tom Gray <tom_gray_intp@yahoo.com>
Cc: iptel@lists.research.bell-labs.com
Subject: Re: CPL -- Two questions about is future
In-Reply-To: <19990322145430.7904.rocketmail@send102.yahoomail.com>
References: <19990322145430.7904.rocketmail@send102.yahoomail.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

On Mon, March 22 1999, "Tom Gray" wrote to "iptel@lists.research.bell-labs.com, tom_gray_intp@yahoo.com" saying:

> (1)
> Reading the CPL documents, I get the impression is that only the first
> part of a larger envisaged cpapbility is being described.
> 
> (1a)
>  What are the currrent features that being planned for specificaiton
> in CPL?
> 
> (1b)
> How do these fit into the larger planned capability?

There are a huge number of things that could, in principle, be controlled by
a CPL: automated voice-response systems, which need mid-call information
about the media stream; complicated call-queueing and call-distribution,
where the distribution point needs to know about end systems' state, and
events in one call trigger behavior in another call; end-system API control,
with distinctive ringing and the like; and so on and so forth.  Some of
these were mentioned in the examples section of Requirements draft, but many
more could be invented.

The number of possibilities is so vast, though, that I thought it would be
better to define the simple, basic functionality -- those behaviors which
can be expressed as "proxy", "reject", and "redirect" of a single call setup
-- in a scheme which would be extensible enough to allow all the other
functionality to be included.  As long as we have a general idea of the
sorts of more complex features we want to add in the future, I think this
will be workable.

In general, I think the additional action nodes would consist of in-call
nodes, which come from an out-arrow of proxy, as mentioned in the current
CPL draft; inter-call nodes, "queue" and "distribute"; and end-system nodes,
such as "alert".  The details of these remain to be worked out, but I think
this is roughly the level of abstraction the CPL should use.
 
> (2)
> In the requirements document, mention is made of the feature
> interaction problem of siganlling ambiguity. The example given is the
> classic 'call waiting' and 'three way calling' interaction in which
> the user may be confused by since the hook switch signal given may
> result in two incompatible actions - answering the call waiting party
> or going to conference dial tone to add another party to a three way
> call. The answer given  in the requirement document is to refer to the
> explicit nature of signalling in internet telephony protocols as
> resolving this ambiguity.
> 
> This seems to miss the  point of the interaction. The interation
> problem does not arise here internal to the protocols. There is no
> indeterminacy which results in unpredictable system operation. Rather
> the interaction occurs because of race condition between the user
> becoming aware of call waiting and his actions. The interaction
> manifiests itself at the user in his confusion.
> 
> We must recall that some people will be using ordinary telephones to
> access IPTel servcie and the hook switch ambiguity will still cause
> confusion. Secondly even if someone is using a special set which has
> more explicit signals, the problem still arises since:
> 
> a) keys or soft keys must be provided and controlled which indicate
> the possibiliites at any point in the  call. Thus the software which
> controls the set must handle the interaction and worse yet must be
> aware  of the state of the call in the server. A whole host of new
> interaction bugs lie await here.
> 
> b) since it is user confusion which is the manisfistation of the
> undesired interaction. The user must be trained on the use of the
> multiple possibilities presented by such special set so that one cause
> of confusion is not subsitituted for another.
> 
> In brief, I do no think that the  explicit signalling in IPTEL
> protocols has much to add to the solution to this interaction. 

Well, clearly there is still a potential problem here, but it isn't a
*signalling* problem; it's a *user interface* problem.  Whereas traditional
analog telephones have this one signal, "flash", with which they communicate
to their servers, there's no concept of anything like this with SIP or
H.323.  (I'm not sure what's happening with Megaco, so I won't address that
here.)  Instead, there's no way for an IPTel device to confuse a "change
current call connection" signal with a "begin to add additional party"
signal (not to imply that the actual signals would be precisely like this,
but hopefully you understand the idea).

From the point of view of an IPTel system, a traditional telephone isn't an
end system; the end system is the traditional telephone combined with its
IPTel gateway.  Any ambiguity in the signalling between the different parts
of the end system is entirely the end system's problem.  Similarly, an
Ethernet phone which looks like and emulates a traditional telephone needs
to send the appropriate IPTel signalling information, whether or not it's
trying to use "flash switchhook" as its user interface for these events.

I don't understand your point about the client knowing "the state of the
call in the server".  Server don't, can't, and shouldn't know all the
information about the state of clients; this is inherent in the IPTel
architecture.  Servers have to compensate for this.  What scenario are you
envisioning here?

-- 
Jonathan Lennox
lennox@cs.columbia.edu

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar 23 02:07:00 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA18018
	for <iptel-archive@odin.ietf.org>; Tue, 23 Mar 1999 02:07:00 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C13AB52B7; Mon, 22 Mar 1999 23:33:30 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 594AC52C2; Mon, 22 Mar 1999 23:17:06 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F711E1.7B3E9B8D@dnrc.bell-labs.com>
Date: Mon, 22 Mar 1999 23:00:33 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Tom Gray <tom_gray_intp@yahoo.com>
Cc: iptel@lists.research.bell-labs.com,
        Christian Huitema <huitema@research.bellcore.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <19990322150736.25759.rocketmail@send103.yahoomail.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

It has been discussed. Its largely outside of our scope, since GLP is an
interdomain protocol. But, nothing prevents an enterprise from using it
internally. I suspect our eventual representation for phone numbers will
be flexible enough to allow internal numbers (although not things like
HELP unless you translate it to the number equivalent - 4357).

-Jonathan R.

Tom Gray wrote:
> 
> Hi
> 
> Has the issue of a private network been considered here along with the
> issues of tenanting? I'm new so please excuse me if this has been
> considered. A server may be used to service several enterprises say in
> a shared office building. Each may well want their help desk to be
> HELP or 4357, the president to be 1111, etc. They will certainly not
> agree to have special tenant suffixes applied. Additionally it will
> probably be a requirement to extend this  over multiple servers with
> overlapping sets of tenants at each server.
> 
> Tom Gray
> 
> _________________________________________________________
> DO YOU YAHOO!?
> Get your free @yahoo.com address at http://mail.yahoo.com
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar 23 02:23:32 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA18073
	for <iptel-archive@odin.ietf.org>; Tue, 23 Mar 1999 02:23:31 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C832552BB; Mon, 22 Mar 1999 23:50:37 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B8C9452C9; Mon, 22 Mar 1999 23:34:37 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F71817.D44860C1@dnrc.bell-labs.com>
Date: Mon, 22 Mar 1999 23:27:03 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: Matt Squire <msquire@nortelnetworks.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990319080424.00860100@bl-mail1.corpeast.baynetworks.com> <36F52CB5.199214D2@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein Salama wrote:
> 
> I prefer calling it capacity multiplexing, not capacity partitioning.
> 
> I don't think load balancing conflicts with capacity multiplexing as
> long as you don't do load balancing between two path leading to the same
> egress gateway. This is easy to avoid by comparing the egress
> administrative domains in the AS_PATH attribute of TBGP (or its
> equivalent) for the various alternate routes. If they are identical for
> two routes A and B, no load balancing should be done among A and B.

If there is an AS_PATH; if there is aggregation, this gets collected
into an AS_SET, no? In this case, one cannot tell which domain the
egress is actually in. In fact, an aggregate can represent a whole bunch
of gateways in a whole bunch of domains, each with different capacities.
I'm not convinced we need to get really worried about this; the capacity
metric of an aggregate can never be anything more than a "hint" in any
case, as Matt pointed out. As the route gets propagated more and more,
the attribute has less and less real meaning.

As far as the unit itself, can we agree on the purposely-vague "number
of calls" as a unit?

> The problem you're describing above disappears if we adopt the
> semi-dynamic route advertising/withdrawal mechanism I proposed at the
> IPTEL meeting. Just a refresher. We don't want to advertise the
> available resources on a route, because it is a dynamic metric. However,
> if the utilized resources on a given route as a fraction of the total
> capacity exceed a given high water mark, we withdraw that route, and
> send an explicit Update message removing it from all call routing
> tables. Then when the utilized resources go below a certain low water
> mark, we start re-advertising the route once again. No dynamic metrics
> involved.

Well, one thing at a time. I think your proposal is orthogonal to the
capacity metric itself. I suspect there's going to be a long discussion
around this proposal (I have some concerns), but I'd rather like to
tackle the basic attributes we can at least agree we need, first.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar 23 06:04:17 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA19314
	for <iptel-archive@odin.ietf.org>; Tue, 23 Mar 1999 06:04:15 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B0A6352AA; Mon, 22 Mar 1999 22:41:33 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0957552BB; Mon, 22 Mar 1999 22:41:32 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F70D4F.C048F572@dnrc.bell-labs.com>
Date: Mon, 22 Mar 1999 22:41:03 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Jonathan Lennox <lennox@conrail.cs.columbia.edu>
Cc: Tom Gray <tom_gray_intp@yahoo.com>, iptel@lists.research.bell-labs.com
Subject: Re: CPL -- Two questions about is future
References: <19990322145430.7904.rocketmail@send102.yahoomail.com> <199903222226.RAA48174@conrail.cs.columbia.edu>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Jonathan Lennox wrote:
> 
> > (1)
> > Reading the CPL documents, I get the impression is that only the first
> > part of a larger envisaged cpapbility is being described.

Yes.

> >
> > (1a)
> >  What are the currrent features that being planned for specificaiton
> > in CPL?

Predominantly call-routing related features, along the lines of
call-redirect based on X, where X can be caller, time of day, and other
call parameters. There is also call reject, call forward, and email
notification, also based on the same parameters. You should check out
the actual CPL document, which describes these in some detail.

> >
> > (1b)
> > How do these fit into the larger planned capability?
> 
> The number of possibilities is so vast, though, that I thought it would be
> better to define the simple, basic functionality -- those behaviors which
> can be expressed as "proxy", "reject", and "redirect" of a single call setup
> -- in a scheme which would be extensible enough to allow all the other
> functionality to be included.  As long as we have a general idea of the
> sorts of more complex features we want to add in the future, I think this
> will be workable.

KISS (keep it simple stupid) is *THE* primary engineering goal for
CPLv1. This stuff is fairly new, and we need to do some basic, yet
useful stuff first to get operational experience.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar 23 09:33:37 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA20814
	for <iptel-archive@odin.ietf.org>; Tue, 23 Mar 1999 09:33:37 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5961A52BF; Mon, 22 Mar 1999 22:59:11 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B8B7F52BB; Mon, 22 Mar 1999 22:59:09 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F70E73.B5EA8B37@dnrc.bell-labs.com>
Date: Mon, 22 Mar 1999 22:45:55 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Dean Willis <Dean.Willis@MCI.COM>
Cc: Scott Petrack <scott.petrack@metatel.com>,
        Christian Huitema <huitema@research.bellcore.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <000001be74a7$974fc360$aaa22ca6@dwillispc3.mcit.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Dean Willis wrote:
> 
> > Yup. Its not clear to me yet how to represent such provider
> > ID's. In one
> > way, they can (like the "scope" attribute we've been discussing), be
> > consider part of the address itself. So, if some gateway can route 800
> > numbers belonging to Bell Atlantic (who is provider 3, for example),
> > they might advertise the route:
> >
> 
> Some phrases to consider before using the provider-ID for aggregation:
> 
> 1) Dual Homed Network: i.e., This number is serviced by GTE and MCI --
> can't do this today because of provider aggregation.

This problem needs to be solved with an intelligent database mechanism
in front of GLP. Basically, one would take +1 555-1234, and obtain two
provider ID's - one for GTE, and one for MCI. THen, GLP contains routes
to tell you how best to reach each. The entity which did the translation
can then decide which provider to connect to (also outside of GLP).

> 
> 2) Number Portability (local AND 800): A number once serviced by carrier
> A that is now serviced exclusively by carrier B. LNP as we know and love
> it as a relic of provider aggregation.

Same story as above. 

I think Dean's point is that the set of numbers aggregated into a
provider ID is a complex thing. Agreed. My point is that, given a
translation service which can work this stuff out, it might be nice to
be able to route to a provider ID. 

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar 23 17:09:58 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA27817
	for <iptel-archive@odin.ietf.org>; Tue, 23 Mar 1999 17:09:58 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 50B2752C3; Tue, 23 Mar 1999 17:05:05 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8CB9E52C6; Tue, 23 Mar 1999 17:05:04 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F80C69.C96744D4@lucent.com>
Date: Tue, 23 Mar 1999 16:49:29 -0500
From: Binshi Cao <bcao@lucent.com>
Organization: Lucent Technologies
X-Mailer: Mozilla 4.06 [en]C-EMS-1.4  (Win95; U)
MIME-Version: 1.0
To: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: Looking for large scale VoIP Equipment
References: <000401be6fde$0cfa36a0$70182ca6@vai.mci201.com>
Content-Type: multipart/mixed; boundary="------------D83DF62AAE2025EC78ED360C"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.
--------------D83DF62AAE2025EC78ED360C
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Do you talk to Lucent Technologies for their PacketStar VoIP products: Gateway
1000, Gatekeeper 1000, etc.?

Hope they will satisfy your needs.

Binshi Cao.


villuzzi wrote:
> 
> I am looking to either migrate then replace or upgrade large scale PBX
> equipment with VoIP capable gear. Any one have vendor information for large
> scale VoIP capable equipment with feature rich PBX like capabilities?
> 
> Thanks,
> 
> Vince
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
--------------D83DF62AAE2025EC78ED360C
Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf"
Content-Description: Card for Binshi Cao
Content-Disposition: attachment; filename="vcard.vcf"
Content-Transfer-Encoding: 7bit

begin:          vcard
fn:             Binshi Cao
n:              Cao;Binshi
org:            Lucent Technologies
adr:            2C-122;;600, mountain Av.;Murray Hill;NJ;07974;U.S.A.
email;internet: bcao@lucent.com
title:          Member of Technical Staff
tel;work:       (908) 582 5824
tel;fax:        (908) 582 6729
tel;home:       (908) 541 9774
x-mozilla-cpt:  ;0
x-mozilla-html: FALSE
version:        2.1
end:            vcard


--------------D83DF62AAE2025EC78ED360C--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar 23 22:52:12 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA09465
	for <iptel-archive@odin.ietf.org>; Tue, 23 Mar 1999 22:52:11 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A9DE252C5; Tue, 23 Mar 1999 22:47:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 092A652C7; Tue, 23 Mar 1999 22:47:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F8600C.758FD3ED@cisco.com>
Date: Tue, 23 Mar 1999 19:46:20 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: Christian Huitema <huitema@research.bellcore.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <Pine.LNX.4.04.9903111856190.6794-100000@petrack.metatel.com> <36ECA53F.7F8A06F9@research.bellcore.com> <36F1C0CC.80073FA6@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Jonathan Rosenberg wrote:
> 
> Christian Huitema wrote:
> >
> > Scott Petrack wrote:
> > >
> > > The fact that this situation is a mess in the PSTN does not mean that it
> > > must be a mess in GLP. And the fact that the "scope" of routing
> > > numbers in the PSTN is the "provider ID" does not
> > > mean that this is the only scope that we should allow within GLP.
> > > Certainly, 1-700 numbers will absolutely require a scope which is a
> > > "provider id", but 800 numbers might not. For example, my company owns a
> > > certain 800 number, and it will resolve to our office telephone number no
> > > matter what provider we happen to choose to give us service. Nothing in
> > > GLP will require that the scope will be "provider id."
> >
> > I think that this all go back to a two phases approach for telephony
> > routing:
> >
> > 1) take a dip in a mapping directory to understand the "provider" for
> > the
> >    phone number, which will result in a <provider, number> pair -- maybe
> >    with a specific provider that says "e164"
> >
> > 2) route based on the pair.
> 
> I think there is an important point here. To date, we have discussed two
> models for handling number translations:
> 
> 1. we do it in the IP side (perhaps via enum), before GLP, so that GLP
> advertises routable numbers only
> 2. we do it in the PSTN side, so that GLP needs to advertise "names"
> like 800 numbers
> 
> What Christian is really saying is that there is a third possibility:
> 
> 1. a partial translation is done on the IP side, yielding a provider ID
> or something like that. GLP also propagates provider ID's, so the call
> is routed there. The final translation is done inside the PSTN.

Does the above discussion apply to 800 numbers only? Otherwise, we no
longer need route selection, because we won't do route selection among
routes being advertised by different providers.

Hussein

> 
> I think its a valid requirement to support this. Comments?
> 
> -Jonathan R.
> 
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX: (732) 834-5379                         Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Mar 23 23:47:48 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA10424
	for <iptel-archive@odin.ietf.org>; Tue, 23 Mar 1999 23:47:48 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1EC4B52C2; Tue, 23 Mar 1999 23:43:22 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8393A52C8; Tue, 23 Mar 1999 23:43:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F86D4B.1833F7BD@dnrc.bell-labs.com>
Date: Tue, 23 Mar 1999 23:42:51 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: Christian Huitema <huitema@research.bellcore.com>,
        Scott Petrack <scott.petrack@metatel.com>,
        iptel <iptel@lists.research.bell-labs.com>
Subject: Re: GLP, number portability and wireless subscribers
References: <Pine.LNX.4.04.9903111856190.6794-100000@petrack.metatel.com> <36ECA53F.7F8A06F9@research.bellcore.com> <36F1C0CC.80073FA6@dnrc.bell-labs.com> <36F8600C.758FD3ED@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> > What Christian is really saying is that there is a third possibility:
> >
> > 1. a partial translation is done on the IP side, yielding a provider ID
> > or something like that. GLP also propagates provider ID's, so the call
> > is routed there. The final translation is done inside the PSTN.
> 
> Does the above discussion apply to 800 numbers only? Otherwise, we no
> longer need route selection, because we won't do route selection among
> routes being advertised by different providers.

Different notion of providers. We use this term loosely a lot and it has
gotten us into trouble. There are two providers here:

1. the gateway provider,
2. the telco which owns this 800 (or other) number - like Bell Atlantic
(number provider)

So, the translation would indicate which number provider to use. Since
an arbitrary number of gateways can reach this number provider, there
still needs to be a selection among gateway providers. So, directly
answering your question, this partial translation could be done for
non-800 numbers (such as LNP numbers), but we still need GLP to find the
gateway.


-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 06:29:52 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA19592
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 06:29:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id ECDE952C7; Wed, 24 Mar 1999 06:25:13 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4292552C8; Wed, 24 Mar 1999 06:25:12 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
X-Lotus-FromDomain: HSSBLR
From: amittal@hss.hns.com
To: iptel@lists.research.bell-labs.com, h323implementors@imtc.org
Message-ID: <6525673E.003EB50F.00@sampark.hss.hns.com>
Date: Wed, 24 Mar 1999 16:58:50 +0530
Subject: Question on Directory Services
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hi

 I am interested in understanding use of directory services in H.323 call
setup. What is the directory standard of choice here -- LDAP, or ILS (is
ILS a standard at all?)? Also, is this directory supposed to be accessed
when the gatekeeper gets an ARQ or a Setup message? Since the directory
would typically be remote, this can have undesirable effects on call setup
times; also, I believe in case of ILS, the termination point and gatekeeper
ought to be using the same ILS (is there a protocol between ILSs to share
data they have?). Or is the basic intent of the directory services to be
accessed by terminals before the terminals originate a call to the
gatekeeper.

Thanks in advance for any inputs!

Regards
Alok


---
Alok Mittal
Technical Leader, Hughes Software Systems
Tel: (91) 80 228 6390, 91, 92    Fax: (91) 80 228 6393
E-mail: amittal@hss.hns.com



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 06:57:54 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA19841
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 06:57:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 371FD52C8; Wed, 24 Mar 1999 06:53:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8BBA552CA; Wed, 24 Mar 1999 06:53:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <012701be75ec$2ae90b00$c21aa4a4@vipin.wipsys.soft.net>
Reply-To: "Vipin  Palawat" <vipinpal@wipsys.soft.net>
From: "Vipin  Palawat" <vipinpal@wipsys.soft.net>
To: <amittal@hss.hns.com>, <iptel@lists.research.bell-labs.com>,
        <h323implementors@imtc.org>
Subject: Re: Question on Directory Services
Date: Wed, 24 Mar 1999 17:17:13 +0530
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.1
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hello Alok,

What I beleives is that gatekeeper resolves the address (H323,E164,...)and
for that
purpose he has to access some directory server. This address comes either in
ARQ or Setup and then the address is resolved to
proxy the message to the destination endpoint. So anyway GK has to
access the directory server to resolve the addresses and ofcouse it
becomes an overhead in Callsetup.

In most cases I have seen the choice was LDAP server.

Regards
Vipin...

