From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec  1 07:06:33 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 HAA18336
	for <iptel-archive@odin.ietf.org>; Wed, 1 Dec 1999 07:06:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3679752AB; Wed,  1 Dec 1999 07:01:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id AE3DC52BB; Wed,  1 Dec 1999 07:01:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199912011159.GAA14495@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-06.txt
Date: Wed, 01 Dec 1999 06:59:12 -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 Telephony Routing over IP
	Author(s)	: J. Rosenberg, H. Schulzrinne
	Filename	: draft-ietf-iptel-gwloc-framework-06.txt
	Pages		: 24
	Date		: 30-Nov-99
	
This document serves as a framework for Telephony Routing over IP
(TRIP), which supports the discovery and exchange of IP telephony
gateway routing tables between providers. The document defines the
problem of telephony routing exchange, and motivates the need for the
protocol. It presents an architectural framework for TRIP, defines
terminology, specifies the various protocol 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-06.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-06.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-06.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:	<19991130105345.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-iptel-gwloc-framework-06.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-iptel-gwloc-framework-06.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<19991130105345.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  Wed Dec  1 13:56:59 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 NAA10891
	for <iptel-archive@odin.ietf.org>; Wed, 1 Dec 1999 13:56:58 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 75A5652DB; Wed,  1 Dec 1999 13:51:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C84B152DC; Wed,  1 Dec 1999 13:51:28 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Wed, 01 Dec 1999 11:32:11 -0600
From: Henry Sinnreich <henry.sinnreich@wcom.com>
Subject: RE: question on OSP
In-reply-to: <4142706ABBE2D211A1B700805F9FA43B4E9A83@exchange.vipcalling.com>
To: Phil Mutooni <PMutooni@ibasis.net>, Philippe.Terrien@alcatel.fr,
        "'Stephen Thomas'" <stephen.thomas@transnexus.com>
Cc: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>,
        "'Gerry Reilly'" <gerry@projects.telecity.co.uk>,
        "'IPTEL IETF Mailing List'" <iptel@lists.research.bell-labs.com>,
        "'Matt Squire'" <msquire@nortelnetworks.com>
Message-id: <NBBBIIJFOKPMFOOILMBKOEOKEOAA.henry.sinnreich@wcom.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Content-type: text/plain;	charset="Windows-1252"
Content-transfer-encoding: 7bit
Importance: Normal
X-Priority: 3 (Normal)
X-MSMail-priority: Normal
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Page 50 shows the "ETSI Protocol Stack". There is no such thing. These (IP,
TCP, SSL/TLS, HTTP) ) are all IETF protocols and the IETF has the
intellectual property rights and change control. I believe the same applies
to W3C web protocols, XML, etc.

Henry

> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Phil
> Mutooni
> Sent: Tuesday, November 30, 1999 4:42 PM
> To: 'Philippe.Terrien@alcatel.fr'; 'Stephen Thomas'
> Cc: 'Jonathan Rosenberg'; 'Gerry Reilly'; 'IPTEL IETF Mailing List';
> 'Matt Squire'
> Subject: RE: question on OSP
>
>
> Attached is a paper that I wrote early this year that compares
> the pros and
> cons of
> the OSP vs the ITU-T H.225 Annex G. It appeared in the June issue
> of Billing
> World.
>
> Philip Mutooni
> iBasis, Inc.
> w: http://www.ibasis.net
> t: 781.505.7523
> f: 781.505.7300
>
> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of
> Philippe.Terrien@alcatel.fr
> Sent: Tuesday, November 30, 1999 6:12 AM
> To: Stephen Thomas
> Cc: Jonathan Rosenberg; Gerry Reilly; IPTEL IETF Mailing List; Matt
> Squire
> Subject: Re: question on OSP
>
>
>
>
> The convergence between OSP v2 and H.225 v3 Annex G doesn't
> appear obvious,
> there seems some major differences between the two approaches. Below is a
> rough comparison I sketched.
>
> System Model
> ---------------------
> - Different administrative domain organisations are supported in the H.225
> Annex G, such as: Hierarchical, Distributed or Full mesh,
> ClearingHouse,...organisation. H.225 Annex G doesn't mandate a specific
> organisation among, or within, administrative
> domains. An architecture with Border Elements (BEs) for communication
> between administrative domains is defined. The protocol is fundamentally
> peer to peer.
>
> - OSP seems to be LIMITED to the centralised ClearingHouse administrative
> domain organisation. The architecture is fundamentally
> client-server. There
> appear NO WAY FOR direct relationships  between administrative domains
> (hiearchical, distributed or full
> meshed...).
>
> Message definition, format and encoding
> ------------------------------------------------------------
> Message fields and headers are quite different in both cases. The messages
> are encoded using ASN.1 for the H.225 Annex G, while OSP exchanges textual
> information using the XML for message structuring.
>
> The set of message types are also different in both cases. Today
> the current
> scope of OSP is inter-Domain pricing, authorization and usage
> exchange, and
> not inter-domain call routing.
>
>
> P.S. the  H.225 Annex G specifications can be found at the following URLs
>  ftp://standard.pictel.com/avc-site/9910_Red/   the name of the
> specification is TD-027.
>
>
> Philippe Terrien
> Alcatel SRD
> 10 rue Latecoere BP 57
> 78141 Velizy Cedex  France
>
>
>
>
>
> "Stephen Thomas" <stephen.thomas@transnexus.com> on 29/11/99 20:17:44
>
>
>
>  To:      Philippe TERRIEN/FR/ALCATEL@ALCATEL, "Jonathan
>           Rosenberg" <jdrosen@dynamicsoft.com>
>
>  cc:      "Gerry Reilly" <gerry@projects.telecity.co.uk>,
>           "IPTEL IETF Mailing List"
>           <iptel@lists.research.bell-labs.com>, "Matt Squire"
>           <msquire@nortelnetworks.com>
>
>
>
>  Subject: Re: new subscriber question on OSP
>
>
>
>
>
> Although there is some overlap between Annex G and OSP, there are
> some important philosophical differences. Annex G is, by it's
> very definition, an extension of the H.323 protocol suite. Like
> the rest of H.323, Annex G is fundamentally a peer-to-peer
> protocol. OSP, on the other hand, is independent of call signaling
> protocols. It is also architected as a client-server protocol.
> In some ways, OSP is quite close to SIP. It is text encoded
> (actually XML) and relies on standard Internet protocols (HTTP
> and TLS) for transport. A more detailed analysis would also show
> some functionality differences, but they are relatively modest.
>
> As for status, OSP was approved as an ETSI Technical Specification
> in December 1998. ETSI standards are available to the public at no
> charge, though ETSI does require you to register before downloading
> a final copy. (Thus I cannot provide a simple URL.) The ETSI
> web site is www.etsi.org.
>
> Stephen
>
>
>
> ----- Original Message -----
> From: <Philippe.Terrien@alcatel.fr>
> To: "Jonathan Rosenberg" <jdrosen@dynamicsoft.com>
> Cc: "Stephen Thomas" <stephen.thomas@transnexus.com>; "Gerry Reilly"
> <gerry@projects.telecity.co.uk>; "IPTEL IETF Mailing List"
> <iptel@lists.research.bell-labs.com>; "Matt Squire"
> <msquire@nortelnetworks.com>
> Sent: 25 November 1999 6:17 AM
> Subject: Re: new subscriber question on OSP
> |
> |
> | I have an additional question.
> |
> | How OSP is positionned, compared with the ITU-T H.225 Annex G standard
> which provides an architecture, protocol and operations to solve
> inter-domain routing ? My feeling is that they are competitors.
> | What is the rationale, scope, and current status of OSP ?
> |
> | Attached below the H.225 annex G for those who are not aware o this
> standard
> | (See attached file: TD-27.zip)
> |
> |
> ||
> | Jonathan Rosenberg <jdrosen@dynamicsoft.com> on 23/11/99 22:35:29
> |
> ||  To:      Stephen Thomas <stephen.thomas@transnexus.com>
> |
> |  cc:      Gerry Reilly <gerry@projects.telecity.co.uk>, IPTEL
> |           IETF Mailing List
> |           <iptel@lists.research.bell-labs.com>, Matt Squire
> |           <msquire@nortelnetworks.com>(bcc: Philippe
> |           TERRIEN/FR/ALCATEL)
> |
> |
> |
> |  Subject: Re: new subscriber question on OSP
> ||
> |
> | Thanks for the clarification, Stephen.
> |
> | Based on this, I see their roles as complementary. An OSP server needs
> | some way to build up the routing table it uses to answer OSP queries.
> | TRIP fits this need perfectly. In fact, OSP is effectively what is
> | called "the front end" in the TRIP framework document; its the way by
> | which TRIP databases are accessed. OSP is a query protocol, and TRIP is
> | more like a routing protocol.
> |
> | -Jonathan R.
> |
> | Stephen Thomas wrote:
> | >
> | > Matt has some of the essentials exactly right, but OSP
> | > can (optionally) provide call routing information as
> | > well as authorization. A typical application would
> | > work something like this:
> | >
> | >   Server X in domain 1 needs to locate a peer (e.g.
> | >     a terminating gateway to complete a call to the
> | >     PSTN).
> | >
> | >   Server X uses OSP (as a client) to query a third
> | >     party for the call routing and authorization
> | >     information.
> | >
> | >   Server X recieves the IP address of Server Y and
> | >     an authorization token that will grant it access
> | >     to the peer (sort of like Kerberos).
> | >
> | >   Server X sets up a the call to server Y (e.g. SIP
> | >     INVITE), including the authorization token
> | >     (e.g. as per draft-thomas-mime-osp-token-00.txt).
> | >
> | >   After the call is released, servers X and Y use OSP
> | >     to report usage information.
> | >
> | > To Matt's specific questions:
> | >
> | >   OSP Servers can route based on E.164 numbers as well as other
> | >     identifiers (e.g. email addresses).
> | >
> | >   Routing policies and route selection are up to the OSP
> | >     Server.
> | >
> | > At a high enough level, there is overlap between TRIP and
> | > OSP. OSP does things other than call routing as well, but
> | > where the overlap exists, I would characterize the
> | > difference as primarily one of philosophy. TRIP is
> | > a distributed protocol like IP routing protocols (BGP,
> | > and also RIP, OSPF, ...). OSP is a hierarchical protocol
> | > like directory service protocols (DNS, LDAP, ...)
> | >
> | > Stephen
> | >
> | > ----- Original Message -----
> | > From: "Matt Squire" <msquire@nortelnetworks.com>
> | > To: "Gerry Reilly" <gerry@projects.telecity.co.uk>;
> | > <iptel@lists.research.bell-labs.com>
> | > Sent: 23 November 1999 10:32 AM
> | > Subject: Re: new subscriber question on OSP
> | >
> | > |
> | > | My understanding of OSP was that it was basically an
> authorization and
> | > | accounting protocol - server X in domain 1 can request access to
> server Y
> | > | in domain 2, transmit usage records, etc.  I know its a bit more
> | > | complicated than this, but this is my one sentence summary.  OSP is,
> | > | generally, much more comparable to the work going on in AAA
> than here.
> I
> | > | didn't think it was designed to route signalling messages.
> | > |
> | > | So I'm not aware of the route selection capabilities of OSP, in
> particular
> | > | whether it makes routing decisions based upon E164 numbers,
> what kind
> of
> | > | routing policies and route selection it supports, etc.  To
> the extent
> that
> | > | OSP does these things, there may indeed be overlap.  Just don't know
> | > enough
> | > | about OSP to answer definitively.
> | > |
> | > |
> | > | At 03:11 PM 11/22/99 +0000, Gerry Reilly wrote:
> | > | >Hi
> | > | >
> | > | >Please forgive this newbee question.
> | > | >
> | > | >I've been reading through the TBGP, SCSP & TRIP drafts with great
> | > | >interest:-) As a relative newcomer to the VoIP area, I'm
> not entirely
> | > | >clear as to where the Open Settlements Protocol fits into
> the model,
> if
> | > at
> | > | >all.
> | > | >
> | > | >OSP seems to claim to be designed for inter-domain route
> selection as
> | > well
> | > | >as facilitating billing.
> | > | >Would this protocol supplement TRIP, or am I missing the point?
> | > | >
> | > | >Kind Regards
> | > | >
> | > | >Gerry Reilly
> | > | >
> | > | >
> | > | >
> | > | >---------
> | > | >This message came from the IETF IPTEL Working Group Mailing List.
> | > | >
> | > |
> | > | Matt Squire
> | > | Nortel Networks
> | > | innovations.lab
> | > | msquire@nortelnetworks.com
> | > | (919)992-9048
> | > |
> | > | ---------
> | > | This message came from the IETF IPTEL Working Group Mailing List.
> | > |
> | >
> | > ---------
> | > This message came from the IETF IPTEL Working Group Mailing List.
> |
> | --
> | Jonathan D. Rosenberg                       200 Executive Drive
> | Chief Scientist                             Suite 120
> | dynamicsoft                                 West Orange, NJ 07052
> | jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
> | http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
> | http://www.dynamicsoft.com
> |
> | ---------
> | 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.
>
>


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec  1 23:44:10 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 XAA02386
	for <iptel-archive@odin.ietf.org>; Wed, 1 Dec 1999 23:44:09 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C88A852BB; Wed,  1 Dec 1999 23:39:13 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2C01952C4; Wed,  1 Dec 1999 23:39:13 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <4142706ABBE2D211A1B700805F9FA43B4E9A9B@exchange.vipcalling.com>
From: Phil Mutooni <PMutooni@ibasis.net>
To: "'Henry Sinnreich'" <henry.sinnreich@wcom.com>,
        Phil Mutooni <PMutooni@ibasis.net>, Philippe.Terrien@alcatel.fr,
        "'Stephen Thomas'" <stephen.thomas@transnexus.com>
Cc: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>,
        "'Gerry Reilly'" <gerry@projects.telecity.co.uk>,
        "'IPTEL IETF Mailing List'" <iptel@lists.research.bell-labs.com>,
        "'Matt Squire'" <msquire@nortelnetworks.com>
Subject: RE: question on OSP
Date: Wed, 1 Dec 1999 17:44:38 -0500 
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

What you have pointed out is correct - there is a factual error in the
titling of that text box - this was an addition by the technical editors at 
the magazine.

-----Original Message-----
From: Henry Sinnreich [mailto:henry.sinnreich@wcom.com]
Sent: Wednesday, December 01, 1999 12:32 PM
To: Phil Mutooni; Philippe.Terrien@alcatel.fr; 'Stephen Thomas'
Cc: 'Jonathan Rosenberg'; 'Gerry Reilly'; 'IPTEL IETF Mailing List';
'Matt Squire'
Subject: RE: question on OSP


Page 50 shows the "ETSI Protocol Stack". There is no such thing. These (IP,
TCP, SSL/TLS, HTTP) ) are all IETF protocols and the IETF has the
intellectual property rights and change control. I believe the same applies
to W3C web protocols, XML, etc.

Henry

> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of Phil
> Mutooni
> Sent: Tuesday, November 30, 1999 4:42 PM
> To: 'Philippe.Terrien@alcatel.fr'; 'Stephen Thomas'
> Cc: 'Jonathan Rosenberg'; 'Gerry Reilly'; 'IPTEL IETF Mailing List';
> 'Matt Squire'
> Subject: RE: question on OSP
>
>
> Attached is a paper that I wrote early this year that compares
> the pros and
> cons of
> the OSP vs the ITU-T H.225 Annex G. It appeared in the June issue
> of Billing
> World.
>
> Philip Mutooni
> iBasis, Inc.
> w: http://www.ibasis.net
> t: 781.505.7523
> f: 781.505.7300
>
> -----Original Message-----
> From: owner-iptel@lists.research.bell-labs.com
> [mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of
> Philippe.Terrien@alcatel.fr
> Sent: Tuesday, November 30, 1999 6:12 AM
> To: Stephen Thomas
> Cc: Jonathan Rosenberg; Gerry Reilly; IPTEL IETF Mailing List; Matt
> Squire
> Subject: Re: question on OSP
>
>
>
>
> The convergence between OSP v2 and H.225 v3 Annex G doesn't
> appear obvious,
> there seems some major differences between the two approaches. Below is a
> rough comparison I sketched.
>
> System Model
> ---------------------
> - Different administrative domain organisations are supported in the H.225
> Annex G, such as: Hierarchical, Distributed or Full mesh,
> ClearingHouse,...organisation. H.225 Annex G doesn't mandate a specific
> organisation among, or within, administrative
> domains. An architecture with Border Elements (BEs) for communication
> between administrative domains is defined. The protocol is fundamentally
> peer to peer.
>
> - OSP seems to be LIMITED to the centralised ClearingHouse administrative
> domain organisation. The architecture is fundamentally
> client-server. There
> appear NO WAY FOR direct relationships  between administrative domains
> (hiearchical, distributed or full
> meshed...).
>
> Message definition, format and encoding
> ------------------------------------------------------------
> Message fields and headers are quite different in both cases. The messages
> are encoded using ASN.1 for the H.225 Annex G, while OSP exchanges textual
> information using the XML for message structuring.
>
> The set of message types are also different in both cases. Today
> the current
> scope of OSP is inter-Domain pricing, authorization and usage
> exchange, and
> not inter-domain call routing.
>
>
> P.S. the  H.225 Annex G specifications can be found at the following URLs
>  ftp://standard.pictel.com/avc-site/9910_Red/   the name of the
> specification is TD-027.
>
>
> Philippe Terrien
> Alcatel SRD
> 10 rue Latecoere BP 57
> 78141 Velizy Cedex  France
>
>
>
>
>
> "Stephen Thomas" <stephen.thomas@transnexus.com> on 29/11/99 20:17:44
>
>
>
>  To:      Philippe TERRIEN/FR/ALCATEL@ALCATEL, "Jonathan
>           Rosenberg" <jdrosen@dynamicsoft.com>
>
>  cc:      "Gerry Reilly" <gerry@projects.telecity.co.uk>,
>           "IPTEL IETF Mailing List"
>           <iptel@lists.research.bell-labs.com>, "Matt Squire"
>           <msquire@nortelnetworks.com>
>
>
>
>  Subject: Re: new subscriber question on OSP
>
>
>
>
>
> Although there is some overlap between Annex G and OSP, there are
> some important philosophical differences. Annex G is, by it's
> very definition, an extension of the H.323 protocol suite. Like
> the rest of H.323, Annex G is fundamentally a peer-to-peer
> protocol. OSP, on the other hand, is independent of call signaling
> protocols. It is also architected as a client-server protocol.
> In some ways, OSP is quite close to SIP. It is text encoded
> (actually XML) and relies on standard Internet protocols (HTTP
> and TLS) for transport. A more detailed analysis would also show
> some functionality differences, but they are relatively modest.
>
> As for status, OSP was approved as an ETSI Technical Specification
> in December 1998. ETSI standards are available to the public at no
> charge, though ETSI does require you to register before downloading
> a final copy. (Thus I cannot provide a simple URL.) The ETSI
> web site is www.etsi.org.
>
> Stephen
>
>
>
> ----- Original Message -----
> From: <Philippe.Terrien@alcatel.fr>
> To: "Jonathan Rosenberg" <jdrosen@dynamicsoft.com>
> Cc: "Stephen Thomas" <stephen.thomas@transnexus.com>; "Gerry Reilly"
> <gerry@projects.telecity.co.uk>; "IPTEL IETF Mailing List"
> <iptel@lists.research.bell-labs.com>; "Matt Squire"
> <msquire@nortelnetworks.com>
> Sent: 25 November 1999 6:17 AM
> Subject: Re: new subscriber question on OSP
> |
> |
> | I have an additional question.
> |
> | How OSP is positionned, compared with the ITU-T H.225 Annex G standard
> which provides an architecture, protocol and operations to solve
> inter-domain routing ? My feeling is that they are competitors.
> | What is the rationale, scope, and current status of OSP ?
> |
> | Attached below the H.225 annex G for those who are not aware o this
> standard
> | (See attached file: TD-27.zip)
> |
> |
> ||
> | Jonathan Rosenberg <jdrosen@dynamicsoft.com> on 23/11/99 22:35:29
> |
> ||  To:      Stephen Thomas <stephen.thomas@transnexus.com>
> |
> |  cc:      Gerry Reilly <gerry@projects.telecity.co.uk>, IPTEL
> |           IETF Mailing List
> |           <iptel@lists.research.bell-labs.com>, Matt Squire
> |           <msquire@nortelnetworks.com>(bcc: Philippe
> |           TERRIEN/FR/ALCATEL)
> |
> |
> |
> |  Subject: Re: new subscriber question on OSP
> ||
> |
> | Thanks for the clarification, Stephen.
> |
> | Based on this, I see their roles as complementary. An OSP server needs
> | some way to build up the routing table it uses to answer OSP queries.
> | TRIP fits this need perfectly. In fact, OSP is effectively what is
> | called "the front end" in the TRIP framework document; its the way by
> | which TRIP databases are accessed. OSP is a query protocol, and TRIP is
> | more like a routing protocol.
> |
> | -Jonathan R.
> |
> | Stephen Thomas wrote:
> | >
> | > Matt has some of the essentials exactly right, but OSP
> | > can (optionally) provide call routing information as
> | > well as authorization. A typical application would
> | > work something like this:
> | >
> | >   Server X in domain 1 needs to locate a peer (e.g.
> | >     a terminating gateway to complete a call to the
> | >     PSTN).
> | >
> | >   Server X uses OSP (as a client) to query a third
> | >     party for the call routing and authorization
> | >     information.
> | >
> | >   Server X recieves the IP address of Server Y and
> | >     an authorization token that will grant it access
> | >     to the peer (sort of like Kerberos).
> | >
> | >   Server X sets up a the call to server Y (e.g. SIP
> | >     INVITE), including the authorization token
> | >     (e.g. as per draft-thomas-mime-osp-token-00.txt).
> | >
> | >   After the call is released, servers X and Y use OSP
> | >     to report usage information.
> | >
> | > To Matt's specific questions:
> | >
> | >   OSP Servers can route based on E.164 numbers as well as other
> | >     identifiers (e.g. email addresses).
> | >
> | >   Routing policies and route selection are up to the OSP
> | >     Server.
> | >
> | > At a high enough level, there is overlap between TRIP and
> | > OSP. OSP does things other than call routing as well, but
> | > where the overlap exists, I would characterize the
> | > difference as primarily one of philosophy. TRIP is
> | > a distributed protocol like IP routing protocols (BGP,
> | > and also RIP, OSPF, ...). OSP is a hierarchical protocol
> | > like directory service protocols (DNS, LDAP, ...)
> | >
> | > Stephen
> | >
> | > ----- Original Message -----
> | > From: "Matt Squire" <msquire@nortelnetworks.com>
> | > To: "Gerry Reilly" <gerry@projects.telecity.co.uk>;
> | > <iptel@lists.research.bell-labs.com>
> | > Sent: 23 November 1999 10:32 AM
> | > Subject: Re: new subscriber question on OSP
> | >
> | > |
> | > | My understanding of OSP was that it was basically an
> authorization and
> | > | accounting protocol - server X in domain 1 can request access to
> server Y
> | > | in domain 2, transmit usage records, etc.  I know its a bit more
> | > | complicated than this, but this is my one sentence summary.  OSP is,
> | > | generally, much more comparable to the work going on in AAA
> than here.
> I
> | > | didn't think it was designed to route signalling messages.
> | > |
> | > | So I'm not aware of the route selection capabilities of OSP, in
> particular
> | > | whether it makes routing decisions based upon E164 numbers,
> what kind
> of
> | > | routing policies and route selection it supports, etc.  To
> the extent
> that
> | > | OSP does these things, there may indeed be overlap.  Just don't know
> | > enough
> | > | about OSP to answer definitively.
> | > |
> | > |
> | > | At 03:11 PM 11/22/99 +0000, Gerry Reilly wrote:
> | > | >Hi
> | > | >
> | > | >Please forgive this newbee question.
> | > | >
> | > | >I've been reading through the TBGP, SCSP & TRIP drafts with great
> | > | >interest:-) As a relative newcomer to the VoIP area, I'm
> not entirely
> | > | >clear as to where the Open Settlements Protocol fits into
> the model,
> if
> | > at
> | > | >all.
> | > | >
> | > | >OSP seems to claim to be designed for inter-domain route
> selection as
> | > well
> | > | >as facilitating billing.
> | > | >Would this protocol supplement TRIP, or am I missing the point?
> | > | >
> | > | >Kind Regards
> | > | >
> | > | >Gerry Reilly
> | > | >
> | > | >
> | > | >
> | > | >---------
> | > | >This message came from the IETF IPTEL Working Group Mailing List.
> | > | >
> | > |
> | > | Matt Squire
> | > | Nortel Networks
> | > | innovations.lab
> | > | msquire@nortelnetworks.com
> | > | (919)992-9048
> | > |
> | > | ---------
> | > | This message came from the IETF IPTEL Working Group Mailing List.
> | > |
> | >
> | > ---------
> | > This message came from the IETF IPTEL Working Group Mailing List.
> |
> | --
> | Jonathan D. Rosenberg                       200 Executive Drive
> | Chief Scientist                             Suite 120
> | dynamicsoft                                 West Orange, NJ 07052
> | jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
> | http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
> | http://www.dynamicsoft.com
> |
> | ---------
> | 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.
>
>

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Dec  3 01:21:41 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 BAA05111
	for <iptel-archive@odin.ietf.org>; Fri, 3 Dec 1999 01:21:41 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2867952E5; Fri,  3 Dec 1999 01:15:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id AB72452E4; Fri,  3 Dec 1999 01:15:25 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C3BE91B@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: TRIP
Date: Thu, 2 Dec 1999 20:25:01 -0600 
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

The telephone number ranges used for routing decisions need to consider the
impacts of number portability when routing calls to the SCNs where number
portability is supported.  For example, the subscriber number (e.g., dialed
number) and the network routing number (e.g., the Location Routing Number or
LRN) are in two different domains in the North America.  In Europe, the
routing prefix plays a similar role as the LRN (one difference is that the
LRN identifies a specific switch while the routing prefix may identify a
switch or just a network).

If TRIP is used to pass routing information, it should be decided whether
the "telephone number" is the subscriber number or the LRN/routing prefix.
My opinion is that it should be the LRN/routing prefix.  This will keep the
routing information static rather than dynamic (e.g., changes every time a
number is ported).  Any other thoughts?

For those of you who do not know NeuStar, it is the North American Numbering
Plan Administrator (NANPA), and the Number Portability Administration Center
(NPAC) for both US and Canada.


James Yu
Principal Technical Industry Liaison
NeuStar Inc.
1120 Vermont Avenue, NW,
Suite 550
Washington, D.C. 20005
USA
+1-202-533-2814 (B)
+1-202-533-2975 (Fax)
+1-202-256-1200 (Mobile)
james.yu@neustar.com


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Fri Dec  3 01:29:46 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 BAA09414
	for <iptel-archive@odin.ietf.org>; Fri, 3 Dec 1999 01:29:46 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7ED6052E2; Fri,  3 Dec 1999 01:15:38 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id CB3C852DD; Fri,  3 Dec 1999 01:15:32 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C3BE91C@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: TRIP
Date: Thu, 2 Dec 1999 22:30:57 -0600 
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

In the previous message, I mentioned LRN and routing prefix.  Since routing
prefixes (used in Europe) do not overlap with the subscriber telephone
numbers, it is not an issue.  The routing prefixes can be part of the
"telephone number" ranges as well as the "subscriber telephone numbers" to
be used as routing information in TRIP.

The LRN is actually a NANP (North American) number which is ten-digit long.
So the "LRN" in our discussion should include the country code 1 for
international telephone numbers.

James Yu
Principal Technical Industry Liaison
NeuStar Inc.
1120 Vermont Avenue, NW,
Suite 550
Washington, D.C. 20005
USA
+1-202-533-2814 (B)
+1-202-533-2975 (Fax)
+1-202-256-1200 (Mobile)
james.yu@neustar.com


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Dec  4 00:26:53 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 AAA21523
	for <iptel-archive@odin.ietf.org>; Sat, 4 Dec 1999 00:26:53 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C912252D5; Sat,  4 Dec 1999 00:18:32 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F02B652D6; Sat,  4 Dec 1999 00:18:30 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <3848A4B7.3B9BDC65@dynamicsoft.com>
Date: Sat, 04 Dec 1999 00:20:55 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: Dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: James Yu <james.yu@neustar.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Re: TRIP
References: <ED88182BFF78D211A4D800A0C9E9435C3BE91B@dc02.npac.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



James Yu wrote:
> 
> The telephone number ranges used for routing decisions need to consider the
> impacts of number portability when routing calls to the SCNs where number
> portability is supported.  For example, the subscriber number (e.g., dialed
> number) and the network routing number (e.g., the Location Routing Number or
> LRN) are in two different domains in the North America.  In Europe, the
> routing prefix plays a similar role as the LRN (one difference is that the
> LRN identifies a specific switch while the routing prefix may identify a
> switch or just a network).
> 
> If TRIP is used to pass routing information, it should be decided whether
> the "telephone number" is the subscriber number or the LRN/routing prefix.
> My opinion is that it should be the LRN/routing prefix.  This will keep the
> routing information static rather than dynamic (e.g., changes every time a
> number is ported).  Any other thoughts?

I have a question. In your other message, you mention that the LRNs
identify switches, and that they do overlap with the subscriber numbers.
Thats actually OK, so long as the overlap is such that the LRN is for
the switch serving those subscribers whose numbers the switch overlaps.
In other words, if a switch has an LRN that is +1 212 455, that OK so
long as this switch serves the subscribers with numbers +1 212 455-0000
to +1 212 455-9999. The reason that this is OK is that a longest prefix
match on the LRN or a subscriber would still get you the right gateway.
So, my question is: is this the case? Or, is it the case that the LRN
has no relationship with the subscriber numbers it overlaps. 

-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Dec  5 09:20:16 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 JAA20796
	for <iptel-archive@odin.ietf.org>; Sun, 5 Dec 1999 09:20:16 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 852BC52AB; Sun,  5 Dec 1999 09:15:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id EAB6252D6; Sun,  5 Dec 1999 09:15:27 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Sun, 5 Dec 1999 08:49:53 -0500 (EST)
From: Scott Petrack <scott.petrack@metatel.com>
X-Sender: scott.petrack@adsl-151-203-17-31.bellatlantic.net
To: Dean Willis <dean.willis@wcom.com>
Cc: iptel@lists.research.bell-labs.com, enum@ietf.org
Subject: Re: [ENUM] Dynamics -- was RE: Human readable Route descriptors
In-Reply-To: <000501bf2c78$a2a0ad40$9f168082@fh.ietf.innovationslab.net>
Message-ID: <Pine.LNX.4.10.9912050835250.4803-100000@adsl-151-203-17-31.bellatlantic.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Catching up on almost a month of email over the next 2 days...