________________________________________________
Have  a  Nice  Day!!!

Vipin  Palawat
Wipro  Infotech :Enterprise Solutions (Telecom Division)
K-312, 5th Block , Koramangala,
Bangalore - 560 095 (India)
Ph: 91-80-5538301 Extn:2431
mailto:vipinpal@wipsys.soft.net

-----Original Message-----
From: amittal@hss.hns.com <amittal@hss.hns.com>
To: iptel@lists.research.bell-labs.com <iptel@lists.research.bell-labs.com>;
h323implementors@imtc.org <h323implementors@imtc.org>
Date: Wednesday, March 24, 1999 5:04 PM
Subject: Question on Directory Services


Hi

I am interested in understanding use of directory services in H.323 call
setup. What is the directory standard of choice here -- LDAP, or ILS (is
ILS a standard at all?)? Also, is this directory supposed to be accessed
when the gatekeeper gets an ARQ or a Setup message? Since the directory
would typically be remote, this can have undesirable effects on call setup
times; also, I believe in case of ILS, the termination point and gatekeeper
ought to be using the same ILS (is there a protocol between ILSs to share
data they have?). Or is the basic intent of the directory services to be
accessed by terminals before the terminals originate a call to the
gatekeeper.

Thanks in advance for any inputs!

Regards
Alok


---
Alok Mittal
Technical Leader, Hughes Software Systems
Tel: (91) 80 228 6390, 91, 92    Fax: (91) 80 228 6393
E-mail: amittal@hss.hns.com



---------
This message came from the IETF IPTEL Working Group Mailing List.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 07:08:04 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA19920
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 07:08:04 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 71DB052CB; Wed, 24 Mar 1999 07:01:26 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B90FA52C2; Wed, 24 Mar 1999 07:01:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <c=NL%a=400net%p=att-unisource%l=UNETACSHOOXPO-990324115712Z-7208@unetxgw1.att-unisource.com>
From: "Buchanan, Paul" <Paul.Buchanan@att-unisource.com>
To: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>,
        "'h323implementors@imtc.org'" <h323implementors@imtc.org>,
        "'amittal@hss.hns.com'" <amittal@hss.hns.com>
Subject: RE: Question on Directory Services
Date: Wed, 24 Mar 1999 12:57:12 +0100
X-Mailer:  Microsoft Exchange Server Internet Mail Connector Version 4.0.995.52
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

As a subset to this question are there any developments with regard to
Active Directory  (aka DEN)??


Paul Buchanan (MD)

Office: 0031 (0) 23 569 7561
Mobile: 0044 (0) 411 414071
Network Architect - Voice over IP
paul.buchanan@att-unisource.com
Paul.Buchanan@cwcom.net
http://www.att-unisource.net/


>----------
>From: 	amittal@hss.hns.com[SMTP:amittal@hss.hns.com]
>Sent: 	24 March 1999 12:28
>To: 	iptel@lists.research.bell-labs.com; h323implementors@imtc.org
>Subject: 	Question on Directory Services
>
>Hi
>
> I am interested in understanding use of directory services in H.323 call
>setup. What is the directory standard of choice here -- LDAP, or ILS (is
>ILS a standard at all?)? Also, is this directory supposed to be accessed
>when the gatekeeper gets an ARQ or a Setup message? Since the directory
>would typically be remote, this can have undesirable effects on call setup
>times; also, I believe in case of ILS, the termination point and gatekeeper
>ought to be using the same ILS (is there a protocol between ILSs to share
>data they have?). Or is the basic intent of the directory services to be
>accessed by terminals before the terminals originate a call to the
>gatekeeper.
>
>Thanks in advance for any inputs!
>
>Regards
>Alok
>
>
>---
>Alok Mittal
>Technical Leader, Hughes Software Systems
>Tel: (91) 80 228 6390, 91, 92    Fax: (91) 80 228 6393
>E-mail: amittal@hss.hns.com
>
>
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.
>

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 08:40:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA20994
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 08:40:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A53F052CA; Wed, 24 Mar 1999 08:35:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 15DB152CD; Wed, 24 Mar 1999 08:35:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Wed, 24 Mar 1999 08:29:22 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>, hsalama@cisco.com
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: data issue 4: capacity
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>
>As far as the unit itself, can we agree on the purposely-vague "number
>of calls" as a unit?

Sounds as good as anything else I've heard.

>
>> The problem you're describing above disappears if we adopt the
>> semi-dynamic route advertising/withdrawal mechanism I proposed at the
>> IPTEL meeting. Just a refresher. We don't want to advertise the
>> available resources on a route, because it is a dynamic metric. However,
>> if the utilized resources on a given route as a fraction of the total
>> capacity exceed a given high water mark, we withdraw that route, and
>> send an explicit Update message removing it from all call routing
>> tables. Then when the utilized resources go below a certain low water
>> mark, we start re-advertising the route once again. No dynamic metrics
>> involved.
>
>Well, one thing at a time. I think your proposal is orthogonal to the
>capacity metric itself. I suspect there's going to be a long discussion
>around this proposal (I have some concerns), but I'd rather like to
>tackle the basic attributes we can at least agree we need, first.
>

Which brings us back to the basic question of what we mean by capacity, or
more importantly, how this attribute is treated by LSs (which of course
implicitly answers the first question). 

 1) when propagating a route, should the capacity attribute remain
unchanged or should an LS be permitted to alter its value (ex. call
multiplexing/partitioning)
 2) should this behavior be different depending if the LS inserts a new
next-hop or propagates the same next-hop
 3) how should this attribute be treated with regards to
aggregation/deaggregation?

These are the kinds of questions we're going to have to handle for every
attribute we define and relate to whether an attribute is a function of the
egress gw(s), of the advertisement's path, or of the call-routing path.

In basic BGP, all attributes are attributes of the path.  In GLP, we seem
to be talking about three different kinds of attributes.  Do we classify
capacity as a property of the gateways, of the routed path to the gateways,
or of the routing advertisement's path?  I'm assuming the latter
alternative is not desired by anyone, but the first two seem viable.  

I think Hussein's capacity might be a capacity of the routed path, where
Jonathan's might be of the egress gateways.  Do we need a multiple capacity
attributes? 

In a related note, I think we might need two PATH attributes.  For lack of
anything better, I'll call them CALL_ROUTED_PATH and ADVERTISEMENT_PATH.
The ADVERTISEMENT_PATH would describe the path of propagation of the call
route, and the CALL_ROUTED_PATH would describe the path that a call using
this routing object would take.  Due to the fact that an LS need not insert
a new next-hop, these paths are likely to be different.  The
ADVERTISEMENT_PATH can be used for things like loop-detection, and the
CALL_ROUTED_PATH seems more useful for route selection since it shows the
actual route for potential calls.  

- Matt



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 11:22:13 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA24747
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 11:22:13 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 65CC952CF; Wed, 24 Mar 1999 11:07:38 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5DA9752D0; Wed, 24 Mar 1999 11:07:35 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Wed, 24 Mar 1999 16:55:00 +0100 (MET)
X-Sender: tm@mailhost.fokus.gmd.de
Message-Id: <v0302092ab31ec9ab6623@[193.175.135.110]>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
To: tsg11q8@itu.int, tsg11jrg@itu.int, icintpc@eurescom.de, comswtc@GMU.EDU,
        tina99pc@kestrel.hpl.hp.com, telecom@omg.org, tina-in-wg@tinac.com,
        pint@lists.research.bell-labs.com, iptel@lists.research.bell-labs.com,
        agcore@bxl.dg13.cec.be, agclub@bxl.dg13.cec.be
From: Thomas Magedanz <magedanz@fokus.gmd.de>
Subject: Call for Papers - Intelligent Networks in the New Millenium
Cc: igorf@holta.ho.lucent.com, horn@nortelnetworks.com,
        jvisser@nortelnetworks.com, venieris@theseas.softlab.ece.ntua.gr,
        Karmouch@site.uottawa.ca, gavras@eurescom.de, fogarty@eurescom.de,
        ralf.rieken@oen.siemens.de, Frank.Burghardt@sietec.de,
        max.sevcik@scn.de, straeteh@rc.bel.alcatel.be,
        zuidwegj@rc.bel.alcatel.be, joechester@compuserve.com,
        christian.chabernaud@vz.cit.alcatel.fr, jay.hilton@telops.gte.com,
        SCHOLTFJ@telkom.co.za, schulzk@tzd.telekom.de, EUSNILM@am1.ericsson.se,
        rm@broadcom.ie, trigila@fub.it, Kimmo.Raatikainen@cs.Helsinki.fi,
        lorenzo.faglia@italtel.it, Fabrizio.Zizza@italtel.it,
        jdrosen@dnrc.bell-labs.com, lwc@roke.co.uk, abrusilovsky@lucent.com,
        couturie@alcatel.fr, Didier.Guy@cnet.francetelecom.fr,
        carlo.licciardi@cselt.it, bruno.chatras@cnet.francetelecom.fr,
        huitema@research.bellcore.com, stadler@ctr.columbia.edu,
        raymond.forbes@marconicomms.com, bruno.chatras@cnet.francetelecom.fr,
        kennedy@eurescom.de, gavras@eurescom.de, kimmo.kakko@tkk.tele.fi
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Ladies and Gentlemen!

I think this CfP may be of interest for you, as it may be a target
for dissemination of your results.

Please distribute it also to colleagues who may be interested.
Thank you.


[Apologies if you receive multiple copies of this announcement.]
*******************************************************************

Call for Papers

IEEE Communications Magazine
Feature Topic "Intelligent Networks in the New Millenium"
scheduled for June 2000

Guest Editors: R. Glitho and T. Magedanz

The Intelligent Network (IN) is an architectural framework developed in the
80s for the rapid and uniform creation, deployment, provision and management
of advanced telecommunication services. The IN relies on the separation of
call controll and service control to ease the deployment of new services.
The international standard bodies, namely ITU-T, ETSI and ANSI, introduced
the concept of IN capability sets (CSs) which allows for the stepwise
extension of IN capabilities, and cost reduction in IN deployment.

The last IEEE Communications Magazine feature topic on IN dates back to 1993,
when the first CS was finalized. IN has evolved quite a lot since then; today
work on CS-3 is nearly completed and work on CS-4 as been started. This
evolution is driven on the one hand by the application of IN principles
(for a unification of service creation and provision) to new bearer networks,
such as mobile networks, brdband networks and most recently the internet.
On the other hand new technologies from the IT world, such as CORBA, new
object-oriented service architectures, such as TINA, and most recently agent
technologies, have raised the question of interworking and migration/evolution.

The goal of this feature topic is to present tutorial oriented articles and
status papers of completed work, which describe the recent evolution and
sketch the future of IN. Therefore we solicit papers related to the
following subjects:

- IN standards evolution (i.e., CS-3, and CS-4)
- IN application to mobile networks (e.g., CAMEL, WIN, UMTS, IMT-2000)
- IN application to internet telephony and internet services
- IN and new technologies (e.g., CORBA, agents)
- IN interworking with alternative emerging architectures (e.g., TINA)

Authors should submit their papers following the style requirements of
the IEEE Communicatons Magazine in electronic form (PDF preferred) to
the following Guest Editors:

Dr. Thomas Magedanz                   Roch Glitho
IKV++ GmbH                            Ericsson Research
Kurfuerstendamm 173-174               8400 Decarie blvd.
D-10707 Berlin, Germany               Town of Mount Royal H4P 2N2
Email: magedanz@ikv.de                Email:lmcrocl@LMC.ericsson.se


Important Dates:

Deadline for submissions:    September 1st 1999
Acceptance notification:     December 1st 1999
Final manuscript due date:   February 1st 2000
Publication date:            June 2000

*******************************************************************


Best regards,

Dr. Thomas Magedanz

IKV++ GmbH, Kurfuerstendamm 173-174, D-10707 Berlin, Germany
Fon: + 49 171 172 70 70              Email: magedanz@ikv.de
Fax: + 49 30-881-90-21     



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 18:57:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA06000
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 18:57:53 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 494B252CE; Wed, 24 Mar 1999 18:53:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A511252D0; Wed, 24 Mar 1999 18:53:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <E32BD2062DAFD1119CFD0000F804F31702C04910@cosexch003.mcit.com>
From: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>
To: "'Hanlin_Fang@3com.com'" <Hanlin_Fang@3com.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: RE: question about H.323/SIP and MDCP/MGCP
Date: Wed, 24 Mar 1999 23:52:08 -0000
Importance: high
X-Priority: 1
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BE7651.53C29B26"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BE7651.53C29B26
Content-Type: text/plain

Hanlin,

As I understand them -

1) the MDCP/MGCP protocols control devices and provide bridging between the
SS7 world and the network world

2) We could use one or the other of the protocols for both internal and
external signaling if they are extended to allow the use of scripts, maybe
like VXML or Perl or ?

3) MGCP can do call setup, but the MCG is always in the call stream for the
duration of the call, making scalability an issue.  It is good for
connection to the PSTN.

These questions are the basis for the current holy wars raging on which
protocol to use for what purpose.

Ralph

> -----Original Message-----
> From:	Hanlin_Fang@3com.com [SMTP:Hanlin_Fang@3com.com]
> Sent:	Wednesday, March 24, 1999 4:17 PM
> To:	list iptel
> Subject:	question about H.323/SIP and MDCP/MGCP
> 
> 
> 
> I have couple of questions about both signalling protocols in VoIP.
>    Why we need separate protocols for internal signalling (MDCP/MGCP) and
>    external signalling (H.323/SIP)? Can we use H.323/SIP or MDCP/MGCP for
> both
>    external and internal signalling?
>    If not, why? By my understanding, MDCP/MGCP have nothing to do with
>    end-to-end call setup, even there is callID field in their connection
> setup
>    messages. Is this right? Any other issues?
>    Why H.323/SIP can do it either?
> 
> Any comments are appreciated!
> 
> Hanlin
> 
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

------_=_NextPart_001_01BE7651.53C29B26
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2448.0">
<TITLE>RE: question about H.323/SIP and MDCP/MGCP</TITLE>
</HEAD>
<BODY>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Hanlin,</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">As I understand them -</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">1) the MDCP/MGCP protocols =
control devices and provide bridging between the SS7 world and the =
network world</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">2) We could use one or the =
other of the protocols for both internal and external signaling if they =
are extended to allow the use of scripts, maybe like VXML or Perl or =
?</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">3) MGCP can do call setup, =
but the MCG is always in the call stream for the duration of the call, =
making scalability an issue.&nbsp; It is good for connection to the =
PSTN.</FONT></P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">These questions are the basis =
for the current holy wars raging on which protocol to use for what =
purpose.</FONT>
</P>

<P><FONT COLOR=3D"#0000FF" FACE=3D"Arial">Ralph</FONT>
</P>
<UL>
<P><FONT SIZE=3D1 FACE=3D"Arial">-----Original Message-----</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">From:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Hanlin_Fang@3com.com =
[SMTP:Hanlin_Fang@3com.com]</FONT>
<BR><B><FONT SIZE=3D1 FACE=3D"Arial">Sent:&nbsp;&nbsp;</FONT></B> <FONT =
SIZE=3D1 FACE=3D"Arial">Wednesday, March 24, 1999 4:17 PM</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">To:&nbsp;&nbsp;&nbsp;&nbsp;</FONT></B> <FONT SIZE=3D1 =
FACE=3D"Arial">list iptel</FONT>
<BR><B><FONT SIZE=3D1 =
FACE=3D"Arial">Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT>=
</B> <FONT SIZE=3D1 FACE=3D"Arial">question about H.323/SIP and =
MDCP/MGCP</FONT>
</P>
<BR>
<BR>

<P><FONT SIZE=3D2 FACE=3D"Courier New">I have couple of questions about =
both signalling protocols in VoIP.</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; Why we need =
separate protocols for internal signalling (MDCP/MGCP) and</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; external =
signalling (H.323/SIP)? Can we use H.323/SIP or MDCP/MGCP for =
both</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; external and =
internal signalling?</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; If not, why? By my =
understanding, MDCP/MGCP have nothing to do with</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; end-to-end call =
setup, even there is callID field in their connection setup</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; messages. Is this =
right? Any other issues?</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; Why H.323/SIP can =
do it either?</FONT>
</P>

<P><FONT SIZE=3D2 FACE=3D"Courier New">Any comments are =
appreciated!</FONT>
</P>

<P><FONT SIZE=3D2 FACE=3D"Courier New">Hanlin</FONT>
</P>
<BR>
<BR>

<P><FONT SIZE=3D2 FACE=3D"Courier New">---------</FONT>
<BR><FONT SIZE=3D2 FACE=3D"Courier New">This message came from the IETF =
IPTEL Working Group Mailing List.</FONT>
</P>
</UL>
</BODY>
</HTML>
------_=_NextPart_001_01BE7651.53C29B26--

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 19:02:11 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA06274
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 19:02:10 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A04C352CC; Wed, 24 Mar 1999 18:21:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 10F7452CE; Wed, 24 Mar 1999 18:21:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Hanlin_Fang@3com.com
X-Lotus-FromDomain: 3COM
To: list iptel <iptel@lists.research.bell-labs.com>
Message-ID: <8825673E.0080342E.00@hqoutbound.ops.3com.com>
Date: Wed, 24 Mar 1999 15:17:14 -0800
Subject: question about H.323/SIP and MDCP/MGCP
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



I have couple of questions about both signalling protocols in VoIP.
   Why we need separate protocols for internal signalling (MDCP/MGCP) and
   external signalling (H.323/SIP)? Can we use H.323/SIP or MDCP/MGCP for both
   external and internal signalling?
   If not, why? By my understanding, MDCP/MGCP have nothing to do with
   end-to-end call setup, even there is callID field in their connection setup
   messages. Is this right? Any other issues?
   Why H.323/SIP can do it either?

Any comments are appreciated!

Hanlin



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 19:08:49 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA06723
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 19:08:49 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 4D35C52D0; Wed, 24 Mar 1999 18:57:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1A5E652D2; Wed, 24 Mar 1999 18:57:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <02fb01be7651$d6bb9a00$7d2544ab@glock.cisco.com>
From: "Stephen Sprunk" <ssprunk@cisco.com>
To: <Hanlin_Fang@3com.com>, "list iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: question about H.323/SIP and MDCP/MGCP
Date: Wed, 24 Mar 1999 17:55:44 -0600
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.1
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

MGCP/MDCP is (in theory) much simpler to implement, and therefore better
suited for mass-production dumb devices like desktop feature phones or
high-density media gateways; those dumb devices would then be controlled by
a intelligent SIP/H.323 device.  You could implement H.323 or SIP end-to-end
if you wished.

Stephen

     |          |         Stephen Sprunk, K5SSS, CCIE #3723
    :|:        :|:        NSA, Network Consulting Engineer
   :|||:      :|||:       14875 Landmark Blvd #400; Dallas, TX
.:|||||||:..:|||||||:.    Pager: 800-365-4578 / 800-901-6078
C I S C O S Y S T E M S   Email: ssprunk@cisco.com

-----Original Message-----
From: Hanlin_Fang@3com.com <Hanlin_Fang@3com.com>
To: list iptel <iptel@lists.research.bell-labs.com>
Date: Wednesday, March 24, 1999 17:23
Subject: question about H.323/SIP and MDCP/MGCP


>
>
>I have couple of questions about both signalling protocols in VoIP.
>   Why we need separate protocols for internal signalling (MDCP/MGCP) and
>   external signalling (H.323/SIP)? Can we use H.323/SIP or MDCP/MGCP for
both
>   external and internal signalling?
>   If not, why? By my understanding, MDCP/MGCP have nothing to do with
>   end-to-end call setup, even there is callID field in their connection
setup
>   messages. Is this right? Any other issues?
>   Why H.323/SIP can do it either?
>
>Any comments are appreciated!
>
>Hanlin
>
>
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 22:42:32 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA13762
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 22:42:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 15F3152C7; Wed, 24 Mar 1999 22:37:55 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6FA4652D2; Wed, 24 Mar 1999 22:37:54 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F9AF74.F360B23A@dnrc.bell-labs.com>
Date: Wed, 24 Mar 1999 22:37:24 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: hsalama@cisco.com,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Matt Squire wrote:
> 
> Which brings us back to the basic question of what we mean by capacity, or
> more importantly, how this attribute is treated by LSs (which of course
> implicitly answers the first question).
> 
>  1) when propagating a route, should the capacity attribute remain
> unchanged or should an LS be permitted to alter its value (ex. call
> multiplexing/partitioning)

If a route is not aggregated or de-aggregated, but just propagated as
is, I think an intermediate LS shouldn't modify the capacity. Only the
originator of the route should decide how to set it.

>  2) should this behavior be different depending if the LS inserts a new
> next-hop or propagates the same next-hop

I think capacity in this context refers to the gateway, not the routed
path (answering your other issue), so a change in next hop has no impact
on this.

>  3) how should this attribute be treated with regards to
> aggregation/deaggregation?

Thats the $10M question. Some thoughts:

I think the goal of the capacity metric is to indicate the amount of
calls, roughly, that can be handled by the set of gateways represented
by a route (there can be more than one as aggregation occurs). How its
used is at the discretion of the server policy - perhaps for load
balancing, as has been suggested. So, when aggregating two routes:

* if the two routes are two prefixes with no overlap, for example 516*
and 517* getting aggregated to 51, their capacities should add since
they do not refer to the same gateway. So the total number of calls
supported by two gateways is (sort of) the sum of their capacities - see
note below on problems with addition.