On Thu, 11 Nov 1999, Dean Willis wrote:

> Part of my initial adverse reaction to ENUM is my belief that the
> association of IP addresses to phone numbers is an exceptionally dynamic
> relationship, and that this relationship could not be adequately expressed
> in a relatively static and cached DNS.
> 

Yes, but the relationship of phone numbers to URLs is likely to be less
dynamic. So one high-level service that you might need is 
"phone number --> sip: URL". Another might be "phone number --> tel: URL".
It might even be that the sip signalling that ensues as a result of
the first mapping might lead to a SIP redirect to the tel: URL. 

...

> 
> For example -- (all names and numbers fictitious) I as a phone user might
> rent a phone number from a carrier, perhaps 13045551212 from BPI (Big Phone
> Inc). The NANP NS delegates to the 304 area NS, which delegates exchange 555
> to the SBC NS. The SBC NS might contain a record pointing .1212 towards
> sip:3045551212@sip.bpi.net;user=phone. As part of my rental fee for the
> phone number, SBC allows me to either use sip.sbc.net as a registrar, or to
> provision contacts into it using a web interface. In the simplest mode,
> assume I use the registration services. When I turn my phone on, it sends a
> SIP REGISTER to sip.bpi.net on behalf of 3045551212 listing the phone's
> current IP address as a Contact:.
> 
> Calls from the GSTN to my phone resolve in the following way: The call
> enters the GSTN as a sequence of dialed digits. The carrier switching system
> follows the standard PSTN routing to the nearest BPI switch registered as
> handling my exchange (30555). BPI hands the call to an IP gateway.  The
> gateway does a DNS lookup on 3045551212, which resolves to
> 3045551212@sip.bpi.net. The GW then sends an INVITE to sip.bpi.net and is
> proxied or redirected to my phone. Note that nobody ever looked in the TRIP
> routes to resolve the call.
> 
Even in the scenario you give, there is a step: "BPI hands the call to an
IP gateway". In a standards-based world this is what TRIP is used for.
Even if the BPI switch knows that 3045551212 is a BPI number, it might
decide to use TRIP to find a gateway.

> In the event that I place a call to the PSTN, 19153627878, things happen in
> a very different fashion.
> 
> In this sequence, my phone does a DNS lookup on the dialed number. Perhaps
> it does not resolve (It's not an IP or LNP registered phone number). My
> phone then sends an INVITE to sip.bpi.net with a request-URI of
> tel:9153627878, perhaps with some sort of scope specifier. Next sip.bpi.net
> consults its TRIP routing table to find one or more egress gateways and
> either proxies or redirects my call.
> 
> Alternative:
> My phone could be configured to always send my calls to my outgoung proxy
> sip.bpi.net. In this case, it sends the INVITE tel:9153627878. The proxy
> does a DNS lookup, finds that the name does not resolve, consults TRIP, and
> proxies or forwards the call.
> 
> These scenarios work pretty well. But, I haven't been able to figure out a
> scenario for local number portability from within DNS -- theoretcially, one
> of the ENUM possibilities. Anybody want to tackle that? I'm concerned that
> LNP is more like NAT than it is a 302: redirect . . .
> 

There are many different kinds of number portability, and many different
ways to implement each kind. I think that it might be useful to produce an
informative draft with a few examples.

Scott


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Sun Dec  5 15:29:07 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 PAA10611
	for <iptel-archive@odin.ietf.org>; Sun, 5 Dec 1999 15:29:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 712CC52B6; Sun,  5 Dec 1999 15:17:54 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B86CC52DC; Sun,  5 Dec 1999 15:17:53 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Sun, 5 Dec 1999 15:16:35 -0500 (EST)
From: Henning Schulzrinne <hgs@cs.columbia.edu>
Message-Id: <199912052016.PAA18381@bart.cs.columbia.edu>
To: iptel@lists.research.bell-labs.com
Subject: CFP: JCN Special Issue on QoS in IP Networks
List: iptel@lists.research.bell-labs.com
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

         JOURNAL OF COMMUNICATIONS AND NETWORKING (JCN)

 CALL FOR PAPERS - SPECIAL ISSUE ON QoS IN IP NETWORKS
                        JUNE, 2000

A Special Issue of JCN dedicated to the realization of QoS-sensitive
services in IP networks will be published in June, 2000.  It will be
Guest Edited by Prof. Henning Schulzrinne of Columbia University,
Prof. Hideo Miyahara of Osaka University, and Prof. Luigi Fratta of
the University of Milan.  The topics include but are not limited to:

Differentiated Services
Integrated Services
MPLS
Real-time multicast
IP network traffic engineering
Performance measurement and evaluation
Pricing and billing
User perception of QOS
QOS policy management
Adaptive and other new service models
Light-weight reservation protocols and aggregation

Continuing JCN's tradition of fast turnaround together with full peer
reviews, the following schedule has been set:

Dec. 31, 1999   Submit manuscript via web page (see below)

Mar. 31, 2000   First reviews returned to author, revisions returned
                within three weeks

June, 2000      Special Issue published

The guest editors are
Prof. Luigi Fratta  <fratta@elet.polimi.it>
Prof. Hideo Miyahara <miyahara@mercury.nal.ics.es.osaka-u.ac.jp>
Prof. Henning Schulzrinne <hgs@cs.columbia.edu>

Papers should be submitted via the procedure described at

http://www.cs.columbia.edu/~hgs/edas/JCN

Only PostScript and PDF formats are accepted.

Further information about JCN is available at http://JCN.snu.ac.kr. 
JCN is a high-quality quarterly archival journal, published by the
Korean Institute of Communications Sciences with the technical
cosponsorship of the IEEE Communications Society, covering the fields of
Communication Theory and Systems, Wireless Communications, and Networks
and Services.  JCN began publication in March, 1999.


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Dec  6 07:02:31 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 HAA14064
	for <iptel-archive@odin.ietf.org>; Mon, 6 Dec 1999 07:02:31 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 68C5152DB; Mon,  6 Dec 1999 06:57:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2742352DC; Mon,  6 Dec 1999 06:57:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <00b201bf3fe3$2bc4ff40$0a0a000a@Ninety.ninety.de>
From: "Christian Kruetzfeldt" <ckruetze@foni.net>
To: <iptel@lists.research.bell-labs.com>
References: <ED88182BFF78D211A4D800A0C9E9435C3BE91B@dc02.npac.com>
Subject: Re: TRIP
Date: Mon, 6 Dec 1999 13:12:28 +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 5.00.2919.5600
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.5600
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Hello,

it is not quit true what you where saying about Europe. Most european
countries use routing prefixes, but not all!
Danmark uses no routing prefixes. All phone numbers in danmark are 8 digits
long, and they don't say anything about the location.

So how should routing be done in such enviroments ?
As far as I can see it, it will not work with the LRN/routing prefix, since
there is no routing prefix, or did I miss something ?

Christian Kruetzfeldt

James Yu wrote
> The telephone number ranges used for routing decisions need to consider
the
> impacts of number portability when routing calls to the SCNs where number
> portability is supported.  For example, the subscriber number (e.g.,
dialed
> number) and the network routing number (e.g., the Location Routing Number
or
> LRN) are in two different domains in the North America.  In Europe, the
> routing prefix plays a similar role as the LRN (one difference is that the
> LRN identifies a specific switch while the routing prefix may identify a
> switch or just a network).
>
> If TRIP is used to pass routing information, it should be decided whether
> the "telephone number" is the subscriber number or the LRN/routing prefix.
> My opinion is that it should be the LRN/routing prefix.  This will keep
the
> routing information static rather than dynamic (e.g., changes every time a
> number is ported).  Any other thoughts?
>



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


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Dec  6 10:46:15 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 KAA12311
	for <iptel-archive@odin.ietf.org>; Mon, 6 Dec 1999 10:46:14 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 60DC852AB; Mon,  6 Dec 1999 10:40:27 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BB55A52B6; Mon,  6 Dec 1999 10:40:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C39507A@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Sun, 5 Dec 1999 23:02:48 -0600 
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

Jonathan,

I'll explain using an example.  Switch A of local exchange carrier (LEC,
e.g., the local telehpone company instead of the long distance or
international telephone companies) X serves 212-455-0000 to 212-455-9999
(I'll ignore the country code for the time being because number portability
is supported within the U.S. or Canada; therefore, the LRN discussed in all
the number portability documents is a North American Numbering Plan Number)
before number portability is supported.  Switch B of local exchange carrier
Y serves 212-466-0000 to 212-466-9999.  Now due to the need to support
number portability, Switch A is given a LRN of 212-455-0000 (only the first
6 digits of the LRN are used for routing within the SCN in the U.S. or
Canada).  Switch B is given a LRN of 212-466-0000.  We'll call these two
LRNs LRNA and LRNB respectively.  Assume that one customer whose number is
212-455-1234 switches from LEC X to LEC Y.  There is a process among LEC X,
LEC Y and NeuStar (as the Number Portability Administration Center or NPAC)
to confirm the transfer and uppdate all the Number Poratbility Databases
(NPDBs) about the change.  Basically, the LRN associated with 212-455-1234
in the NPDBs is changed from LRNA to LRNB.  Please note that this customer's
home/business phone would be pysically connected with Switch B instead of
Switch A after the phone number is "ported."

North American Numbering Plan numbers have the format of NPA+NXX+XXXX where
NPA (Numbering Plan Area) is the area code and has the format of NXX (The
area codes used to have the format of N0/1X where the second digit can be
either 0 or 1. That restriction has been removed to allow more area codes to
be assigned.). N is any digit from 2 to 9 and X is any digit from 0 to 9.
The NXX after NPA is called "office code."  Normally, a switch is assigned
the whole block of NPA+NXX (e.g., 212-455-0000 to 212-455-9999).  Now
"number pooling" will allow several LECs to share one NPA+NXX.  That is, a
LEC will be assigned, for example, 212-477-1 (block from 212-477-1000 to
212-477-1999) and another LEC assigned 212-477-2 (block from 212-477-2000 to
212-477-2999).  We can talk about "number pooling" later if you or others
have questions.  The focus of this discussion is on LRN and number
portability.

When any number in a NPA+NXX block is ported out to another switch, this
particular NPA+NXX is marked as a portable NPA+NXX.  Calls to numbers in the
portable NPA+NXX blocks will trigger NPDB queries.  This is to avoid NPDB
queries on all calls.

Assume someone in a different state (e.g., California) in the U.S. dials
1-212-455-1234.  The originating switch in California will deliver the call
to the caller's presubscribed interexchange carrier (IEC, e.g., long
distance telephone company).  The IEC performs a NPDB query with the called
party number because 212-455 is a portable NPA+NXX.  The response will
contain the LRN assocaited with 212-455-1234 which is LRNB.  Certainly, for
numbers that have not been ported out of 212-455, the same "called party
number" is returned in the response because the NPDB presently does not
store information about the "non-ported" numbers.  After receiving the
response, the IEC switch sets an indicator in the ISUP Initial Address
Message (IAM) to indicate that NPDB has been dipped.  This will prevent
subsequent NPDB dips from happening at any subsequent switches.  For a
ported number, the dialed number, 212-455-1234, will be carried in a ISUP
Generic Address Parameter (GAP) since the called party number parameter now
carries the LRNB.  All the subsequent switches route the call by analyzing
the first six digits of the called party number.  The call will be routed to
Switch B where it retrieves the dialed number, 212-455-1234, from the GAP in
the received IAM.  

For a ported number, the called party number in the IAM contains the LRN
associated with the serving switch of the ported number.  For a non-ported
number, the called party number is the susbcriber number.  In either case,
the first six digits of the subscriber number or LRN are the same.  

If the TRIP information contains the reachability of subscriber number
ranges, A GW that can reach Switch A but not Switch B will have to indicate
that it can reach +1-212-455-0000 to +1-212-455-1233 and
+1-212-455-1235-+1-212-455-9999 (if only one number has been ported out of
Switch A).  The problem of this is that many ported-out numbers need to be
singled out from the reachable phone number ranges ranges.  This is a minor
problem when comparing to the bigger problem of updating the TRIP
information every time a number changes its assoicated LRN after been ported
to a new LEC.

For calls routed to the North America, it would be much simpler for a GW to
indicate the switches (e.g., LRNs (plus the country code 1)) it can reach
becuase the information is static.

James

-----Original Message-----
From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]
Sent: Saturday, December 04, 1999 12:21 AM
To: James Yu
Cc: 'iptel@lists.research.bell-labs.com'
Subject: Re: TRIP




James Yu wrote:
> 
> The telephone number ranges used for routing decisions need to consider
the
> impacts of number portability when routing calls to the SCNs where number
> portability is supported.  For example, the subscriber number (e.g.,
dialed
> number) and the network routing number (e.g., the Location Routing Number
or
> LRN) are in two different domains in the North America.  In Europe, the
> routing prefix plays a similar role as the LRN (one difference is that the
> LRN identifies a specific switch while the routing prefix may identify a
> switch or just a network).
> 
> If TRIP is used to pass routing information, it should be decided whether
> the "telephone number" is the subscriber number or the LRN/routing prefix.
> My opinion is that it should be the LRN/routing prefix.  This will keep
the
> routing information static rather than dynamic (e.g., changes every time a
> number is ported).  Any other thoughts?

I have a question. In your other message, you mention that the LRNs
identify switches, and that they do overlap with the subscriber numbers.
Thats actually OK, so long as the overlap is such that the LRN is for
the switch serving those subscribers whose numbers the switch overlaps.
In other words, if a switch has an LRN that is +1 212 455, that OK so
long as this switch serves the subscribers with numbers +1 212 455-0000
to +1 212 455-9999. The reason that this is OK is that a longest prefix
match on the LRN or a subscriber would still get you the right gateway.
So, my question is: is this the case? Or, is it the case that the LRN
has no relationship with the subscriber numbers it overlaps.
 

-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From szhang@cogentel.com  Mon Dec  6 16:32:14 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 QAA14162
	for <iptel-archive@odin.ietf.org>; Mon, 6 Dec 1999 16:32:14 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id F29E352D6; Mon,  6 Dec 1999 16:27:11 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: from scummy.research.bell-labs.com (research.bell-labs.com [135.104.2.10])
	by lists.research.bell-labs.com (Postfix) with SMTP id AFEA952D5
	for <iptel-outgoing-local@paperless.dnrc.bell-labs.com>; Mon,  6 Dec 1999 16:27:06 -0500 (EST)
Received: from dusty.research.bell-labs.com ([135.104.2.7]) by scummy; Mon Dec  6 16:25:19 EST 1999
Received: from gull.prod.itd.earthlink.net ([207.217.121.85]) by dusty; Mon Dec  6 16:23:51 EST 1999
Received: from cogentel.com (dialup-209.244.101.47.NewYork2.Level3.net [209.244.101.47])
	by gull.prod.itd.earthlink.net (8.9.3/8.9.3) with ESMTP id NAA29468;
	Mon, 6 Dec 1999 13:25:13 -0800 (PST)
Message-ID: <384C29BE.F2D682F6@cogentel.com>
Date: Mon, 06 Dec 1999 16:25:18 -0500
From: Siuling Zhang <szhang@cogentel.com>
X-Mailer: Mozilla 4.5 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: iptel-outgoing-local@paperless.dnrc.bell-labs.com,
        iptel-local@paperless.dnrc.bell-labs.com,
        iptel@lists.research.bell-labs.co
Subject: unsubscribe
References: <1B0DEBA1C1BAD211B4B70008C7B98A770115AB@server.bgs.sk>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

unsubscribe


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Dec  6 19:34:03 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 TAA05198
	for <iptel-archive@odin.ietf.org>; Mon, 6 Dec 1999 19:34:03 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 93E9052DB; Mon,  6 Dec 1999 19:29:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 09F7552DC; Mon,  6 Dec 1999 19:29:27 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <384C5593.87AA223E@dynamicsoft.com>
Date: Mon, 06 Dec 1999 19:32:19 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: Dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: James Yu <james.yu@neustar.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Re: TRIP
References: <ED88182BFF78D211A4D800A0C9E9435C39507A@dc02.npac.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

James,

Thanks for the excellent description of the process. Comments below:


James Yu wrote:
> 
> Jonathan,
> 
> North American Numbering Plan numbers have the format of NPA+NXX+XXXX where
> NPA (Numbering Plan Area) is the area code and has the format of NXX (The
> area codes used to have the format of N0/1X where the second digit can be
> either 0 or 1. That restriction has been removed to allow more area codes to
> be assigned.). N is any digit from 2 to 9 and X is any digit from 0 to 9.
> The NXX after NPA is called "office code."  Normally, a switch is assigned
> the whole block of NPA+NXX (e.g., 212-455-0000 to 212-455-9999).  Now
> "number pooling" will allow several LECs to share one NPA+NXX.  That is, a
> LEC will be assigned, for example, 212-477-1 (block from 212-477-1000 to
> 212-477-1999) and another LEC assigned 212-477-2 (block from 212-477-2000 to
> 212-477-2999).  We can talk about "number pooling" later if you or others
> have questions.  The focus of this discussion is on LRN and number
> portability.

Hmm, sounds familiar.... CIDR?

> If the TRIP information contains the reachability of subscriber number
> ranges, A GW that can reach Switch A but not Switch B will have to indicate
> that it can reach +1-212-455-0000 to +1-212-455-1233 and
> +1-212-455-1235-+1-212-455-9999 (if only one number has been ported out of
> Switch A).  The problem of this is that many ported-out numbers need to be
> singled out from the reachable phone number ranges ranges.  This is a minor
> problem when comparing to the bigger problem of updating the TRIP
> information every time a number changes its assoicated LRN after been ported
> to a new LEC.

I think TRIP should carry subscriber numbers, but that the LNP dip
should happen before the TRIP database is accessed. So, returning to
your example, a gateway reaching only switch A would advertise
+1-212-455* and a gateway reaching only switch B +1-212-466*. A PC users
calls +1-212-455-1234. The INVITE is sent to its local proxy, which
first does an LNP dip. It translates this to an LRN of +1-212-456, which
is looked up in TRIP. The longest prefix match (which is an exact match
for the LRN) is found. The INVITE is then sent to the next hop towards
that LRN as learned through TRIP. The original dialed number will be
carried in the To field. 

Note that we could also allow switch B to advertise +1-212-455-1234 in
TRIP. If the LNP dip is not done, the proxy will pick this because of
longest prefix matching, and the call still completes. However, I would
recommend doing this, since it will result in a completely flat TRIP
routing table.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Dec  6 20:01:49 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 UAA17426
	for <iptel-archive@odin.ietf.org>; Mon, 6 Dec 1999 20:01:49 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C115552DC; Mon,  6 Dec 1999 19:57:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 51CA352B6; Mon,  6 Dec 1999 19:57:27 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <384C5B05.8268C93D@cisco.com>
Date: Mon, 06 Dec 1999 16:55:33 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Cc: James Yu <james.yu@neustar.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Re: TRIP
References: <ED88182BFF78D211A4D800A0C9E9435C39507A@dc02.npac.com> <384C5593.87AA223E@dynamicsoft.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:
> 
> James,
> 
> Thanks for the excellent description of the process. Comments below:
> 
> James Yu wrote:
> >
> > Jonathan,
> >
> > North American Numbering Plan numbers have the format of NPA+NXX+XXXX where
> > NPA (Numbering Plan Area) is the area code and has the format of NXX (The
> > area codes used to have the format of N0/1X where the second digit can be
> > either 0 or 1. That restriction has been removed to allow more area codes to
> > be assigned.). N is any digit from 2 to 9 and X is any digit from 0 to 9.
> > The NXX after NPA is called "office code."  Normally, a switch is assigned
> > the whole block of NPA+NXX (e.g., 212-455-0000 to 212-455-9999).  Now
> > "number pooling" will allow several LECs to share one NPA+NXX.  That is, a
> > LEC will be assigned, for example, 212-477-1 (block from 212-477-1000 to
> > 212-477-1999) and another LEC assigned 212-477-2 (block from 212-477-2000 to
> > 212-477-2999).  We can talk about "number pooling" later if you or others
> > have questions.  The focus of this discussion is on LRN and number
> > portability.
> 
> Hmm, sounds familiar.... CIDR?
> 
> > If the TRIP information contains the reachability of subscriber number
> > ranges, A GW that can reach Switch A but not Switch B will have to indicate
> > that it can reach +1-212-455-0000 to +1-212-455-1233 and
> > +1-212-455-1235-+1-212-455-9999 (if only one number has been ported out of
> > Switch A).  The problem of this is that many ported-out numbers need to be
> > singled out from the reachable phone number ranges ranges.  This is a minor
> > problem when comparing to the bigger problem of updating the TRIP
> > information every time a number changes its assoicated LRN after been ported
> > to a new LEC.
> 
> I think TRIP should carry subscriber numbers, but that the LNP dip
> should happen before the TRIP database is accessed. So, returning to
> your example, a gateway reaching only switch A would advertise
> +1-212-455* and a gateway reaching only switch B +1-212-466*. A PC users
> calls +1-212-455-1234. The INVITE is sent to its local proxy, which
> first does an LNP dip. 

The N-1 agreement in classic telephony makes the (N-1)th network (i.e.
the next to last network), on the path towards the destination,
responsible for making the LNP database query. I wonder what's the
reason for this? Wouldn't executing the LNP database query at the
originating network have resulted in more efficient routing?    


> It translates this to an LRN of +1-212-456, which
> is looked up in TRIP. The longest prefix match (which is an exact match
> for the LRN) is found. The INVITE is then sent to the next hop towards
> that LRN as learned through TRIP. The original dialed number will be
> carried in the To field.
> 

Looks like a SIP extension is required to indicate to other signaling
points on the call path that an LNP query has already been done for this
destination, in order to avoid multiple LNP queries for the same call.
Is such a flag already there?

Hussein


> Note that we could also allow switch B to advertise +1-212-455-1234 in
> TRIP. If the LNP dip is not done, the proxy will pick this because of
> longest prefix matching, and the call still completes. However, I would
> recommend doing this, since it will result in a completely flat TRIP
> routing table.
> 
> -Jonathan R.
> 
> --
> Jonathan D. Rosenberg                       200 Executive Drive
> Chief Scientist                             Suite 120
> dynamicsoft                                 West Orange, NJ 07052
> jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
> http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
> http://www.dynamicsoft.com
> 
> ---------
> 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 Dec  7 14:41:35 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 OAA13676
	for <iptel-archive@odin.ietf.org>; Tue, 7 Dec 1999 14:41:34 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id E4DBE52C4; Tue,  7 Dec 1999 14:35:31 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3D8B152D5; Tue,  7 Dec 1999 14:35:30 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <89866A7EC9C0D1119C950000F8BCC2820157C00D@zrtpd00n.us.nortel.com>
From: "John Mainwaring" <crm312a@nortelnetworks.com>
To: "'hsalama@cisco.com'" <hsalama@cisco.com>,
        Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Cc: James Yu <james.yu@neustar.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Tue, 7 Dec 1999 14:32:33 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: multipart/alternative;
              boundary="----_=_NextPart_001_01BF40E9.D77E53A2"
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_01BF40E9.D77E53A2
Content-Type: text/plain;
	charset="windows-1252"



|-----Original Message-----
|From: Hussein F. Salama [mailto:hsalama@cisco.com]
|Sent: Monday, December 06, 1999 7:56 PM
|To: Jonathan Rosenberg
|Cc: James Yu; 'iptel@lists.research.bell-labs.com'
|Subject: Re: TRIP
|
|
|
|
|Jonathan Rosenberg wrote:
|> 
|> I think TRIP should carry subscriber numbers, but that the LNP dip
|> should happen before the TRIP database is accessed. So, returning to
|> your example, a gateway reaching only switch A would advertise
|> +1-212-455* and a gateway reaching only switch B 
|+1-212-466*. A PC users
|> calls +1-212-455-1234. The INVITE is sent to its local proxy, which
|> first does an LNP dip. 
|
|The N-1 agreement in classic telephony makes the (N-1)th network (i.e.
|the next to last network), on the path towards the destination,
|responsible for making the LNP database query. I wonder what's the
|reason for this? Wouldn't executing the LNP database query at the
|originating network have resulted in more efficient routing?    

Unfortunately "the" NPDB is distributed around the country.  Nobody
particularly wanted to pay for it, and there may still be issues about who
pays for dips to the database.  Moreover, each switch has to store
information about which numbers it regards as portable.  Prior to LNP, all
that a switch in California might need to keep about area code 212 was a
route list of trunks heading east, with no breakdown by NXX.  There are
practical issues with increasing the scope of the data so that every switch
in North America knows the portability status of every possible 6-digit
NPA-NXX combination, related both to the size of the resulting data file and
the number of places the information has to come from.

Since the classic telephone network is inherently distributed, it made very
good sense to defer the NP query until the call came under the control of a
switch "within" the zone of portability of the called number.  (The term
"within" must be used in a logical rather than a physical sense.)

|> It translates this to an LRN of +1-212-456, which
|> is looked up in TRIP. The longest prefix match (which is an 
|exact match
|> for the LRN) is found. The INVITE is then sent to the next 
|hop towards
|> that LRN as learned through TRIP. The original dialed number will be
|> carried in the To field.
|> 
|
|Looks like a SIP extension is required to indicate to other signaling
|points on the call path that an LNP query has already been 
|done for this
|destination, in order to avoid multiple LNP queries for the same call.
|Is such a flag already there?

You might also want to consider the possible value of defering the NP query
as late as possible in the signaling path.  Memory is cheap, but filling it
with unreliable information may be less useful than leaving it unallocated.

|
|Hussein
|
|
|> Note that we could also allow switch B to advertise 
|+1-212-455-1234 in
|> TRIP. If the LNP dip is not done, the proxy will pick this because of
|> longest prefix matching, and the call still completes. 
|However, I would
|> recommend doing this, since it will result in a completely flat TRIP
|> routing table.

I assume that's a typo for "recommend NOT doing this".  In the same way, it
may be best not to attempt to maintain full portability information or do
LNP queries at source for all calls, no matter where they're bound.

|> 
|> -Jonathan R.
|> 
|> --
|> Jonathan D. Rosenberg                       200 Executive Drive
|> Chief Scientist                             Suite 120
|> dynamicsoft                                 West Orange, NJ 07052
|> jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
|> http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
|> http://www.dynamicsoft.com
|> 
|> ---------
|> 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.

John Mainwaring
Nortel Networks
crm312a@nortelnetworks.com

------_=_NextPart_001_01BF40E9.D77E53A2
Content-Type: text/html;
	charset="windows-1252"
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=3Dwindows-1252">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2448.0">
<TITLE>RE: TRIP</TITLE>
</HEAD>
<BODY>
<BR>
<BR>

<P><FONT SIZE=3D2>|-----Original Message-----</FONT>
<BR><FONT SIZE=3D2>|From: Hussein F. Salama [<A =
HREF=3D"mailto:hsalama@cisco.com">mailto:hsalama@cisco.com</A>]</FONT>
<BR><FONT SIZE=3D2>|Sent: Monday, December 06, 1999 7:56 PM</FONT>
<BR><FONT SIZE=3D2>|To: Jonathan Rosenberg</FONT>
<BR><FONT SIZE=3D2>|Cc: James Yu; =
'iptel@lists.research.bell-labs.com'</FONT>
<BR><FONT SIZE=3D2>|Subject: Re: TRIP</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|Jonathan Rosenberg wrote:</FONT>
<BR><FONT SIZE=3D2>|&gt; </FONT>
<BR><FONT SIZE=3D2>|&gt; I think TRIP should carry subscriber numbers, =
but that the LNP dip</FONT>
<BR><FONT SIZE=3D2>|&gt; should happen before the TRIP database is =
accessed. So, returning to</FONT>
<BR><FONT SIZE=3D2>|&gt; your example, a gateway reaching only switch A =
would advertise</FONT>
<BR><FONT SIZE=3D2>|&gt; +1-212-455* and a gateway reaching only switch =
B </FONT>
<BR><FONT SIZE=3D2>|+1-212-466*. A PC users</FONT>
<BR><FONT SIZE=3D2>|&gt; calls +1-212-455-1234. The INVITE is sent to =
its local proxy, which</FONT>
<BR><FONT SIZE=3D2>|&gt; first does an LNP dip. </FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|The N-1 agreement in classic telephony makes the =
(N-1)th network (i.e.</FONT>
<BR><FONT SIZE=3D2>|the next to last network), on the path towards the =
destination,</FONT>
<BR><FONT SIZE=3D2>|responsible for making the LNP database query. I =
wonder what's the</FONT>
<BR><FONT SIZE=3D2>|reason for this? Wouldn't executing the LNP =
database query at the</FONT>
<BR><FONT SIZE=3D2>|originating network have resulted in more efficient =
routing?&nbsp;&nbsp;&nbsp; </FONT>
</P>

<P><FONT SIZE=3D2>Unfortunately &quot;the&quot; NPDB is distributed =
around the country.&nbsp; Nobody particularly wanted to pay for it, and =
there may still be issues about who pays for dips to the =
database.&nbsp; Moreover, each switch has to store information about =
which numbers it regards as portable.&nbsp; Prior to LNP, all that a =
switch in California might need to keep about area code 212 was a route =
list of trunks heading east, with no breakdown by NXX.&nbsp; There are =
practical issues with increasing the scope of the data so that every =
switch in North America knows the portability status of every possible =
6-digit NPA-NXX combination, related both to the size of the resulting =
data file and the number of places the information has to come =
from.</FONT></P>

<P><FONT SIZE=3D2>Since the classic telephone network is inherently =
distributed, it made very good sense to defer the NP query until the =
call came under the control of a switch &quot;within&quot; the zone of =
portability of the called number.&nbsp; (The term &quot;within&quot; =
must be used in a logical rather than a physical sense.)</FONT></P>

<P><FONT SIZE=3D2>|&gt; It translates this to an LRN of +1-212-456, =
which</FONT>
<BR><FONT SIZE=3D2>|&gt; is looked up in TRIP. The longest prefix match =
(which is an </FONT>
<BR><FONT SIZE=3D2>|exact match</FONT>
<BR><FONT SIZE=3D2>|&gt; for the LRN) is found. The INVITE is then sent =
to the next </FONT>
<BR><FONT SIZE=3D2>|hop towards</FONT>
<BR><FONT SIZE=3D2>|&gt; that LRN as learned through TRIP. The original =
dialed number will be</FONT>
<BR><FONT SIZE=3D2>|&gt; carried in the To field.</FONT>
<BR><FONT SIZE=3D2>|&gt; </FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|Looks like a SIP extension is required to indicate =
to other signaling</FONT>
<BR><FONT SIZE=3D2>|points on the call path that an LNP query has =
already been </FONT>
<BR><FONT SIZE=3D2>|done for this</FONT>
<BR><FONT SIZE=3D2>|destination, in order to avoid multiple LNP queries =
for the same call.</FONT>
<BR><FONT SIZE=3D2>|Is such a flag already there?</FONT>
</P>

<P><FONT SIZE=3D2>You might also want to consider the possible value of =
defering the NP query as late as possible in the signaling path.&nbsp; =
Memory is cheap, but filling it with unreliable information may be less =
useful than leaving it unallocated.</FONT></P>

<P><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|Hussein</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|&gt; Note that we could also allow switch B to =
advertise </FONT>
<BR><FONT SIZE=3D2>|+1-212-455-1234 in</FONT>
<BR><FONT SIZE=3D2>|&gt; TRIP. If the LNP dip is not done, the proxy =
will pick this because of</FONT>
<BR><FONT SIZE=3D2>|&gt; longest prefix matching, and the call still =
completes. </FONT>
<BR><FONT SIZE=3D2>|However, I would</FONT>
<BR><FONT SIZE=3D2>|&gt; recommend doing this, since it will result in =
a completely flat TRIP</FONT>
<BR><FONT SIZE=3D2>|&gt; routing table.</FONT>
</P>

<P><FONT SIZE=3D2>I assume that's a typo for &quot;recommend NOT doing =
this&quot;.&nbsp; In the same way, it may be best not to attempt to =
maintain full portability information or do LNP queries at source for =
all calls, no matter where they're bound.</FONT></P>

<P><FONT SIZE=3D2>|&gt; </FONT>
<BR><FONT SIZE=3D2>|&gt; -Jonathan R.</FONT>
<BR><FONT SIZE=3D2>|&gt; </FONT>
<BR><FONT SIZE=3D2>|&gt; --</FONT>
<BR><FONT SIZE=3D2>|&gt; Jonathan D. =
Rosenberg&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =
200 Executive Drive</FONT>
<BR><FONT SIZE=3D2>|&gt; Chief =
Scientist&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Suite 120</FONT>
<BR><FONT SIZE=3D2>|&gt; =
dynamicsoft&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; West =
Orange, NJ 07052</FONT>
<BR><FONT SIZE=3D2>|&gt; =
jdrosen@dynamicsoft.com&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&=
nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =
FAX:&nbsp;&nbsp; (732) 741-4778</FONT>
<BR><FONT SIZE=3D2>|&gt; <A =
HREF=3D"http://www.cs.columbia.edu/~jdrosen" =
TARGET=3D"_blank">http://www.cs.columbia.edu/~jdrosen</A>&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; PHONE: (732) 741-7244</FONT>
<BR><FONT SIZE=3D2>|&gt; <A HREF=3D"http://www.dynamicsoft.com" =
TARGET=3D"_blank">http://www.dynamicsoft.com</A></FONT>
<BR><FONT SIZE=3D2>|&gt; </FONT>
<BR><FONT SIZE=3D2>|&gt; ---------</FONT>
<BR><FONT SIZE=3D2>|&gt; This message came from the IETF IPTEL Working =
Group Mailing List.</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|-- </FONT>
<BR><FONT SIZE=3D2>|Hussein F. Salama</FONT>
<BR><FONT SIZE=3D2>|Cisco Systems</FONT>
<BR><FONT SIZE=3D2>|Mail Stop SJC6/3, 170 W. Tasman Drive, San Jose, CA =
95134</FONT>
<BR><FONT SIZE=3D2>|Voice: +1 (408) 527-7147, Fax: +1 (408) =
527-1714</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|---------</FONT>
<BR><FONT SIZE=3D2>|This message came from the IETF IPTEL Working Group =
Mailing List.</FONT>
</P>

<P><FONT SIZE=3D2>John Mainwaring</FONT>
<BR><FONT SIZE=3D2>Nortel Networks</FONT>
<BR><FONT SIZE=3D2>crm312a@nortelnetworks.com</FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01BF40E9.D77E53A2--

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec  7 14:54:21 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 OAA19673
	for <iptel-archive@odin.ietf.org>; Tue, 7 Dec 1999 14:54:20 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DF90452D6; Tue,  7 Dec 1999 14:47:31 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3AA2A52D5; Tue,  7 Dec 1999 14:47:31 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <384D6529.57FEF7F4@dynamicsoft.com>
Date: Tue, 07 Dec 1999 14:51:05 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: Dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: John Mainwaring <crm312a@nortelnetworks.com>
Cc: "'hsalama@cisco.com'" <hsalama@cisco.com>, James Yu <james.yu@neustar.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Re: TRIP
References: <89866A7EC9C0D1119C950000F8BCC2820157C00D@zrtpd00n.us.nortel.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 Mainwaring wrote:
> 
> |> Note that we could also allow switch B to advertise
> |+1-212-455-1234 in
> |> TRIP. If the LNP dip is not done, the proxy will pick this because
> of
> |> longest prefix matching, and the call still completes.
> |However, I would
> |> recommend doing this, since it will result in a completely flat
> TRIP
> |> routing table.
> 
> I assume that's a typo for "recommend NOT doing this".  In the same
> way, it may be best not to attempt to maintain full portability
> information or do LNP queries at source for all calls, no matter where
> they're bound.

Yes, that was a typo. I do not recommend injecting each ported number
into TRIP.
I'm less sure about deferring the LNP query to N-1. In the SIP
architecture,
at least, each proxy is responsible for doing a translation function
(called the
location service), which is similar in concept to an LNP database query.
I'll also
note that I have been told that the LNP dip being done at the N-1
carrier is
only a north american thing, and that in Europe, its done differently.


-Jonathan R.

-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec  7 15:22:04 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 PAA02469
	for <iptel-archive@odin.ietf.org>; Tue, 7 Dec 1999 15:22:02 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5464952D5; Tue,  7 Dec 1999 15:17:27 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BEA0752DC; Tue,  7 Dec 1999 15:17:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <384D6C2C.60CBC002@dynamicsoft.com>
Date: Tue, 07 Dec 1999 15:21:00 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: Dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: James Yu <james.yu@neustar.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: Re: TRIP
References: <ED88182BFF78D211A4D800A0C9E9435C3950D4@dc02.npac.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



James Yu wrote:
> 
> I think TRIP should carry subscriber numbers, but that the LNP dip
> should happen before the TRIP database is accessed. So, returning to
> your example, a gateway reaching only switch A would advertise
> +1-212-455* and a gateway reaching only switch B +1-212-466*. A PC users
> calls +1-212-455-1234. The INVITE is sent to its local proxy, which
> first does an LNP dip. It translates this to an LRN of +1-212-456, which
> is looked up in TRIP. The longest prefix match (which is an exact match
> for the LRN) is found. The INVITE is then sent to the next hop towards
> that LRN as learned through TRIP. The original dialed number will be
> carried in the To field.
> 
> ****** In your example, if LRN is used to determine the next hop, then TRIP
> carries the LRN information instead of subscriber numbers.  LRN is used for
> routing within SCN in North America.  The subscriber numbers are for people
> to dial (plus the dialing plan) to reach those subscribers.

Its LRNs in the sense that an LNP dip should be done before TRIP
databases are accessed, but its subscriber numbers in the sense that a
gateway is allowed (although its not recommended) to insert actual
subscriber numbers, or prefixes of any length really, into TRIP. The
format and structure of LRNs as you describe is the same, no?

-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec  7 16:12:38 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 QAA27008
	for <iptel-archive@odin.ietf.org>; Tue, 7 Dec 1999 16:12:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6FD3152DB; Tue,  7 Dec 1999 16:07:28 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BF07152DE; Tue,  7 Dec 1999 16:07:27 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <89866A7EC9C0D1119C950000F8BCC2820157C00E@zrtpd00n.us.nortel.com>
From: "John Mainwaring" <crm312a@nortelnetworks.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>
Cc: "'hsalama@cisco.com'" <hsalama@cisco.com>, James Yu <james.yu@neustar.com>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Tue, 7 Dec 1999 16:02:16 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: multipart/alternative;
              boundary="----_=_NextPart_001_01BF40F6.5EBEB152"
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_01BF40F6.5EBEB152
Content-Type: text/plain;
	charset="windows-1252"

|From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]

|I'm less sure about deferring the LNP query to N-1. In the SIP
|architecture,
|at least, each proxy is responsible for doing a translation function
|(called the
|location service), which is similar in concept to an LNP 
|database query.

I understand that SIP is different from the tandem network of circuit
switches.  I think SIP will require the originator to know LNP information,
but I'm not sure that existing collection of SCPs accessed by the TCAP/SS7
signaling system is the place to get it.

|I'll also
|note that I have been told that the LNP dip being done at the N-1
|carrier is
|only a north american thing, and that in Europe, its done differently.

Of course.

Actually, last I heard the Europeans were still considering Release to
Pivot, which might be a plausible model for SIP.  Initiate to the default
code owner first, and allow for a "refuse with redirection" or something.
It's got its weaknesses, which include putting a burden on the donor of the
ported number and relying on an uninvolved third party to set up a call, but
it does keep the supporting infrastructure simple.

Another interesting working model is wireless roaming.  It is capable of
dynamically redirecting you anywhere (at least within a continent or so) and
it works at the level of individual telephone numbers.  Of course a single
wireless operating company owns all the numbers in an NPA NXX, which avoids
the major problem LNP was designed to create.  (No, that's not a typo, it's
editorial comment.)

Probably it's more of an administrative problem than a technical one.  It's
so easy to fall into the trap of thinking there are technical solutions to
administrative problems, although technical solutions can certainly help
perpetuate administrative problems.

John Mainwaring 
Nortel Networks 
crm312a@nortelnetworks.com 

------_=_NextPart_001_01BF40F6.5EBEB152
Content-Type: text/html;
	charset="windows-1252"
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=3Dwindows-1252">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2448.0">
<TITLE>RE: TRIP</TITLE>
</HEAD>
<BODY>

<P><FONT SIZE=3D2>|From: Jonathan Rosenberg [<A =
HREF=3D"mailto:jdrosen@dynamicsoft.com">mailto:jdrosen@dynamicsoft.com</=
A>]</FONT>
</P>

<P><FONT SIZE=3D2>|I'm less sure about deferring the LNP query to N-1. =
In the SIP</FONT>
<BR><FONT SIZE=3D2>|architecture,</FONT>
<BR><FONT SIZE=3D2>|at least, each proxy is responsible for doing a =
translation function</FONT>
<BR><FONT SIZE=3D2>|(called the</FONT>
<BR><FONT SIZE=3D2>|location service), which is similar in concept to =
an LNP </FONT>
<BR><FONT SIZE=3D2>|database query.</FONT>
</P>

<P><FONT SIZE=3D2>I understand that SIP is different from the tandem =
network of circuit switches.&nbsp; I think SIP will require the =
originator to know LNP information, but I'm not sure that existing =
collection of SCPs accessed by the TCAP/SS7 signaling system is the =
place to get it.</FONT></P>

<P><FONT SIZE=3D2>|I'll also</FONT>
<BR><FONT SIZE=3D2>|note that I have been told that the LNP dip being =
done at the N-1</FONT>
<BR><FONT SIZE=3D2>|carrier is</FONT>
<BR><FONT SIZE=3D2>|only a north american thing, and that in Europe, =
its done differently.</FONT>
</P>

<P><FONT SIZE=3D2>Of course.</FONT>
</P>

<P><FONT SIZE=3D2>Actually, last I heard the Europeans were still =
considering Release to Pivot, which might be a plausible model for =
SIP.&nbsp; Initiate to the default code owner first, and allow for a =
&quot;refuse with redirection&quot; or something.&nbsp; It's got its =
weaknesses, which include putting a burden on the donor of the ported =
number and relying on an uninvolved third party to set up a call, but =
it does keep the supporting infrastructure simple.</FONT></P>

<P><FONT SIZE=3D2>Another interesting working model is wireless =
roaming.&nbsp; It is capable of dynamically redirecting you anywhere =
(at least within a continent or so) and it works at the level of =
individual telephone numbers.&nbsp; Of course a single wireless =
operating company owns all the numbers in an NPA NXX, which avoids the =
major problem LNP was designed to create.&nbsp; (No, that's not a typo, =
it's editorial comment.)</FONT></P>

<P><FONT SIZE=3D2>Probably it's more of an administrative problem than =
a technical one.&nbsp; It's so easy to fall into the trap of thinking =
there are technical solutions to administrative problems, although =
technical solutions can certainly help perpetuate administrative =
problems.</FONT></P>

<P><FONT SIZE=3D2>John Mainwaring </FONT>
<BR><FONT SIZE=3D2>Nortel Networks </FONT>
<BR><FONT SIZE=3D2>crm312a@nortelnetworks.com </FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01BF40F6.5EBEB152--

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec  7 16:24:24 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 QAA03055
	for <iptel-archive@odin.ietf.org>; Tue, 7 Dec 1999 16:24:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A908F52DE; Tue,  7 Dec 1999 16:19:29 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1C75A52DF; Tue,  7 Dec 1999 16:19:29 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <4FBEA8857476D311A03300204840E1CF0684DD@whq-msgusr-02.fore.com>
From: "Rosen, Brian" <brosen@fore.com>
To: "'John Mainwaring'" <crm312a@nortelnetworks.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Tue, 7 Dec 1999 16:17:56 -0500 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01BF40F8.4BE581E4"
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_01BF40F8.4BE581E4
Content-Type: text/plain;
	charset="windows-1252"

No need to get the data from the SCPs.  The DNS can be fed from the NPAC
directly.
 
Brian

-----Original Message-----
From: John Mainwaring [mailto:crm312a@nortelnetworks.com]
Sent: Tuesday, December 07, 1999 4:02 PM
To: 'Jonathan Rosenberg'
Cc: 'hsalama@cisco.com'; James Yu; 'iptel@lists.research.bell-labs.com'
Subject: RE: TRIP



|From: Jonathan Rosenberg [ mailto:jdrosen@dynamicsoft.com
<mailto:jdrosen@dynamicsoft.com> ] 

|I'm less sure about deferring the LNP query to N-1. In the SIP 
|architecture, 
|at least, each proxy is responsible for doing a translation function 
|(called the 
|location service), which is similar in concept to an LNP 
|database query. 

I understand that SIP is different from the tandem network of circuit
switches.  I think SIP will require the originator to know LNP information,
but I'm not sure that existing collection of SCPs accessed by the TCAP/SS7
signaling system is the place to get it.

|I'll also 
|note that I have been told that the LNP dip being done at the N-1 
|carrier is 
|only a north american thing, and that in Europe, its done differently. 

Of course. 

Actually, last I heard the Europeans were still considering Release to
Pivot, which might be a plausible model for SIP.  Initiate to the default
code owner first, and allow for a "refuse with redirection" or something.
It's got its weaknesses, which include putting a burden on the donor of the
ported number and relying on an uninvolved third party to set up a call, but
it does keep the supporting infrastructure simple.

Another interesting working model is wireless roaming.  It is capable of
dynamically redirecting you anywhere (at least within a continent or so) and
it works at the level of individual telephone numbers.  Of course a single
wireless operating company owns all the numbers in an NPA NXX, which avoids
the major problem LNP was designed to create.  (No, that's not a typo, it's
editorial comment.)

Probably it's more of an administrative problem than a technical one.  It's
so easy to fall into the trap of thinking there are technical solutions to
administrative problems, although technical solutions can certainly help
perpetuate administrative problems.

John Mainwaring 
Nortel Networks 
crm312a@nortelnetworks.com 


------_=_NextPart_001_01BF40F8.4BE581E4
Content-Type: text/html;
	charset="windows-1252"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=windows-1252">
<TITLE>RE: TRIP</TITLE>

<META content="MSHTML 5.00.2614.3401" name=GENERATOR></HEAD>
<BODY>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN class=010281221-07121999>No 
need to get the data from the SCPs.&nbsp; The DNS can be fed from the 
NPAC</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN 
class=010281221-07121999>directly.</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN 
class=010281221-07121999></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN 
class=010281221-07121999>Brian</SPAN></FONT></DIV>
<BLOCKQUOTE 
style="BORDER-LEFT: #0000ff 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px">
  <DIV align=left class=OutlookMessageHeader dir=ltr><FONT face=Tahoma 
  size=2>-----Original Message-----<BR><B>From:</B> John Mainwaring 
  [mailto:crm312a@nortelnetworks.com]<BR><B>Sent:</B> Tuesday, December 07, 1999 
  4:02 PM<BR><B>To:</B> 'Jonathan Rosenberg'<BR><B>Cc:</B> 'hsalama@cisco.com'; 
  James Yu; 'iptel@lists.research.bell-labs.com'<BR><B>Subject:</B> RE: 
  TRIP<BR><BR></DIV></FONT>
  <P><FONT size=2>|From: Jonathan Rosenberg [<A 
  href="mailto:jdrosen@dynamicsoft.com">mailto:jdrosen@dynamicsoft.com</A>]</FONT> 
  </P>
  <P><FONT size=2>|I'm less sure about deferring the LNP query to N-1. In the 
  SIP</FONT> <BR><FONT size=2>|architecture,</FONT> <BR><FONT size=2>|at least, 
  each proxy is responsible for doing a translation function</FONT> <BR><FONT 
  size=2>|(called the</FONT> <BR><FONT size=2>|location service), which is 
  similar in concept to an LNP </FONT><BR><FONT size=2>|database query.</FONT> 
  </P>
  <P><FONT size=2>I understand that SIP is different from the tandem network of 
  circuit switches.&nbsp; I think SIP will require the originator to know LNP 
  information, but I'm not sure that existing collection of SCPs accessed by the 
  TCAP/SS7 signaling system is the place to get it.</FONT></P>
  <P><FONT size=2>|I'll also</FONT> <BR><FONT size=2>|note that I have been told 
  that the LNP dip being done at the N-1</FONT> <BR><FONT size=2>|carrier 
  is</FONT> <BR><FONT size=2>|only a north american thing, and that in Europe, 
  its done differently.</FONT> </P>
  <P><FONT size=2>Of course.</FONT> </P>
  <P><FONT size=2>Actually, last I heard the Europeans were still considering 
  Release to Pivot, which might be a plausible model for SIP.&nbsp; Initiate to 
  the default code owner first, and allow for a "refuse with redirection" or 
  something.&nbsp; It's got its weaknesses, which include putting a burden on 
  the donor of the ported number and relying on an uninvolved third party to set 
  up a call, but it does keep the supporting infrastructure simple.</FONT></P>
  <P><FONT size=2>Another interesting working model is wireless roaming.&nbsp; 
  It is capable of dynamically redirecting you anywhere (at least within a 
  continent or so) and it works at the level of individual telephone 
  numbers.&nbsp; Of course a single wireless operating company owns all the 
  numbers in an NPA NXX, which avoids the major problem LNP was designed to 
  create.&nbsp; (No, that's not a typo, it's editorial comment.)</FONT></P>
  <P><FONT size=2>Probably it's more of an administrative problem than a 
  technical one.&nbsp; It's so easy to fall into the trap of thinking there are 
  technical solutions to administrative problems, although technical solutions 
  can certainly help perpetuate administrative problems.</FONT></P>
  <P><FONT size=2>John Mainwaring </FONT><BR><FONT size=2>Nortel Networks 
  </FONT><BR><FONT size=2>crm312a@nortelnetworks.com 
</FONT></P></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01BF40F8.4BE581E4--

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec  8 09:56:38 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 JAA28038
	for <iptel-archive@odin.ietf.org>; Wed, 8 Dec 1999 09:56:37 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 2223152E6; Wed,  8 Dec 1999 09:47:17 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id CB50E52E5; Wed,  8 Dec 1999 09:47:10 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C3950D4@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Tue, 7 Dec 1999 10:13:39 -0600 
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

Jonathan, 

Please see the comments/questions below.

James

-----Original Message-----
From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]
Sent: Monday, December 06, 1999 7:32 PM
To: James Yu
Cc: 'iptel@lists.research.bell-labs.com'
Subject: Re: TRIP


James,

Thanks for the excellent description of the process. Comments below:


James Yu wrote:
> 
> Jonathan,
> 
> North American Numbering Plan numbers have the format of NPA+NXX+XXXX
where
> NPA (Numbering Plan Area) is the area code and has the format of NXX (The
> area codes used to have the format of N0/1X where the second digit can be
> either 0 or 1. That restriction has been removed to allow more area codes
to
> be assigned.). N is any digit from 2 to 9 and X is any digit from 0 to 9.
> The NXX after NPA is called "office code."  Normally, a switch is assigned
> the whole block of NPA+NXX (e.g., 212-455-0000 to 212-455-9999).  Now
> "number pooling" will allow several LECs to share one NPA+NXX.  That is, a
> LEC will be assigned, for example, 212-477-1 (block from 212-477-1000 to
> 212-477-1999) and another LEC assigned 212-477-2 (block from 212-477-2000
to
> 212-477-2999).  We can talk about "number pooling" later if you or others
> have questions.  The focus of this discussion is on LRN and number
> portability.

Hmm, sounds familiar.... CIDR?

What is CIDR?

> If the TRIP information contains the reachability of subscriber number
> ranges, A GW that can reach Switch A but not Switch B will have to
indicate
> that it can reach +1-212-455-0000 to +1-212-455-1233 and
> +1-212-455-1235-+1-212-455-9999 (if only one number has been ported out of
> Switch A).  The problem of this is that many ported-out numbers need to be
> singled out from the reachable phone number ranges ranges.  This is a
minor
> problem when comparing to the bigger problem of updating the TRIP
> information every time a number changes its assoicated LRN after been
ported
> to a new LEC.

I think TRIP should carry subscriber numbers, but that the LNP dip
should happen before the TRIP database is accessed. So, returning to
your example, a gateway reaching only switch A would advertise
+1-212-455* and a gateway reaching only switch B +1-212-466*. A PC users
calls +1-212-455-1234. The INVITE is sent to its local proxy, which
first does an LNP dip. It translates this to an LRN of +1-212-456, which
is looked up in TRIP. The longest prefix match (which is an exact match
for the LRN) is found. The INVITE is then sent to the next hop towards
that LRN as learned through TRIP. The original dialed number will be
carried in the To field. 

****** In your example, if LRN is used to determine the next hop, then TRIP
carries the LRN information instead of subscriber numbers.  LRN is used for
routing within SCN in North America.  The subscriber numbers are for people
to dial (plus the dialing plan) to reach those subscribers.

Note that we could also allow switch B to advertise +1-212-455-1234 in
TRIP. If the LNP dip is not done, the proxy will pick this because of
longest prefix matching, and the call still completes. However, I would
recommend doing this, since it will result in a completely flat TRIP
routing table.

****** This is what I try to avoid because number portability will require
that the advertisement be changed everytime a number is ported to a new LEC.


-Jonathan R.

-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec  8 10:04:25 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 KAA02595
	for <iptel-archive@odin.ietf.org>; Wed, 8 Dec 1999 10:04:25 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id DD45B52D6; Wed,  8 Dec 1999 09:47:38 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4852952DF; Wed,  8 Dec 1999 09:47:22 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <EE88182BFF78D211A4D800A0C9E9435CD70604@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'hsalama@cisco.com'" <hsalama@cisco.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Tue, 7 Dec 1999 10:19:45 -0600 
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

Hussein,

Please see my comments below (although you addressed your questions to
Jonathan).

James

-----Original Message-----
From: Hussein F. Salama [mailto:hsalama@cisco.com]
Sent: Monday, December 06, 1999 7:56 PM
To: Jonathan Rosenberg
Cc: James Yu; 'iptel@lists.research.bell-labs.com'
Subject: Re: TRIP




Jonathan Rosenberg wrote:
> 
> James,
> 
> Thanks for the excellent description of the process. Comments below:
> 
> James Yu wrote:
> >
> > Jonathan,
> >
> > North American Numbering Plan numbers have the format of NPA+NXX+XXXX
where
> > NPA (Numbering Plan Area) is the area code and has the format of NXX
(The
> > area codes used to have the format of N0/1X where the second digit can
be
> > either 0 or 1. That restriction has been removed to allow more area
codes to
> > be assigned.). N is any digit from 2 to 9 and X is any digit from 0 to
9.
> > The NXX after NPA is called "office code."  Normally, a switch is
assigned
> > the whole block of NPA+NXX (e.g., 212-455-0000 to 212-455-9999).  Now
> > "number pooling" will allow several LECs to share one NPA+NXX.  That is,
a
> > LEC will be assigned, for example, 212-477-1 (block from 212-477-1000 to
> > 212-477-1999) and another LEC assigned 212-477-2 (block from
212-477-2000 to
> > 212-477-2999).  We can talk about "number pooling" later if you or
others
> > have questions.  The focus of this discussion is on LRN and number
> > portability.
> 
> Hmm, sounds familiar.... CIDR?
> 
> > If the TRIP information contains the reachability of subscriber number
> > ranges, A GW that can reach Switch A but not Switch B will have to
indicate
> > that it can reach +1-212-455-0000 to +1-212-455-1233 and
> > +1-212-455-1235-+1-212-455-9999 (if only one number has been ported out
of
> > Switch A).  The problem of this is that many ported-out numbers need to
be
> > singled out from the reachable phone number ranges ranges.  This is a
minor
> > problem when comparing to the bigger problem of updating the TRIP
> > information every time a number changes its assoicated LRN after been
ported
> > to a new LEC.
> 
> I think TRIP should carry subscriber numbers, but that the LNP dip
> should happen before the TRIP database is accessed. So, returning to
> your example, a gateway reaching only switch A would advertise
> +1-212-455* and a gateway reaching only switch B +1-212-466*. A PC users
> calls +1-212-455-1234. The INVITE is sent to its local proxy, which
> first does an LNP dip. 

The N-1 agreement in classic telephony makes the (N-1)th network (i.e.
the next to last network), on the path towards the destination,
responsible for making the LNP database query. I wonder what's the
reason for this? Wouldn't executing the LNP database query at the
originating network have resulted in more efficient routing? 

****** In the U.S. the N-1 network is usually the IEC network.  The reason
that N-1 network is to perform the NPDB query is because the caller is
served by his/her long-distance carrier for this call.  That is, the IEC
will bill the caller (sometimes the billing is done through the LEC, but
from the provider-customer relationship, the IEC is the service provide for
long distance service) and pay the originating and terminating networks
(local exchange carriers) for providing "access."  If the originating
network performs the NPDB dip, it will want to be paid (by the IEC) to
perform that function.  That is why the originating network is not required
to perform the NPDB for long distance calls unless the IEC requests it to do
so.  

If a call terminated at a cellular/PCS network in the U.S., the N-1 network
(the second to the last network) would be the terminating local exchange
carrier that is connected to the terminating cellular/PCS network if the IEC
does not terminate the call to cellular/PCS network directly over direct
trunk connection between the two networks. There won't be any routing
efficiency whether the originating or IEC network performs the NPDB query
for long distance calls.  When location portability is supported in the
future (across the Local Access and Transport Area (LATA) boundary), the
originating network would be the one to perform the NPDB queries for all the
calls because it cannot be determined whether a call is a long distance or
local call before NPDB query is done.

Please note that the internet in many cases will be the "N-1" network for
VoIP for SCN-terminated calls where the calls are originated from within the
internet or from SCN via internet.
******


> It translates this to an LRN of +1-212-456, which
> is looked up in TRIP. The longest prefix match (which is an exact match
> for the LRN) is found. The INVITE is then sent to the next hop towards
> that LRN as learned through TRIP. The original dialed number will be
> carried in the To field.
> 

Looks like a SIP extension is required to indicate to other signaling
points on the call path that an LNP query has already been done for this
destination, in order to avoid multiple LNP queries for the same call.
Is such a flag already there?

****** If ISUP is carried over IP, the information would be carried in the
encapsulated ISUP message.  Otherwise, SIP may need to be enhanced to carry
not only the dip indicator but also the LRN (in addition to the dialed
number).  I don't know whether a flag is already there. ******




Hussein


> Note that we could also allow switch B to advertise +1-212-455-1234 in
> TRIP. If the LNP dip is not done, the proxy will pick this because of
> longest prefix matching, and the call still completes. However, I would
> recommend doing this, since it will result in a completely flat TRIP
> routing table.
> 
> -Jonathan R.
> 
> --
> Jonathan D. Rosenberg                       200 Executive Drive
> Chief Scientist                             Suite 120
> dynamicsoft                                 West Orange, NJ 07052
> jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
> http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
> http://www.dynamicsoft.com
> 
> ---------
> 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  Wed Dec  8 10:13:44 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 KAA07608
	for <iptel-archive@odin.ietf.org>; Wed, 8 Dec 1999 10:13:43 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id CB44752DB; Wed,  8 Dec 1999 09:50:57 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id B280752EC; Wed,  8 Dec 1999 09:48:21 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <01BF40D2.E4733AD0.stsang@research.telcordia.com>
From: Simon Tsang <stsang@research.telcordia.com>
Reply-To: "stsang@research.telcordia.com" <stsang@research.telcordia.com>
To: "'Rosen, Brian'" <brosen@fore.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Tue, 7 Dec 1999 16:48:29 -0500
Organization: Telcordia Technologies
X-Mailer: Microsoft Internet E-mail/MAPI - 8.0.0.4211
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

As John noted earlier, things in Europe are different from in the USA. 
 There are no NPACs in Europe.  Ported numbers are held either in SCPs or 
in the local exchange switches in the donor networks.  Unfortunately for 
these cases, it means that it will be difficult to load DNS and keep it 
synchronised with ported number data due to the large number or places 
where this data exists.

Legislation may force carriers in Europe and the rest of the world to 
follow the US NP model.  Until then, the Release to Pivot idea (which John 
described in his earlier message) is probably the cleanest technical 
solution to the problem in these areas.  I agree this really is more an 
administrative than technical problem.

Simon
Service Integration Research
Telcordia Technologies
stsang@telcordia.com

-----Original Message-----
From:	Rosen, Brian [SMTP:brosen@fore.com]
Sent:	Tuesday, December 07, 1999 4:18 PM
To:	'John Mainwaring'
Cc:	'iptel@lists.research.bell-labs.com'
Subject:	RE: TRIP
No need to get the data from the SCPs.  The DNS can be fed from the NPAC
directly.