* if there are overlaps in the prefixes, but the LS_PATH or LS_SET
(equivalent of AS_PATH/AS_SET, but for Location Servers) don't overlap,
its also safe to add the capacities, since they don't refer to the same
gateway.

* if there is partial overlap, so with prefix P1 and P2, P1 intersection
P2 = P3, with the size of P3 less than both P1 and P2, you want to come
up with a number less than the sum of the capacities, since there might
be overlap.

* if there is full overlap, so that P1 is a subset of P2, you may want
to do something much less than the sum as the aggregated capacity.


In fact, even for non-overlapping routes, addition doesn't work.
Consider the following example. Gateway 1 serves a single number, +1
555-1212, and has a capacity of 100. Gateway 2 serves all the US, +1 *,
and has a capacity of 1. These prefixes can be aggregated into +1 *.
But, the capacity isn't really 100+1 = 101, since you can't actually
support 101 calls to anywhere in the US. I think you need to take the
size of the prefix into account when adding. So, if prefix 1 (P1) has
size S1 and capacity C1, and prefix P2 has size S2 with capacity C2, the
capacity of the aggregate is:

C3 = (C1 * S1  + C2 * S2)/(S1 + S2)

Of course, how does one measure size? I think one can do it similar to
how it would be done in IP - the number of addresses within the prefix
(whether or not all the addresses within the prefix exist or not). Its a
bit harder with phone numbers since the addresses are variable length. I
think a relative measure is possible, though.

I'm not sure yet on how this would be done when the prefixes overlap,
but I think its doable.


> In basic BGP, all attributes are attributes of the path.  In GLP, we seem
> to be talking about three different kinds of attributes.  Do we classify
> capacity as a property of the gateways, of the routed path to the gateways,
> or of the routing advertisement's path?  I'm assuming the latter
> alternative is not desired by anyone, but the first two seem viable.

As I mention above, I think its a gateway metric. Its usually determined
by size of the trunk to PSTN, DSP resources, and things that are gateway
specific. The call routing path (set of LS servers, *not* the IP path) 
may have limits, but I suspect they are substantially higher than
gateway limits, and it seems easiest to ignore them.


> In a related note, I think we might need two PATH attributes.  For lack of
> anything better, I'll call them CALL_ROUTED_PATH and ADVERTISEMENT_PATH.
> The ADVERTISEMENT_PATH would describe the path of propagation of the call
> route, and the CALL_ROUTED_PATH would describe the path that a call using
> this routing object would take.  Due to the fact that an LS need not insert
> a new next-hop, these paths are likely to be different.  The
> ADVERTISEMENT_PATH can be used for things like loop-detection, and the
> CALL_ROUTED_PATH seems more useful for route selection since it shows the
> actual route for potential calls.

Why do you need the actual route? In normal routing, you only know the
next hop, not the whole routed path. Also, with aggregation, you won't
know the advertisement *path*, but rather the advertisement *set*.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 23:00:41 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA13963
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 23:00:41 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6329B52D3; Wed, 24 Mar 1999 22:55:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B7BDD52D4; Wed, 24 Mar 1999 22:55:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <001901be7672$a55c6e00$c21aa4a4@vipin.wipsys.soft.net>
Reply-To: "Vipin  Palawat" <vipinpal@wipsys.soft.net>
From: "Vipin  Palawat" <vipinpal@wipsys.soft.net>
To: <Hanlin_Fang@3com.com>, "list iptel" <iptel@lists.research.bell-labs.com>
Subject: Re: question about H.323/SIP and MDCP/MGCP
Date: Thu, 25 Mar 1999 09:20:30 +0530
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.1
X-MIMEOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hello Hanlin,

I think you are slightly confused with H.323/SIP, MGCP/MDCP.

In fact SIP is a proposed alternative to H.323 due to its light weight
nature and sort of HTML message flows.
Also the call setup procedures in SIP are simpler as compared to
H.323.

On the other hand MGCP is basically suited for Call agent and has
some plus points in comparison with H.323.

But you can have a complete VOIP product suite with H.323 itself.

Regards
Vipin


________________________________________________
Have  a  Nice  Day!!!

Vipin  Palawat
Wipro  Infotech :Enterprise Solutions (Telecom Division)
K-312, 5th Block , Koramangala,
Bangalore - 560 095 (India)
Ph: 91-80-5538301 Extn:2431
mailto:vipinpal@wipsys.soft.net

-----Original Message-----
From: Hanlin_Fang@3com.com <Hanlin_Fang@3com.com>
To: list iptel <iptel@lists.research.bell-labs.com>
Date: Thursday, March 25, 1999 5:00 AM
Subject: question about H.323/SIP and MDCP/MGCP




I have couple of questions about both signalling protocols in VoIP.
   Why we need separate protocols for internal signalling (MDCP/MGCP) and
   external signalling (H.323/SIP)? Can we use H.323/SIP or MDCP/MGCP for
both
   external and internal signalling?
   If not, why? By my understanding, MDCP/MGCP have nothing to do with
   end-to-end call setup, even there is callID field in their connection
setup
   messages. Is this right? Any other issues?
   Why H.323/SIP can do it either?

Any comments are appreciated!

Hanlin



---------
This message came from the IETF IPTEL Working Group Mailing List.



---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 24 23:07:51 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA14085
	for <iptel-archive@odin.ietf.org>; Wed, 24 Mar 1999 23:07:50 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id AAE5952D5; Wed, 24 Mar 1999 22:55:37 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 221BB52D2; Wed, 24 Mar 1999 22:55:31 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36F9B345.20F31E@cisco.com>
Date: Wed, 24 Mar 1999 19:53:41 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Matt Squire wrote:
> 
> >
> >As far as the unit itself, can we agree on the purposely-vague "number
> >of calls" as a unit?
> 
> Sounds as good as anything else I've heard.
> 
> >
> >> The problem you're describing above disappears if we adopt the
> >> semi-dynamic route advertising/withdrawal mechanism I proposed at the
> >> IPTEL meeting. Just a refresher. We don't want to advertise the
> >> available resources on a route, because it is a dynamic metric. However,
> >> if the utilized resources on a given route as a fraction of the total
> >> capacity exceed a given high water mark, we withdraw that route, and
> >> send an explicit Update message removing it from all call routing
> >> tables. Then when the utilized resources go below a certain low water
> >> mark, we start re-advertising the route once again. No dynamic metrics
> >> involved.
> >
> >Well, one thing at a time. I think your proposal is orthogonal to the
> >capacity metric itself. I suspect there's going to be a long discussion
> >around this proposal (I have some concerns), but I'd rather like to
> >tackle the basic attributes we can at least agree we need, first.
> >
> 
> Which brings us back to the basic question of what we mean by capacity, or
> more importantly, how this attribute is treated by LSs (which of course
> implicitly answers the first question).
> 
>  1) when propagating a route, should the capacity attribute remain
> unchanged or should an LS be permitted to alter its value (ex. call
> multiplexing/partitioning)
>  2) should this behavior be different depending if the LS inserts a new
> next-hop or propagates the same next-hop
>  3) how should this attribute be treated with regards to
> aggregation/deaggregation?
> 
> These are the kinds of questions we're going to have to handle for every
> attribute we define and relate to whether an attribute is a function of the
> egress gw(s), of the advertisement's path, or of the call-routing path.
> 
> In basic BGP, all attributes are attributes of the path.  In GLP, we seem
> to be talking about three different kinds of attributes.  Do we classify
> capacity as a property of the gateways, of the routed path to the gateways,
> or of the routing advertisement's path?  I'm assuming the latter
> alternative is not desired by anyone, but the first two seem viable.
> 
> I think Hussein's capacity might be a capacity of the routed path, where
> Jonathan's might be of the egress gateways.  Do we need a multiple capacity
> attributes?

Yes, either multiple attributes or none at all. 

> 
> In a related note, I think we might need two PATH attributes.  For lack of
> anything better, I'll call them CALL_ROUTED_PATH and ADVERTISEMENT_PATH.
> The ADVERTISEMENT_PATH would describe the path of propagation of the call
> route, and the CALL_ROUTED_PATH would describe the path that a call using
> this routing object would take.  Due to the fact that an LS need not insert
> a new next-hop, these paths are likely to be different.  The
> ADVERTISEMENT_PATH can be used for things like loop-detection, and the
> CALL_ROUTED_PATH seems more useful for route selection since it shows the
> actual route for potential calls.

The ADVERTISEMENT_PATH is identical to BGP's AS_PATH. Right?

I think I understand what you mean by the CALL_ROUTED_PATH, but I don't
see how you'll be able to compute it. For example:

          gw1----AD1----AD2----AD3----AD4----AD5----gw2
        1408*     |                           |
                  |                           |
                  |-----AD6----AD7----AD8-----|
                         z      y      x

Assume the call routing AD topology is identical to the layer 3 AS
topology, i.e., AS=AD in the above picture. Assume also that for call
routing AD5 prefers routes advertised by AD4 over those advertised by
AD8, while for layer 3 routing AD5 prefers routes advertised by AD8 over
those advertised by AD4. Assume also the LSs at all AD do not change the
next hop attribute for any call routes they advertise. gw1 is part of
AD1 and gw2 is part of AD5.

As a result at AD5:
 - The call route to 1408* will have gw1 as the next hop and the
   ADVERTISEMENT_PATH will be {AD1, AD2, AD3, AD4}.
 - The layer 3 route towards gw1 will next hop x (some router in AD8)
and the
   AS_PATH will be {AD1, AD6, AD7, AD8}.

Now when a call arrives at gw2 for 14089876543, gw2 routes it directly
to gw1.  The call route has a single hop gw2 (AD5) --> gw1 (AD1). The
layer 3 route, on the other hand will have multiple hops:
     gw2 (AD5) --> x (AD8) --> y (AD7) --> z (AD6) --> gw1 (AD1)

How can call routing guess/compute such a CALL_ROUTED_PATH? The
situation is even worth when define different boundaries for the call
routing ADs than those of the layer 3 ASs. Assuming that layer 3 routing
and call routing are completely independent, I don't see how call
routing can capture the characteristics of a layer 3 route. Unless:
	- the AD boundaries overlap (or almost overlap) with those of the
          corresponding ASs, and
	- a proxy is inserted as the next hop at each intermediate AD on
the                ADVERTISEMENT_PATH. 
The result will be ADVERTISEMENT_PATH = CALL_ROUTED_PATH.

How do you define call route? Is it the signaling route, or the media
route, or the signaling+media route? I think it is more important to do
route selection based on the media route not the signaling route,
because it's the one which will carry the bulk of the traffic. That's
when a media proxy or a signaling+media proxy comes in handy.

Hussein

>     
> - Matt
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 01:50:50 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA16233
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 01:50:49 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7414052D2; Thu, 25 Mar 1999 01:45:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C89CB52D4; Thu, 25 Mar 1999 01:45:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Dean Willis" <Dean.Willis@MCI.COM>
To: <Hanlin_Fang@3com.com>,
        "'list iptel'" <iptel@lists.research.bell-labs.com>
Subject: RE: question about H.323/SIP and MDCP/MGCP
Date: Thu, 25 Mar 1999 00:42:22 -0600
Message-ID: <001001be768a$a3693ca0$7e8bfea9@dwillispc.mcit.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0
In-Reply-To: <8825673E.0080342E.00@hqoutbound.ops.3com.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit


Actually, it is quite conceivable to use SIP for everything, you just
have to morph the call model into something the telco kiddies don't
understand and it makes their heads hurt. Trust me on this one, I spend
a lot of time explaining the concepts to serious bellheads, and it's not
easy. They think "big switch in network controls all stupid endpoints".

Even with MxCP, my thoughts are to use it only as a "local bus" protocol
between a gateway controller (read "SIP User Agent") and a bunch of
gateway ports. Even here, it begins to make less sense as the scope of
MxCP increases -- eventually, the code size/complexity of MxCP will be
greater than that required to run SIP on the gateway port, eliminating
the basic cost-reduction goal of the original project (any of you
developers have a code base to compare?).

I think the people who want to use MxCP to "control" handsets on
somebody's desk are way off-base. I assert that no informed user wants
their Internet device to be slaved to a carrier's Class-X switch (or
SCP). A peer-to-peer model is far more palatable for the general case.
Furthermore, selection and application of various third-party call
features should be at the end-user's discretion, not Ma Bell's (insert
favorite legacy telco perjorative here) and slave devices tend to
require support from the Big Switch to bring in third-party services.

Let's keep the network stupid and the end devices smart!

--
dean

Disclaim: These are my own opinions. Some of my coworkers disagree
greatly. Some of them even discuss topics on this list . . .


-----Original Message-----
From: owner-iptel@lists.research.bell-labs.com
[mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of
Hanlin_Fang@3com.com
Sent: Wednesday, March 24, 1999 5:17 PM
To: list iptel
Subject: question about H.323/SIP and MDCP/MGCP




I have couple of questions about both signalling protocols in VoIP.
   Why we need separate protocols for internal signalling (MDCP/MGCP)
and
   external signalling (H.323/SIP)? Can we use H.323/SIP or MDCP/MGCP
for both
   external and internal signalling?
   If not, why? By my understanding, MDCP/MGCP have nothing to do with
   end-to-end call setup, even there is callID field in their connection
setup
   messages. Is this right? Any other issues?
   Why H.323/SIP can do it either?

Any comments are appreciated!

Hanlin



---------
This message came from the IETF IPTEL Working Group Mailing List.


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 07:30:02 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA24642
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 07:30:01 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 182ED52AA; Thu, 25 Mar 1999 07:25:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 74E5952B7; Thu, 25 Mar 1999 07:25:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <99Mar25.141638ist.113799@firewall.radvision.rad.co.il>
From: Ami Amir <amir@radvision.rad.co.il>
To: Dean Willis <Dean.Willis@MCI.COM>, Hanlin_Fang@3com.com,
        "'list iptel'" <iptel@lists.research.bell-labs.com>
Subject: RE: question about H.323/SIP and MDCP/MGCP
Date: Thu, 25 Mar 1999 14:24:58 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Dean,

Well, as an old time H.323 person. When I see what is happening to SIP
(which was positioned as a simple alternative to H.323) - the same argument
can be put forward - that once the real functions are provided for, SIP
tends to become closer and closer to H.323.

;-)

Ami Amir

	-----Original Message-----
	From:	Dean Willis [SMTP:Dean.Willis@MCI.COM]
	Sent:	Thursday, March 25, 1999 8:42 AM
	To:	Hanlin_Fang@3com.com; 'list iptel'
	Subject:	RE: question about H.323/SIP and MDCP/MGCP


	Actually, it is quite conceivable to use SIP for everything, you
just
	have to morph the call model into something the telco kiddies don't
	understand and it makes their heads hurt. Trust me on this one, I
spend
	a lot of time explaining the concepts to serious bellheads, and it's
not
	easy. They think "big switch in network controls all stupid
endpoints".

	Even with MxCP, my thoughts are to use it only as a "local bus"
protocol
	between a gateway controller (read "SIP User Agent") and a bunch of
	gateway ports. Even here, it begins to make less sense as the scope
of
	MxCP increases -- eventually, the code size/complexity of MxCP will
be
	greater than that required to run SIP on the gateway port,
eliminating
	the basic cost-reduction goal of the original project (any of you
	developers have a code base to compare?).

	I think the people who want to use MxCP to "control" handsets on
	somebody's desk are way off-base. I assert that no informed user
wants
	their Internet device to be slaved to a carrier's Class-X switch (or
	SCP). A peer-to-peer model is far more palatable for the general
case.
	Furthermore, selection and application of various third-party call
	features should be at the end-user's discretion, not Ma Bell's
(insert
	favorite legacy telco perjorative here) and slave devices tend to
	require support from the Big Switch to bring in third-party
services.

	Let's keep the network stupid and the end devices smart!

	--
	dean

	Disclaim: These are my own opinions. Some of my coworkers disagree
	greatly. Some of them even discuss topics on this list . . .


	-----Original Message-----
	From: owner-iptel@lists.research.bell-labs.com
	[mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of
	Hanlin_Fang@3com.com
	Sent: Wednesday, March 24, 1999 5:17 PM
	To: list iptel
	Subject: question about H.323/SIP and MDCP/MGCP




	I have couple of questions about both signalling protocols in VoIP.
	   Why we need separate protocols for internal signalling
(MDCP/MGCP)
	and
	   external signalling (H.323/SIP)? Can we use H.323/SIP or
MDCP/MGCP
	for both
	   external and internal signalling?
	   If not, why? By my understanding, MDCP/MGCP have nothing to do
with
	   end-to-end call setup, even there is callID field in their
connection
	setup
	   messages. Is this right? Any other issues?
	   Why H.323/SIP can do it either?

	Any comments are appreciated!

	Hanlin



	---------
	This message came from the IETF IPTEL Working Group Mailing List.


	---------
	This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 08:01:55 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA25062
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 08:01:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id CCCB252BB; Thu, 25 Mar 1999 07:57:20 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1FC8152BF; Thu, 25 Mar 1999 07:57:20 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990325075255.0087f320@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 25 Mar 1999 07:52:58 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: data issue 4: capacity
Cc: hsalama@cisco.com,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>
>> In a related note, I think we might need two PATH attributes.  For lack of
>> anything better, I'll call them CALL_ROUTED_PATH and ADVERTISEMENT_PATH.
>> The ADVERTISEMENT_PATH would describe the path of propagation of the call
>> route, and the CALL_ROUTED_PATH would describe the path that a call using
>> this routing object would take.  Due to the fact that an LS need not insert
>> a new next-hop, these paths are likely to be different.  The
>> ADVERTISEMENT_PATH can be used for things like loop-detection, and the
>> CALL_ROUTED_PATH seems more useful for route selection since it shows the
>> actual route for potential calls.
>
>Why do you need the actual route? In normal routing, you only know the
>next hop, not the whole routed path. Also, with aggregation, you won't
>know the advertisement *path*, but rather the advertisement *set*.
>

The potential use is for route selection.  Although the 'cost/price'
attribute is probably a more important criteria, one could also use
'signalling-route-length' (ie the number of signalling hops to destination)
as a secondary criteria.  

I'm not convinced of the utility yet myself, but I wanted to put it out
there.  

- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 08:12:52 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA25145
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 08:12:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DA37552BF; Thu, 25 Mar 1999 08:03:31 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id DE55C52C3; Thu, 25 Mar 1999 08:03:30 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990325075827.007f26d0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 25 Mar 1999 07:58:30 -0500
To: hsalama@cisco.com
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: data issue 4: capacity
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>
>I think I understand what you mean by the CALL_ROUTED_PATH, but I don't
>see how you'll be able to compute it. For example:

This attribute was suggested to capture the call-route, ie sequence of
signalling hops, not the ip route.  This seems as easy to compute as our
AS_PATH equivalent.  The forwarding rules are the almost the same, but you
only modify the attribute if you modify the next-hop.  This is contrasted
with AS_PATH==ADVERTISEMENT_PATH that is modified by every LS regardless of
whether they insert themselves into a route.

- Matt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 09:50:09 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA26948
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 09:50:09 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 780B052C5; Thu, 25 Mar 1999 09:45:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D76C852C7; Thu, 25 Mar 1999 09:45:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990325094050.0087bc20@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Thu, 25 Mar 1999 09:40:52 -0500
To: hsalama@cisco.com
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: data issue 4: capacity
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk


>
>How do you define call route? Is it the signaling route, or the media
>route, or the signaling+media route? I think it is more important to do
>route selection based on the media route not the signaling route,
>because it's the one which will carry the bulk of the traffic. That's
>when a media proxy or a signaling+media proxy comes in handy.
>

I agree it would be better, but I don't see it as a possibility in the
general scenario.
- Matt

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 10:14:18 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA27718
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 10:14:18 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 55B5C52B7; Thu, 25 Mar 1999 10:09:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B816C52C3; Thu, 25 Mar 1999 10:09:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <4.1.19990325080048.0094edc0@orinoco.cisco.com>
X-Sender: rtbell@orinoco.cisco.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1 
X-Priority: 1 (Highest)
Date: Thu, 25 Mar 1999 08:07:13 -0700
To: "Hoefelmeyer, Ralph (MCI)" <Ralph.Hoefelmeyer@mci.com>,
        "'Hanlin_Fang@3com.com'" <Hanlin_Fang@3com.com>,
        list iptel <iptel@lists.research.bell-labs.com>
From: Bob Bell <rtbell@cisco.com>
Subject: RE: question about H.323/SIP and MDCP/MGCP
In-Reply-To: <E32BD2062DAFD1119CFD0000F804F31702C04910@cosexch003.mcit.c
 om>
Mime-Version: 1.0
Content-Type: multipart/alternative;
	boundary="=====================_1034297==_.ALT"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

--=====================_1034297==_.ALT
Content-Type: text/plain; charset="us-ascii"

Just a comment on this response.

First, the MGCP/MDCP or Megaco protocol controls devices which includes ss7
gateways but is not limited to same.

Second, scripting is a desirable feature but is not as yet defined in the
protocol under development. It is wanted, but not yet available.

Third, the MGC is only in the call stream from the call control perspective not
the information flow (RTP streams).
At 04:52 PM 3/24/99 , Hoefelmeyer, Ralph (MCI) wrote: 

>
> Hanlin, 
>
> As I understand them - 
>
> 1) the MDCP/MGCP protocols control devices and provide bridging between the
> SS7 world and the network world 
>
> 2) We could use one or the other of the protocols for both internal and
> external signaling if they are extended to allow the use of scripts, maybe
> like VXML or Perl or ?
>
> 3) MGCP can do call setup, but the MCG is always in the call stream for the
> duration of the call, making scalability an issue.  It is good for connection
> to the PSTN.
>
> These questions are the basis for the current holy wars raging on which
> protocol to use for what purpose. 
>
> Ralph  
> -----Original Message-----  From:   Hanlin_Fang@3com.com
> [SMTP:Hanlin_Fang@3com.com]  Sent:   Wednesday, March 24, 1999 4:17 PM
To:   
>  list iptel  Subject:        question about H.323/SIP and MDCP/MGCP 
>
>
>
> I have couple of questions about both signalling protocols in VoIP.
Why we
> need separate protocols for internal signalling (MDCP/MGCP) and     external
> signalling (H.323/SIP)? Can we use H.323/SIP or MDCP/MGCP for both    
> external and internal signalling?     If not, why? By my understanding,
> MDCP/MGCP have nothing to do with     end-to-end call setup, even there is
> callID field in their connection setup     messages. Is this right? Any other
> issues?     Why H.323/SIP can do it either? 
>
> Any comments are appreciated! 
>
> Hanlin 
>
>
>
> ---------  This message came from the IETF IPTEL Working Group Mailing List. 



Bob Bell
Distinguished Member of the Technical Staff
Cisco Systems Inc.
801-294-3034(v)
801-294-3023(f)
--=====================_1034297==_.ALT
Content-Type: text/html; charset="us-ascii"

<html>
Just a comment on this response.<br>
<br>
First, the MGCP/MDCP or Megaco protocol controls devices which includes
ss7 gateways but is not limited to same.<br>
<br>
Second, scripting is a desirable feature but is not as yet defined in the
protocol under development. It is wanted, but not yet available.<br>
<br>
Third, the MGC is only in the call stream from the call control
perspective not the information flow (RTP streams).<br>
At 04:52 PM 3/24/99 , Hoefelmeyer, Ralph (MCI) wrote: <br>
<br>
<font color="#0000FF"><blockquote type=cite cite>Hanlin,</font> <br>
<br>
<font color="#0000FF">As I understand them -</font> <br>
<br>
<font color="#0000FF">1) the MDCP/MGCP protocols control devices and
provide bridging between the SS7 world and the network world</font> 
<br>
<br>
<font color="#0000FF">2) We could use one or the other of the protocols
for both internal and external signaling if they are extended to allow
the use of scripts, maybe like VXML or Perl or ?<br>
</font><br>
3) MGCP can do call setup, but the MCG is always in the call stream for
the duration of the call, making scalability an issue.&nbsp; It is good
for connection to the PSTN.<br>
<br>
These questions are the basis for the current holy wars raging on which
protocol to use for what purpose. <br>
<br>
<font color="#0000FF">Ralph</font> <font size=1>
<ul>
-----Original Message-----</font> <font size=1><b>
From:&nbsp; </font></b> <font size=1>Hanlin_Fang@3com.com
[SMTP:Hanlin_Fang@3com.com]</font> <font size=1><b>
Sent:&nbsp; </font></b> <font size=1>Wednesday, March 24, 1999 4:17
PM</font> <font size=1><b>
To:&nbsp;&nbsp;&nbsp; </font></b> <font size=1>list iptel</font>
<font size=1><b>
Subject:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </font></b>
<font size=1>question about H.323/SIP and MDCP/MGCP</font> <br>
<br>
<br>
<br>
<font face="Courier New, Courier" size=2>
I have couple of questions about both signalling protocols in
VoIP.</font> <font face="Courier New, Courier" size=2>
&nbsp;&nbsp; Why we need separate protocols for internal signalling
(MDCP/MGCP) and</font> <font face="Courier New, Courier" size=2>
&nbsp;&nbsp; external signalling (H.323/SIP)? Can we use H.323/SIP or
MDCP/MGCP for both</font> <font face="Courier New, Courier" size=2>
&nbsp;&nbsp; external and internal signalling?</font>
<font face="Courier New, Courier" size=2>
&nbsp;&nbsp; If not, why? By my understanding, MDCP/MGCP have nothing to
do with</font> <font face="Courier New, Courier" size=2>
&nbsp;&nbsp; end-to-end call setup, even there is callID field in their
connection setup</font> <font face="Courier New, Courier" size=2>
&nbsp;&nbsp; messages. Is this right? Any other issues?</font>
<font face="Courier New, Courier" size=2>
&nbsp;&nbsp; Why H.323/SIP can do it either?</font> <br>
<br>
<font face="Courier New, Courier" size=2>
Any comments are appreciated!</font> <br>
<br>
<font face="Courier New, Courier" size=2>
Hanlin</font> <br>
<br>
<br>
<br>
<font face="Courier New, Courier" size=2>
---------</font> <font face="Courier New, Courier" size=2>
This message came from the IETF IPTEL Working Group Mailing List.</font> 
</ul></blockquote><br>
<br>

<font size=5 color="#0000FF"><b>Bob Bell<br>
</font></b>Distinguished Member of the Technical Staff<br>
Cisco Systems Inc.<br>
801-294-3034(v)<br>
801-294-3023(f)</html>

--=====================_1034297==_.ALT--


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 13:49:00 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA01359
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 13:48:59 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 42F2452C3; Thu, 25 Mar 1999 12:49:27 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A76E652C8; Thu, 25 Mar 1999 12:49:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199903251758.OAA10649@interino.feedback.net.ar>
Date: Thu, 25 Feb 1999 14:55:32 -0300
From: Buenos Aires New Travel <info@newtravel.com.ar>
Subject: Tarifas Aereas y Semana Santa
Reply-To: info@newtravel.com.ar
To: iptel@lists.research.bell-labs.com
X-Mailer: dbMail v1.32 by Mach5 Software. Please report spamming to dbspam@mach5.com.
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

,

SEMANA SANTA BLOQUEOS ESPECIALES CABOTAJE E INTERNACIONAL

Ante nada le pedimos disculpas por este mail no solicitado, pensamos que es
de su interes, de no querer seguir recibiendo esta informacion, por favor
contestar a este mensaje con OUT en el Subject. Le comentamos que su
dirección es obtenida por medio de una lista de distribución o recomendación de un
conocido.
__________________________________________________________________________

Precios de Pasajes Aereos convenientes:

Santiago de Chile: British Airways $ 130.- Aerolineas Argentinas $ 159.-
Montevideo: $ 85.- Aerolineas Argentinas - Pluna - Lapa
San Pablo: $ 222.- Canadian
Rio de Janeiro: $ 277.- Vasp
Miami: $ 549.- Aerolineas Argentinas vuelo diurno, $ 649.- Vuelo Nocturno
New York: $ 649.- Aerolineas Argentinas
Los Angeles: $ 899.- American Airlines
Lima: $ 349.- Aerolineas Argentinas
Madrid: $ 799.- SpanAir
Londres: $ 849.- Aerolineas Argentinas / $ 899.- british Airways
Paris: $ 849.- Air France
Punta del Este: $ 146.- Aerolineas Argentinas / Pluna

Consulte por el itinerario que necesite, hoteleria, autmoviles, traslados.
Servicio de atencion a empresas. Entrega de documentacion a domicilio.
____________________________________________________________________
Buenos Aires New Travel - Esmeralda 1269 - 1007 - Capital Federal
4-312-5004  //  4313-2490 Lineas Rotativas -  E-mail: bnt@newtravel.com.ar
____________________________________________________________________

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 17:05:49 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA04229
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 17:05:49 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 29C8C52C7; Thu, 25 Mar 1999 16:35:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8EA6E52C8; Thu, 25 Mar 1999 16:35:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <87830212CCB1D011881300A0245175F5715A79@VANTS02>
From: Jerry Skene <Jerry.Skene@tellabs.com>
To: SGCP mailing list <sgcp@bellcore.com>,
        "'megaco@baynetworks.com'" <megaco@baynetworks.com>,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: Call for papers - NCF 99
Date: Thu, 25 Mar 1999 16:32:52 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.1960.3)
Content-Type: text/plain
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

I am chairing a session titled "Standards and Interoperability Update"
being organized by the International Engineering Consortium
(http://www.iec.org) at the National Communications Forum 1999 (NCF
'99), http://www.iec.org/events/1999/ncf/ . 

This session is to be held as part of the Carrier IP Telephony ComForum,
an educational session held during NCF '99. NCF '99 is to be held in
Chicago, and runs from 24 - 28 October, 1999.

I plan to ensure that this Comforum is an educational event, and to give
the attendees, who have paid good money for it, the opportunity to learn
about the current situation in the standards arena as it relates to VOIP
standards. 

If you are interested in presenting a paper at this event, please send
me a brief abstract as soon as possible.

Sincerely;
_______________________________________________
Jerry Skene             |  Tellabs, Inc.
Director,               |  45085 University Drive
International Standards |  Ashburn, VA 20147.2745
tel:+1 703 724-7302     |  URL: www.tellabs.com
fax: +1 703 724-7202    |  email:jerry.skene@tellabs.com

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 22:07:04 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA12379
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 22:07:04 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 59CD252C1; Thu, 25 Mar 1999 21:08:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B8D2252CA; Thu, 25 Mar 1999 21:08:27 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FAEBE3.775E3AC@dnrc.bell-labs.com>
Date: Thu, 25 Mar 1999 21:07:31 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: Matt Squire <msquire@nortelnetworks.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com> <36F9B345.20F31E@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> > I think Hussein's capacity might be a capacity of the routed path, where
> > Jonathan's might be of the egress gateways.  Do we need a multiple capacity
> > attributes?
> 
> Yes, either multiple attributes or none at all.

I would generally think, based on the KISS principle, one attribute is
the best choice.


> The ADVERTISEMENT_PATH is identical to BGP's AS_PATH. Right?
> 
> I think I understand what you mean by the CALL_ROUTED_PATH, but I don't
> see how you'll be able to compute it. For example:

No; this is not my understanding. We have agreed that there is no
relationship between IP telephony administrative domains and the layer 3
AS topologies. The "path" we talk about in GLP is for signaling messages
between location servers, and has nothing to do with either the layer 3
path between location servers, and certainly nothing to do with the
media path.

My understanding of Matt's ADVERTISEMENT_PATH vs. CALL_ROUTED_PATH is
the following:


GW1------LS1--------LS2-------LS3


GW1 propagates a route to LS1, which goes to LS2, then to LS3. The
ADVERTISEMENT_PATH is {GW1, LS1, LS2} when it arrives at LS3. Now, lets
assume that GW1 lists itself as the next hop (for signaling, remember),
LS1 does not change this next hop, but LS2 places itself as the next hop
before propagating the route to LS3. Now, if LS3 sets up a call to GW1,
the signaling messages go from LS3 to LS2 to GW1. So, call
CALL_ROUTED_PATH is {GW1, LS2} when the route arrives at LS3.


> How do you define call route? Is it the signaling route, or the media
> route, or the signaling+media route? I think it is more important to do
> route selection based on the media route not the signaling route,
> because it's the one which will carry the bulk of the traffic. That's
> when a media proxy or a signaling+media proxy comes in handy.

Since there is a total separation of media path/layer 3 topology from
signaling/GLP topology, there's little we can do here.

In any case, I'm very uneasy about these signaling+media proxies. We
have a perfectly fine layer 3 routing topology. We need GLP to help us
find our way, signaling wise, to gateways. But, once we've got that, we
know the IP address of the gateway. Why send the media through proxies?
Its inefficient, routing wise, and makes the scalability of these boxes
orders more complex.

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Mar 25 23:39:50 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA15372
	for <iptel-archive@odin.ietf.org>; Thu, 25 Mar 1999 23:39:50 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E8FA552CA; Thu, 25 Mar 1999 23:35:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5181852CB; Thu, 25 Mar 1999 23:35:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FB0E4A.EC9ACDA5@cisco.com>
Date: Thu, 25 Mar 1999 20:34:18 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com> <36F9B345.20F31E@cisco.com> <36FAEBE3.775E3AC@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Jonathan Rosenberg wrote:
> 
> Hussein F. Salama wrote:
> >
> > > I think Hussein's capacity might be a capacity of the routed path, where
> > > Jonathan's might be of the egress gateways.  Do we need a multiple capacity
> > > attributes?
> >
> > Yes, either multiple attributes or none at all.
> 
> I would generally think, based on the KISS principle, one attribute is
> the best choice.

Except that we have two types of capacity (egress gateway capacity and
call route capacity) which IMHO are equally important and I don't see a
way for combining them into a single attribute.

> 
> > The ADVERTISEMENT_PATH is identical to BGP's AS_PATH. Right?
> >
> > I think I understand what you mean by the CALL_ROUTED_PATH, but I don't
> > see how you'll be able to compute it. For example:
> 
> No; this is not my understanding. We have agreed that there is no
> relationship between IP telephony administrative domains and the layer 3
> AS topologies. The "path" we talk about in GLP is for signaling messages
> between location servers, and has nothing to do with either the layer 3
> path between location servers, and certainly nothing to do with the
> media path.
> 
> My understanding of Matt's ADVERTISEMENT_PATH vs. CALL_ROUTED_PATH is
> the following:
> 
> GW1------LS1--------LS2-------LS3
> 
> GW1 propagates a route to LS1, which goes to LS2, then to LS3. The
> ADVERTISEMENT_PATH is {GW1, LS1, LS2} when it arrives at LS3. Now, lets
> assume that GW1 lists itself as the next hop (for signaling, remember),
> LS1 does not change this next hop, but LS2 places itself as the next hop
> before propagating the route to LS3. Now, if LS3 sets up a call to GW1,
> the signaling messages go from LS3 to LS2 to GW1. So, call
> CALL_ROUTED_PATH is {GW1, LS2} when the route arrives at LS3.

I'm not sure if GW1 should be included in the ADVERTISEMENT_PATH and the
CALL_ROUTED_PATH, but that's a minor thing.

> 
> > How do you define call route? Is it the signaling route, or the media
> > route, or the signaling+media route? I think it is more important to do
> > route selection based on the media route not the signaling route,
> > because it's the one which will carry the bulk of the traffic. That's
> > when a media proxy or a signaling+media proxy comes in handy.
> 
> Since there is a total separation of media path/layer 3 topology from
> signaling/GLP topology, there's little we can do here.

We can force both signaling and media to follow the same path. 

I may be missing something, so let me ask a basic question: What's the
use of doing multi-hop call signaling if, at the end, we won't have any
control over the media path? 

In your above example, when using signaling only proxies, call signaling
from an endpoint, EP, in LS3's domain to GW1 has to hop through LS2, but
the media streams of the same call may not flow through LS2's domain at
all. 

Here are some possible services that an intermediate proxy may offer for
a call:
- Control over capability negotiation and media channels
- QoS for the media
- Policy
- Security
- Accounting, Billing

If the media doesn't flow through LS2's domain, which of the above
services can LS2 provide for this call? And even if it can provide any
of these services, why would it care to be involved in the call at all?
Are there other services I am missing?

> 
> In any case, I'm very uneasy about these signaling+media proxies. We
> have a perfectly fine layer 3 routing topology. We need GLP to help us
> find our way, signaling wise, to gateways. 

Yes, we need GLP, but it does not need to be a multi-hop call routing
protocol unless we want to control the media path through the IP
network. The signaling proxy at LS2, and intermediate signaling hops in
general, are not necessary for gateway location. A single hop call
routing protocol or a directory-based solution are sufficient for
locating gateways.

> But, once we've got that, we
> know the IP address of the gateway. Why send the media through proxies?
> Its inefficient, routing wise, and makes the scalability of these boxes
> orders more complex.

Based on my experience with Cisco's H.323 signaling+media proxy, I claim
that it is quite efficient, routing wise. AND it can provide most of the
services listed above.

Hussein

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 26 00:23:38 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA15764
	for <iptel-archive@odin.ietf.org>; Fri, 26 Mar 1999 00:23:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D0A8552C8; Fri, 26 Mar 1999 00:15:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2709652CC; Fri, 26 Mar 1999 00:15:27 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FB17C3.E9903F4D@dnrc.bell-labs.com>
Date: Fri, 26 Mar 1999 00:14:43 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com> <36F9B345.20F31E@cisco.com> <36FAEBE3.775E3AC@dnrc.bell-labs.com> <36FB0E4A.EC9ACDA5@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> > I would generally think, based on the KISS principle, one attribute is
> > the best choice.
> 
> Except that we have two types of capacity (egress gateway capacity and
> call route capacity) which IMHO are equally important and I don't see a
> way for combining them into a single attribute.

Again, let me make sure we're talking about the same thing. For call
route capacity, we're talking about the capacity of the location servers
along the signaling path, not the media path. My comments apply to this
definition. My argument was that the gateway capacity is generally much
less (in terms of calls it can handle) than an LS, so LS capacity was of
little value.

> > > How do you define call route? Is it the signaling route, or the media
> > > route, or the signaling+media route? I think it is more important to do
> > > route selection based on the media route not the signaling route,
> > > because it's the one which will carry the bulk of the traffic. That's
> > > when a media proxy or a signaling+media proxy comes in handy.
> >
> > Since there is a total separation of media path/layer 3 topology from
> > signaling/GLP topology, there's little we can do here.
> 
> We can force both signaling and media to follow the same path.
> 
> I may be missing something, so let me ask a basic question: What's the
> use of doing multi-hop call signaling if, at the end, we won't have any
> control over the media path?
> 
> In your above example, when using signaling only proxies, call signaling
> from an endpoint, EP, in LS3's domain to GW1 has to hop through LS2, but
> the media streams of the same call may not flow through LS2's domain at
> all.
> 
> Here are some possible services that an intermediate proxy may offer for
> a call:
> - Control over capability negotiation and media channels
> - QoS for the media
> - Policy
> - Security
> - Accounting, Billing
> 
> If the media doesn't flow through LS2's domain, which of the above
> services can LS2 provide for this call? And even if it can provide any
> of these services, why would it care to be involved in the call at all?
> Are there other services I am missing?

Most of the services you are talking about don't require the media to
flow through the LS. Certainly control over capability negotiation does
not (in H.323, gatekeeper routed H.245 is sufficient for this, and in
SIP a proxy could rewrite SDP if it felt it needed to, for some reason),
same with policy, security, accounting, and billing. QoS for the media
is handled by things like rsvp and diffserv, not the call signaling in
any case. For both gatekeeper routed signaling in H.323, and proxy
invite's in SIP, the media flows regularly between endpoints, not
through the gatekeepers or sip servers. Yet, people seem to be able to
bill, account, maintain control, etc. 

> 
> >
> > In any case, I'm very uneasy about these signaling+media proxies. We
> > have a perfectly fine layer 3 routing topology. We need GLP to help us
> > find our way, signaling wise, to gateways.
> 
> Yes, we need GLP, but it does not need to be a multi-hop call routing
> protocol unless we want to control the media path through the IP
> network. The signaling proxy at LS2, and intermediate signaling hops in
> general, are not necessary for gateway location. A single hop call
> routing protocol or a directory-based solution are sufficient for
> locating gateways.

I'm not so sure. We had lengthy discussions on this list regarding
directory services for GLP. It doesn't work. Many gateways can service a
particular number; the right one to use depends on the agreements
between providers. This kind of environment cannot be supported in a
global directory solution, like DNS. Your alternative, a single hop call
routing protocol, doesn't scale all that well. It would require gateway
providers to all have bilateral agreements with each other. Would you
advocate regular IP routing protocols that support only one hop?

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 26 09:49:32 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA26693
	for <iptel-archive@odin.ietf.org>; Fri, 26 Mar 1999 09:49:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 520F952BB; Fri, 26 Mar 1999 09:43:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7CDA652BF; Fri, 26 Mar 1999 09:43:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Sailaja_Talagadadeevi@3com.com
X-Lotus-FromDomain: 3COM
To: hsalama@cisco.com
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Message-ID: <88256740.0050CA4A.00@hqoutbound.ops.3com.com>
Date: Fri, 26 Mar 1999 08:48:20 -0600
Subject: Re: data issue 4: capacity
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk




How do the H450 supplementary services work in H.323 signaling+media proxy
scenario??

sailaja




"Hussein F. Salama" <hsalama@cisco.com> on 03/25/99 10:34:18 PM

Please respond to hsalama@cisco.com

To:   Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
cc:   IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com> (Sailaja
      Talagadadeevi/MW/US/3Com)
Subject:  Re: data issue 4: capacity






Jonathan Rosenberg wrote:
>
> Hussein F. Salama wrote:
> >
> > > I think Hussein's capacity might be a capacity of the routed path, where
> > > Jonathan's might be of the egress gateways.  Do we need a multiple
capacity
> > > attributes?
> >
> > Yes, either multiple attributes or none at all.
>
> I would generally think, based on the KISS principle, one attribute is
> the best choice.