Brian

-----Original Message-----
From: John Mainwaring [mailto:crm312a@nortelnetworks.com]
Sent: Tuesday, December 07, 1999 4:02 PM
To: 'Jonathan Rosenberg'
Cc: 'hsalama@cisco.com'; James Yu; 'iptel@lists.research.bell-labs.com'
Subject: RE: TRIP



|From: Jonathan Rosenberg [ mailto:jdrosen@dynamicsoft.com
<mailto:jdrosen@dynamicsoft.com> ]

|I'm less sure about deferring the LNP query to N-1. In the SIP
|architecture,
|at least, each proxy is responsible for doing a translation function
|(called the
|location service), which is similar in concept to an LNP
|database query.

I understand that SIP is different from the tandem network of circuit
switches.  I think SIP will require the originator to know LNP information,
but I'm not sure that existing collection of SCPs accessed by the TCAP/SS7
signaling system is the place to get it.

|I'll also
|note that I have been told that the LNP dip being done at the N-1
|carrier is
|only a north american thing, and that in Europe, its done differently.

Of course.

Actually, last I heard the Europeans were still considering Release to
Pivot, which might be a plausible model for SIP.  Initiate to the default
code owner first, and allow for a "refuse with redirection" or something.
It's got its weaknesses, which include putting a burden on the donor of the
ported number and relying on an uninvolved third party to set up a call, 
but
it does keep the supporting infrastructure simple.

Another interesting working model is wireless roaming.  It is capable of
dynamically redirecting you anywhere (at least within a continent or so) 
and
it works at the level of individual telephone numbers.  Of course a single
wireless operating company owns all the numbers in an NPA NXX, which avoids
the major problem LNP was designed to create.  (No, that's not a typo, it's
editorial comment.)

Probably it's more of an administrative problem than a technical one.  It's
so easy to fall into the trap of thinking there are technical solutions to
administrative problems, although technical solutions can certainly help
perpetuate administrative problems.

John Mainwaring
Nortel Networks
crm312a@nortelnetworks.com

 << File: ATT00002.txt; charset = windows-1252 >>  << File: ATT00003.html; 
charset = windows-1252 >> 


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec  8 11:32:15 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 LAA19489
	for <iptel-archive@odin.ietf.org>; Wed, 8 Dec 1999 11:32:15 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id EB4B352DE; Wed,  8 Dec 1999 09:50:58 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7F30052E0; Wed,  8 Dec 1999 09:48:27 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C39515B@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: John Mainwaring <crm312a@nortelnetworks.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Wed, 8 Dec 1999 02:32:06 -0600 
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
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by ietf.org id LAA19489

 

-----Original Message-----
From: John Mainwaring [mailto:crm312a@nortelnetworks.com]
Sent: Tuesday, December 07, 1999 4:02 PM
To: 'Jonathan Rosenberg'
Cc: 'hsalama@cisco.com'; James Yu; 'iptel@lists.research.bell-labs.com'
Subject: RE: TRIP



|From: Jonathan Rosenberg [ mailto:jdrosen@dynamicsoft.com
<mailto:jdrosen@dynamicsoft.com> ] 

|I'm less sure about deferring the LNP query to N-1. In the SIP 
|architecture, 
|at least, each proxy is responsible for doing a translation function 
|(called the 
|location service), which is similar in concept to an LNP 
|database query. 

I understand that SIP is different from the tandem network of circuit
switches.  I think SIP will require the originator to know LNP information,
but I'm not sure that existing collection of SCPs accessed by the TCAP/SS7
signaling system is the place to get it.

 *********  For SCN to SCN via IP or IP to SCN calls, the internet is the
N-1 network (although many networks may be involved in the IP-domain
routing), the entity that performs the LNP query via TCAP/SS7 over IP or
real-time or non-real-time LNP-like mechanism via DNS) can be the
originating GW/GK/SIP server or an intermediate GW/GK/SIP server but won't
be the terminating GW/GK/SIP server because it may be too late to perform
LNP query to find out that it does not reach the serving switch identified
by the LRN.  There are certainly many issues if the originating GW/GK/SIP is
to perform the LNP query.  For example, several protocols can be used
between a switch and a NPDB.  In Europe, I heard that there are
country-specific variations to the switch-NPDB interface protocol.  For the
originating GW/GK/SIP server to query all the NPDBs worldwide would be a big
burden.  

One possible solution is that a standardized query can be launched by the
entity which will be mapped by an interworking like entity to
country-specific NPDB query to reach the existing NPDBs.  Another potential
solution is to have NPDBs in the internet that support the standardized IP
entity-to-NPDB interface.  Certainly this does not rule out that the
existing NPDB from being enhanced to support this new IP-domain interface.
*************

 |I'll also 
|note that I have been told that the LNP dip being done at the N-1 
|carrier is 
|only a north american thing, and that in Europe, its done differently. 

Of course. 

Actually, last I heard the Europeans were still considering Release to
Pivot, which might be a plausible model for SIP.  Initiate to the default
code owner first, and allow for a "refuse with redirection" or something. 
It's got its weaknesses, which include putting a burden on the donor of the
ported number and relying on an uninvolved third party to set up a call, but
it does keep the supporting infrastructure simple. 

******* Release to Pivot will not work in the U.S. because the FCC has ruled
out any solution that relies on the donor network to complete the call. 
Also, the "donor" switch and the "pivot" switch at least must be upgraded to
support this solution.  If the switch does not store the new LRN/routing
prefix information for a ported-out number, a database dip must be performed
in order to provide the LRN/routing prefix in the ISUP Release message.  If
a "donor" switch stores the information, the provisioning/administration of
the information is not a simple process. **********

Another interesting working model is wireless roaming.  It is capable of
dynamically redirecting you anywhere (at least within a continent or so) and
it works at the level of individual telephone numbers.  Of course a single
wireless operating company owns all the numbers in an NPA NXX, which avoids
the major problem LNP was designed to create.  (No, that's not a typo, it's
editorial comment.) 

Probably it's more of an administrative problem than a technical one.  It's
so easy to fall into the trap of thinking there are technical solutions to
administrative problems, although technical solutions can certainly help
perpetuate administrative problems. 

******** FCC has granted a delay for the wireless carriers in the U.S. to
support local number portability (LNP) because the new entrants - PCS
carriers are too busy in deploying their networks.  LNP is intended to help
the new entrants to compete with the incumbent carriers.  So if the new
entrants ask for a delay, FCC certainly has no reason to push wireless LNP. 
Also there are issues on roaming where the roaming markets may not be
required to support LNP (LNP is to be supported in the top 100 major
Statistical Trading Areas in phases).  But I believe that the delay is
temporary because FCC's goal may be nation-wide number portability which
allows the subscriber to switch from wireline service provider to a wireless
service provider and move from one location to another location.  The LRN
method will allow that with minimum system upgrades.  The major issue about
location portability is the call charing/rating.  If flat rate is used for
all toll calls (e.g., 10 cents per minute regardless of the distance between
the calling party and called party), location portability can be supported
easily. 

            Wireles is a similar model if we view the "donor switch" as the
"home" system" and the "serving switch" as the "vistied system."  However,
there are differences.  The "home system" in the wireless model is the
service provider of the "roamer."  It authorizes the visited system to
provide the services to the roamer and guarantees the payment even if the
subscriber does not pay his/her bill.  Also, Mobile Application Part (MAP)
is used between the visited and home systems to register the "roamer" and to
get a temporary routing number to route the call to the visited system. 
Either the call must always be routed through the home system to reach the
visited system or a query to the home system must be performed first to
perform optimal routing to the visited system directly.  NPDB approach
actually allows the query to be done (for optimal routing) without the
involvement of the "donor" network.  *************

 John Mainwaring 
Nortel Networks 
crm312a@nortelnetworks.com 


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec  8 11:41:40 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 LAA24660
	for <iptel-archive@odin.ietf.org>; Wed, 8 Dec 1999 11:41:38 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id A358952DF; Wed,  8 Dec 1999 09:54:23 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id ECC7E52AB; Wed,  8 Dec 1999 09:48:36 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C39515D@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Wed, 8 Dec 1999 02:32:10 -0600 
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



-----Original Message-----
From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]
Sent: Tuesday, December 07, 1999 3:21 PM
To: James Yu
Cc: 'iptel@lists.research.bell-labs.com'
Subject: Re: TRIP




James Yu wrote:
> 
> I think TRIP should carry subscriber numbers, but that the LNP dip
> should happen before the TRIP database is accessed. So, returning to
> your example, a gateway reaching only switch A would advertise
> +1-212-455* and a gateway reaching only switch B +1-212-466*. A PC users
> calls +1-212-455-1234. The INVITE is sent to its local proxy, which
> first does an LNP dip. It translates this to an LRN of +1-212-456, which
> is looked up in TRIP. The longest prefix match (which is an exact match
> for the LRN) is found. The INVITE is then sent to the next hop towards
> that LRN as learned through TRIP. The original dialed number will be
> carried in the To field.
> 
> ****** In your example, if LRN is used to determine the next hop, then
TRIP
> carries the LRN information instead of subscriber numbers.  LRN is used
for
> routing within SCN in North America.  The subscriber numbers are for
people
> to dial (plus the dialing plan) to reach those subscribers.

Its LRNs in the sense that an LNP dip should be done before TRIP
databases are accessed, but its subscriber numbers in the sense that a
gateway is allowed (although its not recommended) to insert actual
subscriber numbers, or prefixes of any length really, into TRIP. The
format and structure of LRNs as you describe is the same, no?

******* Under LNP, the routing in the network (in the North America) is
based on LRN whether NPDB dip is performed or not.  This is why the call is
routed to the "donor" switch if NPDB is done for a ported-out number.  TRIP
can certainly pass individual subscriber numbers, but the administratoin of
numbers being ported in and out would make it messy to update the TRIP
information in the associated location servers.

James **********

-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec  8 12:01:22 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 MAA04850
	for <iptel-archive@odin.ietf.org>; Wed, 8 Dec 1999 12:01:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 880AA52F7; Wed,  8 Dec 1999 11:25:53 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7DACC52F6; Wed,  8 Dec 1999 11:25:50 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <89866A7EC9C0D1119C950000F8BCC2820157C012@zrtpd00n.us.nortel.com>
From: "John Mainwaring" <crm312a@nortelnetworks.com>
To: "'James Yu'" <james.yu@neustar.com>
Cc: "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Wed, 8 Dec 1999 10:45:28 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: multipart/alternative;
              boundary="----_=_NextPart_001_01BF4193.46F2C43A"
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_01BF4193.46F2C43A
Content-Type: text/plain;
	charset="iso-8859-1"

Generally I agree with your remarks.

Specifically, I agree that the destination gateway is too late to do the
query, unless you take an RTP-like approach.  In effect, the default
destination gateway is a proxy for the donor switch.  By the time you get
there, it's a bit late to do a query.  On the other hand, you probably don't
want to push the query all the way back to the originating gateway.  That
would require either that each gateway have knowledge of portability
information for every NPA-NXX in existence or that an LNP or equivalent
query be launched on every call.

I think you are suggesting that LNP queries should be done at intermediate
points that are "close enough" to the destination to be expected to know
about the portability status of the destination.  If so, I agree with you.

I also agree with your suggestion that there may be a need for specialized
NPDB nodes in the IP network.  I'm not sure what form it should take, but
the existing NPDBs are not only engineered for the PSTN, they are a cost
center in the PSTN, and I understand that there are charges associated with
using them.  It may be appropriate to use them, or it may be better to
explore other solutions.  The point is that it is necessary to do more than
wave our hands and say "An NPDB query happens somewhere."

You are right that the FCC ruled out Release to Pivot or any solution that
relies on the donor network to complete a call in the PSTN.  I doubt if they
were even aware of the possibility of VoIP at the time they made that
ruling.  There are significant ongoing costs associated with NPDBs.  It
seems appropriate to consider other architectures, especially since there is
so little legacy encumbrance in SIP right now.  It's probably true that an
appropriately administered NPDB (loaded directly from the NPAC, as Brian
suggested) is a more robust solution than anything that relies on the donor.
My question is whether an RTP solution might be sufficiently reliable to be
acceptable in practice and sufficiently cheap to be attractive.  As you
pointed out in connection with wireless number portability, the FCC doesn't
always impose the same burdens in new areas as it does in the legacy
environment.


-----Original Message-----
From: James Yu [mailto:james.yu@neustar.com]
Sent: Wednesday, December 08, 1999 3:32 AM
To: Mainwaring, John [NC2:3B30-I:EXCH]
Cc: 'iptel@lists.research.bell-labs.com'
Subject: RE: TRIP


 

-----Original Message-----
From: John Mainwaring [mailto:crm312a@nortelnetworks.com]
Sent: Tuesday, December 07, 1999 4:02 PM
To: 'Jonathan Rosenberg'
Cc: 'hsalama@cisco.com'; James Yu; 'iptel@lists.research.bell-labs.com'
Subject: RE: TRIP



|From: Jonathan Rosenberg [ mailto:jdrosen@dynamicsoft.com
<mailto:jdrosen@dynamicsoft.com> ] 

|I'm less sure about deferring the LNP query to N-1. In the SIP 
|architecture, 
|at least, each proxy is responsible for doing a translation function 
|(called the 
|location service), which is similar in concept to an LNP 
|database query. 

I understand that SIP is different from the tandem network of circuit
switches.  I think SIP will require the originator to know LNP information,
but I'm not sure that existing collection of SCPs accessed by the TCAP/SS7
signaling system is the place to get it.

 *********  For SCN to SCN via IP or IP to SCN calls, the internet is the
N-1 network (although many networks may be involved in the IP-domain
routing), the entity that performs the LNP query via TCAP/SS7 over IP or
real-time or non-real-time LNP-like mechanism via DNS) can be the
originating GW/GK/SIP server or an intermediate GW/GK/SIP server but won't
be the terminating GW/GK/SIP server because it may be too late to perform
LNP query to find out that it does not reach the serving switch identified
by the LRN.  There are certainly many issues if the originating GW/GK/SIP is
to perform the LNP query.  For example, several protocols can be used
between a switch and a NPDB.  In Europe, I heard that there are
country-specific variations to the switch-NPDB interface protocol.  For the
originating GW/GK/SIP server to query all the NPDBs worldwide would be a big
burden.  

One possible solution is that a standardized query can be launched by the
entity which will be mapped by an interworking like entity to
country-specific NPDB query to reach the existing NPDBs.  Another potential
solution is to have NPDBs in the internet that support the standardized IP
entity-to-NPDB interface.  Certainly this does not rule out that the
existing NPDB from being enhanced to support this new IP-domain interface.
*************

 |I'll also 
|note that I have been told that the LNP dip being done at the N-1 
|carrier is 
|only a north american thing, and that in Europe, its done differently. 

Of course. 

Actually, last I heard the Europeans were still considering Release to
Pivot, which might be a plausible model for SIP.  Initiate to the default
code owner first, and allow for a "refuse with redirection" or something. 
It's got its weaknesses, which include putting a burden on the donor of the
ported number and relying on an uninvolved third party to set up a call, but
it does keep the supporting infrastructure simple. 

******* Release to Pivot will not work in the U.S. because the FCC has ruled
out any solution that relies on the donor network to complete the call. 
Also, the "donor" switch and the "pivot" switch at least must be upgraded to
support this solution.  If the switch does not store the new LRN/routing
prefix information for a ported-out number, a database dip must be performed
in order to provide the LRN/routing prefix in the ISUP Release message.  If
a "donor" switch stores the information, the provisioning/administration of
the information is not a simple process. **********

Another interesting working model is wireless roaming.  It is capable of
dynamically redirecting you anywhere (at least within a continent or so) and
it works at the level of individual telephone numbers.  Of course a single
wireless operating company owns all the numbers in an NPA NXX, which avoids
the major problem LNP was designed to create.  (No, that's not a typo, it's
editorial comment.) 

Probably it's more of an administrative problem than a technical one.  It's
so easy to fall into the trap of thinking there are technical solutions to
administrative problems, although technical solutions can certainly help
perpetuate administrative problems. 

******** FCC has granted a delay for the wireless carriers in the U.S. to
support local number portability (LNP) because the new entrants - PCS
carriers are too busy in deploying their networks.  LNP is intended to help
the new entrants to compete with the incumbent carriers.  So if the new
entrants ask for a delay, FCC certainly has no reason to push wireless LNP. 
Also there are issues on roaming where the roaming markets may not be
required to support LNP (LNP is to be supported in the top 100 major
Statistical Trading Areas in phases).  But I believe that the delay is
temporary because FCC's goal may be nation-wide number portability which
allows the subscriber to switch from wireline service provider to a wireless
service provider and move from one location to another location.  The LRN
method will allow that with minimum system upgrades.  The major issue about
location portability is the call charing/rating.  If flat rate is used for
all toll calls (e.g., 10 cents per minute regardless of the distance between
the calling party and called party), location portability can be supported
easily. 

            Wireles is a similar model if we view the "donor switch" as the
"home" system" and the "serving switch" as the "vistied system."  However,
there are differences.  The "home system" in the wireless model is the
service provider of the "roamer."  It authorizes the visited system to
provide the services to the roamer and guarantees the payment even if the
subscriber does not pay his/her bill.  Also, Mobile Application Part (MAP)
is used between the visited and home systems to register the "roamer" and to
get a temporary routing number to route the call to the visited system. 
Either the call must always be routed through the home system to reach the
visited system or a query to the home system must be performed first to
perform optimal routing to the visited system directly.  NPDB approach
actually allows the query to be done (for optimal routing) without the
involvement of the "donor" network.  *************

 John Mainwaring 
Nortel Networks 
crm312a@nortelnetworks.com 


------_=_NextPart_001_01BF4193.46F2C43A
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 =
5.5.2448.0">
<TITLE>RE: TRIP</TITLE>
</HEAD>
<BODY>

<P><FONT SIZE=3D2>Generally I agree with your remarks.</FONT>
</P>

<P><FONT SIZE=3D2>Specifically, I agree that the destination gateway is =
too late to do the query, unless you take an RTP-like approach.&nbsp; =
In effect, the default destination gateway is a proxy for the donor =
switch.&nbsp; By the time you get there, it's a bit late to do a =
query.&nbsp; On the other hand, you probably don't want to push the =
query all the way back to the originating gateway.&nbsp; That would =
require either that each gateway have knowledge of portability =
information for every NPA-NXX in existence or that an LNP or equivalent =
query be launched on every call.</FONT></P>

<P><FONT SIZE=3D2>I think you are suggesting that LNP queries should be =
done at intermediate points that are &quot;close enough&quot; to the =
destination to be expected to know about the portability status of the =
destination.&nbsp; If so, I agree with you.</FONT></P>

<P><FONT SIZE=3D2>I also agree with your suggestion that there may be a =
need for specialized NPDB nodes in the IP network.&nbsp; I'm not sure =
what form it should take, but the existing NPDBs are not only =
engineered for the PSTN, they are a cost center in the PSTN, and I =
understand that there are charges associated with using them.&nbsp; It =
may be appropriate to use them, or it may be better to explore other =
solutions.&nbsp; The point is that it is necessary to do more than wave =
our hands and say &quot;An NPDB query happens =
somewhere.&quot;</FONT></P>

<P><FONT SIZE=3D2>You are right that the FCC ruled out Release to Pivot =
or any solution that relies on the donor network to complete a call in =
the PSTN.&nbsp; I doubt if they were even aware of the possibility of =
VoIP at the time they made that ruling.&nbsp; There are significant =
ongoing costs associated with NPDBs.&nbsp; It seems appropriate to =
consider other architectures, especially since there is so little =
legacy encumbrance in SIP right now.&nbsp; It's probably true that an =
appropriately administered NPDB (loaded directly from the NPAC, as =
Brian suggested) is a more robust solution than anything that relies on =
the donor.&nbsp; My question is whether an RTP solution might be =
sufficiently reliable to be acceptable in practice and sufficiently =
cheap to be attractive.&nbsp; As you pointed out in connection with =
wireless number portability, the FCC doesn't always impose the same =
burdens in new areas as it does in the legacy environment.</FONT></P>
<BR>

<P><FONT SIZE=3D2>-----Original Message-----</FONT>
<BR><FONT SIZE=3D2>From: James Yu [<A =
HREF=3D"mailto:james.yu@neustar.com">mailto:james.yu@neustar.com</A>]</F=
ONT>
<BR><FONT SIZE=3D2>Sent: Wednesday, December 08, 1999 3:32 AM</FONT>
<BR><FONT SIZE=3D2>To: Mainwaring, John [NC2:3B30-I:EXCH]</FONT>
<BR><FONT SIZE=3D2>Cc: 'iptel@lists.research.bell-labs.com'</FONT>
<BR><FONT SIZE=3D2>Subject: RE: TRIP</FONT>
</P>
<BR>

<P><FONT SIZE=3D2>=A0</FONT>
</P>

<P><FONT SIZE=3D2>-----Original Message-----</FONT>
<BR><FONT SIZE=3D2>From: John Mainwaring [<A =
HREF=3D"mailto:crm312a@nortelnetworks.com">mailto:crm312a@nortelnetworks=
.com</A>]</FONT>
<BR><FONT SIZE=3D2>Sent: Tuesday, December 07, 1999 4:02 PM</FONT>
<BR><FONT SIZE=3D2>To: 'Jonathan Rosenberg'</FONT>
<BR><FONT SIZE=3D2>Cc: 'hsalama@cisco.com'; James Yu; =
'iptel@lists.research.bell-labs.com'</FONT>
<BR><FONT SIZE=3D2>Subject: RE: TRIP</FONT>
</P>
<BR>
<BR>

<P><FONT SIZE=3D2>|From: Jonathan Rosenberg [ <A =
HREF=3D"mailto:jdrosen@dynamicsoft.com">mailto:jdrosen@dynamicsoft.com</=
A></FONT>
<BR><FONT SIZE=3D2>&lt;<A =
HREF=3D"mailto:jdrosen@dynamicsoft.com">mailto:jdrosen@dynamicsoft.com</=
A>&gt; ] </FONT>
</P>

<P><FONT SIZE=3D2>|I'm less sure about deferring the LNP query to N-1. =
In the SIP </FONT>
<BR><FONT SIZE=3D2>|architecture, </FONT>
<BR><FONT SIZE=3D2>|at least, each proxy is responsible for doing a =
translation function </FONT>
<BR><FONT SIZE=3D2>|(called the </FONT>
<BR><FONT SIZE=3D2>|location service), which is similar in concept to =
an LNP </FONT>
<BR><FONT SIZE=3D2>|database query. </FONT>
</P>

<P><FONT SIZE=3D2>I understand that SIP is different from the tandem =
network of circuit</FONT>
<BR><FONT SIZE=3D2>switches.=A0 I think SIP will require the originator =
to know LNP information,</FONT>
<BR><FONT SIZE=3D2>but I'm not sure that existing collection of SCPs =
accessed by the TCAP/SS7</FONT>
<BR><FONT SIZE=3D2>signaling system is the place to get it.</FONT>
</P>

<P><FONT SIZE=3D2>=A0*********=A0 For SCN to SCN via IP or IP to SCN =
calls, the internet is the</FONT>
<BR><FONT SIZE=3D2>N-1 network (although many networks may be involved =
in the IP-domain</FONT>
<BR><FONT SIZE=3D2>routing), the entity that performs the LNP query via =
TCAP/SS7 over IP or</FONT>
<BR><FONT SIZE=3D2>real-time or non-real-time LNP-like mechanism via =
DNS) can be the</FONT>
<BR><FONT SIZE=3D2>originating GW/GK/SIP server or an intermediate =
GW/GK/SIP server but won't</FONT>
<BR><FONT SIZE=3D2>be the terminating GW/GK/SIP server because it may =
be too late to perform</FONT>
<BR><FONT SIZE=3D2>LNP query to find out that it does not reach the =
serving switch identified</FONT>
<BR><FONT SIZE=3D2>by the LRN.=A0 There are certainly many issues if =
the originating GW/GK/SIP is</FONT>
<BR><FONT SIZE=3D2>to perform the LNP query.=A0 For example, several =
protocols can be used</FONT>
<BR><FONT SIZE=3D2>between a switch and a NPDB.=A0 In Europe, I heard =
that there are</FONT>
<BR><FONT SIZE=3D2>country-specific variations to the switch-NPDB =
interface protocol.=A0 For the</FONT>
<BR><FONT SIZE=3D2>originating GW/GK/SIP server to query all the NPDBs =
worldwide would be a big</FONT>
<BR><FONT SIZE=3D2>burden.=A0 </FONT>
</P>

<P><FONT SIZE=3D2>One possible solution is that a standardized query =
can be launched by the</FONT>
<BR><FONT SIZE=3D2>entity which will be mapped by an interworking like =
entity to</FONT>
<BR><FONT SIZE=3D2>country-specific NPDB query to reach the existing =
NPDBs.=A0 Another potential</FONT>
<BR><FONT SIZE=3D2>solution is to have NPDBs in the internet that =
support the standardized IP</FONT>
<BR><FONT SIZE=3D2>entity-to-NPDB interface.=A0 Certainly this does not =
rule out that the</FONT>
<BR><FONT SIZE=3D2>existing NPDB from being enhanced to support this =
new IP-domain interface.</FONT>
<BR><FONT SIZE=3D2>*************</FONT>
</P>

<P><FONT SIZE=3D2>=A0|I'll also </FONT>
<BR><FONT SIZE=3D2>|note that I have been told that the LNP dip being =
done at the N-1 </FONT>
<BR><FONT SIZE=3D2>|carrier is </FONT>
<BR><FONT SIZE=3D2>|only a north american thing, and that in Europe, =
its done differently. </FONT>
</P>

<P><FONT SIZE=3D2>Of course. </FONT>
</P>

<P><FONT SIZE=3D2>Actually, last I heard the Europeans were still =
considering Release to</FONT>
<BR><FONT SIZE=3D2>Pivot, which might be a plausible model for SIP.=A0 =
Initiate to the default</FONT>
<BR><FONT SIZE=3D2>code owner first, and allow for a &quot;refuse with =
redirection&quot; or something.=A0</FONT>
<BR><FONT SIZE=3D2>It's got its weaknesses, which include putting a =
burden on the donor of the</FONT>
<BR><FONT SIZE=3D2>ported number and relying on an uninvolved third =
party to set up a call, but</FONT>
<BR><FONT SIZE=3D2>it does keep the supporting infrastructure =
simple.=A0</FONT>
</P>

<P><FONT SIZE=3D2>******* Release to Pivot will not work in the U.S. =
because the FCC has ruled</FONT>
<BR><FONT SIZE=3D2>out any solution that relies on the donor network to =
complete the call.=A0</FONT>
<BR><FONT SIZE=3D2>Also, the &quot;donor&quot; switch and the =
&quot;pivot&quot; switch at least must be upgraded to</FONT>
<BR><FONT SIZE=3D2>support this solution.=A0 If the switch does not =
store the new LRN/routing</FONT>
<BR><FONT SIZE=3D2>prefix information for a ported-out number, a =
database dip must be performed</FONT>
<BR><FONT SIZE=3D2>in order to provide the LRN/routing prefix in the =
ISUP Release message.=A0 If</FONT>
<BR><FONT SIZE=3D2>a &quot;donor&quot; switch stores the information, =
the provisioning/administration of</FONT>
<BR><FONT SIZE=3D2>the information is not a simple process. =
**********</FONT>
</P>

<P><FONT SIZE=3D2>Another interesting working model is wireless =
roaming.=A0 It is capable of</FONT>
<BR><FONT SIZE=3D2>dynamically redirecting you anywhere (at least =
within a continent or so) and</FONT>
<BR><FONT SIZE=3D2>it works at the level of individual telephone =
numbers.=A0 Of course a single</FONT>
<BR><FONT SIZE=3D2>wireless operating company owns all the numbers in =
an NPA NXX, which avoids</FONT>
<BR><FONT SIZE=3D2>the major problem LNP was designed to create.=A0 =
(No, that's not a typo, it's</FONT>
<BR><FONT SIZE=3D2>editorial comment.)=A0</FONT>
</P>

<P><FONT SIZE=3D2>Probably it's more of an administrative problem than =
a technical one.=A0 It's</FONT>
<BR><FONT SIZE=3D2>so easy to fall into the trap of thinking there are =
technical solutions to</FONT>
<BR><FONT SIZE=3D2>administrative problems, although technical =
solutions can certainly help</FONT>
<BR><FONT SIZE=3D2>perpetuate administrative problems.=A0</FONT>
</P>

<P><FONT SIZE=3D2>******** FCC has granted a delay for the wireless =
carriers in the U.S. to</FONT>
<BR><FONT SIZE=3D2>support local number portability (LNP) because the =
new entrants - PCS</FONT>
<BR><FONT SIZE=3D2>carriers are too busy in deploying their =
networks.=A0 LNP is intended to help</FONT>
<BR><FONT SIZE=3D2>the new entrants to compete with the incumbent =
carriers.=A0 So if the new</FONT>
<BR><FONT SIZE=3D2>entrants ask for a delay, FCC certainly has no =
reason to push wireless LNP.=A0</FONT>
<BR><FONT SIZE=3D2>Also there are issues on roaming where the roaming =
markets may not be</FONT>
<BR><FONT SIZE=3D2>required to support LNP (LNP is to be supported in =
the top 100 major</FONT>
<BR><FONT SIZE=3D2>Statistical Trading Areas in phases).=A0 But I =
believe that the delay is</FONT>
<BR><FONT SIZE=3D2>temporary because FCC's goal may be nation-wide =
number portability which</FONT>
<BR><FONT SIZE=3D2>allows the subscriber to switch from wireline =
service provider to a wireless</FONT>
<BR><FONT SIZE=3D2>service provider and move from one location to =
another location.=A0 The LRN</FONT>
<BR><FONT SIZE=3D2>method will allow that with minimum system =
upgrades.=A0 The major issue about</FONT>
<BR><FONT SIZE=3D2>location portability is the call charing/rating.=A0 =
If flat rate is used for</FONT>
<BR><FONT SIZE=3D2>all toll calls (e.g., 10 cents per minute regardless =
of the distance between</FONT>
<BR><FONT SIZE=3D2>the calling party and called party), location =
portability can be supported</FONT>
<BR><FONT SIZE=3D2>easily. </FONT>
</P>

<P><FONT SIZE=3D2>=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 Wireles is a =
similar model if we view the &quot;donor switch&quot; as the</FONT>
<BR><FONT SIZE=3D2>&quot;home&quot; system&quot; and the &quot;serving =
switch&quot; as the &quot;vistied system.&quot;=A0 However,</FONT>
<BR><FONT SIZE=3D2>there are differences.=A0 The &quot;home =
system&quot; in the wireless model is the</FONT>
<BR><FONT SIZE=3D2>service provider of the &quot;roamer.&quot;=A0 It =
authorizes the visited system to</FONT>
<BR><FONT SIZE=3D2>provide the services to the roamer and guarantees =
the payment even if the</FONT>
<BR><FONT SIZE=3D2>subscriber does not pay his/her bill.=A0 Also, =
Mobile Application Part (MAP)</FONT>
<BR><FONT SIZE=3D2>is used between the visited and home systems to =
register the &quot;roamer&quot; and to</FONT>
<BR><FONT SIZE=3D2>get a temporary routing number to route the call to =
the visited system.=A0</FONT>
<BR><FONT SIZE=3D2>Either the call must always be routed through the =
home system to reach the</FONT>
<BR><FONT SIZE=3D2>visited system or a query to the home system must be =
performed first to</FONT>
<BR><FONT SIZE=3D2>perform optimal routing to the visited system =
directly.=A0 NPDB approach</FONT>
<BR><FONT SIZE=3D2>actually allows the query to be done (for optimal =
routing) without the</FONT>
<BR><FONT SIZE=3D2>involvement of the &quot;donor&quot; network.=A0 =
*************</FONT>
</P>

<P><FONT SIZE=3D2>=A0John Mainwaring </FONT>
<BR><FONT SIZE=3D2>Nortel Networks </FONT>
<BR><FONT SIZE=3D2>crm312a@nortelnetworks.com </FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01BF4193.46F2C43A--

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec  8 15:22:56 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 PAA23188
	for <iptel-archive@odin.ietf.org>; Wed, 8 Dec 1999 15:22:55 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 003FD52D5; Wed,  8 Dec 1999 15:17:31 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 4F79152DC; Wed,  8 Dec 1999 15:17:30 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Mark D. Foster" <mark.foster@npac.com>
To: "'John Mainwaring'" <crm312a@nortelnetworks.com>
Cc: <iptel@lists.research.bell-labs.com>, "'James Yu'" <james.yu@neustar.com>
Subject: RE: TRIP
Date: Wed, 8 Dec 1999 15:14:44 -0500
Message-Id: <000901bf41b8$e9e21900$0864a8c0@loudoun.com>
Mime-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_000A_01BF418F.010C1100"
X-Priority: 3 (Normal)
X-Msmail-Priority: Normal
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <89866A7EC9C0D1119C950000F8BCC2820157C012@zrtpd00n.us.nortel.com>
X-Mimeole: Produced By Microsoft MimeOLE V5.00.2314.1300
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

This is a multi-part message in MIME format.