Except that we have two types of capacity (egress gateway capacity and
call route capacity) which IMHO are equally important and I don't see a
way for combining them into a single attribute.

>
> > The ADVERTISEMENT_PATH is identical to BGP's AS_PATH. Right?
> >
> > I think I understand what you mean by the CALL_ROUTED_PATH, but I don't
> > see how you'll be able to compute it. For example:
>
> No; this is not my understanding. We have agreed that there is no
> relationship between IP telephony administrative domains and the layer 3
> AS topologies. The "path" we talk about in GLP is for signaling messages
> between location servers, and has nothing to do with either the layer 3
> path between location servers, and certainly nothing to do with the
> media path.
>
> My understanding of Matt's ADVERTISEMENT_PATH vs. CALL_ROUTED_PATH is
> the following:
>
> GW1------LS1--------LS2-------LS3
>
> GW1 propagates a route to LS1, which goes to LS2, then to LS3. The
> ADVERTISEMENT_PATH is {GW1, LS1, LS2} when it arrives at LS3. Now, lets
> assume that GW1 lists itself as the next hop (for signaling, remember),
> LS1 does not change this next hop, but LS2 places itself as the next hop
> before propagating the route to LS3. Now, if LS3 sets up a call to GW1,
> the signaling messages go from LS3 to LS2 to GW1. So, call
> CALL_ROUTED_PATH is {GW1, LS2} when the route arrives at LS3.

I'm not sure if GW1 should be included in the ADVERTISEMENT_PATH and the
CALL_ROUTED_PATH, but that's a minor thing.

>
> > How do you define call route? Is it the signaling route, or the media
> > route, or the signaling+media route? I think it is more important to do
> > route selection based on the media route not the signaling route,
> > because it's the one which will carry the bulk of the traffic. That's
> > when a media proxy or a signaling+media proxy comes in handy.
>
> Since there is a total separation of media path/layer 3 topology from
> signaling/GLP topology, there's little we can do here.

We can force both signaling and media to follow the same path.

I may be missing something, so let me ask a basic question: What's the
use of doing multi-hop call signaling if, at the end, we won't have any
control over the media path?

In your above example, when using signaling only proxies, call signaling
from an endpoint, EP, in LS3's domain to GW1 has to hop through LS2, but
the media streams of the same call may not flow through LS2's domain at
all.

Here are some possible services that an intermediate proxy may offer for
a call:
- Control over capability negotiation and media channels
- QoS for the media
- Policy
- Security
- Accounting, Billing

If the media doesn't flow through LS2's domain, which of the above
services can LS2 provide for this call? And even if it can provide any
of these services, why would it care to be involved in the call at all?
Are there other services I am missing?

>
> In any case, I'm very uneasy about these signaling+media proxies. We
> have a perfectly fine layer 3 routing topology. We need GLP to help us
> find our way, signaling wise, to gateways.

Yes, we need GLP, but it does not need to be a multi-hop call routing
protocol unless we want to control the media path through the IP
network. The signaling proxy at LS2, and intermediate signaling hops in
general, are not necessary for gateway location. A single hop call
routing protocol or a directory-based solution are sufficient for
locating gateways.

> But, once we've got that, we
> know the IP address of the gateway. Why send the media through proxies?
> Its inefficient, routing wise, and makes the scalability of these boxes
> orders more complex.

Based on my experience with Cisco's H.323 signaling+media proxy, I claim
that it is quite efficient, routing wise. AND it can provide most of the
services listed above.

Hussein

--
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-1714

---------
This message came from the IETF IPTEL Working Group Mailing List.








---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 26 10:05:41 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA27033
	for <iptel-archive@odin.ietf.org>; Fri, 26 Mar 1999 10:05:41 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 486FC52AA; Fri, 26 Mar 1999 09:19:19 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id AFEE552BB; Fri, 26 Mar 1999 09:19:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990326091608.0089d860@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Fri, 26 Mar 1999 09:16:09 -0500
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: data issue 4: capacity
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 10:37 PM 3/24/99 -0500, Jonathan Rosenberg wrote:
>Matt Squire wrote:
>> 
>> Which brings us back to the basic question of what we mean by capacity, or
>> more importantly, how this attribute is treated by LSs (which of course
>> implicitly answers the first question).
>> 
>>  2) should this behavior be different depending if the LS inserts a new
>> next-hop or propagates the same next-hop
>
>I think capacity in this context refers to the gateway, not the routed
>path (answering your other issue), so a change in next hop has no impact
>on this.
>

I'd like to clarify the term 'aggregation' in our context.  
  - Is it aggregation if a LS takes a +1408 route for SIP and a +1408 route
for H323 and combines them into a single route for +1408 and SIP/H323.  
  - Is is aggregation if a LS takes a +1408 route for SIP and another +1408
route for SIP and advertises a single +1408 route for SIP with itself as
the next-hop, but it wants to perform load-balancing betwen the two routes. 
In the second case, the next-hop changes, but the destination and
signalling protocol do not, but I think the capacity should.  I don't know
if I view the second as aggregation, but I think you might.  

>>  3) how should this attribute be treated with regards to
>> aggregation/deaggregation?
>
>Thats the $10M question. Some thoughts:
>
>I think the goal of the capacity metric is to indicate the amount of
>calls, roughly, that can be handled by the set of gateways represented
>by a route (there can be more than one as aggregation occurs). How its
>used is at the discretion of the server policy - perhaps for load
>balancing, as has been suggested. So, when aggregating two routes:
>
>* if the two routes are two prefixes with no overlap, for example 516*
>and 517* getting aggregated to 51, their capacities should add since
>they do not refer to the same gateway. So the total number of calls
>supported by two gateways is (sort of) the sum of their capacities - see
>note below on problems with addition.
>

I think that multiple non-overlapping routes could still refer to the same
gateway, and that some kind of inspection of the path/set is still required
(as you mention below).  

>* if there are overlaps in the prefixes, but the LS_PATH or LS_SET
>(equivalent of AS_PATH/AS_SET, but for Location Servers) don't overlap,
>its also safe to add the capacities, since they don't refer to the same
>gateway.
>

By overlap, do you mean that they don't share any egress domains?  I think
we could ignore overlap "in the middle" if capacity is an egress attribute.  

>I think you need to take the
>size of the prefix into account when adding. So, if prefix 1 (P1) has
>size S1 and capacity C1, and prefix P2 has size S2 with capacity C2, the
>capacity of the aggregate is:
>
>C3 = (C1 * S1  + C2 * S2)/(S1 + S2)
>
>Of course, how does one measure size? I think one can do it similar to
>how it would be done in IP - the number of addresses within the prefix
>(whether or not all the addresses within the prefix exist or not). Its a
>bit harder with phone numbers since the addresses are variable length. I
>think a relative measure is possible, though.
>
>I'm not sure yet on how this would be done when the prefixes overlap,
>but I think its doable.
>

This is certainly different:)

So, in effect, the unit for capacity becomes "average number of calls per
phone number".  Although the calculations would project a uniform
distribution of capacity across all destinations, which certainly isn't the
case, it does seem to be more interesting than most other suggestions.  I
need to think about this more. 

- Matt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 26 17:05:40 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA02917
	for <iptel-archive@odin.ietf.org>; Fri, 26 Mar 1999 17:05:40 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 9733952C3; Fri, 26 Mar 1999 15:36:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 17BE452BF; Fri, 26 Mar 1999 15:36:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FBEC76.715A1113@cisco.com>
Date: Fri, 26 Mar 1999 12:22:14 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com> <36F9B345.20F31E@cisco.com> <36FAEBE3.775E3AC@dnrc.bell-labs.com> <36FB0E4A.EC9ACDA5@cisco.com> <36FB17C3.E9903F4D@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Jonathan Rosenberg wrote:
> 
> Hussein F. Salama wrote:
> >
> >
> > Here are some possible services that an intermediate proxy may offer for
> > a call:
> > - Control over capability negotiation and media channels
> > - QoS for the media
> > - Policy
> > - Security
> > - Accounting, Billing
> >
> > If the media doesn't flow through LS2's domain, which of the above
> > services can LS2 provide for this call? And even if it can provide any
> > of these services, why would it care to be involved in the call at all?
> > Are there other services I am missing?
> 
> Most of the services you are talking about don't require the media to
> flow through the LS. Certainly control over capability negotiation does
> not (in H.323, gatekeeper routed H.245 is sufficient for this, and in
> SIP a proxy could rewrite SDP if it felt it needed to, for some reason),
> same with policy, security, accounting, and billing. 

You're missing my point. Assume call signaling goes through an ingress
gateway in AD1, an intermediate signaling only proxy in AD2, and an
egress gateway in AD3. However, the media streams for that call do not
pass through AD2 at all. Assume the follow the path AD1-AD4-AD5-AD3.
Let's use H.323 as an example. The gatekeeper in AD1 may be interested
in controlling the capabilities because the ingress gateway is in its
domain and the media streams flow through that domain. Same thing
applies to the gatekeeper in the egress domain, AD3. But why would a
gatekeeper in AD2 care for being involved in, and routing this call? I
know it can, be I don't think it care. Since the layer 3 route doesn't
go through AD2.

What if the gatekeepers in AD4 and AD5 want to control the capabilities
(codec selection) for  above call? I think it's their right to do that
since the voice streams will ultimately flow through their domains.
Would they be able to do that? Not when separating signaling from media.

The same applies for accounting. What if the gatekeepers in AD4 and AD5
want to do accounting for  above call? I think it's their right to
account for voice traffic traversing their domain. Would they be able to
do that?

> QoS for the media
> is handled by things like rsvp and diffserv, not the call signaling in
> any case. 

What if different intermediate domains on the media path of a call want
to provide different levels of QoS to same media stream? Let's use
RSVP/Intserv as an example. What if AD4 wants to provide controlled load
service for the audio streams while AD5 prefers to provide guaranteed
QoS?

> For both gatekeeper routed signaling in H.323, and proxy
> invite's in SIP, the media flows regularly between endpoints, not
> through the gatekeepers or sip servers. Yet, people seem to be able to
> bill, account, maintain control, etc.

See my above comment on accounting. So far people are doing accounting,
billing, ... etc at the ingress and egress domains only. This will
change as the volume of VoIP traffic increases.

> 
> >
> > >
> > > In any case, I'm very uneasy about these signaling+media proxies. We
> > > have a perfectly fine layer 3 routing topology. We need GLP to help us
> > > find our way, signaling wise, to gateways.
> >
> > Yes, we need GLP, but it does not need to be a multi-hop call routing
> > protocol unless we want to control the media path through the IP
> > network. The signaling proxy at LS2, and intermediate signaling hops in
> > general, are not necessary for gateway location. A single hop call
> > routing protocol or a directory-based solution are sufficient for
> > locating gateways.
> 
> I'm not so sure. We had lengthy discussions on this list regarding
> directory services for GLP. It doesn't work. Many gateways can service a
> particular number; the right one to use depends on the agreements
> between providers. This kind of environment cannot be supported in a
> global directory solution, like DNS. Your alternative, a single hop call
> routing protocol, doesn't scale all that well. It would require gateway
> providers to all have bilateral agreements with each other. Would you
> advocate regular IP routing protocols that support only one hop?

I didn't use "single hop call routing" in the right context.What I meant
"single hop inter domain call signaling." This can be done using either
the BGP-based GLP proposal or the SCSP-based proposal, and never
changing the NEXT_HOP attribute when an update is propagated from one
domain to another. The NEXT_HOP will always point directly to the egress
domain.

Hussein


> 
> -Jonathan R.
> --
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX: (732) 834-5379                         Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: +1 (408) 527-7147, Fax: +1 (408) 527-1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 26 22:23:28 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA11796
	for <iptel-archive@odin.ietf.org>; Fri, 26 Mar 1999 22:23:27 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id AB85C52B7; Fri, 26 Mar 1999 22:17:36 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1777452C7; Fri, 26 Mar 1999 22:17:36 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FC4DB5.C26A25F9@dnrc.bell-labs.com>
Date: Fri, 26 Mar 1999 22:17:09 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com> <36F9B345.20F31E@cisco.com> <36FAEBE3.775E3AC@dnrc.bell-labs.com> <36FB0E4A.EC9ACDA5@cisco.com> <36FB17C3.E9903F4D@dnrc.bell-labs.com> <36FBEC76.715A1113@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F. Salama wrote:
> 
> > Most of the services you are talking about don't require the media to
> > flow through the LS. Certainly control over capability negotiation does
> > not (in H.323, gatekeeper routed H.245 is sufficient for this, and in
> > SIP a proxy could rewrite SDP if it felt it needed to, for some reason),
> > same with policy, security, accounting, and billing.
> 
> You're missing my point. Assume call signaling goes through an ingress
> gateway in AD1, an intermediate signaling only proxy in AD2, and an
> egress gateway in AD3. However, the media streams for that call do not
> pass through AD2 at all. Assume the follow the path AD1-AD4-AD5-AD3.
> Let's use H.323 as an example. The gatekeeper in AD1 may be interested
> in controlling the capabilities because the ingress gateway is in its
> domain and the media streams flow through that domain. Same thing
> applies to the gatekeeper in the egress domain, AD3. But why would a
> gatekeeper in AD2 care for being involved in, and routing this call? I
> know it can, be I don't think it care. Since the layer 3 route doesn't
> go through AD2.

A gatekeeper in AD2 would be interested since GLP effectively introduces
a "subcontracting relationship" among the various providers. AD1 allows
AD2 to use its gateways. By propagating this information to AD3, AD2 can
appear to be its own gateway provider, when in fact its using gateways
from AD1 to provide a more complete service. So, AD2 may want to see the
signaling messages for call accounting purposes.

Same situation arises in building a house. One big contractor interfaces
with you for the job, and they in turn subcontract some work out to
other contactors, who may further subcontract that out. Invoices for
work still pass through all of them, since they each provide the
value-add of "aggregation of service".

> 
> What if the gatekeepers in AD4 and AD5 want to control the capabilities
> (codec selection) for  above call? I think it's their right to do that
> since the voice streams will ultimately flow through their domains.
> Would they be able to do that? Not when separating signaling from media.

Well, this has not happened yet at all for any other IP application.
When I browse a web page, a hit a server in some domain AD1. I'm in AD2,
and I have a local proxy. The actual traffic flows through many
intermediate transit providers. To date, these providers have had no say
in whether I download the text or glitzy-image version of a page, or
whether I use gif or jpeg images. 

I do believe that an ISP should have control over what data leaves and
enters its network, but that this functionality lives at layer 3.
Diffserv and other things are how you control this. What you are
proposing requires us to overlay an independent network of proxies and
servers for every single application. One of the fundamental engineering
assumptions in the Internet is that it is application unaware, and that
apps live in the end systems. It is this model that has allowed new apps
(like the web) to come into being.

> 
> The same applies for accounting. What if the gatekeepers in AD4 and AD5
> want to do accounting for  above call? I think it's their right to
> account for voice traffic traversing their domain. Would they be able to
> do that?

Accounting would be done in border routers (i.e., layer 3), using the
diffserv model. The voice traffic (or any other high priority, delay
sensitive traffic) would be marked at the ingress domain. Each border
router could account for the amount of this kind of traffic, and then
remark. So, you get the kind of accounting you want, but at layer 3.

> 
> > QoS for the media
> > is handled by things like rsvp and diffserv, not the call signaling in
> > any case.
> 
> What if different intermediate domains on the media path of a call want
> to provide different levels of QoS to same media stream? Let's use
> RSVP/Intserv as an example. What if AD4 wants to provide controlled load
> service for the audio streams while AD5 prefers to provide guaranteed
> QoS?

So, let them do it. Normally RSVP is used to ask for only one of the
two, but if for some reason a domain wanted to do something different,
it could perhaps rewrite the RSVP messages at the border routers, or
something like that. Certainly one domain can do diffserv, and another
rsvp. Check out draft-ietf-diffserv-rsvp. In either case, gatekeepers
and sip servers are not involved.


> 
> > For both gatekeeper routed signaling in H.323, and proxy
> > invite's in SIP, the media flows regularly between endpoints, not
> > through the gatekeepers or sip servers. Yet, people seem to be able to
> > bill, account, maintain control, etc.
> 
> See my above comment on accounting. So far people are doing accounting,
> billing, ... etc at the ingress and egress domains only. This will
> change as the volume of VoIP traffic increases.

I agree we need accounting for voice traffic in intermediate domains.
Thats with diffserv border routers are for.

> > I'm not so sure. We had lengthy discussions on this list regarding
> > directory services for GLP. It doesn't work. Many gateways can service a
> > particular number; the right one to use depends on the agreements
> > between providers. This kind of environment cannot be supported in a
> > global directory solution, like DNS. Your alternative, a single hop call
> > routing protocol, doesn't scale all that well. It would require gateway
> > providers to all have bilateral agreements with each other. Would you
> > advocate regular IP routing protocols that support only one hop?
> 
> I didn't use "single hop call routing" in the right context.What I meant
> "single hop inter domain call signaling." This can be done using either
> the BGP-based GLP proposal or the SCSP-based proposal, and never
> changing the NEXT_HOP attribute when an update is propagated from one
> domain to another. The NEXT_HOP will always point directly to the egress
> domain.

I see, sorry for my misunderstanding. I think I address your point
already above. Intermediate LS's may want to be on the signaling path
for accounting (call usage, mind you, not voice level accounting). If
they don't want this, then you are correct. Single-hop signaling is just
fine. As I said, we have a perfectly nice underlying transport network
that allows us to go from anywhere to anywhere else.

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 26 22:35:25 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA13633
	for <iptel-archive@odin.ietf.org>; Fri, 26 Mar 1999 22:35:25 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0EA3952C7; Fri, 26 Mar 1999 22:29:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6D6BB52CC; Fri, 26 Mar 1999 22:29:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <AF62C6C0EC3DD211929900805FEA97CF6080DA@wltg-msg01.mccaw-stg.com>
From: "Hutton, Charles" <charles.hutton@attws.com>
To: hsalama@cisco.com
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: RE: data issue 4: capacity
Date: Fri, 26 Mar 1999 19:27:40 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2232.9)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

I think the implication here is that tandeming (transcoding) of codecs would
/ could be performed in AD4 and AD5. Unacceptable voice quality will quickly
result with G.723.1 and G.729 if they are tandemed; subjective measurements
show noticeable degradation with 1 tandeming and by 3 tandemings it is a
struggle to understand completely.

Also I would think that the audio coding format would be a don't-care to AD4
and AD5.