------=_NextPart_000_000A_01BF418F.010C1100
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit

RE: TRIPWhile the FCC did not specify in 95-116 any particular signaling or
implementation standards for LNP, they did establish a number principles
with which any implementation must comply, whether wireline or wireless.  As
examples, they did say that LRN met their requirements, and that RTP and QoR
did not.

A core requirement was that a carrier who is serving ported numbers need not
be reliant on any other carrier (especially the donor network) for
completing calls, whether for call transport/routing or for signaling.
That's not to say that a carrier couldn't voluntarily opt to use another
carrier or the donor for queries or call routing, but the key here is
voluntarily.  Again, this requirement was imposed on all LNP implementations
in the US regardless of the network technology employed.

Similar requirements were adopted by the CRTC for Canada, and as I
understand it, by regulatory and/or industry bodies in some other countries
with IN-based NP (e.g. Belgium, Spain).

As various edicts from the FCC stand now, the only form of IP telephony
exempt from these rules is where the called party uses a general purpose
computer (e.g. PC) as the telephony terminal, self-installed with telephony
terminal software (e.g. net2phone).  This use is considered an "enhanced
service" by the FCC, and therefore exempt.  All other forms of IP telephony
(e.g. cable set-up box, DSL modem, IP-based wireless terminal, IP trunk to
PBX) apparently must comply with existing rules for common carrier services.

That at least some forms of Iptel must comply with the non-reliance
requirement in at least some countries suggests that the SIP, TRIP, enum,
etc., must be flexible enough to allow implementations that will satisfy
these requirements, and still be interoperable with those that don't.

In the US and Canada, the key to this was the N-1 trigger placement policy.
This was simply a guideline on which networks should be prepared to dip
calls placed to which areas in order to not force calls to be re-routed at
the donor network (termination routing).  Key to having flexibility on where
triggers are placed, resulting in NPDB queries, is having a form of
post-query signaling that makes the query results available to all
downstream networks.

The key data elements in ANSI ISUP are: the LNP-query-done indicator (bit in
the FCI); placing a routing number returned from the NPDB in the CdPN (LRN
or whatever), and moving the original dialed number to a GAP.  Subsequent
call routing is done on the routing number, while not loosing the dialed
portable subscriber number.  Given the need for SIP to interoperate with
ANSI ISUP, there should be some consideration of implementing these
semantics.

If SIP captured these fields, so as to provide a post-dip form of downstream
signaling, and TRIP were designed to build route tables based on
administratively assigned number ranges (which, by definition would make
calls to routing numbers [LRN] routable) , and enum did not require DNS
queries to be answered by donor networks, then the existing regulatory
requirements for LNP could be met.

R, Mark

      Mark D. Foster
      CTO
      TEL: +1 202 533 2800, or 2702
      1120 Vermont Ave NW, Suite 550
     EMAIL: mark.foster@neustar.com
      NeuStar, Inc.
      FAX: +1 202 533 2975
      Washington, DC 20005




  -----Original Message-----
  From: owner-iptel@lists.research.bell-labs.com
[mailto:owner-iptel@lists.research.bell-labs.com]On Behalf Of John
Mainwaring
  Sent: Wednesday, December 08, 1999 10:45 AM
  To: 'James Yu'
  Cc: 'iptel@lists.research.bell-labs.com'
  Subject: RE: TRIP


  Generally I agree with your remarks.

  Specifically, I agree that the destination gateway is too late to do the
query, unless you take an RTP-like approach.  In effect, the default
destination gateway is a proxy for the donor switch.  By the time you get
there, it's a bit late to do a query.  On the other hand, you probably don't
want to push the query all the way back to the originating gateway.  That
would require either that each gateway have knowledge of portability
information for every NPA-NXX in existence or that an LNP or equivalent
query be launched on every call.

  I think you are suggesting that LNP queries should be done at intermediate
points that are "close enough" to the destination to be expected to know
about the portability status of the destination.  If so, I agree with you.

  I also agree with your suggestion that there may be a need for specialized
NPDB nodes in the IP network.  I'm not sure what form it should take, but
the existing NPDBs are not only engineered for the PSTN, they are a cost
center in the PSTN, and I understand that there are charges associated with
using them.  It may be appropriate to use them, or it may be better to
explore other solutions.  The point is that it is necessary to do more than
wave our hands and say "An NPDB query happens somewhere."

  You are right that the FCC ruled out Release to Pivot or any solution that
relies on the donor network to complete a call in the PSTN.  I doubt if they
were even aware of the possibility of VoIP at the time they made that
ruling.  There are significant ongoing costs associated with NPDBs.  It
seems appropriate to consider other architectures, especially since there is
so little legacy encumbrance in SIP right now.  It's probably true that an
appropriately administered NPDB (loaded directly from the NPAC, as Brian
suggested) is a more robust solution than anything that relies on the donor.
My question is whether an RTP solution might be sufficiently reliable to be
acceptable in practice and sufficiently cheap to be attractive.  As you
pointed out in connection with wireless number portability, the FCC doesn't
always impose the same burdens in new areas as it does in the legacy
environment.



  -----Original Message-----
  From: James Yu [mailto:james.yu@neustar.com]
  Sent: Wednesday, December 08, 1999 3:32 AM
  To: Mainwaring, John [NC2:3B30-I:EXCH]
  Cc: 'iptel@lists.research.bell-labs.com'
  Subject: RE: TRIP





  -----Original Message-----
  From: John Mainwaring [mailto:crm312a@nortelnetworks.com]
  Sent: Tuesday, December 07, 1999 4:02 PM
  To: 'Jonathan Rosenberg'
  Cc: 'hsalama@cisco.com'; James Yu; 'iptel@lists.research.bell-labs.com'
  Subject: RE: TRIP




  |From: Jonathan Rosenberg [ mailto:jdrosen@dynamicsoft.com
  <mailto:jdrosen@dynamicsoft.com> ]

  |I'm less sure about deferring the LNP query to N-1. In the SIP
  |architecture,
  |at least, each proxy is responsible for doing a translation function
  |(called the
  |location service), which is similar in concept to an LNP
  |database query.

  I understand that SIP is different from the tandem network of circuit
  switches.  I think SIP will require the originator to know LNP
information,
  but I'm not sure that existing collection of SCPs accessed by the TCAP/SS7
  signaling system is the place to get it.

   *********  For SCN to SCN via IP or IP to SCN calls, the internet is the
  N-1 network (although many networks may be involved in the IP-domain
  routing), the entity that performs the LNP query via TCAP/SS7 over IP or
  real-time or non-real-time LNP-like mechanism via DNS) can be the
  originating GW/GK/SIP server or an intermediate GW/GK/SIP server but won't
  be the terminating GW/GK/SIP server because it may be too late to perform
  LNP query to find out that it does not reach the serving switch identified
  by the LRN.  There are certainly many issues if the originating GW/GK/SIP
is
  to perform the LNP query.  For example, several protocols can be used
  between a switch and a NPDB.  In Europe, I heard that there are
  country-specific variations to the switch-NPDB interface protocol.  For
the
  originating GW/GK/SIP server to query all the NPDBs worldwide would be a
big
  burden.

  One possible solution is that a standardized query can be launched by the
  entity which will be mapped by an interworking like entity to
  country-specific NPDB query to reach the existing NPDBs.  Another
potential
  solution is to have NPDBs in the internet that support the standardized IP
  entity-to-NPDB interface.  Certainly this does not rule out that the
  existing NPDB from being enhanced to support this new IP-domain interface.
  *************

   |I'll also
  |note that I have been told that the LNP dip being done at the N-1
  |carrier is
  |only a north american thing, and that in Europe, its done differently.

  Of course.

  Actually, last I heard the Europeans were still considering Release to
  Pivot, which might be a plausible model for SIP.  Initiate to the default
  code owner first, and allow for a "refuse with redirection" or something.
  It's got its weaknesses, which include putting a burden on the donor of
the
  ported number and relying on an uninvolved third party to set up a call,
but
  it does keep the supporting infrastructure simple.

  ******* Release to Pivot will not work in the U.S. because the FCC has
ruled
  out any solution that relies on the donor network to complete the call.
  Also, the "donor" switch and the "pivot" switch at least must be upgraded
to
  support this solution.  If the switch does not store the new LRN/routing
  prefix information for a ported-out number, a database dip must be
performed
  in order to provide the LRN/routing prefix in the ISUP Release message.
If
  a "donor" switch stores the information, the provisioning/administration
of
  the information is not a simple process. **********

  Another interesting working model is wireless roaming.  It is capable of
  dynamically redirecting you anywhere (at least within a continent or so)
and
  it works at the level of individual telephone numbers.  Of course a single
  wireless operating company owns all the numbers in an NPA NXX, which
avoids
  the major problem LNP was designed to create.  (No, that's not a typo,
it's
  editorial comment.)

  Probably it's more of an administrative problem than a technical one.
It's
  so easy to fall into the trap of thinking there are technical solutions to
  administrative problems, although technical solutions can certainly help
  perpetuate administrative problems.

  ******** FCC has granted a delay for the wireless carriers in the U.S. to
  support local number portability (LNP) because the new entrants - PCS
  carriers are too busy in deploying their networks.  LNP is intended to
help
  the new entrants to compete with the incumbent carriers.  So if the new
  entrants ask for a delay, FCC certainly has no reason to push wireless
LNP.
  Also there are issues on roaming where the roaming markets may not be
  required to support LNP (LNP is to be supported in the top 100 major
  Statistical Trading Areas in phases).  But I believe that the delay is
  temporary because FCC's goal may be nation-wide number portability which
  allows the subscriber to switch from wireline service provider to a
wireless
  service provider and move from one location to another location.  The LRN
  method will allow that with minimum system upgrades.  The major issue
about
  location portability is the call charing/rating.  If flat rate is used for
  all toll calls (e.g., 10 cents per minute regardless of the distance
between
  the calling party and called party), location portability can be supported
  easily.

              Wireles is a similar model if we view the "donor switch" as
the
  "home" system" and the "serving switch" as the "vistied system."  However,
  there are differences.  The "home system" in the wireless model is the
  service provider of the "roamer."  It authorizes the visited system to
  provide the services to the roamer and guarantees the payment even if the
  subscriber does not pay his/her bill.  Also, Mobile Application Part (MAP)
  is used between the visited and home systems to register the "roamer" and
to
  get a temporary routing number to route the call to the visited system.
  Either the call must always be routed through the home system to reach the
  visited system or a query to the home system must be performed first to
  perform optimal routing to the visited system directly.  NPDB approach
  actually allows the query to be done (for optimal routing) without the
  involvement of the "donor" network.  *************

   John Mainwaring
  Nortel Networks
  crm312a@nortelnetworks.com


------=_NextPart_000_000A_01BF418F.010C1100
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Diso-8859-1">
<TITLE>RE: TRIP</TITLE>

<META content=3D"MSHTML 5.00.2722.2800" name=3DGENERATOR></HEAD>
<BODY>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN =
class=3D008484118-08121999>While=20
the FCC did not specify in 95-116 any particular signaling or =
implementation=20
standards for LNP, they did establish a number principles&nbsp;with =
which any=20
implementation must comply, whether wireline or wireless.&nbsp;&nbsp;As=20
examples, they did say that LRN met their requirements, and that RTP and =
QoR did=20
not.</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN =
class=3D008484118-08121999>A core=20
requirement was that a carrier who is serving ported numbers need not be =
reliant=20
on any other carrier (especially the donor network) for completing =
calls,=20
whether for call transport/routing or for signaling.&nbsp;&nbsp;That's =
not to=20
say that a carrier couldn't voluntarily opt to use another carrier or =
the donor=20
for queries or call routing, but the key here is voluntarily.&nbsp; =
Again, this=20
requirement was imposed on all LNP implementations in the US regardless =
of the=20
network technology employed.</SPAN></FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999>Similar requirements were adopted by the CRTC =
for=20
Canada, and as I understand it,&nbsp;by regulatory and/or industry =
bodies in=20
some other countries with IN-based&nbsp;NP (e.g. Belgium,=20
Spain).</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999>As&nbsp;various edicts from the FCC stand =
now, the only=20
form of IP telephony exempt from these rules is where the called party =
uses a=20
general purpose computer (e.g. PC) as the telephony terminal, =
self-installed=20
with telephony terminal software (e.g. net2phone).&nbsp; This use is =
considered=20
an "enhanced service" by the FCC, and therefore exempt.&nbsp; All other =
forms of=20
IP telephony (e.g. cable set-up box, DSL modem, IP-based wireless =
terminal, IP=20
trunk to PBX) apparently must comply with existing rules for common =
carrier=20
services.</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN =
class=3D008484118-08121999>That=20
at least some forms of Iptel must comply with the non-reliance =
requirement in at=20
least some countries suggests that the SIP, TRIP, enum, etc., must be =
flexible=20
enough to allow&nbsp;implementations that will satisfy these =
requirements, and=20
still be interoperable with those that don't.</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN =
class=3D008484118-08121999>In the=20
US and Canada, the key to this was the N-1 trigger placement =
policy.&nbsp; This=20
was simply a guideline on which networks should be prepared to dip calls =
placed=20
to which areas in order to not force calls to be re-routed at the donor =
network=20
(termination routing).&nbsp; Key to having flexibility on where triggers =
are=20
placed, resulting in NPDB queries, is having a form of post-query =
signaling that=20
makes the query results&nbsp;available to all downstream=20
networks.</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999>The&nbsp;key data elements in ANSI ISUP are: =
the=20
LNP-query-done indicator (bit in the FCI); placing a routing number =
returned=20
from the NPDB in the CdPN (LRN or whatever), and moving the original =
dialed=20
number to a GAP.&nbsp; Subsequent call routing is done on the routing =
number,=20
while not loosing the dialed portable subscriber number.&nbsp; Given the =
need=20
for SIP to interoperate with ANSI ISUP, there should be some =
consideration of=20
implementing these semantics.</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN =
class=3D008484118-08121999>If SIP=20
captured these fields, so as to provide a post-dip form of downstream =
signaling,=20
and TRIP were designed to build route tables based on administratively =
assigned=20
number ranges (which, by definition would make calls to routing numbers =
[LRN]=20
routable) , and enum did not require DNS queries to be answered by donor =

networks, then the existing regulatory requirements for LNP could be=20
met.</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN =
class=3D008484118-08121999>R,=20
Mark</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D008484118-08121999></SPAN></FONT>&nbsp;</DIV>
<DIV class=3DSection1>
<TABLE bgColor=3D#cccccc border=3D1 cellPadding=3D0 cellSpacing=3D0=20
style=3D"BACKGROUND: #cccccc; BORDER-BOTTOM: windowtext 0.5pt solid; =
BORDER-LEFT: windowtext 0.5pt solid; BORDER-RIGHT: windowtext 0.5pt =
solid; BORDER-TOP: windowtext 0.5pt solid; mso-cellspacing: 0in; =
mso-table-layout-alt: fixed; mso-padding-alt: 5.25pt 5.25pt 5.25pt =
5.25pt">
  <TBODY>
  <TR style=3D"HEIGHT: 17.4pt">
    <TD=20
    style=3D"BORDER-BOTTOM: windowtext 0.5pt solid; BORDER-LEFT: =
windowtext 0.5pt solid; BORDER-RIGHT: windowtext 0.5pt solid; =
BORDER-TOP: windowtext 0.5pt solid; HEIGHT: 17.4pt; PADDING-BOTTOM: =
5.25pt; PADDING-LEFT: 5.25pt; PADDING-RIGHT: 5.25pt; PADDING-TOP: =
5.25pt; WIDTH: 165.95pt"=20
    vAlign=3Dtop width=3D221>
      <P><!doctype HTML PUBLIC><B><SPAN=20
      style=3D"FONT-SIZE: 10pt; mso-bidi-font-size: 12.0pt">Mark D.=20
      Foster</SPAN></B><SPAN=20
      style=3D"FONT-SIZE: 10pt; mso-bidi-font-size: =
12.0pt"><BR>CTO<BR>TEL: +1 202=20
      533 2800, or 2702<BR>1120 Vermont Ave NW, Suite 550<?xml:namespace =
prefix=20
      =3D o ns =3D "urn:schemas-microsoft-com:office:office"=20
    /><o:p></o:p></SPAN></P></TD>
    <TD=20
    style=3D"BORDER-BOTTOM: windowtext 0.5pt solid; BORDER-LEFT: =
windowtext 0.5pt solid; BORDER-RIGHT: windowtext 0.5pt solid; =
BORDER-TOP: windowtext 0.5pt solid; HEIGHT: 17.4pt; PADDING-BOTTOM: =
5.25pt; PADDING-LEFT: 5.25pt; PADDING-RIGHT: 5.25pt; PADDING-TOP: =
5.25pt; WIDTH: 194.8pt"=20
    vAlign=3Dtop width=3D260>
      <P><SPAN style=3D"FONT-SIZE: 10pt; mso-bidi-font-size: =
12.0pt">EMAIL: <I><A=20
      =
href=3D"mailto:mark.foster@npac.com">mark.foster@neustar.com</A></I><BR>N=
euStar,=20
      Inc.<BR>FAX: +1 202 533 2975<BR>Washington, DC=20
    20005<o:p></o:p></SPAN></P></TD></TR></TBODY></TABLE>
<P class=3DMsoNormal>&nbsp;<o:p></o:p></P></DIV>
<BLOCKQUOTE=20
style=3D"BORDER-LEFT: #0000ff 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: =
0px; PADDING-LEFT: 5px">
  <DIV align=3Dleft class=3DOutlookMessageHeader dir=3Dltr><FONT =
face=3DTahoma=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>John=20
  Mainwaring<BR><B>Sent:</B> Wednesday, December 08, 1999 10:45 =
AM<BR><B>To:</B>=20
  'James Yu'<BR><B>Cc:</B>=20
  'iptel@lists.research.bell-labs.com'<BR><B>Subject:</B> RE:=20
  TRIP<BR><BR></DIV></FONT>
  <P><FONT size=3D2>Generally I agree with your remarks.</FONT> </P>
  <P><FONT size=3D2>Specifically, I agree that the destination gateway =
is too late=20
  to do the query, unless you take an RTP-like approach.&nbsp; In =
effect, the=20
  default destination gateway is a proxy for the donor switch.&nbsp; By =
the time=20
  you get there, it's a bit late to do a query.&nbsp; On the other hand, =
you=20
  probably don't want to push the query all the way back to the =
originating=20
  gateway.&nbsp; That would require either that each gateway have =
knowledge of=20
  portability information for every NPA-NXX in existence or that an LNP =
or=20
  equivalent query be launched on every call.</FONT></P>
  <P><FONT size=3D2>I think you are suggesting that LNP queries should =
be done at=20
  intermediate points that are "close enough" to the destination to be =
expected=20
  to know about the portability status of the destination.&nbsp; If so, =
I agree=20
  with you.</FONT></P>
  <P><FONT size=3D2>I also agree with your suggestion that there may be =
a need for=20
  specialized NPDB nodes in the IP network.&nbsp; I'm not sure what form =
it=20
  should take, but the existing NPDBs are not only engineered for the =
PSTN, they=20
  are a cost center in the PSTN, and I understand that there are charges =

  associated with using them.&nbsp; It may be appropriate to use them, =
or it may=20
  be better to explore other solutions.&nbsp; The point is that it is =
necessary=20
  to do more than wave our hands and say "An NPDB query happens=20
  somewhere."</FONT></P>
  <P><FONT size=3D2>You are right that the FCC ruled out Release to =
Pivot or any=20
  solution that relies on the donor network to complete a call in the=20
  PSTN.&nbsp; I doubt if they were even aware of the possibility of VoIP =
at the=20
  time they made that ruling.&nbsp; There are significant ongoing costs=20
  associated with NPDBs.&nbsp; It seems appropriate to consider other=20
  architectures, especially since there is so little legacy encumbrance =
in SIP=20
  right now.&nbsp; It's probably true that an appropriately administered =
NPDB=20
  (loaded directly from the NPAC, as Brian suggested) is a more robust =
solution=20
  than anything that relies on the donor.&nbsp; My question is whether =
an RTP=20
  solution might be sufficiently reliable to be acceptable in practice =
and=20
  sufficiently cheap to be attractive.&nbsp; As you pointed out in =
connection=20
  with wireless number portability, the FCC doesn't always impose the =
same=20
  burdens in new areas as it does in the legacy =
environment.</FONT></P><BR>
  <P><FONT size=3D2>-----Original Message-----</FONT> <BR><FONT =
size=3D2>From: James=20
  Yu [<A=20
  =
href=3D"mailto:james.yu@neustar.com">mailto:james.yu@neustar.com</A>]</FO=
NT>=20
  <BR><FONT size=3D2>Sent: Wednesday, December 08, 1999 3:32 AM</FONT> =
<BR><FONT=20
  size=3D2>To: Mainwaring, John [NC2:3B30-I:EXCH]</FONT> <BR><FONT =
size=3D2>Cc:=20
  'iptel@lists.research.bell-labs.com'</FONT> <BR><FONT =
size=3D2>Subject: RE:=20
  TRIP</FONT> </P><BR>
  <P><FONT size=3D2>&nbsp;</FONT> </P>
  <P><FONT size=3D2>-----Original Message-----</FONT> <BR><FONT =
size=3D2>From: John=20
  Mainwaring [<A=20
  =
href=3D"mailto:crm312a@nortelnetworks.com">mailto:crm312a@nortelnetworks.=
com</A>]</FONT>=20
  <BR><FONT size=3D2>Sent: Tuesday, December 07, 1999 4:02 PM</FONT> =
<BR><FONT=20
  size=3D2>To: 'Jonathan Rosenberg'</FONT> <BR><FONT size=3D2>Cc:=20
  'hsalama@cisco.com'; James Yu; =
'iptel@lists.research.bell-labs.com'</FONT>=20
  <BR><FONT size=3D2>Subject: RE: TRIP</FONT> </P><BR><BR>
  <P><FONT size=3D2>|From: Jonathan Rosenberg [ <A=20
  =
href=3D"mailto:jdrosen@dynamicsoft.com">mailto:jdrosen@dynamicsoft.com</A=
></FONT>=20
  <BR><FONT size=3D2>&lt;<A=20
  =
href=3D"mailto:jdrosen@dynamicsoft.com">mailto:jdrosen@dynamicsoft.com</A=
>&gt; ]=20
  </FONT></P>
  <P><FONT size=3D2>|I'm less sure about deferring the LNP query to N-1. =
In the=20
  SIP </FONT><BR><FONT size=3D2>|architecture, </FONT><BR><FONT =
size=3D2>|at least,=20
  each proxy is responsible for doing a translation function =
</FONT><BR><FONT=20
  size=3D2>|(called the </FONT><BR><FONT size=3D2>|location service), =
which is=20
  similar in concept to an LNP </FONT><BR><FONT size=3D2>|database =
query.=20
  </FONT></P>
  <P><FONT size=3D2>I understand that SIP is different from the tandem =
network of=20
  circuit</FONT> <BR><FONT size=3D2>switches.&nbsp; I think SIP will =
require the=20
  originator to know LNP information,</FONT> <BR><FONT size=3D2>but I'm =
not sure=20
  that existing collection of SCPs accessed by the TCAP/SS7</FONT> =
<BR><FONT=20
  size=3D2>signaling system is the place to get it.</FONT> </P>
  <P><FONT size=3D2>&nbsp;*********&nbsp; For SCN to SCN via IP or IP to =
SCN=20
  calls, the internet is the</FONT> <BR><FONT size=3D2>N-1 network =
(although many=20
  networks may be involved in the IP-domain</FONT> <BR><FONT =
size=3D2>routing),=20
  the entity that performs the LNP query via TCAP/SS7 over IP or</FONT>=20
  <BR><FONT size=3D2>real-time or non-real-time LNP-like mechanism via =
DNS) can be=20
  the</FONT> <BR><FONT size=3D2>originating GW/GK/SIP server or an =
intermediate=20
  GW/GK/SIP server but won't</FONT> <BR><FONT size=3D2>be the =
terminating=20
  GW/GK/SIP server because it may be too late to perform</FONT> =
<BR><FONT=20
  size=3D2>LNP query to find out that it does not reach the serving =
switch=20
  identified</FONT> <BR><FONT size=3D2>by the LRN.&nbsp; There are =
certainly many=20
  issues if the originating GW/GK/SIP is</FONT> <BR><FONT size=3D2>to =
perform the=20
  LNP query.&nbsp; For example, several protocols can be used</FONT> =
<BR><FONT=20
  size=3D2>between a switch and a NPDB.&nbsp; In Europe, I heard that =
there=20
  are</FONT> <BR><FONT size=3D2>country-specific variations to the =
switch-NPDB=20
  interface protocol.&nbsp; For the</FONT> <BR><FONT =
size=3D2>originating=20
  GW/GK/SIP server to query all the NPDBs worldwide would be a =
big</FONT>=20
  <BR><FONT size=3D2>burden.&nbsp; </FONT></P>
  <P><FONT size=3D2>One possible solution is that a standardized query =
can be=20
  launched by the</FONT> <BR><FONT size=3D2>entity which will be mapped =
by an=20
  interworking like entity to</FONT> <BR><FONT size=3D2>country-specific =
NPDB=20
  query to reach the existing NPDBs.&nbsp; Another potential</FONT> =
<BR><FONT=20
  size=3D2>solution is to have NPDBs in the internet that support the =
standardized=20
  IP</FONT> <BR><FONT size=3D2>entity-to-NPDB interface.&nbsp; Certainly =
this does=20
  not rule out that the</FONT> <BR><FONT size=3D2>existing NPDB from =
being=20
  enhanced to support this new IP-domain interface.</FONT> <BR><FONT=20
  size=3D2>*************</FONT> </P>
  <P><FONT size=3D2>&nbsp;|I'll also </FONT><BR><FONT size=3D2>|note =
that I have=20
  been told that the LNP dip being done at the N-1 </FONT><BR><FONT=20
  size=3D2>|carrier is </FONT><BR><FONT size=3D2>|only a north american =
thing, and=20
  that in Europe, its done differently. </FONT></P>
  <P><FONT size=3D2>Of course. </FONT></P>
  <P><FONT size=3D2>Actually, last I heard the Europeans were still =
considering=20
  Release to</FONT> <BR><FONT size=3D2>Pivot, which might be a plausible =
model for=20
  SIP.&nbsp; Initiate to the default</FONT> <BR><FONT size=3D2>code =
owner first,=20
  and allow for a "refuse with redirection" or something.&nbsp;</FONT> =
<BR><FONT=20
  size=3D2>It's got its weaknesses, which include putting a burden on =
the donor of=20
  the</FONT> <BR><FONT size=3D2>ported number and relying on an =
uninvolved third=20
  party to set up a call, but</FONT> <BR><FONT size=3D2>it does keep the =

  supporting infrastructure simple.&nbsp;</FONT> </P>
  <P><FONT size=3D2>******* Release to Pivot will not work in the U.S. =
because the=20
  FCC has ruled</FONT> <BR><FONT size=3D2>out any solution that relies =
on the=20
  donor network to complete the call.&nbsp;</FONT> <BR><FONT =
size=3D2>Also, the=20
  "donor" switch and the "pivot" switch at least must be upgraded =
to</FONT>=20
  <BR><FONT size=3D2>support this solution.&nbsp; If the switch does not =
store the=20
  new LRN/routing</FONT> <BR><FONT size=3D2>prefix information for a =
ported-out=20
  number, a database dip must be performed</FONT> <BR><FONT size=3D2>in =
order to=20
  provide the LRN/routing prefix in the ISUP Release message.&nbsp; =
If</FONT>=20
  <BR><FONT size=3D2>a "donor" switch stores the information, the=20
  provisioning/administration of</FONT> <BR><FONT size=3D2>the =
information is not=20
  a simple process. **********</FONT> </P>
  <P><FONT size=3D2>Another interesting working model is wireless =
roaming.&nbsp;=20
  It is capable of</FONT> <BR><FONT size=3D2>dynamically redirecting you =
anywhere=20
  (at least within a continent or so) and</FONT> <BR><FONT size=3D2>it =
works at=20
  the level of individual telephone numbers.&nbsp; Of course a =
single</FONT>=20
  <BR><FONT size=3D2>wireless operating company owns all the numbers in =
an NPA=20
  NXX, which avoids</FONT> <BR><FONT size=3D2>the major problem LNP was =
designed=20
  to create.&nbsp; (No, that's not a typo, it's</FONT> <BR><FONT=20
  size=3D2>editorial comment.)&nbsp;</FONT> </P>
  <P><FONT size=3D2>Probably it's more of an administrative problem than =
a=20
  technical one.&nbsp; It's</FONT> <BR><FONT size=3D2>so easy to fall =
into the=20
  trap of thinking there are technical solutions to</FONT> <BR><FONT=20
  size=3D2>administrative problems, although technical solutions can =
certainly=20
  help</FONT> <BR><FONT size=3D2>perpetuate administrative =
problems.&nbsp;</FONT>=20
  </P>
  <P><FONT size=3D2>******** FCC has granted a delay for the wireless =
carriers in=20
  the U.S. to</FONT> <BR><FONT size=3D2>support local number portability =
(LNP)=20
  because the new entrants - PCS</FONT> <BR><FONT size=3D2>carriers are =
too busy=20
  in deploying their networks.&nbsp; LNP is intended to help</FONT> =
<BR><FONT=20
  size=3D2>the new entrants to compete with the incumbent =
carriers.&nbsp; So if=20
  the new</FONT> <BR><FONT size=3D2>entrants ask for a delay, FCC =
certainly has no=20
  reason to push wireless LNP.&nbsp;</FONT> <BR><FONT size=3D2>Also =
there are=20
  issues on roaming where the roaming markets may not be</FONT> =
<BR><FONT=20
  size=3D2>required to support LNP (LNP is to be supported in the top =
100=20
  major</FONT> <BR><FONT size=3D2>Statistical Trading Areas in =
phases).&nbsp; But=20
  I believe that the delay is</FONT> <BR><FONT size=3D2>temporary =
because FCC's=20
  goal may be nation-wide number portability which</FONT> <BR><FONT=20
  size=3D2>allows the subscriber to switch from wireline service =
provider to a=20
  wireless</FONT> <BR><FONT size=3D2>service provider and move from one =
location=20
  to another location.&nbsp; The LRN</FONT> <BR><FONT size=3D2>method =
will allow=20
  that with minimum system upgrades.&nbsp; The major issue about</FONT>=20
  <BR><FONT size=3D2>location portability is the call =
charing/rating.&nbsp; If=20
  flat rate is used for</FONT> <BR><FONT size=3D2>all toll calls (e.g., =
10 cents=20
  per minute regardless of the distance between</FONT> <BR><FONT =
size=3D2>the=20
  calling party and called party), location portability can be =
supported</FONT>=20
  <BR><FONT size=3D2>easily. </FONT></P>
  <P><FONT=20
  =
size=3D2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbs=
p;=20
  Wireles is a similar model if we view the "donor switch" as the</FONT> =