In summary (and hoping I have not misunderstood):

    - we can not tandem in every AD used in a call
    - and luckily, we have no reason to tandem in each AD






> Hussein F. Salama wrote:
> > 
> > > Most of the services you are talking about don't require the media to
> > > flow through the LS. Certainly control over capability negotiation
> does
> > > not (in H.323, gatekeeper routed H.245 is sufficient for this, and in
> > > SIP a proxy could rewrite SDP if it felt it needed to, for some
> reason),
> > > same with policy, security, accounting, and billing.
> > 
> > You're missing my point. Assume call signaling goes through an ingress
> > gateway in AD1, an intermediate signaling only proxy in AD2, and an
> > egress gateway in AD3. However, the media streams for that call do not
> > pass through AD2 at all. Assume the follow the path AD1-AD4-AD5-AD3.
> > Let's use H.323 as an example. The gatekeeper in AD1 may be interested
> > in controlling the capabilities because the ingress gateway is in its
> > domain and the media streams flow through that domain. Same thing
> > applies to the gatekeeper in the egress domain, AD3. But why would a
> > gatekeeper in AD2 care for being involved in, and routing this call? I
> > know it can, be I don't think it care. Since the layer 3 route doesn't
> > go through AD2.
> 
> A gatekeeper in AD2 would be interested since GLP effectively introduces
> a "subcontracting relationship" among the various providers. AD1 allows
> AD2 to use its gateways. By propagating this information to AD3, AD2 can
> appear to be its own gateway provider, when in fact its using gateways
> from AD1 to provide a more complete service. So, AD2 may want to see the
> signaling messages for call accounting purposes.
> 
> Same situation arises in building a house. One big contractor interfaces
> with you for the job, and they in turn subcontract some work out to
> other contactors, who may further subcontract that out. Invoices for
> work still pass through all of them, since they each provide the
> value-add of "aggregation of service".
> 
> > 
> > What if the gatekeepers in AD4 and AD5 want to control the capabilities
> > (codec selection) for  above call? I think it's their right to do that
> > since the voice streams will ultimately flow through their domains.
> > Would they be able to do that? Not when separating signaling from media.
> 
> Well, this has not happened yet at all for any other IP application.
> When I browse a web page, a hit a server in some domain AD1. I'm in AD2,
> and I have a local proxy. The actual traffic flows through many
> intermediate transit providers. To date, these providers have had no say
> in whether I download the text or glitzy-image version of a page, or
> whether I use gif or jpeg images. 
> 
> I do believe that an ISP should have control over what data leaves and
> enters its network, but that this functionality lives at layer 3.
> Diffserv and other things are how you control this. What you are
> proposing requires us to overlay an independent network of proxies and
> servers for every single application. One of the fundamental engineering
> assumptions in the Internet is that it is application unaware, and that
> apps live in the end systems. It is this model that has allowed new apps
> (like the web) to come into being.
> 
> > 
> > The same applies for accounting. What if the gatekeepers in AD4 and AD5
> > want to do accounting for  above call? I think it's their right to
> > account for voice traffic traversing their domain. Would they be able to
> > do that?
> 
> Accounting would be done in border routers (i.e., layer 3), using the
> diffserv model. The voice traffic (or any other high priority, delay
> sensitive traffic) would be marked at the ingress domain. Each border
> router could account for the amount of this kind of traffic, and then
> remark. So, you get the kind of accounting you want, but at layer 3.
> 
> > 
> > > QoS for the media
> > > is handled by things like rsvp and diffserv, not the call signaling in
> > > any case.
> > 
> > What if different intermediate domains on the media path of a call want
> > to provide different levels of QoS to same media stream? Let's use
> > RSVP/Intserv as an example. What if AD4 wants to provide controlled load
> > service for the audio streams while AD5 prefers to provide guaranteed
> > QoS?
> 
> So, let them do it. Normally RSVP is used to ask for only one of the
> two, but if for some reason a domain wanted to do something different,
> it could perhaps rewrite the RSVP messages at the border routers, or
> something like that. Certainly one domain can do diffserv, and another
> rsvp. Check out draft-ietf-diffserv-rsvp. In either case, gatekeepers
> and sip servers are not involved.
> 
> 
> > 
> > > For both gatekeeper routed signaling in H.323, and proxy
> > > invite's in SIP, the media flows regularly between endpoints, not
> > > through the gatekeepers or sip servers. Yet, people seem to be able to
> > > bill, account, maintain control, etc.
> > 
> > See my above comment on accounting. So far people are doing accounting,
> > billing, ... etc at the ingress and egress domains only. This will
> > change as the volume of VoIP traffic increases.
> 
> I agree we need accounting for voice traffic in intermediate domains.
> Thats with diffserv border routers are for.
> 
> > > I'm not so sure. We had lengthy discussions on this list regarding
> > > directory services for GLP. It doesn't work. Many gateways can service
> a
> > > particular number; the right one to use depends on the agreements
> > > between providers. This kind of environment cannot be supported in a
> > > global directory solution, like DNS. Your alternative, a single hop
> call
> > > routing protocol, doesn't scale all that well. It would require
> gateway
> > > providers to all have bilateral agreements with each other. Would you
> > > advocate regular IP routing protocols that support only one hop?
> > 
> > I didn't use "single hop call routing" in the right context.What I meant
> > "single hop inter domain call signaling." This can be done using either
> > the BGP-based GLP proposal or the SCSP-based proposal, and never
> > changing the NEXT_HOP attribute when an update is propagated from one
> > domain to another. The NEXT_HOP will always point directly to the egress
> > domain.
> 
> I see, sorry for my misunderstanding. I think I address your point
> already above. Intermediate LS's may want to be on the signaling path
> for accounting (call usage, mind you, not voice level accounting). If
> they don't want this, then you are correct. Single-hop signaling is just
> fine. As I said, we have a perfectly nice underlying transport network
> that allows us to go from anywhere to anywhere else.
> 
> -Jonathan R.
> -- 
> Jonathan D. Rosenberg                       Lucent Technologies
> Member of Technical Staff                   101 Crawfords Corner Rd.
> High Speed Networks Research                Holmdel, NJ 07733
> FAX: (732) 834-5379                         Rm. 4C-526
> EMAIL: jdrosen@bell-labs.com
> URL: http://www.cs.columbia.edu/~jdrosen
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.
> 

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Mar 26 22:53:50 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA13954
	for <iptel-archive@odin.ietf.org>; Fri, 26 Mar 1999 22:53:49 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1DEF952C5; Fri, 26 Mar 1999 21:56:49 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6E9AF52C7; Fri, 26 Mar 1999 21:56:48 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FC48D5.C531B855@dnrc.bell-labs.com>
Date: Fri, 26 Mar 1999 21:56:21 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990326091608.0089d860@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Matt Squire wrote:
> 
> 
> I'd like to clarify the term 'aggregation' in our context.
>   - Is it aggregation if a LS takes a +1408 route for SIP and a +1408 route
> for H323 and combines them into a single route for +1408 and SIP/H323.

Well, I wasn't really thinking of allowing aggregation when the
signaling protocol attribues were different. This would require us to
define codepoints for all 2^N combinations of N signaling protocols,
kind of messy. But, if we did do this, yes, I would call it aggregation.

>   - Is is aggregation if a LS takes a +1408 route for SIP and another +1408
> route for SIP and advertises a single +1408 route for SIP with itself as
> the next-hop, but it wants to perform load-balancing betwen the two routes.
> In the second case, the next-hop changes, but the destination and
> signalling protocol do not, but I think the capacity should.  I don't know
> if I view the second as aggregation, but I think you might.

Yes, I agree the capacity would change here.


> >* if the two routes are two prefixes with no overlap, for example 516*
> >and 517* getting aggregated to 51, their capacities should add since
> >they do not refer to the same gateway. So the total number of calls
> >supported by two gateways is (sort of) the sum of their capacities - see
> >note below on problems with addition.
> >
> 
> I think that multiple non-overlapping routes could still refer to the same
> gateway, and that some kind of inspection of the path/set is still required
> (as you mention below).

True if a gateway provider advertises different prefixes for the same
gateway.

> 
> >* if there are overlaps in the prefixes, but the LS_PATH or LS_SET
> >(equivalent of AS_PATH/AS_SET, but for Location Servers) don't overlap,
> >its also safe to add the capacities, since they don't refer to the same
> >gateway.
> >
> 
> By overlap, do you mean that they don't share any egress domains?  I think
> we could ignore overlap "in the middle" if capacity is an egress attribute.

True. However, when there is aggregation, the LS_PATH is lost (I
believe), and its converted to a LS_SET, so you couldn't tell whats "in
the middle" as opposed to egress.


> This is certainly different:)
> 
> So, in effect, the unit for capacity becomes "average number of calls per
> phone number".  Although the calculations would project a uniform
> distribution of capacity across all destinations, which certainly isn't the
> case, it does seem to be more interesting than most other suggestions.  I
> need to think about this more.

Yes, its a density. And yes, it does project a uniform density. I also
need to think more about it :)

-Jonathan R.

-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Mar 27 10:51:46 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA24758
	for <iptel-archive@odin.ietf.org>; Sat, 27 Mar 1999 10:51:46 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7E0B752CC; Sat, 27 Mar 1999 10:47:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id CFF9552CE; Sat, 27 Mar 1999 10:47:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.32.19990327104332.008192b0@bl-mail1.corpeast.baynetworks.com>
X-Sender: msquire@bl-mail1.corpeast.baynetworks.com
X-Mailer: Windows Eudora Pro Version 3.0 (32)
Date: Sat, 27 Mar 1999 10:43:36 -0500
To: hsalama@cisco.com, Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Matt Squire <msquire@nortelnetworks.com>
Subject: Re: data issue 4: capacity
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 12:22 PM 3/26/99 -0800, Hussein F. Salama wrote:
>
>
>Jonathan Rosenberg wrote:
>> 
>> Hussein F. Salama wrote:
>> >
>> >
>> > Here are some possible services that an intermediate proxy may offer for
>> > a call:
>> > - Control over capability negotiation and media channels
>> > - QoS for the media
>> > - Policy
>> > - Security
>> > - Accounting, Billing
>> >
>> > If the media doesn't flow through LS2's domain, which of the above
>> > services can LS2 provide for this call? And even if it can provide any
>> > of these services, why would it care to be involved in the call at all?
>> > Are there other services I am missing?
>> 
>> Most of the services you are talking about don't require the media to
>> flow through the LS. Certainly control over capability negotiation does
>> not (in H.323, gatekeeper routed H.245 is sufficient for this, and in
>> SIP a proxy could rewrite SDP if it felt it needed to, for some reason),
>> same with policy, security, accounting, and billing. 
>
>You're missing my point. Assume call signaling goes through an ingress
>gateway in AD1, an intermediate signaling only proxy in AD2, and an
>egress gateway in AD3. However, the media streams for that call do not
>pass through AD2 at all. Assume the follow the path AD1-AD4-AD5-AD3.
>Let's use H.323 as an example. The gatekeeper in AD1 may be interested
>in controlling the capabilities because the ingress gateway is in its
>domain and the media streams flow through that domain. Same thing
>applies to the gatekeeper in the egress domain, AD3. But why would a
>gatekeeper in AD2 care for being involved in, and routing this call? I
>know it can, be I don't think it care. Since the layer 3 route doesn't
>go through AD2.
>
>What if the gatekeepers in AD4 and AD5 want to control the capabilities
>(codec selection) for  above call? I think it's their right to do that
>since the voice streams will ultimately flow through their domains.
>Would they be able to do that? Not when separating signaling from media.
>
>The same applies for accounting. What if the gatekeepers in AD4 and AD5
>want to do accounting for  above call? I think it's their right to
>account for voice traffic traversing their domain. Would they be able to
>do that?
>

There are numerous applications that require quality of service,
accounting, policy, security, etc.  Voice is just one application.  Voice
connections that have chosen a destination based on a e164 address are just
a subset of that application.  Voice connections that have been routed
using GLP are an even smaller subset.  

Solving these larger problems within GLP is not the answer.   Between ASs
(or diffserv domains) these problems must be addressed, but for all data,
not each application individually.  

>> QoS for the media
>> is handled by things like rsvp and diffserv, not the call signaling in
>> any case. 
>
>What if different intermediate domains on the media path of a call want
>to provide different levels of QoS to same media stream? Let's use
>RSVP/Intserv as an example. What if AD4 wants to provide controlled load
>service for the audio streams while AD5 prefers to provide guaranteed
>QoS?
>

QoS routing is not within our scope.  Its a huge problem.  

I think/hope that we still have consensus that the separation of the call
control plane from the L3 routing plane is the right thing to do.  

>> For both gatekeeper routed signaling in H.323, and proxy
>> invite's in SIP, the media flows regularly between endpoints, not
>> through the gatekeepers or sip servers. Yet, people seem to be able to
>> bill, account, maintain control, etc.
>
>See my above comment on accounting. So far people are doing accounting,
>billing, ... etc at the ingress and egress domains only. This will
>change as the volume of VoIP traffic increases.
>

This will change as the volume of traffic increases, as the QoS
requirements of the traffic varies, and as our dependency on the
reliability of that traffic increases.  Whether the data is voice shouldn't
be used as the differentiator.  

- Matt


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 28 03:49:00 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA00932
	for <iptel-archive@odin.ietf.org>; Sun, 28 Mar 1999 03:48:59 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C16C252CF; Sun, 28 Mar 1999 03:43:15 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1F1D852D0; Sun, 28 Mar 1999 03:43:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FDEBFE.8F67438D@cisco.com>
Date: Sun, 28 Mar 1999 00:44:46 -0800
From: Hussein F Salama <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Matt Squire <msquire@nortelnetworks.com>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990327104332.008192b0@bl-mail1.corpeast.baynetworks.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Matt Squire wrote:
> 
> At 12:22 PM 3/26/99 -0800, Hussein F. Salama wrote:
> >
> >
> >Jonathan Rosenberg wrote:
> >>
> >> Hussein F. Salama wrote:
> >> >
> >> >
> >> > Here are some possible services that an intermediate proxy may offer for
> >> > a call:
> >> > - Control over capability negotiation and media channels
> >> > - QoS for the media
> >> > - Policy
> >> > - Security
> >> > - Accounting, Billing
> >> >
> >> > If the media doesn't flow through LS2's domain, which of the above
> >> > services can LS2 provide for this call? And even if it can provide any
> >> > of these services, why would it care to be involved in the call at all?
> >> > Are there other services I am missing?
> >>
> >> Most of the services you are talking about don't require the media to
> >> flow through the LS. Certainly control over capability negotiation does
> >> not (in H.323, gatekeeper routed H.245 is sufficient for this, and in
> >> SIP a proxy could rewrite SDP if it felt it needed to, for some reason),
> >> same with policy, security, accounting, and billing.
> >
> >You're missing my point. Assume call signaling goes through an ingress
> >gateway in AD1, an intermediate signaling only proxy in AD2, and an
> >egress gateway in AD3. However, the media streams for that call do not
> >pass through AD2 at all. Assume the follow the path AD1-AD4-AD5-AD3.
> >Let's use H.323 as an example. The gatekeeper in AD1 may be interested
> >in controlling the capabilities because the ingress gateway is in its
> >domain and the media streams flow through that domain. Same thing
> >applies to the gatekeeper in the egress domain, AD3. But why would a
> >gatekeeper in AD2 care for being involved in, and routing this call? I
> >know it can, be I don't think it care. Since the layer 3 route doesn't
> >go through AD2.
> >
> >What if the gatekeepers in AD4 and AD5 want to control the capabilities
> >(codec selection) for  above call? I think it's their right to do that
> >since the voice streams will ultimately flow through their domains.
> >Would they be able to do that? Not when separating signaling from media.
> >
> >The same applies for accounting. What if the gatekeepers in AD4 and AD5
> >want to do accounting for  above call? I think it's their right to
> >account for voice traffic traversing their domain. Would they be able to
> >do that?
> >
> 
> There are numerous applications that require quality of service,
> accounting, policy, security, etc.  Voice is just one application.  Voice
> connections that have chosen a destination based on a e164 address are just
> a subset of that application.  Voice connections that have been routed
> using GLP are an even smaller subset.
> 
> Solving these larger problems within GLP is not the answer.

Yes, I agree.

> Between ASs
> (or diffserv domains) these problems must be addressed, but for all data,
> not each application individually.

As I mentioned in my last message, some applications can't  be easily
identified by layer 3 devices, and hence the need for assistance from
the application layer.

> 
> >> QoS for the media
> >> is handled by things like rsvp and diffserv, not the call signaling in
> >> any case.
> >
> >What if different intermediate domains on the media path of a call want
> >to provide different levels of QoS to same media stream? Let's use
> >RSVP/Intserv as an example. What if AD4 wants to provide controlled load
> >service for the audio streams while AD5 prefers to provide guaranteed
> >QoS?
> >
> 
> QoS routing is not within our scope.  Its a huge problem.

I didn't propose solving a QoS problem. All I trying to say is that
signaling+media proxies can provide more flexibility when enforcing QoS
policies.

> 
> I think/hope that we still have consensus that the separation of the call
> control plane from the L3 routing plane is the right thing to do.

Yes, it is safer to solve the general case where the call routing
topology is independent of the layer 3 routing topology. However, I am
slowly coming to realize that future deployments of GLP will create a
call routing topology that's identical to the underlying layer 3 routing
topology.

Hussein


> 
> >> For both gatekeeper routed signaling in H.323, and proxy
> >> invite's in SIP, the media flows regularly between endpoints, not
> >> through the gatekeepers or sip servers. Yet, people seem to be able to
> >> bill, account, maintain control, etc.
> >
> >See my above comment on accounting. So far people are doing accounting,
> >billing, ... etc at the ingress and egress domains only. This will
> >change as the volume of VoIP traffic increases.
> >
> 
> This will change as the volume of traffic increases, as the QoS
> requirements of the traffic varies, and as our dependency on the
> reliability of that traffic increases.  Whether the data is voice shouldn't
> be used as the differentiator.
> 
> - Matt
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC-6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: (408) 527 7147, Fax: (408) 527 1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 28 04:07:08 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA01055
	for <iptel-archive@odin.ietf.org>; Sun, 28 Mar 1999 04:07:08 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DFF2952BF; Sun, 28 Mar 1999 02:35:21 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8FC8552CE; Sun, 28 Mar 1999 02:35:19 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FDDC23.9D55AF73@cisco.com>
Date: Sat, 27 Mar 1999 23:37:07 -0800
From: Hussein F Salama <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: "Hutton, Charles" <charles.hutton@attws.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <AF62C6C0EC3DD211929900805FEA97CF6080DA@wltg-msg01.mccaw-stg.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Charles,

I agree with you, doing multiple transcodings is bad. Howerver,
tandemiing does not imply transcoding.

Hussein


"Hutton, Charles" wrote:
> 
> I think the implication here is that tandeming (transcoding) of codecs would
> / could be performed in AD4 and AD5. Unacceptable voice quality will quickly
> result with G.723.1 and G.729 if they are tandemed; subjective measurements
> show noticeable degradation with 1 tandeming and by 3 tandemings it is a
> struggle to understand completely.
> 
> Also I would think that the audio coding format would be a don't-care to AD4
> and AD5.
> 
> In summary (and hoping I have not misunderstood):
> 
>     - we can not tandem in every AD used in a call
>     - and luckily, we have no reason to tandem in each AD
> 
> > Hussein F. Salama wrote:
> > >
> > > > Most of the services you are talking about don't require the media to
> > > > flow through the LS. Certainly control over capability negotiation
> > does
> > > > not (in H.323, gatekeeper routed H.245 is sufficient for this, and in
> > > > SIP a proxy could rewrite SDP if it felt it needed to, for some
> > reason),
> > > > same with policy, security, accounting, and billing.
> > >
> > > You're missing my point. Assume call signaling goes through an ingress
> > > gateway in AD1, an intermediate signaling only proxy in AD2, and an
> > > egress gateway in AD3. However, the media streams for that call do not
> > > pass through AD2 at all. Assume the follow the path AD1-AD4-AD5-AD3.
> > > Let's use H.323 as an example. The gatekeeper in AD1 may be interested
> > > in controlling the capabilities because the ingress gateway is in its
> > > domain and the media streams flow through that domain. Same thing
> > > applies to the gatekeeper in the egress domain, AD3. But why would a
> > > gatekeeper in AD2 care for being involved in, and routing this call? I
> > > know it can, be I don't think it care. Since the layer 3 route doesn't
> > > go through AD2.
> >
> > A gatekeeper in AD2 would be interested since GLP effectively introduces
> > a "subcontracting relationship" among the various providers. AD1 allows
> > AD2 to use its gateways. By propagating this information to AD3, AD2 can
> > appear to be its own gateway provider, when in fact its using gateways
> > from AD1 to provide a more complete service. So, AD2 may want to see the
> > signaling messages for call accounting purposes.
> >
> > Same situation arises in building a house. One big contractor interfaces
> > with you for the job, and they in turn subcontract some work out to
> > other contactors, who may further subcontract that out. Invoices for
> > work still pass through all of them, since they each provide the
> > value-add of "aggregation of service".
> >
> > >
> > > What if the gatekeepers in AD4 and AD5 want to control the capabilities
> > > (codec selection) for  above call? I think it's their right to do that
> > > since the voice streams will ultimately flow through their domains.
> > > Would they be able to do that? Not when separating signaling from media.
> >
> > Well, this has not happened yet at all for any other IP application.
> > When I browse a web page, a hit a server in some domain AD1. I'm in AD2,
> > and I have a local proxy. The actual traffic flows through many
> > intermediate transit providers. To date, these providers have had no say
> > in whether I download the text or glitzy-image version of a page, or
> > whether I use gif or jpeg images.
> >
> > I do believe that an ISP should have control over what data leaves and
> > enters its network, but that this functionality lives at layer 3.
> > Diffserv and other things are how you control this. What you are
> > proposing requires us to overlay an independent network of proxies and
> > servers for every single application. One of the fundamental engineering
> > assumptions in the Internet is that it is application unaware, and that
> > apps live in the end systems. It is this model that has allowed new apps
> > (like the web) to come into being.
> >
> > >
> > > The same applies for accounting. What if the gatekeepers in AD4 and AD5
> > > want to do accounting for  above call? I think it's their right to
> > > account for voice traffic traversing their domain. Would they be able to
> > > do that?
> >
> > Accounting would be done in border routers (i.e., layer 3), using the
> > diffserv model. The voice traffic (or any other high priority, delay
> > sensitive traffic) would be marked at the ingress domain. Each border
> > router could account for the amount of this kind of traffic, and then
> > remark. So, you get the kind of accounting you want, but at layer 3.
> >
> > >
> > > > QoS for the media
> > > > is handled by things like rsvp and diffserv, not the call signaling in
> > > > any case.
> > >
> > > What if different intermediate domains on the media path of a call want
> > > to provide different levels of QoS to same media stream? Let's use
> > > RSVP/Intserv as an example. What if AD4 wants to provide controlled load
> > > service for the audio streams while AD5 prefers to provide guaranteed
> > > QoS?
> >
> > So, let them do it. Normally RSVP is used to ask for only one of the
> > two, but if for some reason a domain wanted to do something different,
> > it could perhaps rewrite the RSVP messages at the border routers, or
> > something like that. Certainly one domain can do diffserv, and another
> > rsvp. Check out draft-ietf-diffserv-rsvp. In either case, gatekeepers
> > and sip servers are not involved.
> >
> >
> > >
> > > > For both gatekeeper routed signaling in H.323, and proxy
> > > > invite's in SIP, the media flows regularly between endpoints, not
> > > > through the gatekeepers or sip servers. Yet, people seem to be able to
> > > > bill, account, maintain control, etc.
> > >
> > > See my above comment on accounting. So far people are doing accounting,
> > > billing, ... etc at the ingress and egress domains only. This will
> > > change as the volume of VoIP traffic increases.
> >
> > I agree we need accounting for voice traffic in intermediate domains.
> > Thats with diffserv border routers are for.
> >
> > > > I'm not so sure. We had lengthy discussions on this list regarding
> > > > directory services for GLP. It doesn't work. Many gateways can service
> > a
> > > > particular number; the right one to use depends on the agreements
> > > > between providers. This kind of environment cannot be supported in a
> > > > global directory solution, like DNS. Your alternative, a single hop
> > call
> > > > routing protocol, doesn't scale all that well. It would require
> > gateway
> > > > providers to all have bilateral agreements with each other. Would you
> > > > advocate regular IP routing protocols that support only one hop?
> > >
> > > I didn't use "single hop call routing" in the right context.What I meant
> > > "single hop inter domain call signaling." This can be done using either
> > > the BGP-based GLP proposal or the SCSP-based proposal, and never
> > > changing the NEXT_HOP attribute when an update is propagated from one
> > > domain to another. The NEXT_HOP will always point directly to the egress
> > > domain.
> >
> > I see, sorry for my misunderstanding. I think I address your point
> > already above. Intermediate LS's may want to be on the signaling path
> > for accounting (call usage, mind you, not voice level accounting). If
> > they don't want this, then you are correct. Single-hop signaling is just
> > fine. As I said, we have a perfectly nice underlying transport network
> > that allows us to go from anywhere to anywhere else.
> >
> > -Jonathan R.
> > --
> > Jonathan D. Rosenberg                       Lucent Technologies
> > Member of Technical Staff                   101 Crawfords Corner Rd.
> > High Speed Networks Research                Holmdel, NJ 07733
> > FAX: (732) 834-5379                         Rm. 4C-526
> > EMAIL: jdrosen@bell-labs.com
> > URL: http://www.cs.columbia.edu/~jdrosen
> >
> > ---------
> > This message came from the IETF IPTEL Working Group Mailing List.
> >