  <BR><FONT size=3D2>"home" system" and the "serving switch" as the =
"vistied=20
  system."&nbsp; However,</FONT> <BR><FONT size=3D2>there are =
differences.&nbsp;=20
  The "home system" in the wireless model is the</FONT> <BR><FONT =
size=3D2>service=20
  provider of the "roamer."&nbsp; It authorizes the visited system =
to</FONT>=20
  <BR><FONT size=3D2>provide the services to the roamer and guarantees =
the payment=20
  even if the</FONT> <BR><FONT size=3D2>subscriber does not pay his/her=20
  bill.&nbsp; Also, Mobile Application Part (MAP)</FONT> <BR><FONT =
size=3D2>is=20
  used between the visited and home systems to register the "roamer" and =

  to</FONT> <BR><FONT size=3D2>get a temporary routing number to route =
the call to=20
  the visited system.&nbsp;</FONT> <BR><FONT size=3D2>Either the call =
must always=20
  be routed through the home system to reach the</FONT> <BR><FONT =
size=3D2>visited=20
  system or a query to the home system must be performed first to</FONT> =

  <BR><FONT size=3D2>perform optimal routing to the visited system =
directly.&nbsp;=20
  NPDB approach</FONT> <BR><FONT size=3D2>actually allows the query to =
be done=20
  (for optimal routing) without the</FONT> <BR><FONT =
size=3D2>involvement of the=20
  "donor" network.&nbsp; *************</FONT> </P>
  <P><FONT size=3D2>&nbsp;John Mainwaring </FONT><BR><FONT =
size=3D2>Nortel Networks=20
  </FONT><BR><FONT size=3D2>crm312a@nortelnetworks.com=20
</FONT></P></BLOCKQUOTE></BODY></HTML>

------=_NextPart_000_000A_01BF418F.010C1100--


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Dec  9 01:15:33 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 BAA25577
	for <iptel-archive@odin.ietf.org>; Thu, 9 Dec 1999 01:15:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 850B152E1; Thu,  9 Dec 1999 01:04:01 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 9CA6E52E3; Thu,  9 Dec 1999 01:04:00 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <384F4720.358162F3@dynamicsoft.com>
Date: Thu, 09 Dec 1999 01:07:28 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: Dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Mark D. Foster" <mark.foster@npac.com>
Cc: "'John Mainwaring'" <crm312a@nortelnetworks.com>,
        iptel@lists.research.bell-labs.com,
        "'James Yu'" <james.yu@neustar.com>, sip@lists.research.bell-labs.com
Subject: Re: TRIP
References: <000901bf41b8$e9e21900$0864a8c0@loudoun.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

Mark Foster wrote:
> That at least some forms of Iptel must comply with the non-reliance requirement in at least some countries suggests that the
> SIP, TRIP, enum, etc., must be flexible enough to allow implementations that will satisfy these requirements, and still be
> interoperable with those that don't.
>  
> In the US and Canada, the key to this was the N-1 trigger placement policy.  This was simply a guideline on which networks
> should be prepared to dip calls placed to which areas in order to not force calls to be re-routed at the donor network
> (termination routing).  Key to having flexibility on where triggers are placed, resulting in NPDB queries, is having a form of
> post-query signaling that makes the query results available to all downstream networks.

I'm not really sure that the N-1 concept works that well in an IP
environment:

1. If you look at DNS as a similar service to LNP on the IP side, the
originating host does a name to address translation right away. This
happens even if the name thats being translated is "far away". This
avoids triangle routing, as has been discussed.

2. The interconnection of SIP servers and providers within the IP world
is not that rigid. I'm not sure how a SIP proxy would indeed determine
that it is the N-1th server in the chain.

In any case, determination of where the NPDB query is done is something
I don't think is within the scope of iptel to determine. All we need to
figure out is what kind of numbers are distributed in TRIP. Near as I
can tell, since LRNs still refer to the prefix of numbers served by a
switch, doing what we were doing all along works just fine. 

>  
> The key data elements in ANSI ISUP are: the LNP-query-done indicator (bit in the FCI); placing a routing number returned
> from the NPDB in the CdPN (LRN or whatever), and moving the original dialed number to a GAP.  Subsequent call routing is
> done on the routing number, while not loosing the dialed portable subscriber number.  Given the need for SIP to interoperate
> with ANSI ISUP, there should be some consideration of implementing these semantics.
>  
> If SIP captured these fields, so as to provide a post-dip form of downstream signaling, and TRIP were designed to build route
> tables based on administratively assigned number ranges (which, by definition would make calls to routing numbers [LRN]
> routable) , and enum did not require DNS queries to be answered by donor networks, then the existing regulatory
> requirements for LNP could be met.

First off, SIP has two fields that identify the destination party. The
To field represents the original number that the caller called, and the
request URI represents the address translated at each hop. Thus, the To
field carries the original dialed number, and the request URI would
carrying the routing number. Now, do we need to somehow indicate that
the LNP dip has been done? I'm still unsure about that. I believe
Henning had suggested that the use of a sip URL (as opposed to a tel
URL) in the request URI would effectively be an indicator of such. That
aside, I think there is a possible confusion if one were to assume an
LRN looked like a normal SIP URL:

INVITE sip:1212344 SIP/2.0
To: sip:14081234567

The problem is related to past discussions on the SIP mailing list
regarding overlap dialing. A terminating gateway would look at
sip:1212344, and know how to route it, but might not know whether this
was a dialed number with more digits coming, or an LRN. Does it matter?
Not being an ISUP expert, I'm not sure. If it does matter, something
like user=lrn within the SIP URL would fix that.

But anyway, thats also not a problem for iptel, although I've cc'd the
sip list on this message.

-Jonathan R.

-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Dec  9 13:59:07 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 NAA05210
	for <iptel-archive@odin.ietf.org>; Thu, 9 Dec 1999 13:59:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BAE9F52DB; Thu,  9 Dec 1999 13:48:02 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 2044D52D5; Thu,  9 Dec 1999 13:47:58 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <384FF91C.DC9E7EE7@cisco.com>
Date: Thu, 09 Dec 1999 10:46:52 -0800
From: "Hussein F. Salama" <hsalama@cisco.com>
Reply-To: hsalama@cisco.com
Organization: Cisco Systems
X-Mailer: Mozilla 4.5 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Cc: "Mark D. Foster" <mark.foster@npac.com>,
        "'John Mainwaring'" <crm312a@nortelnetworks.com>,
        iptel@lists.research.bell-labs.com,
        "'James Yu'" <james.yu@neustar.com>, sip@lists.research.bell-labs.com
Subject: Re: TRIP
References: <000901bf41b8$e9e21900$0864a8c0@loudoun.com> <384F4720.358162F3@dynamicsoft.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:
> 
> Mark Foster wrote:
> > That at least some forms of Iptel must comply with the non-reliance requirement in at least some countries suggests that the
> > SIP, TRIP, enum, etc., must be flexible enough to allow implementations that will satisfy these requirements, and still be
> > interoperable with those that don't.
> >
> > In the US and Canada, the key to this was the N-1 trigger placement policy.  This was simply a guideline on which networks
> > should be prepared to dip calls placed to which areas in order to not force calls to be re-routed at the donor network
> > (termination routing).  Key to having flexibility on where triggers are placed, resulting in NPDB queries, is having a form of
> > post-query signaling that makes the query results available to all downstream networks.
> 
> I'm not really sure that the N-1 concept works that well in an IP
> environment:
> 
> 1. If you look at DNS as a similar service to LNP on the IP side, the
> originating host does a name to address translation right away. This
> happens even if the name thats being translated is "far away". This
> avoids triangle routing, as has been discussed.
> 
> 2. The interconnection of SIP servers and providers within the IP world
> is not that rigid. I'm not sure how a SIP proxy would indeed determine
> that it is the N-1th server in the chain.
> 
> In any case, determination of where the NPDB query is done is something
> I don't think is within the scope of iptel to determine. All we need to
> figure out is what kind of numbers are distributed in TRIP. Near as I
> can tell, since LRNs still refer to the prefix of numbers served by a
> switch, doing what we were doing all along works just fine.
> 

Yes what we're doing is OK, because the LRN is part of an office's
NPA-NXX prefix. But we may have to add a new field to the
ReachableRoutes attribute to mark a prefix as portable. A prefix is
marked as portable if at lease one number has been ported out of this
prefix. 

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  Thu Dec  9 17:22:42 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 RAA18615
	for <iptel-archive@odin.ietf.org>; Thu, 9 Dec 1999 17:22:41 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 8907E52D6; Thu,  9 Dec 1999 17:17:45 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E5AD152D5; Thu,  9 Dec 1999 17:17:44 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <89866A7EC9C0D1119C950000F8BCC2820157C018@zrtpd00n.us.nortel.com>
From: "John Mainwaring" <crm312a@nortelnetworks.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>,
        "Mark D. Foster" <mark.foster@npac.com>
Cc: iptel@lists.research.bell-labs.com, "'James Yu'" <james.yu@neustar.com>,
        sip@lists.research.bell-labs.com
Subject: RE: TRIP
Date: Thu, 9 Dec 1999 17:02:09 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: multipart/alternative;
              boundary="----_=_NextPart_001_01BF4291.155B5E06"
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_01BF4291.155B5E06
Content-Type: text/plain;
	charset="windows-1252"



|-----Original Message-----
|From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]
|
|Mark Foster wrote:
|> In the US and Canada, the key to this was the N-1 trigger 
|placement policy.  This was simply a guideline on which networks
|> should be prepared to dip calls placed to which areas in 
|order to not force calls to be re-routed at the donor network
|> (termination routing).  Key to having flexibility on where 
|triggers are placed, resulting in NPDB queries, is having a form of
|> post-query signaling that makes the query results available 
|to all downstream networks.
|
|I'm not really sure that the N-1 concept works that well in an IP
|environment:

I thinks you're right, and what follows is that a concept that does work
will have to be found.

|1. If you look at DNS as a similar service to LNP on the IP side, the
|originating host does a name to address translation right away. This
|happens even if the name thats being translated is "far away". This
|avoids triangle routing, as has been discussed.
|
|2. The interconnection of SIP servers and providers within the IP world
|is not that rigid. I'm not sure how a SIP proxy would indeed determine
|that it is the N-1th server in the chain.

It isn't really the n-1 server that has to do the NP query, it's the n-1
network, typically an IEC for calls originating ouside a zone of
portability.  The IEC can do the query anywhere in their network that makes
sense.

A server doesn't know it is the n-1th server in any arbitrary chain.  At
most, it knows that it is aware of the portability status of some numbers
and not others.  You have to make sure that you don't try to route directly
to numbers whose portability status you do not know; you should route
through a transit server that does know, or query a database that will give
you the portability status.

|In any case, determination of where the NPDB query is done is something
|I don't think is within the scope of iptel to determine. All we need to
|figure out is what kind of numbers are distributed in TRIP. Near as I
|can tell, since LRNs still refer to the prefix of numbers served by a
|switch, doing what we were doing all along works just fine. 

I agree.  If you have done an NP query, you should use the LRN to query the
TRIP and the result should give you the right proxy.  The LRN is the
surrogate Directory Number for routing purposes.

However, the question of where the NP query can/should be done certainly
affects the overall architecture that TRIP has to fit into.  I think the
discussion below points in a workable direction.
 
|> The key data elements in ANSI ISUP are: the LNP-query-done 
|indicator (bit in the FCI); placing a routing number returned
|> from the NPDB in the CdPN (LRN or whatever), and moving the 
|original dialed number to a GAP.  Subsequent call routing is
|> done on the routing number, while not loosing the dialed 
|portable subscriber number.  Given the need for SIP to interoperate
|> with ANSI ISUP, there should be some consideration of 
|implementing these semantics.
|>  
|> If SIP captured these fields, so as to provide a post-dip 
|form of downstream signaling, and TRIP were designed to build route
|> tables based on administratively assigned number ranges 
|(which, by definition would make calls to routing numbers [LRN]
|> routable) , and enum did not require DNS queries to be 
|answered by donor networks, then the existing regulatory
|> requirements for LNP could be met.
|
|First off, SIP has two fields that identify the destination party. The
|To field represents the original number that the caller called, and the
|request URI represents the address translated at each hop. Thus, the To
|field carries the original dialed number, and the request URI would
|carrying the routing number. Now, do we need to somehow indicate that
|the LNP dip has been done? I'm still unsure about that. I believe
|Henning had suggested that the use of a sip URL (as opposed to a tel
|URL) in the request URI would effectively be an indicator of such. That
|aside, I think there is a possible confusion if one were to assume an
|LRN looked like a normal SIP URL:
|
|INVITE sip:1212344 SIP/2.0
|To: sip:14081234567
|
|The problem is related to past discussions on the SIP mailing list
|regarding overlap dialing. A terminating gateway would look at
|sip:1212344, and know how to route it, but might not know whether this
|was a dialed number with more digits coming, or an LRN. Does it matter?
|Not being an ISUP expert, I'm not sure. If it does matter, something
|like user=lrn within the SIP URL would fix that.

user=lrn is almost right.  If you have queried a portable number and found
that it is not ported, then the DN has the same status as an LRN would have
if the number were ported (i.e. it should be treated as authoritative for
routing).  user=np-queried might be closer to the mark.

It's also necessary to define what the TRIP should do differently depending
on whether the np-queried indication is present.  I must admit I came into
this discussion mainly because I have worked on LNP.  I assume the objective
is not to route unqueried calls to the donor proxy.  I have noted that
originator query may not be practical for nationwide calling and that donor
proxy is too late unless you're prepared to consider RTP.  One answer might
be for the TRIP to provide a response on unqueried numbers that routes to an
intermediate server that does know how to do NP queries for the destination
DN.  No doubt there are other possibilities.  I'm curious to know which
forum would regard this problem as being in scope.

|But anyway, thats also not a problem for iptel, although I've cc'd the
|sip list on this message.
|
|-Jonathan R.
|
|-- 
|Jonathan D. Rosenberg                       200 Executive Drive
|Chief Scientist                             Suite 120 
|dynamicsoft                                 West Orange, NJ 07052
|jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
|http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
|http://www.dynamicsoft.com
|

------_=_NextPart_001_01BF4291.155B5E06
Content-Type: text/html;
	charset="windows-1252"
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=3Dwindows-1252">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2448.0">
<TITLE>RE: TRIP</TITLE>
</HEAD>
<BODY>
<BR>
<BR>

<P><FONT SIZE=3D2>|-----Original Message-----</FONT>
<BR><FONT SIZE=3D2>|From: Jonathan Rosenberg [<A =
HREF=3D"mailto:jdrosen@dynamicsoft.com">mailto:jdrosen@dynamicsoft.com</=
A>]</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|Mark Foster wrote:</FONT>
<BR><FONT SIZE=3D2>|&gt; In the US and Canada, the key to this was the =
N-1 trigger </FONT>
<BR><FONT SIZE=3D2>|placement policy.&nbsp; This was simply a guideline =
on which networks</FONT>
<BR><FONT SIZE=3D2>|&gt; should be prepared to dip calls placed to =
which areas in </FONT>
<BR><FONT SIZE=3D2>|order to not force calls to be re-routed at the =
donor network</FONT>
<BR><FONT SIZE=3D2>|&gt; (termination routing).&nbsp; Key to having =
flexibility on where </FONT>
<BR><FONT SIZE=3D2>|triggers are placed, resulting in NPDB queries, is =
having a form of</FONT>
<BR><FONT SIZE=3D2>|&gt; post-query signaling that makes the query =
results available </FONT>
<BR><FONT SIZE=3D2>|to all downstream networks.</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|I'm not really sure that the N-1 concept works that =
well in an IP</FONT>
<BR><FONT SIZE=3D2>|environment:</FONT>
</P>

<P><FONT SIZE=3D2>I thinks you're right, and what follows is that a =
concept that does work will have to be found.</FONT>
</P>

<P><FONT SIZE=3D2>|1. If you look at DNS as a similar service to LNP on =
the IP side, the</FONT>
<BR><FONT SIZE=3D2>|originating host does a name to address translation =
right away. This</FONT>
<BR><FONT SIZE=3D2>|happens even if the name thats being translated is =
&quot;far away&quot;. This</FONT>
<BR><FONT SIZE=3D2>|avoids triangle routing, as has been =
discussed.</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|2. The interconnection of SIP servers and providers =
within the IP world</FONT>
<BR><FONT SIZE=3D2>|is not that rigid. I'm not sure how a SIP proxy =
would indeed determine</FONT>
<BR><FONT SIZE=3D2>|that it is the N-1th server in the chain.</FONT>
</P>

<P><FONT SIZE=3D2>It isn't really the n-1 server that has to do the NP =
query, it's the n-1 network, typically an IEC for calls originating =
ouside a zone of portability.&nbsp; The IEC can do the query anywhere =
in their network that makes sense.</FONT></P>

<P><FONT SIZE=3D2>A server doesn't know it is the n-1th server in any =
arbitrary chain.&nbsp; At most, it knows that it is aware of the =
portability status of some numbers and not others.&nbsp; You have to =
make sure that you don't try to route directly to numbers whose =
portability status you do not know; you should route through a transit =
server that does know, or query a database that will give you the =
portability status.</FONT></P>

<P><FONT SIZE=3D2>|In any case, determination of where the NPDB query =
is done is something</FONT>
<BR><FONT SIZE=3D2>|I don't think is within the scope of iptel to =
determine. All we need to</FONT>
<BR><FONT SIZE=3D2>|figure out is what kind of numbers are distributed =
in TRIP. Near as I</FONT>
<BR><FONT SIZE=3D2>|can tell, since LRNs still refer to the prefix of =
numbers served by a</FONT>
<BR><FONT SIZE=3D2>|switch, doing what we were doing all along works =
just fine. </FONT>
</P>

<P><FONT SIZE=3D2>I agree.&nbsp; If you have done an NP query, you =
should use the LRN to query the TRIP and the result should give you the =
right proxy.&nbsp; The LRN is the surrogate Directory Number for =
routing purposes.</FONT></P>

<P><FONT SIZE=3D2>However, the question of where the NP query =
can/should be done certainly affects the overall architecture that TRIP =
has to fit into.&nbsp; I think the discussion below points in a =
workable direction.</FONT></P>

<P><FONT SIZE=3D2>&nbsp;</FONT>
<BR><FONT SIZE=3D2>|&gt; The key data elements in ANSI ISUP are: the =
LNP-query-done </FONT>
<BR><FONT SIZE=3D2>|indicator (bit in the FCI); placing a routing =
number returned</FONT>
<BR><FONT SIZE=3D2>|&gt; from the NPDB in the CdPN (LRN or whatever), =
and moving the </FONT>
<BR><FONT SIZE=3D2>|original dialed number to a GAP.&nbsp; Subsequent =
call routing is</FONT>
<BR><FONT SIZE=3D2>|&gt; done on the routing number, while not loosing =
the dialed </FONT>
<BR><FONT SIZE=3D2>|portable subscriber number.&nbsp; Given the need =
for SIP to interoperate</FONT>
<BR><FONT SIZE=3D2>|&gt; with ANSI ISUP, there should be some =
consideration of </FONT>
<BR><FONT SIZE=3D2>|implementing these semantics.</FONT>
<BR><FONT SIZE=3D2>|&gt;&nbsp; </FONT>
<BR><FONT SIZE=3D2>|&gt; If SIP captured these fields, so as to provide =
a post-dip </FONT>
<BR><FONT SIZE=3D2>|form of downstream signaling, and TRIP were =
designed to build route</FONT>
<BR><FONT SIZE=3D2>|&gt; tables based on administratively assigned =
number ranges </FONT>
<BR><FONT SIZE=3D2>|(which, by definition would make calls to routing =
numbers [LRN]</FONT>
<BR><FONT SIZE=3D2>|&gt; routable) , and enum did not require DNS =
queries to be </FONT>
<BR><FONT SIZE=3D2>|answered by donor networks, then the existing =
regulatory</FONT>
<BR><FONT SIZE=3D2>|&gt; requirements for LNP could be met.</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|First off, SIP has two fields that identify the =
destination party. The</FONT>
<BR><FONT SIZE=3D2>|To field represents the original number that the =
caller called, and the</FONT>
<BR><FONT SIZE=3D2>|request URI represents the address translated at =
each hop. Thus, the To</FONT>
<BR><FONT SIZE=3D2>|field carries the original dialed number, and the =
request URI would</FONT>
<BR><FONT SIZE=3D2>|carrying the routing number. Now, do we need to =
somehow indicate that</FONT>
<BR><FONT SIZE=3D2>|the LNP dip has been done? I'm still unsure about =
that. I believe</FONT>
<BR><FONT SIZE=3D2>|Henning had suggested that the use of a sip URL (as =
opposed to a tel</FONT>
<BR><FONT SIZE=3D2>|URL) in the request URI would effectively be an =
indicator of such. That</FONT>
<BR><FONT SIZE=3D2>|aside, I think there is a possible confusion if one =
were to assume an</FONT>
<BR><FONT SIZE=3D2>|LRN looked like a normal SIP URL:</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|INVITE sip:1212344 SIP/2.0</FONT>
<BR><FONT SIZE=3D2>|To: sip:14081234567</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|The problem is related to past discussions on the =
SIP mailing list</FONT>
<BR><FONT SIZE=3D2>|regarding overlap dialing. A terminating gateway =
would look at</FONT>
<BR><FONT SIZE=3D2>|sip:1212344, and know how to route it, but might =
not know whether this</FONT>
<BR><FONT SIZE=3D2>|was a dialed number with more digits coming, or an =
LRN. Does it matter?</FONT>
<BR><FONT SIZE=3D2>|Not being an ISUP expert, I'm not sure. If it does =
matter, something</FONT>
<BR><FONT SIZE=3D2>|like user=3Dlrn within the SIP URL would fix =
that.</FONT>
</P>

<P><FONT SIZE=3D2>user=3Dlrn is almost right.&nbsp; If you have queried =
a portable number and found that it is not ported, then the DN has the =
same status as an LRN would have if the number were ported (i.e. it =
should be treated as authoritative for routing).&nbsp; =
user=3Dnp-queried might be closer to the mark.</FONT></P>

<P><FONT SIZE=3D2>It's also necessary to define what the TRIP should do =
differently depending on whether the np-queried indication is =
present.&nbsp; I must admit I came into this discussion mainly because =
I have worked on LNP.&nbsp; I assume the objective is not to route =
unqueried calls to the donor proxy.&nbsp; I have noted that originator =
query may not be practical for nationwide calling and that donor proxy =
is too late unless you're prepared to consider RTP.&nbsp; One answer =
might be for the TRIP to provide a response on unqueried numbers that =
routes to an intermediate server that does know how to do NP queries =
for the destination DN.&nbsp; No doubt there are other =
possibilities.&nbsp; I'm curious to know which forum would regard this =
problem as being in scope.</FONT></P>

<P><FONT SIZE=3D2>|But anyway, thats also not a problem for iptel, =
although I've cc'd the</FONT>
<BR><FONT SIZE=3D2>|sip list on this message.</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|-Jonathan R.</FONT>
<BR><FONT SIZE=3D2>|</FONT>
<BR><FONT SIZE=3D2>|-- </FONT>
<BR><FONT SIZE=3D2>|Jonathan D. =
Rosenberg&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =
200 Executive Drive</FONT>
<BR><FONT SIZE=3D2>|Chief =
Scientist&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Suite 120 </FONT>
<BR><FONT =
SIZE=3D2>|dynamicsoft&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =
West Orange, NJ 07052</FONT>
<BR><FONT =
SIZE=3D2>|jdrosen@dynamicsoft.com&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp; FAX:&nbsp;&nbsp; (732) 741-4778</FONT>
<BR><FONT SIZE=3D2>|<A HREF=3D"http://www.cs.columbia.edu/~jdrosen" =
TARGET=3D"_blank">http://www.cs.columbia.edu/~jdrosen</A>&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; PHONE: (732) 741-7244</FONT>
<BR><FONT SIZE=3D2>|<A HREF=3D"http://www.dynamicsoft.com" =
TARGET=3D"_blank">http://www.dynamicsoft.com</A></FONT>
<BR><FONT SIZE=3D2>|</FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01BF4291.155B5E06--

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Dec  9 20:53:58 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 UAA01925
	for <iptel-archive@odin.ietf.org>; Thu, 9 Dec 1999 20:53:58 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id EEE5A52D5; Thu,  9 Dec 1999 20:49:27 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 65ACA52DE; Thu,  9 Dec 1999 20:49:26 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <38505BA4.1EE87DF5@cs.columbia.edu>
Date: Thu, 09 Dec 1999 20:47:16 -0500
From: Henning Schulzrinne <schulzrinne@cs.columbia.edu>
Organization: Columbia University, Dept. of Computer Science
X-Mailer: Mozilla 4.7 [en] (X11; U; SunOS 5.7 sun4u)
X-Accept-Language: en
MIME-Version: 1.0
Cc: iptel@lists.research.bell-labs.com
Subject: Re: TRIP
References: <000901bf41b8$e9e21900$0864a8c0@loudoun.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

Btw, 
http://www.fcc.gov/Bureaus/Common_Carrier/Orders/1997/fcc97074pdf.html
has the gory details.
-- 
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  Fri Dec 10 11:23:36 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 LAA26479
	for <iptel-archive@odin.ietf.org>; Fri, 10 Dec 1999 11:23:35 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 3719752E9; Fri, 10 Dec 1999 11:07:09 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 44DC352E8; Fri, 10 Dec 1999 11:07:01 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Thu, 09 Dec 1999 17:36:40 -0500
From: "Mark D. Foster" <mfoster@mdfassoc.com>
Subject: RE: TRIP
In-reply-to: <384F4720.358162F3@dynamicsoft.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>
Cc: "'John Mainwaring'" <crm312a@nortelnetworks.com>,
        iptel@lists.research.bell-labs.com,
        "'James Yu'" <james.yu@neustar.com>, sip@lists.research.bell-labs.com
Message-id: <000b01bf429a$e8c73770$0864a8c0@loudoun.com>
MIME-version: 1.0
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2910.0)
Content-type: text/plain; charset=iso-8859-1
Content-transfer-encoding: 7bit
Importance: Normal
X-MSMail-Priority: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300
X-Priority: 3 (Normal)
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Jonathan Rosenberg wrote:
>
> I'm not really sure that the N-1 concept works that well in an IP
> environment:
>

Totally agree.  N-1 is relevant as an example of the need for flexibility on
where NPDB dips are done.  By varying trigger (dip) placement, one can
implement a wide range of call routing scenarios.  Again, key to trigger
flexibility is a positive query-done indicator so any downstream network
element will behave accordingly.

> 1. If you look at DNS as a similar service to LNP on the IP side, the
> originating host does a name to address translation right away. This
> happens even if the name thats being translated is "far away". This
> avoids triangle routing, as has been discussed.
>

Source routing is the most efficient, but at the expense of having a global
DB mapping service available to originating networks.

N-1 was used in the US to allow local exchange providers to only worry about
hosting that local area's NPDB, because if the call is long-distance it's
expected to be dipped by the IXC.  Frankly, this is all that N-1 was all
about, was relieving the RBOCs of the need to host the entire US NPDB, given
that for regulatory reasons their wireline networks are still LATA-specific.

> 2. The interconnection of SIP servers and providers within
> the IP world
> is not that rigid. I'm not sure how a SIP proxy would indeed determine
> that it is the N-1th server in the chain.
>

Even in the PSTN, a network has no way of knowing they are N-1 prior to
doing a dip.  However, as a heuristic, RBOC switches only load LNP triggers
to dip calls to number prefixes in the same LATA.  And, IXC switches are
expected to dip any non-dipped calls, hence the importance of a positive
query-done indicator.

There could be applications in the IP arena, for example a purely local ITSP
could opt to only dip calls in the NPDB via enum for the local area they
serve, and rely on a downstream transport network to dip everything else.

> In any case, determination of where the NPDB query is done is
> something
> I don't think is within the scope of iptel to determine. All
> we need to
> figure out is what kind of numbers are distributed in TRIP. Near as I
> can tell, since LRNs still refer to the prefix of numbers served by a
> switch, doing what we were doing all along works just fine.
>

Agree.  I think both TRIP's and SIP's role should be to provide the needed
flexibility to allow NPDB queries to occur wherever in the call path
business and regulatory decisions dictate.

>
> First off, SIP has two fields that identify the destination party. The
> To field represents the original number that the caller
> called, and the
> request URI represents the address translated at each hop.
> Thus, the To
> field carries the original dialed number, and the request URI would
> carrying the routing number. Now, do we need to somehow indicate that
> the LNP dip has been done? I'm still unsure about that. I believe
> Henning had suggested that the use of a sip URL (as opposed to a tel
> URL) in the request URI would effectively be an indicator of
> such. That
> aside, I think there is a possible confusion if one were to assume an
> LRN looked like a normal SIP URL:
>
> INVITE sip:1212344 SIP/2.0
> To: sip:14081234567
>
> The problem is related to past discussions on the SIP mailing list
> regarding overlap dialing. A terminating gateway would look at
> sip:1212344, and know how to route it, but might not know whether this
> was a dialed number with more digits coming, or an LRN. Does
> it matter?
> Not being an ISUP expert, I'm not sure. If it does matter, something
> like user=lrn within the SIP URL would fix that.
>

I think it does matter, if for no other reason than to ensure complete
interoperation with ISUP, which could theoretically occur anywhere in the
call path.  For example, if a dip is done but the number is not ported, then
the query-done indicator is still set to suppress downstream triggers even
though the CdPN is left alone.   As long as there's a way to recover each of
the three LNP fields in SIP (query-done, dialed number, and routing number),
gateways should be able to do their thing (e.g. interoperate with ISUP,
TRIP, or NPDB).

> But anyway, thats also not a problem for iptel, although I've cc'd the
> sip list on this message.
>

Yup.

-------------------------------------------------------------
Mark D. Foster                  | EMAIL: mark.foster@neustar.com
CTO                             | NeuStar, Inc.
TEL: +1 202 533 2800/2702       | FAX: +1 202 533 2975
1120 Vermont Ave NW, Suite 550  | Washington, DC 20005


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Dec 13 15:37:06 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 PAA18694
	for <iptel-archive@odin.ietf.org>; Mon, 13 Dec 1999 15:37:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 12CCE52F6; Mon, 13 Dec 1999 15:25:59 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 6DC4652F3; Mon, 13 Dec 1999 15:25:54 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C3BE934@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'Matt Squire'" <msquire@nortelnetworks.com>
Cc: "'sip@lists.research.bell-labs.com'" <sip@lists.research.bell-labs.com>,
        "'enum@ietf.org'" <enum@ietf.org>,
        "'iptel@lists.research.bell-labs.com'" <iptel@lists.research.bell-labs.com>
Subject: RE: [Enum] Nation-Specific ISUP implementations, Overlap Signalin
	g &  Number Portability Database Queries
Date: Mon, 13 Dec 1999 14:17:05 -0600
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

Matt,

It is a good summary of the key points.  You mentioned iptel that I forgot
to cc to.  So I'm adding it to the cc list.

My point is that the WGs should address their respective agendas while
keeping those issues in mind so that the solution(s) developed by each
individual WG can all work together.

James

-----Original Message-----
From: Matt Squire [mailto:msquire@nortelnetworks.com]
Sent: Monday, December 13, 1999 3:01 PM
To: James Yu
Cc: 'sip@lists.research.bell-labs.com'; 'enum@ietf.org'
Subject: Re: [Enum] Nation-Specific ISUP implementations, Overlap
Signaling & Number Portability Database Queries



You certainly seem to pack alot of things into this note.  An attempt to
summarize:
(a) ISUP interworking between gateways
(b) Call routing decisions based on ISUP capabilities
(c) Timing/location of portability query
(d) Complexity of originating switch performing NPDB queries due to
national variants
(e) Overlap signalling and its complexities to call signalling,
particularly international

Here's a 2-bit opinion on some of these things...

(a) ISUP interworking was discusesed for a bit in IPTEL, and it looked
like the consensus was that gateways must be ready to fall back to some
common denominator ISUP capability set if they don't support any common
ISUP variation.

(b) Right now, the call routing portion of the solution (TRIP) is not
considering ISUP variations when making call routing decisions.  It was
discussed, but the consensus appeared to be that falling to a common
international capability set was an acceptable solution.  So it looks
like ISUP variations won't impact the route the call takes to a
destination gateway.

(c) By just reading the notes, it looks like there might be consensus to
perform the portability query before initiating the call and accessing
things like call routing.

(d) I would hope that enum does not require the querier to know all of
the possible national variants to NPDB queries.  Enum should define a
common query mechanism that is independent of national variants.  The
system should be a superset of the national variants and be able to
coordinate the information across nation's boundries.  Without this, the
whole thing seems like very unhelpful.  Enum queries really have to be
international.  

(e) About a month ago there was alot of discussion on overlap signalling
in SIP.  If you haven't read the discussion, you might want to review
it.  There were several proposals on how to handle it.  I don't recall
how well the interaction between overlap signalling and LNP was
discussed, but I recall it being mentioned a few times.  I'm sure others
can provide further elaboration of those discussions.    

 

James Yu wrote:
> 
> The following discussions focus on SCN-IP-SCN calls.
> 
> There are nation-specific variatoins in ISUP implementations.  I don't
know
> whether this issue has been discussed (I'm new to IETF).  When we have
only
> a few GWs, it is fine that those GWs support many ISUP implementations.
But
> when many GWs are deployed, it is not likely that every GW will support
many
> ISUP implementations.  Some may be used only for domestic use.  So it is
> likely that an international call may need to be routed to an
> "international" GW (or more than one) that interworks between two
> nation-specific ISUP implementations.  To fully use the IP network, then
the
> "international" GW will further route the call to a "domestic" GW
> (destination GW) that terminates the call into the SCN.  The call can
> certainly be routed to the destination SCN but the call charge may be
> higher.
> 
> This would seem to be inefficient that the user traffic must be routed
> through additional GW(s).  ISUP interworking probably can be done via a
pure
> signaling conversion/mapping so that the user traffic can be routed
directly
> to the destination GW.
> 
> If we need to perform number portability queries, then a good place would
be
> the destination "international" GW because it is close to the destination
> country and supports the destination country's Number Portability Database
> (NPDB) interface.  In that case, simple ISUP interworking mentioned above
> won't be sufficient.  A NPDB query also needs to be launched. The response
> to the NPDB query will decide the destination GW that should terminate the
> call.
> 
> Due to number portability, the called party's serving SCN switch changes
> when the called user switches to a new service provider.  Whenever that
> happens, the NPDB will reflect the current new serving switch or network
> information (via Location Routing Number in the North America or Routing
> Prefix in the Europe), and the "destination" GW that can reach the serving
> switch or network may change.  So a NPDB query is required to determine
the
> correct GW to terminate the call.
> 
> At ENUM, we have been discussing where to perform the NPDB queries.  The
> originating GW needs to support all nation-specific NPDB interfaces if it
is
> to query those NPDBs.  This would be expensive.  Also there is an issue
> about the encapsulated ISUP message in the SIP if originating GW's
> nationa-specific ISUP message is encapsulated (may be everyone should use
> ITU-T ISUP that is used between international switches for communicating
> among the entities in the IP domain).  The destination GW can support
> specific NPDB interface; however, it may be the wrong GW because it cannot
> reach the serving switch in the SCN that serves the called party.
> 
> Overlap signaling would make the issue more complicate.  If the call
routing
> decision is to route to the "international" GW, it is fine because number
> portability is confined within a national domain.  That "international" GW
> will then wait for additional address information and performs the NPDB
> query, if required.  But if the call is to be routed directly to the
> destination GW, and if the originating GW is the one to perform the NPDB,
> then it has to collect all the address information before it can launch
the
> query.  But the issue is that the originating GW may not know it has
> collected all the address digits.
> 
> It seems that the best entity that knows that it has collected all the
> digits and can perform the NPDB query would be the destination
> "international" GW.
> 
> I have linked several issues (nation-specific ISUP implementations, number
> portability, and user traffic routing efficiency) together that need to be
> addressed as a whole.  There may be other issues.  It is hoped that one or
> more solutions can be identified and finalized.
> 
> _______________________________________________
> enum mailing list
> enum@ietf.org
> http://www.ietf.org/mailman/listinfo/enum

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 14 10:50:33 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 KAA21511
	for <iptel-archive@odin.ietf.org>; Tue, 14 Dec 1999 10:50:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D353752F9; Tue, 14 Dec 1999 10:40:02 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 414A452DC; Tue, 14 Dec 1999 10:39:59 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <3856658E.4F344080@dynamicsoft.com>
Date: Tue, 14 Dec 1999 10:43:10 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: Dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: hsalama@cisco.com
Cc: "Mark D. Foster" <mark.foster@npac.com>,
        "'John Mainwaring'" <crm312a@nortelnetworks.com>,
        iptel@lists.research.bell-labs.com,
        "'James Yu'" <james.yu@neustar.com>, sip@lists.research.bell-labs.com
Subject: Re: TRIP
References: <000901bf41b8$e9e21900$0864a8c0@loudoun.com> <384F4720.358162F3@dynamicsoft.com> <384FF91C.DC9E7EE7@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:
> 
> > In any case, determination of where the NPDB query is done is something
> > I don't think is within the scope of iptel to determine. All we need to
> > figure out is what kind of numbers are distributed in TRIP. Near as I
> > can tell, since LRNs still refer to the prefix of numbers served by a
> > switch, doing what we were doing all along works just fine.
> >
> 
> Yes what we're doing is OK, because the LRN is part of an office's
> NPA-NXX prefix. But we may have to add a new field to the
> ReachableRoutes attribute to mark a prefix as portable. A prefix is
> marked as portable if at lease one number has been ported out of this
> prefix.

Hmm, why is that? Are you basically saying that if a prefix is marked as
portable, an LS must do an LNP query before routing to numbers in that
prefix? Then, if marked as not portable, there is no need for queries?
If so, its an interesting idea. It solves the problem of N-1, which,
from what I understand, was done in part because only carriers near the
destination would need to install triggers to perform queries on
numbers. However, if information on what prefixes are ported is
distributed in TRIP, this information is effectively dynamically
distributed. This means its equally easy for the query to be done at the
first LS as the last.

-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 14 11:10:54 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 LAA22265
	for <iptel-archive@odin.ietf.org>; Tue, 14 Dec 1999 11:10:54 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id D16F652FB; Tue, 14 Dec 1999 11:00:05 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 1C1D952F7; Tue, 14 Dec 1999 11:00:01 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <38566A7A.5D69155B@dynamicsoft.com>
Date: Tue, 14 Dec 1999 11:04:10 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: Dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: John Mainwaring <crm312a@nortelnetworks.com>
Cc: "Mark D. Foster" <mark.foster@npac.com>,
        iptel@lists.research.bell-labs.com,
        "'James Yu'" <james.yu@neustar.com>, sip@lists.research.bell-labs.com
Subject: Re: TRIP
References: <89866A7EC9C0D1119C950000F8BCC2820157C018@zrtpd00n.us.nortel.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 Mainwaring wrote:
> 
> |The problem is related to past discussions on the SIP mailing list
> |regarding overlap dialing. A terminating gateway would look at
> |sip:1212344, and know how to route it, but might not know whether
> this
> |was a dialed number with more digits coming, or an LRN. Does it
> matter?
> |Not being an ISUP expert, I'm not sure. If it does matter, something
> |like user=lrn within the SIP URL would fix that.
> 
> user=lrn is almost right.  If you have queried a portable number and
> found that it is not ported, then the DN has the same status as an LRN
> would have if the number were ported (i.e. it should be treated as
> authoritative for routing).  user=np-queried might be closer to the
> mark.

I wonder if we can't think about this problem in more general terms.

In the SIP model, each proxy owns some piece of the namespace
(engineering.dynamicsoft.com, for example), and is therefore responsible
for the definition of the user space (jdrosen@) within that domain.
Thats why each proxy can, an usually will, perform some kind of URI
translation independent from other proxies. Our underlying assumption
was that usernames only have meaning within their domain. For example,
with sip:jdrosen@dynamicsoft.com, only a dynamicsoft.com proxy knows how
to handle jdrosen. There might be jdrosen's in other domains that aren't
me. However, in the case of telephone numbers, the proxy that owns the
domain does not necessarily own the user space. sip:5551212@gateway.com,
for example, uses a user name that is not "owned" by gateway.com, in the
sense that this name has valid semantics outside of gateway.com. Other
domains, like wcom.com, would know how to handle this number as well.
Its possible that we might even see things besides phone numbers as user
names that are valid across multiple domains. An example is (heaven
forbid), social security numbers (in the US, at least):

sip:123-45-6789@wcom.com;user=ssn

Now, I might imagine that some kind of mobility services might be
supplied for these names, so they might need to be translated also.

I'll call these type of user names "cross domain user names".

With cross domain user names, this is where we run into issues like "has
the number been queried yet". However, I think we might see other
translation mechanisms besides LNP. Soooo, might it not be useful to
define URI parameters that indicate what global translation services
have been applied to a cross domain user name?

As an example:

sip:15551212@cisco.com;user=phone;trans="lnp,gdbs"

where lnp and gdbs are some kind of global translation services that
apply to names of the type phone.

For cross domain user names, we will always need a URI parameter to
indicate the global namespace from which these names are drawn. In the
case of phone numbers, thats user=phone. Our original intent with the
user=phone parameter was to specify only the type of BNF used for the
user part of the URI. But, the more general operation of specifying both
the format and the namespace seems useful.


> 
> It's also necessary to define what the TRIP should do differently
> depending on whether the np-queried indication is present.  I must
> admit I came into this discussion mainly because I have worked on
> LNP.  I assume the objective is not to route unqueried calls to the
> donor proxy. 

Hmm. This is something thats also different from the PSTN. The cost of
routing to the donor proxy in SIP is much less, since proxies can be
stateless. Thus, if a proxy receives a call for some number that is
ported, it stateless proxies the call towards the next destination. The
result is that there is no triangle routing for media (there generally
isn't), nor for subsequent signaling. There is only marginal cost for
the donor provider. There are obviously polticial and business issues;
one may not want to trust or rely upon the donor proxy at all, as is the
case of LNP in the US. But, the general SIP model is that we assume
proxies do route calls; no security measures can prevent a proxy from
dropping messages. Why, in the case of LNP, do we now need to assume
these proxies won't route calls?




 I have noted that originator query may not be practical
> for nationwide calling and that donor proxy is too late unless you're
> prepared to consider RTP. 

The assumption of the impractability of originator queries has
everything to do with the existing structure of LNP. Its mainly because,
AFAIK, the systems are different all over the world. However, if we use
the IP model, with a DNS solution, its no harder to query for a number
on the other side of the planet as it would be for a number next door.

Also, pardon my ignorance, but what is RTP? Many of us on this list know
RTP as the Real Time Transport Protocol (rfc1889) but I don't think
thats what you mean.


 One answer might be for the TRIP to provide
> a response on unqueried numbers that routes to an intermediate server
> that does know how to do NP queries for the destination DN.

TRIP is not a query protocol. Its a routing protocol. This would be a
SIP function, and is already possible. "408 Not Found".

-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 14 12:00:26 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 MAA24047
	for <iptel-archive@odin.ietf.org>; Tue, 14 Dec 1999 12:00:25 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C7AD052FA; Tue, 14 Dec 1999 11:49:55 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 0178E52FC; Tue, 14 Dec 1999 11:49:54 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C3BE93B@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>, hsalama@cisco.com
Cc: iptel@lists.research.bell-labs.com, sip@lists.research.bell-labs.com
Subject: RE: TRIP
Date: Tue, 14 Dec 1999 10:41:41 -0600
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

As indicated by Mark Foster and me that N-1 really does not mean that N-1
network is the best place to perform N-1.  It is simply something like "if
nobody else has performed the NPDB query, you, the second to the last
network, should perform the NPDB query."  This will prevent the call from
being routed to a donor network/switch if the number has been ported out of
the donor switch.  In the U.S., N-1 network implies the long distance
carrier for a long distance call.  For a local call that may involves two
LECs, the originating LEC would be the N-1 network.

One reason is that the "long distance" carrier handles a long distance call
(charges the caller) in the U.S.  If the originating network is to perform
the NPDB query, the long distance carrier has to pay the originating network
for performing the query (through the business agreement).  Another reason
is that in the U.S. NPDB administration is divided into seven regions which
are aligned with the Regional Bell Operating Company (RBOC) regions (e.g.,
NYNEX, Bell Atlantic, US West, etc.).  A NPDB in one region only contains
the directory numbers in that region.  There can be many NPDBs in one
region.  A carrier may decide to deploy its own NPDB or may decide to use
another company's NPDB (paying for NPDB dips).  The originating network in
one region typically only access the NPDB in the same region.  To access a
NPDB in another region requires business agreement between the originating
network and the NPDB and there may be higher SS7 message transport cost.

I agree with the proposal of adding an indication whether a number range has
been ported or not because it would reduce the traffic into the NPDB or an
equivalent entity.  But not all the LSs are required to use the information
to perform the NPDB/DNS query.  The indicator would be useful for an entity
that is to perform the query (e.g., the first LS can route the call to
another LS where query is performed).  As discussed before, the NPDB dip
indicator may need to be carried in the SIP if ISUP is not encapsulated.
That indicator plus this proposed indicator will tell a LS that is to
perform the NPDB-like query whether it should perform such a query.

Since several GWs will be able to terminate SCN-termianted calls, it may be
useful to include antoher indicator that indicates the number of SCN switch
hops between the terminating GW and the serving SCN switch of the called
number if we don't plan to rely on the provisioning process to provide the
information.  The information would allow an LS to determine the GW that
provides the most efficient route to the SCN serving switch of the called
number.  This does not rule out the case where business agreement determines
the the selected GW.  If we want to provide more information, we may need to
indicate the number of non-SS7 signaling routes in addition to the number of
hops.   The intention of the indicator(s) is to provide some indication
about the post dialing delay contributed by the terminating route.  Those
indicators are not needed if contribution to the post dialing delay by the
terminating route is not a concern (e.g., most routes have similar post
dialing delay).

We may also want to consider the pricing information for a route/number
ranges if we don't plan to rely on business agreement to determine the GW.
The pricing info. is about the charge to be paid to the terminating GW for
accepting a call and routing it to the SCN terminating switch.  We may need
two types of pricing info. depending on whether NPDB query has been
performed or not.  Please ignore this if this issue has been debated before
and closed.

James

-----Original Message-----
From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]
Sent: Tuesday, December 14, 1999 10:43 AM
To: hsalama@cisco.com
Cc: Mark D. Foster; 'John Mainwaring';
iptel@lists.research.bell-labs.com; 'James Yu';
sip@lists.research.bell-labs.com
Subject: Re: TRIP




"Hussein F. Salama" wrote:
> 
> > In any case, determination of where the NPDB query is done is something
> > I don't think is within the scope of iptel to determine. All we need to
> > figure out is what kind of numbers are distributed in TRIP. Near as I
> > can tell, since LRNs still refer to the prefix of numbers served by a
> > switch, doing what we were doing all along works just fine.
> >
> 
> Yes what we're doing is OK, because the LRN is part of an office's
> NPA-NXX prefix. But we may have to add a new field to the
> ReachableRoutes attribute to mark a prefix as portable. A prefix is
> marked as portable if at lease one number has been ported out of this
> prefix.

Hmm, why is that? Are you basically saying that if a prefix is marked as
portable, an LS must do an LNP query before routing to numbers in that
prefix? Then, if marked as not portable, there is no need for queries?
If so, its an interesting idea. It solves the problem of N-1, which,
from what I understand, was done in part because only carriers near the
destination would need to install triggers to perform queries on
numbers. However, if information on what prefixes are ported is
distributed in TRIP, this information is effectively dynamically
distributed. This means its equally easy for the query to be done at the
first LS as the last.

-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 14 12:18:51 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 MAA24657
	for <iptel-archive@odin.ietf.org>; Tue, 14 Dec 1999 12:18:50 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6AF6F52FE; Tue, 14 Dec 1999 12:07:49 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8059952FC; Tue, 14 Dec 1999 12:07:44 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <ED88182BFF78D211A4D800A0C9E9435C3BE93C@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>,
        John Mainwaring <crm312a@nortelnetworks.com>
Cc: Mark Foster <mark.foster@neustar.com>, iptel@lists.research.bell-labs.com,
        sip@lists.research.bell-labs.com
Subject: RE: TRIP
Date: Tue, 14 Dec 1999 10:58:31 -0600
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

For overlap signaling, a GW or LS that receives a call with partial address
information may not know that the address info. is complete.  If it launches
a query (DNS or NPDB), the entity that has the "number portability" info.
won't be able to find the info. because the address info. is incomplete.  It
seems that either this LS or GW may wait a little while to see whether there
are subsequent address info. before launching the query, or it can launch a
query (DNS or NPDB-like), receives an error (hope that it is specific such
as incomplete address instead of something ambiguous), and then wait for
additional address info. before timing out.

James
-----Original Message-----
From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]
Sent: Tuesday, December 14, 1999 11:04 AM
To: John Mainwaring
Cc: Mark D. Foster; iptel@lists.research.bell-labs.com; 'James Yu';
sip@lists.research.bell-labs.com
Subject: Re: TRIP




> John Mainwaring wrote:
> 
> |The problem is related to past discussions on the SIP mailing list
> |regarding overlap dialing. A terminating gateway would look at
> |sip:1212344, and know how to route it, but might not know whether
> this
> |was a dialed number with more digits coming, or an LRN. Does it
> matter?
> |Not being an ISUP expert, I'm not sure. If it does matter, something
> |like user=lrn within the SIP URL would fix that.
> 
> user=lrn is almost right.  If you have queried a portable number and
> found that it is not ported, then the DN has the same status as an LRN
> would have if the number were ported (i.e. it should be treated as
> authoritative for routing).  user=np-queried might be closer to the
> mark.

I wonder if we can't think about this problem in more general terms.

In the SIP model, each proxy owns some piece of the namespace
(engineering.dynamicsoft.com, for example), and is therefore responsible
for the definition of the user space (jdrosen@) within that domain.
Thats why each proxy can, an usually will, perform some kind of URI
translation independent from other proxies. Our underlying assumption
was that usernames only have meaning within their domain. For example,
with sip:jdrosen@dynamicsoft.com, only a dynamicsoft.com proxy knows how
to handle jdrosen. There might be jdrosen's in other domains that aren't
me. However, in the case of telephone numbers, the proxy that owns the
domain does not necessarily own the user space. sip:5551212@gateway.com,
for example, uses a user name that is not "owned" by gateway.com, in the
sense that this name has valid semantics outside of gateway.com. Other
domains, like wcom.com, would know how to handle this number as well.
Its possible that we might even see things besides phone numbers as user
names that are valid across multiple domains. An example is (heaven
forbid), social security numbers (in the US, at least):

sip:123-45-6789@wcom.com;user=ssn

Now, I might imagine that some kind of mobility services might be
supplied for these names, so they might need to be translated also.

I'll call these type of user names "cross domain user names".

With cross domain user names, this is where we run into issues like "has
the number been queried yet". However, I think we might see other
translation mechanisms besides LNP. Soooo, might it not be useful to
define URI parameters that indicate what global translation services
have been applied to a cross domain user name?

As an example:

sip:15551212@cisco.com;user=phone;trans="lnp,gdbs"

where lnp and gdbs are some kind of global translation services that
apply to names of the type phone.

For cross domain user names, we will always need a URI parameter to
indicate the global namespace from which these names are drawn. In the
case of phone numbers, thats user=phone. Our original intent with the
user=phone parameter was to specify only the type of BNF used for the
user part of the URI. But, the more general operation of specifying both
the format and the namespace seems useful.


> 
> It's also necessary to define what the TRIP should do differently
> depending on whether the np-queried indication is present.  I must
> admit I came into this discussion mainly because I have worked on
> LNP.  I assume the objective is not to route unqueried calls to the
> donor proxy. 

Hmm. This is something thats also different from the PSTN. The cost of
routing to the donor proxy in SIP is much less, since proxies can be
stateless. Thus, if a proxy receives a call for some number that is
ported, it stateless proxies the call towards the next destination. The
result is that there is no triangle routing for media (there generally
isn't), nor for subsequent signaling. There is only marginal cost for
the donor provider. There are obviously polticial and business issues;
one may not want to trust or rely upon the donor proxy at all, as is the
case of LNP in the US. But, the general SIP model is that we assume
proxies do route calls; no security measures can prevent a proxy from
dropping messages. Why, in the case of LNP, do we now need to assume
these proxies won't route calls?




 I have noted that originator query may not be practical
> for nationwide calling and that donor proxy is too late unless you're
> prepared to consider RTP. 

The assumption of the impractability of originator queries has
everything to do with the existing structure of LNP. Its mainly because,
AFAIK, the systems are different all over the world. However, if we use
the IP model, with a DNS solution, its no harder to query for a number
on the other side of the planet as it would be for a number next door.

Also, pardon my ignorance, but what is RTP? Many of us on this list know
RTP as the Real Time Transport Protocol (rfc1889) but I don't think
thats what you mean.


 One answer might be for the TRIP to provide
> a response on unqueried numbers that routes to an intermediate server
> that does know how to do NP queries for the destination DN.

TRIP is not a query protocol. Its a routing protocol. This would be a
SIP function, and is already possible. "408 Not Found".

-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 14 15:15:17 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 PAA28935
	for <iptel-archive@odin.ietf.org>; Tue, 14 Dec 1999 15:15:11 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 6FF6B52DE; Tue, 14 Dec 1999 15:04:09 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C290552DB; Tue, 14 Dec 1999 15:04:03 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Tue, 14 Dec 1999 13:55:48 -0600
From: Dean Willis <dean.willis@wcom.com>
Subject: URL Generalization for Portability (was TRIP)
In-reply-to: <38566A7A.5D69155B@dynamicsoft.com>
To: Jonathan Rosenberg <jdrosen@dynamicsoft.com>,
        John Mainwaring <crm312a@nortelnetworks.com>
Cc: "Mark D. Foster" <mark.foster@npac.com>,
        iptel@lists.research.bell-labs.com,
        "'James Yu'" <james.yu@neustar.com>, sip@lists.research.bell-labs.com
Message-id: <001e01bf466d$373d3140$d19623a6@mcit.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Content-type: text/plain;	charset="iso-8859-1"
Content-transfer-encoding: 7bit
Importance: Normal
X-Priority: 3 (Normal)
X-MSMail-priority: Normal
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit


Jonathan Rosenburg wrote:
> I wonder if we can't think about this problem in more general terms.

yes, please!


> In the SIP model, each proxy owns some piece of the namespace
> (engineering.dynamicsoft.com, for example), and is therefore responsible
> for the definition of the user space (jdrosen@) within that domain.
> Thats why each proxy can, an usually will, perform some kind of URI
> translation independent from other proxies. Our underlying assumption
> was that usernames only have meaning within their domain. For example,
> with sip:jdrosen@dynamicsoft.com, only a dynamicsoft.com proxy knows how
> to handle jdrosen. There might be jdrosen's in other domains that aren't
> me. However, in the case of telephone numbers, the proxy that owns the
> domain does not necessarily own the user space. sip:5551212@gateway.com,
> for example, uses a user name that is not "owned" by gateway.com, in the
> sense that this name has valid semantics outside of gateway.com. Other
> domains, like wcom.com, would know how to handle this number as well.
> Its possible that we might even see things besides phone numbers as user
> names that are valid across multiple domains. An example is (heaven
> forbid), social security numbers (in the US, at least):

The underlying problem here is differentiating between a phone-number and a
service-name-which-looks-like-a-number. Yes, the user=phone gives a hint on
nature od address, the "+" encodding gives a hint on the scope of address,
but it is still muddy.

There was some discussion of this problem wrt DCS during the November
interim meeting. It was suggested that a cross-domain destination which is a
phone number could be better represented by a tel: URL than by confusing the
scope of the user part of a sip: URL. This simplilifies life greatly.

Something like

	INVITE premiumservice@gateway.carrier.com
	To: tel:5551212;trans=lnp?gdbs

could be used to indicate to gateway.carrier.com that premium gateway
service is requested towards the PSTN location identified by telephone
number 5551212, that LNP and GDBS translation have been applied, and that
the phone number is of local scope for the gateway. Incidentally, I think I
just noticed a minor anomaly in the tel-url grammar -- don't see how to
encode more than one future-extension field . . .