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC-6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: (408) 527 7147, Fax: (408) 527 1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 28 06:07:03 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA01437
	for <iptel-archive@odin.ietf.org>; Sun, 28 Mar 1999 06:07:02 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D5B0A52CD; Sun, 28 Mar 1999 03:09:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2076F52CF; Sun, 28 Mar 1999 03:09:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FDE43F.C941ADCE@cisco.com>
Date: Sun, 28 Mar 1999 00:11:43 -0800
From: Hussein F Salama <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com> <36F9B345.20F31E@cisco.com> <36FAEBE3.775E3AC@dnrc.bell-labs.com> <36FB0E4A.EC9ACDA5@cisco.com> <36FB17C3.E9903F4D@dnrc.bell-labs.com> <36FBEC76.715A1113@cisco.com> <36FC4DB5.C26A25F9@dnrc.bell-labs.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Jonathan Rosenberg wrote:
> 
> Hussein F. Salama wrote:
> >
> > > Most of the services you are talking about don't require the media to
> > > flow through the LS. Certainly control over capability negotiation does
> > > not (in H.323, gatekeeper routed H.245 is sufficient for this, and in
> > > SIP a proxy could rewrite SDP if it felt it needed to, for some reason),
> > > same with policy, security, accounting, and billing.
> >
> > You're missing my point. Assume call signaling goes through an ingress
> > gateway in AD1, an intermediate signaling only proxy in AD2, and an
> > egress gateway in AD3. However, the media streams for that call do not
> > pass through AD2 at all. Assume the follow the path AD1-AD4-AD5-AD3.
> > Let's use H.323 as an example. The gatekeeper in AD1 may be interested
> > in controlling the capabilities because the ingress gateway is in its
> > domain and the media streams flow through that domain. Same thing
> > applies to the gatekeeper in the egress domain, AD3. But why would a
> > gatekeeper in AD2 care for being involved in, and routing this call? I
> > know it can, be I don't think it care. Since the layer 3 route doesn't
> > go through AD2.
> 
> A gatekeeper in AD2 would be interested since GLP effectively introduces
> a "subcontracting relationship" among the various providers. AD1 allows
> AD2 to use its gateways. By propagating this information to AD3, AD2 can
> appear to be its own gateway provider, when in fact its using gateways
> from AD1 to provide a more complete service. So, AD2 may want to see the
> signaling messages for call accounting purposes.
> 
> Same situation arises in building a house. One big contractor interfaces
> with you for the job, and they in turn subcontract some work out to
> other contactors, who may further subcontract that out. Invoices for
> work still pass through all of them, since they each provide the
> value-add of "aggregation of service".

Now I understand your model better.

> 
> >
> > What if the gatekeepers in AD4 and AD5 want to control the capabilities
> > (codec selection) for  above call? I think it's their right to do that
> > since the voice streams will ultimately flow through their domains.
> > Would they be able to do that? Not when separating signaling from media.
> 
> Well, this has not happened yet at all for any other IP application.
> When I browse a web page, a hit a server in some domain AD1. I'm in AD2,
> and I have a local proxy. The actual traffic flows through many
> intermediate transit providers. To date, these providers have had no say
> in whether I download the text or glitzy-image version of a page, or
> whether I use gif or jpeg images.
> 
> I do believe that an ISP should have control over what data leaves and
> enters its network, but that this functionality lives at layer 3.
> Diffserv and other things are how you control this. What you are
> proposing requires us to overlay an independent network of proxies and
> servers for every single application. One of the fundamental engineering
> assumptions in the Internet is that it is application unaware, and that
> apps live in the end systems. It is this model that has allowed new apps
> (like the web) to come into being.

I wish this assumption can hold across multiple administrative domains,
each applying its own policy, in the presence of real-time applications
which use dynamic ports.

> 
> >
> > The same applies for accounting. What if the gatekeepers in AD4 and AD5
> > want to do accounting for  above call? I think it's their right to
> > account for voice traffic traversing their domain. Would they be able to
> > do that?
> 
> Accounting would be done in border routers (i.e., layer 3), using the
> diffserv model. The voice traffic (or any other high priority, delay
> sensitive traffic) would be marked at the ingress domain. Each border
> router could account for the amount of this kind of traffic, and then
> remark. So, you get the kind of accounting you want, but at layer 3.
> 
> >
> > > QoS for the media
> > > is handled by things like rsvp and diffserv, not the call signaling in
> > > any case.
> >
> > What if different intermediate domains on the media path of a call want
> > to provide different levels of QoS to same media stream? Let's use
> > RSVP/Intserv as an example. What if AD4 wants to provide controlled load
> > service for the audio streams while AD5 prefers to provide guaranteed
> > QoS?
> 
> So, let them do it. Normally RSVP is used to ask for only one of the
> two, but if for some reason a domain wanted to do something different,
> it could perhaps rewrite the RSVP messages at the border routers, or
> something like that. Certainly one domain can do diffserv, and another
> rsvp. Check out draft-ietf-diffserv-rsvp. In either case, gatekeepers
> and sip servers are not involved.

A border routers can not decide on its own to rewrite the RSVP messages.
It is an application unaware layer 3 device. Some application has to
instruct the border router to do that. 

The good news is that, in spite of the different models we're
envisioning,  GLP does not care with the intermediate hops on a call
route are signaling-only proxies or are signaling+media proxies.

Hussein


-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC-6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: (408) 527 7147, Fax: (408) 527 1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 28 17:54:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA04647
	for <iptel-archive@odin.ietf.org>; Sun, 28 Mar 1999 17:54:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3A63E52AA; Sun, 28 Mar 1999 17:49:21 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 91F5952C1; Sun, 28 Mar 1999 17:49:20 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990329084836.03ab4f74@pop.ozemail.com.au>
X-Sender: dclowes@pop.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Mon, 29 Mar 1999 08:48:36 +1000
To: "Stephen Sprunk" <ssprunk@cisco.com>
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: Re: question about H.323/SIP and MDCP/MGCP
Cc: <Hanlin_Fang@3com.com>, "list iptel" <iptel@lists.research.bell-labs.com>
In-Reply-To: <02fb01be7651$d6bb9a00$7d2544ab@glock.cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Hanlin,

This response is probably the closest one to date, but your request would
probably be better served on the "megaco" list, where the MGCP/MDCP people
hang out.

SIP and H.323 are overlapping standards at the call-control level, while
MGCP and MDCP are competing standards at the connection control level.

MxCP is targeted at operation _within_ a SIP/H.323 gateway, which is
decomposed into a controller and several media devices. It is possible for
MxCP to establish a media connection between two media devices and produce
what could be regarded as a call, if both media devices were under the
control of a _single_ controller.

To establish a connection between two media devices which were under the
control of _different_ controllers, an exchange of messages is required
between those controllers, to coordinate the connections - this is where
SIP and H.323 come in.

For a gateway that is not decomposed, MxCP is not required. Ideally, at the
SIP/H.323 level, it should not be required to know that MxCP is implemented
in a gateway. SIP and H.323 also operate with endpoints that are not gateways.

Hope this helps.

Douglas

At 17:55 1999-03-24 -0600, Stephen Sprunk wrote:
>MGCP/MDCP is (in theory) much simpler to implement, and therefore better
>suited for mass-production dumb devices like desktop feature phones or
>high-density media gateways; those dumb devices would then be controlled by
>a intelligent SIP/H.323 device.  You could implement H.323 or SIP end-to-end
>if you wished.
>
>Stephen
>
>     |          |         Stephen Sprunk, K5SSS, CCIE #3723
>    :|:        :|:        NSA, Network Consulting Engineer
>   :|||:      :|||:       14875 Landmark Blvd #400; Dallas, TX
>.:|||||||:..:|||||||:.    Pager: 800-365-4578 / 800-901-6078
>C I S C O S Y S T E M S   Email: ssprunk@cisco.com
>
>-----Original Message-----
>From: Hanlin_Fang@3com.com <Hanlin_Fang@3com.com>
>To: list iptel <iptel@lists.research.bell-labs.com>
>Date: Wednesday, March 24, 1999 17:23
>Subject: question about H.323/SIP and MDCP/MGCP
>
>
>>
>>
>>I have couple of questions about both signalling protocols in VoIP.
>>   Why we need separate protocols for internal signalling (MDCP/MGCP) and
>>   external signalling (H.323/SIP)? Can we use H.323/SIP or MDCP/MGCP for
>both
>>   external and internal signalling?
>>   If not, why? By my understanding, MDCP/MGCP have nothing to do with
>>   end-to-end call setup, even there is callID field in their connection
>setup
>>   messages. Is this right? Any other issues?
>>   Why H.323/SIP can do it either?
>>
>>Any comments are appreciated!
>>
>>Hanlin
>>
>>
>>
>>---------
>>This message came from the IETF IPTEL Working Group Mailing List.
>
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.
>
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Mar 28 23:07:07 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA07933
	for <iptel-archive@odin.ietf.org>; Sun, 28 Mar 1999 23:07:07 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1EB3152BB; Sun, 28 Mar 1999 22:39:31 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 762BE52C2; Sun, 28 Mar 1999 22:39:30 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FEF5D7.8F368C54@dnrc.bell-labs.com>
Date: Sun, 28 Mar 1999 22:39:03 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
Organization: Bell Laboratories
X-Mailer: Mozilla 4.05 [en] (Win95; U)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com> <36F9B345.20F31E@cisco.com> <36FAEBE3.775E3AC@dnrc.bell-labs.com> <36FB0E4A.EC9ACDA5@cisco.com> <36FB17C3.E9903F4D@dnrc.bell-labs.com> <36FBEC76.715A1113@cisco.com> <36FC4DB5.C26A25F9@dnrc.bell-labs.com> <36FDE43F.C941ADCE@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F Salama wrote:
> 
> The good news is that, in spite of the different models we're
> envisioning,  GLP does not care with the intermediate hops on a call
> route are signaling-only proxies or are signaling+media proxies.

This is true. However, the reason we came down this road was to address
the issue of the capacity metric. We have three definitions possible:

1. capacity of egress gateway
2. capacity of call signaling path (set of LS's signaling messages
touch)
3. capacity of advertisement path (set of LS's the GLP messages
traverse; a superset of 2)

If you believe that the media should flow through the LS, then (2) is
what you want. If you believe its just signaling, than (1) is what you
want.

In reality, its an artificial distinction. Nothing would actually
prevent an LS from modifying the capacity metric bceause it also acts as
one of these media proxies. So, my proposal is to define it as an egress
gateway capacity, in terms of total calls (or possibly density), and
define aggregation rules based on the assumption they don't reflect
signaling path capacity.

-Jonathan R.
-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX: (732) 834-5379                         Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 29 00:05:09 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA08683
	for <iptel-archive@odin.ietf.org>; Mon, 29 Mar 1999 00:05:09 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0495952C2; Sun, 28 Mar 1999 23:59:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 5BE5552C3; Sun, 28 Mar 1999 23:59:16 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990329145734.006fc820@pop.ozemail.com.au>
X-Sender: dclowes@pop.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Mon, 29 Mar 1999 14:57:34 +1000
To: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: Re: data issue 4: capacity
Cc: hsalama@cisco.com,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
In-Reply-To: <36FEF5D7.8F368C54@dnrc.bell-labs.com>
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com>
 <36F9B345.20F31E@cisco.com>
 <36FAEBE3.775E3AC@dnrc.bell-labs.com>
 <36FB0E4A.EC9ACDA5@cisco.com>
 <36FB17C3.E9903F4D@dnrc.bell-labs.com>
 <36FBEC76.715A1113@cisco.com>
 <36FC4DB5.C26A25F9@dnrc.bell-labs.com>
 <36FDE43F.C941ADCE@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Actually, the metric seems to be, and should be, related to the capacity of
the mechanism which is being propagated.

This may change in meaning as it is propagated, increasing as aggregation
occurs, and decreasing if some other limit is introduced.

If an intermediate LS wants to aggregate three routes to 1-415* that have
capacity of 10, it should propagate a capacity of thirty. Since the three
routes have different signalling addresses, the aggregation should
introduce a common intermediate signalling entity, which can distribute the
requests among the three paths. If the interposed signalling entity has
unlimited call carrying capacity, the LS should propagate the aggregated
capacity of 30.

If an intermediate LS wants to introduce a signalling/media intermediary in
the path, which has a limitation of its own, then it should adjust the
metric to reflect that. If that signalling+media proxy has a carrying
capacity of 3 calls, then that should be the capacity that is propagated,
even though the gateways have greater capacity.

In one of the above, the capacity is 1 below, while in the other it is 2.
They're both valid.

BTW: can you aggregate without introducing an intermediate signalling address?

Douglas

At 22:39 1999-03-28 -0500, Jonathan Rosenberg wrote:
>Hussein F Salama wrote:
>> 
>> The good news is that, in spite of the different models we're
>> envisioning,  GLP does not care with the intermediate hops on a call
>> route are signaling-only proxies or are signaling+media proxies.
>
>This is true. However, the reason we came down this road was to address
>the issue of the capacity metric. We have three definitions possible:
>
>1. capacity of egress gateway
>2. capacity of call signaling path (set of LS's signaling messages
>touch)
>3. capacity of advertisement path (set of LS's the GLP messages
>traverse; a superset of 2)
>
>If you believe that the media should flow through the LS, then (2) is
>what you want. If you believe its just signaling, than (1) is what you
>want.
>
>In reality, its an artificial distinction. Nothing would actually
>prevent an LS from modifying the capacity metric bceause it also acts as
>one of these media proxies. So, my proposal is to define it as an egress
>gateway capacity, in terms of total calls (or possibly density), and
>define aggregation rules based on the assumption they don't reflect
>signaling path capacity.
>
>-Jonathan R.
>-- 
>Jonathan D. Rosenberg                       Lucent Technologies
>Member of Technical Staff                   101 Crawfords Corner Rd.
>High Speed Networks Research                Holmdel, NJ 07733
>FAX: (732) 834-5379                         Rm. 4C-526
>EMAIL: jdrosen@bell-labs.com
>URL: http://www.cs.columbia.edu/~jdrosen
>
>---------
>This message came from the IETF IPTEL Working Group Mailing List.
>
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 29 03:07:10 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA16578
	for <iptel-archive@odin.ietf.org>; Mon, 29 Mar 1999 03:07:10 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 54C9852C3; Mon, 29 Mar 1999 02:27:15 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A67EE52C6; Mon, 29 Mar 1999 02:27:14 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FF2BD4.E9870964@cisco.com>
Date: Sun, 28 Mar 1999 23:29:24 -0800
From: Hussein F Salama <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (Win95; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Douglas Clowes <dclowes@ozemail.com.au>
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com>
	 <36F9B345.20F31E@cisco.com>
	 <36FAEBE3.775E3AC@dnrc.bell-labs.com>
	 <36FB0E4A.EC9ACDA5@cisco.com>
	 <36FB17C3.E9903F4D@dnrc.bell-labs.com>
	 <36FBEC76.715A1113@cisco.com>
	 <36FC4DB5.C26A25F9@dnrc.bell-labs.com>
	 <36FDE43F.C941ADCE@cisco.com> <3.0.3.32.19990329145734.006fc820@pop.ozemail.com.au>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit



Douglas Clowes wrote:
> 
> Actually, the metric seems to be, and should be, related to the capacity of
> the mechanism which is being propagated.
> 
> This may change in meaning as it is propagated, increasing as aggregation
> occurs, and decreasing if some other limit is introduced.
> 
> If an intermediate LS wants to aggregate three routes to 1-415* that have
> capacity of 10, it should propagate a capacity of thirty. Since the three
> routes have different signalling addresses, the aggregation should
> introduce a common intermediate signalling entity, which can distribute the
> requests among the three paths. If the interposed signalling entity has
> unlimited call carrying capacity, the LS should propagate the aggregated
> capacity of 30.
> 
> If an intermediate LS wants to introduce a signalling/media intermediary in
> the path, which has a limitation of its own, then it should adjust the
> metric to reflect that. If that signalling+media proxy has a carrying
> capacity of 3 calls, then that should be the capacity that is propagated,
> even though the gateways have greater capacity.
> 
> In one of the above, the capacity is 1 below, while in the other it is 2.
> They're both valid.

But we have to agree on one of the two, because the capacity will be
propagated inter-domain.

> 
> BTW: can you aggregate without introducing an intermediate signalling address?

Good point, I think an aggregator must change the next hop address.

Hussein


> 
> Douglas
> 
> At 22:39 1999-03-28 -0500, Jonathan Rosenberg wrote:
> >Hussein F Salama wrote:
> >>
> >> The good news is that, in spite of the different models we're
> >> envisioning,  GLP does not care with the intermediate hops on a call
> >> route are signaling-only proxies or are signaling+media proxies.
> >
> >This is true. However, the reason we came down this road was to address
> >the issue of the capacity metric. We have three definitions possible:
> >
> >1. capacity of egress gateway
> >2. capacity of call signaling path (set of LS's signaling messages
> >touch)
> >3. capacity of advertisement path (set of LS's the GLP messages
> >traverse; a superset of 2)
> >
> >If you believe that the media should flow through the LS, then (2) is
> >what you want. If you believe its just signaling, than (1) is what you
> >want.
> >
> >In reality, its an artificial distinction. Nothing would actually
> >prevent an LS from modifying the capacity metric bceause it also acts as
> >one of these media proxies. So, my proposal is to define it as an egress
> >gateway capacity, in terms of total calls (or possibly density), and
> >define aggregation rules based on the assumption they don't reflect
> >signaling path capacity.
> >
> >-Jonathan R.
> >--
> >Jonathan D. Rosenberg                       Lucent Technologies
> >Member of Technical Staff                   101 Crawfords Corner Rd.
> >High Speed Networks Research                Holmdel, NJ 07733
> >FAX: (732) 834-5379                         Rm. 4C-526
> >EMAIL: jdrosen@bell-labs.com
> >URL: http://www.cs.columbia.edu/~jdrosen
> >
> >---------
> >This message came from the IETF IPTEL Working Group Mailing List.
> >
> >
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Hussein F. Salama
Cisco Systems
Mail Stop SJC-6/3, 170 W. Tasman Drive, San Jose, CA 95134
Voice: (408) 527 7147, Fax: (408) 527 1714

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 29 08:05:43 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA17974
	for <iptel-archive@odin.ietf.org>; Mon, 29 Mar 1999 08:05:42 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id B432B52C1; Mon, 29 Mar 1999 07:37:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1478252C8; Mon, 29 Mar 1999 07:37:18 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <3.0.3.32.19990329223550.03a96e28@pop.ozemail.com.au>
X-Sender: dclowes@pop.ozemail.com.au
X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32)
Date: Mon, 29 Mar 1999 22:35:50 +1000
To: hsalama@cisco.com
From: Douglas Clowes <dclowes@ozemail.com.au>
Subject: Re: data issue 4: capacity
Cc: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
In-Reply-To: <36FF2BD4.E9870964@cisco.com>
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com>
 <36F9B345.20F31E@cisco.com>
 <36FAEBE3.775E3AC@dnrc.bell-labs.com>
 <36FB0E4A.EC9ACDA5@cisco.com>
 <36FB17C3.E9903F4D@dnrc.bell-labs.com>
 <36FBEC76.715A1113@cisco.com>
 <36FC4DB5.C26A25F9@dnrc.bell-labs.com>
 <36FDE43F.C941ADCE@cisco.com>
 <3.0.3.32.19990329145734.006fc820@pop.ozemail.com.au>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

At 23:29 1999-03-28 -0800, Hussein F Salama wrote:
>
>
>Douglas Clowes wrote:
>> 

>> In one of the above, the capacity is 1 below, while in the other it is 2.
>> They're both valid.
>
>But we have to agree on one of the two, because the capacity will be
>propagated inter-domain.

I don't think so, because it's still the capacity to carry calls. It's just
a different factor that's limiting that capacity. Whatever is determining
what the limitation is, has to determine what the resultant capacity is.

>> 
>> BTW: can you aggregate without introducing an intermediate signalling
address?
>
>Good point, I think an aggregator must change the next hop address.
>
>Hussein
>
>
>> 
>> Douglas
>> 
>> At 22:39 1999-03-28 -0500, Jonathan Rosenberg wrote:
>> >Hussein F Salama wrote:
>> >>
>> >> The good news is that, in spite of the different models we're
>> >> envisioning,  GLP does not care with the intermediate hops on a call
>> >> route are signaling-only proxies or are signaling+media proxies.
>> >
>> >This is true. However, the reason we came down this road was to address
>> >the issue of the capacity metric. We have three definitions possible:
>> >
>> >1. capacity of egress gateway
>> >2. capacity of call signaling path (set of LS's signaling messages
>> >touch)
>> >3. capacity of advertisement path (set of LS's the GLP messages
>> >traverse; a superset of 2)
>> >
>> >If you believe that the media should flow through the LS, then (2) is
>> >what you want. If you believe its just signaling, than (1) is what you
>> >want.
>> >
>> >In reality, its an artificial distinction. Nothing would actually
>> >prevent an LS from modifying the capacity metric bceause it also acts as
>> >one of these media proxies. So, my proposal is to define it as an egress
>> >gateway capacity, in terms of total calls (or possibly density), and
>> >define aggregation rules based on the assumption they don't reflect
>> >signaling path capacity.
>> >
>> >-Jonathan R.
>> >--
>> >Jonathan D. Rosenberg                       Lucent Technologies
>> >Member of Technical Staff                   101 Crawfords Corner Rd.
>> >High Speed Networks Research                Holmdel, NJ 07733
>> >FAX: (732) 834-5379                         Rm. 4C-526
>> >EMAIL: jdrosen@bell-labs.com
>> >URL: http://www.cs.columbia.edu/~jdrosen
>> >
>> >---------
>> >This message came from the IETF IPTEL Working Group Mailing List.
>> >
>> >
>> 
>> ---------
>> This message came from the IETF IPTEL Working Group Mailing List.
>
>-- 
>Hussein F. Salama
>Cisco Systems
>Mail Stop SJC-6/3, 170 W. Tasman Drive, San Jose, CA 95134
>Voice: (408) 527 7147, Fax: (408) 527 1714
>
>


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Mar 29 14:19:47 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA21841
	for <iptel-archive@odin.ietf.org>; Mon, 29 Mar 1999 14:19:45 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 27DE552CA; Mon, 29 Mar 1999 14:16:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id CD2FA52CB; Mon, 29 Mar 1999 14:16:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <36FF9EAF.660DA5D2@dnrc.bell-labs.com>
Date: Mon, 29 Mar 1999 10:39:27 -0500
From: Jonathan Rosenberg <jdrosen@dnrc.bell-labs.com>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: Douglas Clowes <dclowes@ozemail.com.au>,
        IPTEL IETF Mailing List <iptel@lists.research.bell-labs.com>
Subject: Re: data issue 4: capacity
References: <3.0.32.19990324082921.0079a390@bl-mail1.corpeast.baynetworks.com>
		 <36F9B345.20F31E@cisco.com>
		 <36FAEBE3.775E3AC@dnrc.bell-labs.com>
		 <36FB0E4A.EC9ACDA5@cisco.com>
		 <36FB17C3.E9903F4D@dnrc.bell-labs.com>
		 <36FBEC76.715A1113@cisco.com>
		 <36FC4DB5.C26A25F9@dnrc.bell-labs.com>
		 <36FDE43F.C941ADCE@cisco.com> <3.0.3.32.19990329145734.006fc820@pop.ozemail.com.au> <36FF2BD4.E9870964@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hussein F Salama wrote:
> 
> >
> > BTW: can you aggregate without introducing an intermediate signalling address?
> 
> Good point, I think an aggregator must change the next hop address.

It must do so only if the prefixes it aggregates have different next
hops. In most cases this will be true, but not always.

-Jonathan R.


-- 
Jonathan D. Rosenberg                       Lucent Technologies
Member of Technical Staff                   101 Crawfords Corner Rd.
High Speed Networks Research                Holmdel, NJ 07733
FAX:   (732) 834-5379                       Rm. 4C-526
EMAIL: jdrosen@bell-labs.com
URL: http://www.cs.columbia.edu/~jdrosen

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 31 10:18:01 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA25216
	for <iptel-archive@odin.ietf.org>; Wed, 31 Mar 1999 10:18:01 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2360952CF; Wed, 31 Mar 1999 10:13:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 84A7F52D0; Wed, 31 Mar 1999 10:13:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Sailaja_Talagadadeevi@3com.com
X-Lotus-FromDomain: 3COM
To: Douglas Clowes <dclowes@ozemail.com.au>
Cc: "Stephen Sprunk" <ssprunk@cisco.com>, Hanlin_Fang@3com.com,
        "list iptel" <iptel@lists.research.bell-labs.com>
Message-ID: <88256745.0053745A.00@hqoutbound.ops.3com.com>
Date: Wed, 31 Mar 1999 09:14:47 -0600
Subject: Re: question about H.323/SIP and MDCP/MGCP
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk




Hanlin,


This was part of a message Tim Miller sent over megaco mailing lists. I think
his mail gives reasonal insight into/concerns about SIP/H323 Vs MDCP/MGCP
issues. In adition to to the explanations from Douglas and Stephen this will
give a big picture of where SIP/H323 and MDCP/MGCP fit into VOIP...


"......My view is that one can have an MGC-MG control protocol without ever
mentioning H.323, ETSI TIPHON, or SIP. And if we had restricted our scope
to this as the charter suggests, we would not have all of these requirement
and definition problems proliferating the mail exploder. There seems to be
belief among the H.323 and ETSI TIPHON people that the MGCP backers want to
KILL H.323. Having friends and colleagues among both groups, that simple is
not the case. The truth is that some Network Service Providers or MSO
(cable providers) are not interested in the providing the multimedia
aspects of H.323 or SIP; there service objective to leverage VoIP as an
alternative to providing PSTN-like services using the customers existing
PSTN analog phones. This is the "thin client" model and some of these have
also defined the "rich client" model that would then use H.323 or SIP to
control PCs and enhanced multimedia phones. SIP and H.323 are already
standards for these applications and migrating the MEGACO charter to align
IETF, ITU-T SG16, and ETSI TIPHON so that the MGC-MG protocol is identical,
while a noble goal, may do an injustice to the original goal of producing a
"thin client" functionality needed, both for scalability and cost issues.
Both residential (be they cable modems or xDSL modem) and very large
trunking gateways (a needed network element that is essentially impossible
in an H.323 or SIP model), require a thin-client approach.

Limited focus was clearly demonstrated by the MGCP proponents and the flows
provided good indication of the VoIP features (but none multimedia
features)  that can be provided. Clearly, the MGC could have H.323 and SIP
proxy capability when the deployed network requires it, but the MGC-MG
protocol need not have to carry this baggage. For example, I'm not a
believer that H.245 (and I assume  H.235 for security, etc., etc.) should
ALWAYS be in the MG. I'm not saying the final MGC-MG has to be MGCP, but I
do think the protocol should not call up other protocols that require more
complexity (e.g., H.323 IPR stack issues, etc.) and it should be optimized
to the thin-client model.

I thought the purpose of this protocol was to allow for carrier-class
scalability to trunking and residential gateways requiring the equivalent
of PSTN POTS in a VoIP environment. I think at least some of the network
service providers would like to see gateways that eventually pass the
24-500 line range limits we are achieving with the H.323 model, not to
mention getting the cost per port down to reasonable levels............"




---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 31 10:51:53 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA25576
	for <iptel-archive@odin.ietf.org>; Wed, 31 Mar 1999 10:51:52 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 29F0752D2; Wed, 31 Mar 1999 10:47:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 833CB52D4; Wed, 31 Mar 1999 10:47:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <8EF3FE311AA1D111B07E00A0C92C816F6EE7D4@boca101a.boca.rolm.com>
From: "Callaghan, Robert" <Robert.Callaghan@icn.siemens.com>
To: "'Sailaja_Talagadadeevi@3com.com'" <Sailaja_Talagadadeevi@3com.com>,
        Douglas Clowes <dclowes@ozemail.com.au>
Cc: Stephen Sprunk <ssprunk@cisco.com>, Hanlin_Fang@3com.com,
        list iptel <iptel@lists.research.bell-labs.com>
Subject: RE: question about H.323/SIP and MDCP/MGCP
Date: Wed, 31 Mar 1999 07:44:37 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Sailaja,

If, as you say, it is the desire of some providers to supply an alternative
to PSTN services, then how does the Megaco proposal cover the multimedia
capabilities provided by PSTN using H.320 or H.324?

Bob

------------------------------------------------------------------
Robert Callaghan
Siemens Information & Communication Networks
Tel: +1.561.997.3756	Fax: +1.561.997.3403
Email:	Robert.Callaghan@ICN.Siemens.com
------------------------------------------------------------------


-----Original Message-----
From: Sailaja_Talagadadeevi@3com.com
[mailto:Sailaja_Talagadadeevi@3com.com]
Sent: Wednesday, March 31, 1999 10:15 AM
To: Douglas Clowes
Cc: Stephen Sprunk; Hanlin_Fang@3com.com; list iptel
Subject: Re: question about H.323/SIP and MDCP/MGCP





Hanlin,


This was part of a message Tim Miller sent over megaco mailing lists. I
think
his mail gives reasonal insight into/concerns about SIP/H323 Vs MDCP/MGCP
issues. In adition to to the explanations from Douglas and Stephen this will
give a big picture of where SIP/H323 and MDCP/MGCP fit into VOIP...


"......My view is that one can have an MGC-MG control protocol without ever
mentioning H.323, ETSI TIPHON, or SIP. And if we had restricted our scope
to this as the charter suggests, we would not have all of these requirement
and definition problems proliferating the mail exploder. There seems to be
belief among the H.323 and ETSI TIPHON people that the MGCP backers want to
KILL H.323. Having friends and colleagues among both groups, that simple is
not the case. The truth is that some Network Service Providers or MSO
(cable providers) are not interested in the providing the multimedia
aspects of H.323 or SIP; there service objective to leverage VoIP as an
alternative to providing PSTN-like services using the customers existing
PSTN analog phones. This is the "thin client" model and some of these have
also defined the "rich client" model that would then use H.323 or SIP to
control PCs and enhanced multimedia phones. SIP and H.323 are already
standards for these applications and migrating the MEGACO charter to align
IETF, ITU-T SG16, and ETSI TIPHON so that the MGC-MG protocol is identical,
while a noble goal, may do an injustice to the original goal of producing a
"thin client" functionality needed, both for scalability and cost issues.
Both residential (be they cable modems or xDSL modem) and very large
trunking gateways (a needed network element that is essentially impossible
in an H.323 or SIP model), require a thin-client approach.

Limited focus was clearly demonstrated by the MGCP proponents and the flows
provided good indication of the VoIP features (but none multimedia
features)  that can be provided. Clearly, the MGC could have H.323 and SIP
proxy capability when the deployed network requires it, but the MGC-MG
protocol need not have to carry this baggage. For example, I'm not a
believer that H.245 (and I assume  H.235 for security, etc., etc.) should
ALWAYS be in the MG. I'm not saying the final MGC-MG has to be MGCP, but I
do think the protocol should not call up other protocols that require more
complexity (e.g., H.323 IPR stack issues, etc.) and it should be optimized
to the thin-client model.

I thought the purpose of this protocol was to allow for carrier-class
scalability to trunking and residential gateways requiring the equivalent
of PSTN POTS in a VoIP environment. I think at least some of the network
service providers would like to see gateways that eventually pass the
24-500 line range limits we are achieving with the H.323 model, not to
mention getting the cost per port down to reasonable levels............"




---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 31 13:59:18 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA27777
	for <iptel-archive@odin.ietf.org>; Wed, 31 Mar 1999 13:59:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 67DE452BF; Wed, 31 Mar 1999 13:05:35 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3F98352D4; Wed, 31 Mar 1999 13:05:33 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <19990331180452.27228.rocketmail@web106.yahoomail.com>
Date: Wed, 31 Mar 1999 10:04:52 -0800 (PST)
From: Tom Gray <tom_gray_intp@yahoo.com>
Subject: Quality Measures for IP Telephoney
To: iptel@lists.research.bell-labs.com, tom_gray_intp@yahoo.com
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



Can the list please inform me of any efforts that are going on to
produce a set of quality measures for the performance of IP telephoney.
These would be measures which could take as input a set of network
conditions affecting the packet stream and output an indication of the
psychologically perceived quality of the voice.

Tom Gray



_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com


---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 31 16:46:14 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA01812
	for <iptel-archive@odin.ietf.org>; Wed, 31 Mar 1999 16:46:14 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id EABE452B7; Wed, 31 Mar 1999 15:43:18 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4B6FC52C6; Wed, 31 Mar 1999 15:43:17 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: Sailaja_Talagadadeevi@3com.com
X-Lotus-FromDomain: 3COM
To: "Callaghan, Robert" <Robert.Callaghan@icn.siemens.com>
Cc: list iptel <iptel@lists.research.bell-labs.com>
Message-ID: <88256745.0071ACD2.00@hqoutbound.ops.3com.com>
Date: Wed, 31 Mar 1999 14:47:55 -0600
Subject: RE: question about H.323/SIP and MDCP/MGCP
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk



Robert,
Miller's message captures some of the ideas and concerns about
MGCP/MDCP/H323/SIP protocol placement in VOIP/Multimedia.As he said, MGCP's
origination was based on thin client model that will support carrier class load.
H323/320/324 proxies can be accommodated along with the new protocol (
megaco/MGCP) at MG level or MGC level  if multimedia functionality is necessary.
That does not mean that megaco protocol  has to have multimedia capability.
I guess, to implement SIP/H323 on MG or MGC should be decided by the vendor. If
the requirement is to have plain vanilla VOIP features, there is  not much use
of developing an elixir protocol.
But I believe there is some work going on to have multimedia features
implemented in megaco. H320 means having audio/video/T 120 data on the same
channel. So typically we end up with multiple media streams for the same
call/user.  So the MG and MGC will end up having 3 different contexts (depends
on the design ) or just one single context for all the streams. If I am right,
these issues are not solved yet. But I hope will be addressed  soon.





"Callaghan, Robert" <Robert.Callaghan@icn.siemens.com> on 03/31/99 09:44:37 AM

To:   Sailaja Talagadadeevi/MW/US/3Com, Douglas Clowes <dclowes@ozemail.com.au>
cc:   Stephen Sprunk <ssprunk@cisco.com>, Hanlin Fang/HQ/3Com, list iptel
      <iptel@lists.research.bell-labs.com>
Subject:  RE: question about H.323/SIP and MDCP/MGCP




Sailaja,

If, as you say, it is the desire of some providers to supply an alternative
to PSTN services, then how does the Megaco proposal cover the multimedia
capabilities provided by PSTN using H.320 or H.324?

Bob

------------------------------------------------------------------
Robert Callaghan
Siemens Information & Communication Networks
Tel: +1.561.997.3756     Fax: +1.561.997.3403
Email:    Robert.Callaghan@ICN.Siemens.com
------------------------------------------------------------------


-----Original Message-----
From: Sailaja_Talagadadeevi@3com.com
[mailto:Sailaja_Talagadadeevi@3com.com]
Sent: Wednesday, March 31, 1999 10:15 AM
To: Douglas Clowes
Cc: Stephen Sprunk; Hanlin_Fang@3com.com; list iptel
Subject: Re: question about H.323/SIP and MDCP/MGCP





Hanlin,


This was part of a message Tim Miller sent over megaco mailing lists. I
think
his mail gives reasonal insight into/concerns about SIP/H323 Vs MDCP/MGCP
issues. In adition to to the explanations from Douglas and Stephen this will
give a big picture of where SIP/H323 and MDCP/MGCP fit into VOIP...


"......My view is that one can have an MGC-MG control protocol without ever
mentioning H.323, ETSI TIPHON, or SIP. And if we had restricted our scope
to this as the charter suggests, we would not have all of these requirement
and definition problems proliferating the mail exploder. There seems to be
belief among the H.323 and ETSI TIPHON people that the MGCP backers want to
KILL H.323. Having friends and colleagues among both groups, that simple is
not the case. The truth is that some Network Service Providers or MSO
(cable providers) are not interested in the providing the multimedia
aspects of H.323 or SIP; there service objective to leverage VoIP as an
alternative to providing PSTN-like services using the customers existing
PSTN analog phones. This is the "thin client" model and some of these have
also defined the "rich client" model that would then use H.323 or SIP to
control PCs and enhanced multimedia phones. SIP and H.323 are already
standards for these applications and migrating the MEGACO charter to align
IETF, ITU-T SG16, and ETSI TIPHON so that the MGC-MG protocol is identical,
while a noble goal, may do an injustice to the original goal of producing a
"thin client" functionality needed, both for scalability and cost issues.
Both residential (be they cable modems or xDSL modem) and very large
trunking gateways (a needed network element that is essentially impossible
in an H.323 or SIP model), require a thin-client approach.

Limited focus was clearly demonstrated by the MGCP proponents and the flows
provided good indication of the VoIP features (but none multimedia
features)  that can be provided. Clearly, the MGC could have H.323 and SIP
proxy capability when the deployed network requires it, but the MGC-MG
protocol need not have to carry this baggage. For example, I'm not a
believer that H.245 (and I assume  H.235 for security, etc., etc.) should
ALWAYS be in the MG. I'm not saying the final MGC-MG has to be MGCP, but I
do think the protocol should not call up other protocols that require more
complexity (e.g., H.323 IPR stack issues, etc.) and it should be optimized
to the thin-client model.

I thought the purpose of this protocol was to allow for carrier-class
scalability to trunking and residential gateways requiring the equivalent
of PSTN POTS in a VoIP environment. I think at least some of the network
service providers would like to see gateways that eventually pass the
24-500 line range limits we are achieving with the H.323 model, not to
mention getting the cost per port down to reasonable levels............"




---------
This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.






---------
This message came from the IETF IPTEL Working Group Mailing List.


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Mar 31 22:55:48 1999
Received: from lists.research.bell-labs.com (paperless.dnrc.bell-labs.com [135.180.161.172])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA10053
	for <iptel-archive@odin.ietf.org>; Wed, 31 Mar 1999 22:55:47 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5540652C5; Wed, 31 Mar 1999 22:51:16 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B4ACF52C8; Wed, 31 Mar 1999 22:51:15 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <C51ED84B6F47D211917A0000F8BCBD1101002299@zcard00g.ca.nortel.com>
From: "Tom-PT Taylor" <taylor@nortelnetworks.com>
To: iptel@lists.research.bell-labs.com
Subject: RE: Quality Measures for IP Telephoney
Date: Wed, 31 Mar 1999 21:45:23 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: text/plain
X-Orig: <taylor@americasm01.nt.com>
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

For this you should look at the work being done in ITU-T Study Group 12 and
in ETSI TIPHON.

> -----Original Message-----
> From:	Tom Gray [SMTP:tom_gray_intp@yahoo.com]
> Sent:	Wednesday, March 31, 1999 1:05 PM
> To:	iptel@lists.research.bell-labs.com; tom_gray_intp@yahoo.com
> Subject:	Quality Measures for IP Telephoney
> 
> 
> 
> Can the list please inform me of any efforts that are going on to
> produce a set of quality measures for the performance of IP telephoney.
> These would be measures which could take as input a set of network
> conditions affecting the packet stream and output an indication of the
> psychologically perceived quality of the voice.
> 
> Tom Gray
> 
> 
> 
> _________________________________________________________
> Do You Yahoo!?
> Get your free @yahoo.com address at http://mail.yahoo.com
> 
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

---------
This message came from the IETF IPTEL Working Group Mailing List.