--
Dean


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 14 15:53:23 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 PAA00128
	for <iptel-archive@odin.ietf.org>; Tue, 14 Dec 1999 15:53:21 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5E42352F7; Tue, 14 Dec 1999 15:40:03 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id C16F352DB; Tue, 14 Dec 1999 15:39:57 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <89866A7EC9C0D1119C950000F8BCC2820157C02B@zrtpd00n.us.nortel.com>
From: "John Mainwaring" <crm312a@nortelnetworks.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>
Cc: "Mark D. Foster" <mark.foster@npac.com>,
        iptel@lists.research.bell-labs.com,
        "'James Yu'" <james.yu@neustar.com>, sip@lists.research.bell-labs.com
Subject: RE: TRIP
Date: Tue, 14 Dec 1999 15:38:03 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Type: multipart/alternative;
              boundary="----_=_NextPart_001_01BF4673.2B7063F4"
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_01BF4673.2B7063F4
Content-Type: text/plain;
	charset="windows-1252"



|-----Original Message-----
|From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]
|
|> John Mainwaring wrote:
|> 
|> |The problem is related to past discussions on the SIP mailing list
|> |regarding overlap dialing. A terminating gateway would look at
|> |sip:1212344, and know how to route it, but might not know whether
|> this
|> |was a dialed number with more digits coming, or an LRN. Does it
|> matter?
|> |Not being an ISUP expert, I'm not sure. If it does matter, something
|> |like user=lrn within the SIP URL would fix that.
|> 
|> user=lrn is almost right.  If you have queried a portable number and
|> found that it is not ported, then the DN has the same status 
|as an LRN
|> would have if the number were ported (i.e. it should be treated as
|> authoritative for routing).  user=np-queried might be closer to the
|> mark.
|
|I wonder if we can't think about this problem in more general terms.
|
|In the SIP model, each proxy owns some piece of the namespace
|(engineering.dynamicsoft.com, for example), and is therefore 
|responsible
|for the definition of the user space (jdrosen@) within that domain.
|Thats why each proxy can, an usually will, perform some kind of URI
|translation independent from other proxies. Our underlying assumption
|was that usernames only have meaning within their domain. For example,
|with sip:jdrosen@dynamicsoft.com, only a dynamicsoft.com proxy 
|knows how
|to handle jdrosen. There might be jdrosen's in other domains 
|that aren't
|me. However, in the case of telephone numbers, the proxy that owns the
|domain does not necessarily own the user space. 
|sip:5551212@gateway.com,
|for example, uses a user name that is not "owned" by 
|gateway.com, in the
|sense that this name has valid semantics outside of gateway.com. Other
|domains, like wcom.com, would know how to handle this number as well.
|Its possible that we might even see things besides phone 
|numbers as user
|names that are valid across multiple domains. An example is (heaven
|forbid), social security numbers (in the US, at least):
|
|sip:123-45-6789@wcom.com;user=ssn
|
|Now, I might imagine that some kind of mobility services might be
|supplied for these names, so they might need to be translated also.
|
|I'll call these type of user names "cross domain user names".
|
|With cross domain user names, this is where we run into issues 
|like "has
|the number been queried yet". However, I think we might see other
|translation mechanisms besides LNP. Soooo, might it not be useful to
|define URI parameters that indicate what global translation services
|have been applied to a cross domain user name?
|
|As an example:
|
|sip:15551212@cisco.com;user=phone;trans="lnp,gdbs"
|
|where lnp and gdbs are some kind of global translation services that
|apply to names of the type phone.
|
|For cross domain user names, we will always need a URI parameter to
|indicate the global namespace from which these names are drawn. In the
|case of phone numbers, thats user=phone. Our original intent with the
|user=phone parameter was to specify only the type of BNF used for the
|user part of the URI. But, the more general operation of 
|specifying both
|the format and the namespace seems useful.

I think the problem is a bit different.  919-555-1212 uniquely identifies a
user within the North American Numbering Plan.  It's just that with LNP it
can no longer locate the user.  If someone wants to use
919-555-1212@vanity.org as an email address and someone else uses
919-555-1212@workplace.com, that's fine, but has nothing to do with
telephone numbers.  I think I want to be able to place a call when all I
know is 919-555-1212.

Without LNP, it would have been fairly easy to  discover that 919-555
implied bellsouth, so the number could have been given as
919-555-1212@bellsouth.com.  Now it is possible that some numbers from
919-555 might be ported to othertel.com.  Othertel might the native owner of
919-444-xxxx numbers and might choose to host 919-555-1313 on their 919-444
exchange.  In that case, the LRN for 919-555-1313 would be 919-444.  Once I
have done an LNP query, I might want to use something like
919-555-1212@919-555.bellsouth.com for a number that has not been ported and
919-555-1313@919-444.othertel.com for a number that has been ported to
othertel.

I can't shorten the 10 digit number in the user name, because of things like
NPA overlays and switches that host multiple office codes.  I need the full
address to select the right user as well as the LRN to route to the right
place. 

If a SIP user wants to call someone who uses a PSTN phone, the system needs
to deal with certain inconvenient aspects of the PSTN. A notation like
919-555-1212@919-555 or 919-555-1313@919-444 might be useful to indicate
that an LNP query has been done, and that authoritative routing information
is now available.  Appending text-based names such as
919-555-1212@919-555@onetel.com or 919-555-1313@919-444@anothertel.com might
even allow subsequent routing to take place without further concessions to
PSTN quirkiness.

Of course this system could also be expanded to mobility services; routing
for 919-555-1313 could be set to 212-666-1414@212-666.bigappletel.com in the
daytime and 919-555-1313@919-444.othertel.com in the evening if you went in
for that kind of commuting.

|> It's also necessary to define what the TRIP should do differently
|> depending on whether the np-queried indication is present.  I must
|> admit I came into this discussion mainly because I have worked on
|> LNP.  I assume the objective is not to route unqueried calls to the
|> donor proxy. 
|
|Hmm. This is something thats also different from the PSTN. The cost of
|routing to the donor proxy in SIP is much less, since proxies can be
|stateless. Thus, if a proxy receives a call for some number that is
|ported, it stateless proxies the call towards the next destination. The
|result is that there is no triangle routing for media (there generally
|isn't), nor for subsequent signaling. There is only marginal cost for
|the donor provider. There are obviously polticial and business issues;
|one may not want to trust or rely upon the donor proxy at all, 
|as is the
|case of LNP in the US. But, the general SIP model is that we assume
|proxies do route calls; no security measures can prevent a proxy from
|dropping messages. Why, in the case of LNP, do we now need to assume
|these proxies won't route calls?
|
|
|
|
| I have noted that originator query may not be practical
|> for nationwide calling and that donor proxy is too late unless you're
|> prepared to consider RTP. 
|
|The assumption of the impractability of originator queries has
|everything to do with the existing structure of LNP. Its 
|mainly because,
|AFAIK, the systems are different all over the world. However, if we use
|the IP model, with a DNS solution, its no harder to query for a number
|on the other side of the planet as it would be for a number next door.

No, it's not quite as simple as that.  The system is designed around the
strengths and weaknesses of the PSTN.  The the architecture of the pstn
makes it easy to route calls on partial information, but switches in the
pstn perform better if they do as few database queries as possible.
Switches work fine if they have no idea of the portability status of most
numbers; they are only set up to query nearby numbers.  A switch doesn't
need authoritative data that a number is not portable.  It just needs to be
set up to query that (local) subset of portable numbers which it would
misroute if it didn't query them.

The LNP databases are distributed because of the geographic restrictions on
Bell Operating Companies.  Each RBOC only has to worry about its own data,
and only has to route queries within its own SS7 signalling network.

There are actually two levels of database, the SCPs that actually handle
queries and the NPAC (?) which processes administrative change requests and
downloads the SCPs.  There could be several SCPs owned by different
operating companies that contain the same data; each operating company would
route queries to its own SCP.  The NPAC service is provided by NeuStar,
formerly Lockheed Martin, and Mark Foster of Neustar has given us some
excellent information on LNP within this thread.

It seems unlikely that IP telephony would use the existing SCPs for queries,
or that it would be based on SS7 TCAP messaging.  I don't have a clear idea
of what databases there would be for IP telephony, but they would certainly
have to arrange to download data from the NPAC.  There are costs associated
with the databases and with keeping the LNP information up to date.

If you want to do source routing, it will certainly take a non-trivial
effort to make North American or even world wide LNP information accessible.
There would be a case for aggregating it with whatever centralized routing
servers there are, or for arranging for the destination's proxy to return
information about ported numbers instead of simply refusing the invitation.

|Also, pardon my ignorance, but what is RTP? Many of us on this 
|list know
|RTP as the Real Time Transport Protocol (rfc1889) but I don't think
|thats what you mean.

In that context, RTP = Release To Pivot; switch A routes a call to switch B
which may provide some service and then releases the call back to A (the
pivot switch) with instructions to reroute the call to destination C.
Sorry, I thought I had expanded it earlier.

Is there any TLA in our businsess that only has one expansion?  What does
TLA mean apart from Three Letter Acronym?

| One answer might be for the TRIP to provide
|> a response on unqueried numbers that routes to an intermediate server
|> that does know how to do NP queries for the destination DN.
|
|TRIP is not a query protocol. Its a routing protocol. This would be a
|SIP function, and is already possible. "408 Not Found".

But why was it not found?  In the telephony e.164 numbering domain prior to
LNP, "not found" meant "does not exist".  LNP required switches to add
"might be somewhere else" logic.  It would be desirable for a routing agent
to say something more informative than "not found", if  it had any
additional information.

------_=_NextPart_001_01BF4673.2B7063F4
Content-Type: text/html;
	charset="windows-1252"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=windows-1252">
<META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2448.0">
<TITLE>RE: TRIP</TITLE>
</HEAD>
<BODY>
<BR>
<BR>

<P><FONT SIZE=2>|-----Original Message-----</FONT>
<BR><FONT SIZE=2>|From: Jonathan Rosenberg [<A HREF="mailto:jdrosen@dynamicsoft.com">mailto:jdrosen@dynamicsoft.com</A>]</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|&gt; John Mainwaring wrote:</FONT>
<BR><FONT SIZE=2>|&gt; </FONT>
<BR><FONT SIZE=2>|&gt; |The problem is related to past discussions on the SIP mailing list</FONT>
<BR><FONT SIZE=2>|&gt; |regarding overlap dialing. A terminating gateway would look at</FONT>
<BR><FONT SIZE=2>|&gt; |sip:1212344, and know how to route it, but might not know whether</FONT>
<BR><FONT SIZE=2>|&gt; this</FONT>
<BR><FONT SIZE=2>|&gt; |was a dialed number with more digits coming, or an LRN. Does it</FONT>
<BR><FONT SIZE=2>|&gt; matter?</FONT>
<BR><FONT SIZE=2>|&gt; |Not being an ISUP expert, I'm not sure. If it does matter, something</FONT>
<BR><FONT SIZE=2>|&gt; |like user=lrn within the SIP URL would fix that.</FONT>
<BR><FONT SIZE=2>|&gt; </FONT>
<BR><FONT SIZE=2>|&gt; user=lrn is almost right.&nbsp; If you have queried a portable number and</FONT>
<BR><FONT SIZE=2>|&gt; found that it is not ported, then the DN has the same status </FONT>
<BR><FONT SIZE=2>|as an LRN</FONT>
<BR><FONT SIZE=2>|&gt; would have if the number were ported (i.e. it should be treated as</FONT>
<BR><FONT SIZE=2>|&gt; authoritative for routing).&nbsp; user=np-queried might be closer to the</FONT>
<BR><FONT SIZE=2>|&gt; mark.</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|I wonder if we can't think about this problem in more general terms.</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|In the SIP model, each proxy owns some piece of the namespace</FONT>
<BR><FONT SIZE=2>|(engineering.dynamicsoft.com, for example), and is therefore </FONT>
<BR><FONT SIZE=2>|responsible</FONT>
<BR><FONT SIZE=2>|for the definition of the user space (jdrosen@) within that domain.</FONT>
<BR><FONT SIZE=2>|Thats why each proxy can, an usually will, perform some kind of URI</FONT>
<BR><FONT SIZE=2>|translation independent from other proxies. Our underlying assumption</FONT>
<BR><FONT SIZE=2>|was that usernames only have meaning within their domain. For example,</FONT>
<BR><FONT SIZE=2>|with sip:jdrosen@dynamicsoft.com, only a dynamicsoft.com proxy </FONT>
<BR><FONT SIZE=2>|knows how</FONT>
<BR><FONT SIZE=2>|to handle jdrosen. There might be jdrosen's in other domains </FONT>
<BR><FONT SIZE=2>|that aren't</FONT>
<BR><FONT SIZE=2>|me. However, in the case of telephone numbers, the proxy that owns the</FONT>
<BR><FONT SIZE=2>|domain does not necessarily own the user space. </FONT>
<BR><FONT SIZE=2>|sip:5551212@gateway.com,</FONT>
<BR><FONT SIZE=2>|for example, uses a user name that is not &quot;owned&quot; by </FONT>
<BR><FONT SIZE=2>|gateway.com, in the</FONT>
<BR><FONT SIZE=2>|sense that this name has valid semantics outside of gateway.com. Other</FONT>
<BR><FONT SIZE=2>|domains, like wcom.com, would know how to handle this number as well.</FONT>
<BR><FONT SIZE=2>|Its possible that we might even see things besides phone </FONT>
<BR><FONT SIZE=2>|numbers as user</FONT>
<BR><FONT SIZE=2>|names that are valid across multiple domains. An example is (heaven</FONT>
<BR><FONT SIZE=2>|forbid), social security numbers (in the US, at least):</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|sip:123-45-6789@wcom.com;user=ssn</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|Now, I might imagine that some kind of mobility services might be</FONT>
<BR><FONT SIZE=2>|supplied for these names, so they might need to be translated also.</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|I'll call these type of user names &quot;cross domain user names&quot;.</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|With cross domain user names, this is where we run into issues </FONT>
<BR><FONT SIZE=2>|like &quot;has</FONT>
<BR><FONT SIZE=2>|the number been queried yet&quot;. However, I think we might see other</FONT>
<BR><FONT SIZE=2>|translation mechanisms besides LNP. Soooo, might it not be useful to</FONT>
<BR><FONT SIZE=2>|define URI parameters that indicate what global translation services</FONT>
<BR><FONT SIZE=2>|have been applied to a cross domain user name?</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|As an example:</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|sip:15551212@cisco.com;user=phone;trans=&quot;lnp,gdbs&quot;</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|where lnp and gdbs are some kind of global translation services that</FONT>
<BR><FONT SIZE=2>|apply to names of the type phone.</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|For cross domain user names, we will always need a URI parameter to</FONT>
<BR><FONT SIZE=2>|indicate the global namespace from which these names are drawn. In the</FONT>
<BR><FONT SIZE=2>|case of phone numbers, thats user=phone. Our original intent with the</FONT>
<BR><FONT SIZE=2>|user=phone parameter was to specify only the type of BNF used for the</FONT>
<BR><FONT SIZE=2>|user part of the URI. But, the more general operation of </FONT>
<BR><FONT SIZE=2>|specifying both</FONT>
<BR><FONT SIZE=2>|the format and the namespace seems useful.</FONT>
</P>

<P><FONT SIZE=2>I think the problem is a bit different.&nbsp; 919-555-1212 uniquely identifies a user within the North American Numbering Plan.&nbsp; It's just that with LNP it can no longer locate the user.&nbsp; If someone wants to use 919-555-1212@vanity.org as an email address and someone else uses 919-555-1212@workplace.com, that's fine, but has nothing to do with telephone numbers.&nbsp; I think I want to be able to place a call when all I know is 919-555-1212.</FONT></P>

<P><FONT SIZE=2>Without LNP, it would have been fairly easy to&nbsp; discover that 919-555 implied bellsouth, so the number could have been given as 919-555-1212@bellsouth.com.&nbsp; Now it is possible that some numbers from 919-555 might be ported to othertel.com.&nbsp; Othertel might the native owner of 919-444-xxxx numbers and might choose to host 919-555-1313 on their 919-444 exchange.&nbsp; In that case, the LRN for 919-555-1313 would be 919-444.&nbsp; Once I have done an LNP query, I might want to use something like 919-555-1212@919-555.bellsouth.com for a number that has not been ported and&nbsp; 919-555-1313@919-444.othertel.com for a number that has been ported to othertel.</FONT></P>

<P><FONT SIZE=2>I can't shorten the 10 digit number in the user name, because of things like NPA overlays and switches that host multiple office codes.&nbsp; I need the full address to select the right user as well as the LRN to route to the right place. </FONT></P>

<P><FONT SIZE=2>If a SIP user wants to call someone who uses a PSTN phone, the system needs to deal with certain inconvenient aspects of the PSTN. A notation like 919-555-1212@919-555 or 919-555-1313@919-444 might be useful to indicate that an LNP query has been done, and that authoritative routing information is now available.&nbsp; Appending text-based names such as 919-555-1212@919-555@onetel.com or 919-555-1313@919-444@anothertel.com might even allow subsequent routing to take place without further concessions to PSTN quirkiness.</FONT></P>

<P><FONT SIZE=2>Of course this system could also be expanded to mobility services; routing for 919-555-1313 could be set to 212-666-1414@212-666.bigappletel.com in the daytime and 919-555-1313@919-444.othertel.com in the evening if you went in for that kind of commuting.</FONT></P>

<P><FONT SIZE=2>|&gt; It's also necessary to define what the TRIP should do differently</FONT>
<BR><FONT SIZE=2>|&gt; depending on whether the np-queried indication is present.&nbsp; I must</FONT>
<BR><FONT SIZE=2>|&gt; admit I came into this discussion mainly because I have worked on</FONT>
<BR><FONT SIZE=2>|&gt; LNP.&nbsp; I assume the objective is not to route unqueried calls to the</FONT>
<BR><FONT SIZE=2>|&gt; donor proxy. </FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|Hmm. This is something thats also different from the PSTN. The cost of</FONT>
<BR><FONT SIZE=2>|routing to the donor proxy in SIP is much less, since proxies can be</FONT>
<BR><FONT SIZE=2>|stateless. Thus, if a proxy receives a call for some number that is</FONT>
<BR><FONT SIZE=2>|ported, it stateless proxies the call towards the next destination. The</FONT>
<BR><FONT SIZE=2>|result is that there is no triangle routing for media (there generally</FONT>
<BR><FONT SIZE=2>|isn't), nor for subsequent signaling. There is only marginal cost for</FONT>
<BR><FONT SIZE=2>|the donor provider. There are obviously polticial and business issues;</FONT>
<BR><FONT SIZE=2>|one may not want to trust or rely upon the donor proxy at all, </FONT>
<BR><FONT SIZE=2>|as is the</FONT>
<BR><FONT SIZE=2>|case of LNP in the US. But, the general SIP model is that we assume</FONT>
<BR><FONT SIZE=2>|proxies do route calls; no security measures can prevent a proxy from</FONT>
<BR><FONT SIZE=2>|dropping messages. Why, in the case of LNP, do we now need to assume</FONT>
<BR><FONT SIZE=2>|these proxies won't route calls?</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>| I have noted that originator query may not be practical</FONT>
<BR><FONT SIZE=2>|&gt; for nationwide calling and that donor proxy is too late unless you're</FONT>
<BR><FONT SIZE=2>|&gt; prepared to consider RTP. </FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|The assumption of the impractability of originator queries has</FONT>
<BR><FONT SIZE=2>|everything to do with the existing structure of LNP. Its </FONT>
<BR><FONT SIZE=2>|mainly because,</FONT>
<BR><FONT SIZE=2>|AFAIK, the systems are different all over the world. However, if we use</FONT>
<BR><FONT SIZE=2>|the IP model, with a DNS solution, its no harder to query for a number</FONT>
<BR><FONT SIZE=2>|on the other side of the planet as it would be for a number next door.</FONT>
</P>

<P><FONT SIZE=2>No, it's not quite as simple as that.&nbsp; The system is designed around the strengths and weaknesses of the PSTN.&nbsp; The the architecture of the pstn makes it easy to route calls on partial information, but switches in the pstn perform better if they do as few database queries as possible.&nbsp; Switches work fine if they have no idea of the portability status of most numbers; they are only set up to query nearby numbers.&nbsp; A switch doesn't need authoritative data that a number is not portable.&nbsp; It just needs to be set up to query that (local) subset of portable numbers which it would misroute if it didn't query them.</FONT></P>

<P><FONT SIZE=2>The LNP databases are distributed because of the geographic restrictions on Bell Operating Companies.&nbsp; Each RBOC only has to worry about its own data, and only has to route queries within its own SS7 signalling network.</FONT></P>

<P><FONT SIZE=2>There are actually two levels of database, the SCPs that actually handle queries and the NPAC (?) which processes administrative change requests and downloads the SCPs.&nbsp; There could be several SCPs owned by different operating companies that contain the same data; each operating company would route queries to its own SCP.&nbsp; The NPAC service is provided by NeuStar, formerly Lockheed Martin, and Mark Foster of Neustar has given us some excellent information on LNP within this thread.</FONT></P>

<P><FONT SIZE=2>It seems unlikely that IP telephony would use the existing SCPs for queries, or that it would be based on SS7 TCAP messaging.&nbsp; I don't have a clear idea of what databases there would be for IP telephony, but they would certainly have to arrange to download data from the NPAC.&nbsp; There are costs associated with the databases and with keeping the LNP information up to date.</FONT></P>

<P><FONT SIZE=2>If you want to do source routing, it will certainly take a non-trivial effort to make North American or even world wide LNP information accessible.&nbsp; There would be a case for aggregating it with whatever centralized routing servers there are, or for arranging for the destination's proxy to return information about ported numbers instead of simply refusing the invitation.</FONT></P>

<P><FONT SIZE=2>|Also, pardon my ignorance, but what is RTP? Many of us on this </FONT>
<BR><FONT SIZE=2>|list know</FONT>
<BR><FONT SIZE=2>|RTP as the Real Time Transport Protocol (rfc1889) but I don't think</FONT>
<BR><FONT SIZE=2>|thats what you mean.</FONT>
</P>

<P><FONT SIZE=2>In that context, RTP = Release To Pivot; switch A routes a call to switch B which may provide some service and then releases the call back to A (the pivot switch) with instructions to reroute the call to destination C.&nbsp; Sorry, I thought I had expanded it earlier.</FONT></P>

<P><FONT SIZE=2>Is there any TLA in our businsess that only has one expansion?&nbsp; What does TLA mean apart from Three Letter Acronym?</FONT>
</P>

<P><FONT SIZE=2>| One answer might be for the TRIP to provide</FONT>
<BR><FONT SIZE=2>|&gt; a response on unqueried numbers that routes to an intermediate server</FONT>
<BR><FONT SIZE=2>|&gt; that does know how to do NP queries for the destination DN.</FONT>
<BR><FONT SIZE=2>|</FONT>
<BR><FONT SIZE=2>|TRIP is not a query protocol. Its a routing protocol. This would be a</FONT>
<BR><FONT SIZE=2>|SIP function, and is already possible. &quot;408 Not Found&quot;.</FONT>
</P>

<P><FONT SIZE=2>But why was it not found?&nbsp; In the telephony e.164 numbering domain prior to LNP, &quot;not found&quot; meant &quot;does not exist&quot;.&nbsp; LNP required switches to add &quot;might be somewhere else&quot; logic.&nbsp; It would be desirable for a routing agent to say something more informative than &quot;not found&quot;, if&nbsp; it had any additional information.</FONT></P>

</BODY>
</HTML>
------_=_NextPart_001_01BF4673.2B7063F4--

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec 15 01:12:12 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 BAA15408
	for <iptel-archive@odin.ietf.org>; Wed, 15 Dec 1999 01:12:12 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id C7D6052FF; Wed, 15 Dec 1999 01:01:53 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 09D285301; Wed, 15 Dec 1999 01:01:52 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Tue, 14 Dec 1999 23:56:32 -0600
From: Dean Willis <dean.willis@wcom.com>
Subject: RE: URL Generalization for Portability (was TRIP)
In-reply-to: <199912142224.QAA06833@b04a24.exu.ericsson.se>
To: "Adam B. Roach" <Adam.Roach@Ericsson.com>
Cc: iptel@lists.research.bell-labs.com, sip@lists.research.bell-labs.com
Message-id: <001b01bf46c1$23940980$d19623a6@directlink.net>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Content-type: text/plain;	charset="iso-8859-1"
Content-transfer-encoding: 7bit
Importance: Normal
X-Priority: 3 (Normal)
X-MSMail-priority: Normal
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit


Adam wrote:
> Dean blurted
> >Something like
> >
> >	INVITE premiumservice@gateway.carrier.com
> >	To: tel:5551212;trans=lnp?gdbs
> >
> >could be used to indicate to gateway.carrier.com that premium gateway
> >service is requested towards the PSTN location identified by telephone
> >number 5551212, that LNP and GDBS translation have been applied, and that
> >the phone number is of local scope for the gateway.
>
> Since the To: field isn't allowed to change during the transaction,
> aren't you going to run into problems if some service along the way
> determines that the called number needs to change (as the result of
> LNP, 800/900 lookup, etc)?

Right you are. That should teach me not to come up with examples while
squatting on one of those interminable telco conference calls.

But I still think using a tel: url gives a cleaner semantic than
sip:number@gateway;user=phone.

I was talking with some guys today about scope and context, and I'm inclined
to think we need the nature of address and numbering plan indicator fields
out of some ISUP variants. Scope appears to be global, local, or private,
with some need to represent local and private within some context like area
code or private numbering plan.

So, for example

tel:+19727294162 is a global number with unspecified context,
tel:9727294162 is a local number
tel:7724162;phone-context=mcivnet34 is a private number with context in the
MCI private numbering plan #34
tel:4512345687724162;phone-context=mciswtrunk is a private number, in
numbering plan 45, relative to switch 123 and trunk group 45678
. . . and whatever other sick private variants one might come up with


--
Dean


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec 15 01:51:32 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 BAA18476
	for <iptel-archive@odin.ietf.org>; Wed, 15 Dec 1999 01:51:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 5955A5304; Wed, 15 Dec 1999 01:47:06 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 8AA355300; Wed, 15 Dec 1999 01:47:03 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199912142224.QAA06833@b04a24.exu.ericsson.se>
Subject: Re: URL Generalization for Portability (was TRIP)
To: dean.willis@wcom.com (Dean Willis)
Date: Tue, 14 Dec 1999 16:24:06 -0600 (CST)
Cc: jdrosen@dynamicsoft.com, crm312a@nortelnetworks.com, mark.foster@npac.com,
        iptel@lists.research.bell-labs.com, james.yu@neustar.com,
        sip@lists.research.bell-labs.com
In-Reply-To: <001e01bf466d$373d3140$d19623a6@mcit.com> from "Dean Willis" at Dec 14, 99 01:55:48 pm
From: "Adam B. Roach" <Adam.Roach@Ericsson.com>
X-Mailer: ELM [version 2.4 PL24]
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

>There was some discussion of this problem wrt DCS during the November
>interim meeting. It was suggested that a cross-domain destination which is a
>phone number could be better represented by a tel: URL than by confusing the
>scope of the user part of a sip: URL. This simplilifies life greatly.
>
>Something like
>
>	INVITE premiumservice@gateway.carrier.com
>	To: tel:5551212;trans=lnp?gdbs
>
>could be used to indicate to gateway.carrier.com that premium gateway
>service is requested towards the PSTN location identified by telephone
>number 5551212, that LNP and GDBS translation have been applied, and that
>the phone number is of local scope for the gateway.

Since the To: field isn't allowed to change during the transaction,
aren't you going to run into problems if some service along the way
determines that the called number needs to change (as the result of
LNP, 800/900 lookup, etc)?

For example; let's say that one of the services that I have out in the
network is a "reachable from anywhere" speed dialing list. It maps,
for example, "4" as a telephone number to "2145551212". So I launch
an invite to my local proxy:

  INVITE sip:4@proxy.spleen.com SIP/2.0
  To: tel:4

It goes through my service node:

  INVITE sip:4@usersvc-34987.spleen.com SIP/2.0
  To: tel:4

And he wants to send it on to your gateway for premium service, after
translating the 4 into "2145551212."

It's my interpretation that, under the model you're proposing,
I'd need to send along

  INVITE sip:premiumservice@gateway.carrier.com SIP/2.0
  To: tel:2145551212

But that would break the transaction, since the To: field has changed.

I understand the distinction between destination and service
activation that you're trying to make, but your example seems
unworkable. What am I missing?

-- 
Adam Roach, Ericsson Inc. |  Ph: +1 972 583 7594 | 1010 E. Arapaho, MS L-04
adam.roach@ericsson.com   | Fax: +1 972 669 0154 | Richardson, TX 75081 USA

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec 15 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 VAA21642
	for <iptel-archive@odin.ietf.org>; Wed, 15 Dec 1999 21:20:14 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1920C5307; Wed, 15 Dec 1999 21:09:59 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 3A1C65309; Wed, 15 Dec 1999 21:09:58 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Date: Wed, 15 Dec 1999 15:39:45 -0500
From: "Mark D. Foster" <mfoster@mdfassoc.com>
Subject: RE: TRIP
In-reply-to: <38566A7A.5D69155B@dynamicsoft.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>,
        "'John Mainwaring'" <crm312a@nortelnetworks.com>
Cc: iptel@lists.research.bell-labs.com, "'James Yu'" <james.yu@neustar.com>,
        sip@lists.research.bell-labs.com
Message-id: <001201bf476a$6c72edc0$0864a8c0@loudoun.com>
MIME-version: 1.0
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2910.0)
Content-type: text/plain; charset=iso-8859-1
Content-transfer-encoding: 7bit
Importance: Normal
X-MSMail-Priority: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300
X-Priority: 3 (Normal)
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit

Jonathan Rosenberg wrote:


> ...snip...
> As an example:
>
> sip:15551212@cisco.com;user=phone;trans="lnp,gdbs"
>
> where lnp and gdbs are some kind of global translation services that
> apply to names of the type phone.
>

As long as the "To" field is E.164 (local to E.164 translation presumed to
have occured before egressing the originating network) and is always
preserved, I think this could work.


> ...snip...
> Hmm. This is something thats also different from the PSTN. The cost of
> routing to the donor proxy in SIP is much less, since proxies can be
> stateless. Thus, if a proxy receives a call for some number that is
> ported, it stateless proxies the call towards the next
> destination. The
> result is that there is no triangle routing for media (there generally
> isn't), nor for subsequent signaling. There is only marginal cost for
> the donor provider. There are obviously polticial and business issues;
> one may not want to trust or rely upon the donor proxy at
> all, as is the
> case of LNP in the US. But, the general SIP model is that we assume
> proxies do route calls; no security measures can prevent a proxy from
> dropping messages. Why, in the case of LNP, do we now need to assume
> these proxies won't route calls?
>

I don't think there's a need to change this assumption.  Actually, ensuring
the donor's proxy will re-route is critical, as they are considered the
default route for ported numbers.  Again, for regulatory and business
reasons we shouldn't force networks to ping the donor's proxy, lest it snack
of the deprecated (at least in the US & Canada) QoR/RTP approach.

>
>  I have noted that originator query may not be practical
> > for nationwide calling and that donor proxy is too late
> unless you're
> > prepared to consider RTP.
>
> The assumption of the impractability of originator queries has
> everything to do with the existing structure of LNP. Its
> mainly because,
> AFAIK, the systems are different all over the world. However,
> if we use
> the IP model, with a DNS solution, its no harder to query for a number
> on the other side of the planet as it would be for a number next door.
>
> Also, pardon my ignorance, but what is RTP? Many of us on
> this list know
> RTP as the Real Time Transport Protocol (rfc1889) but I don't think
> thats what you mean.
>

I think RTP was meant to refer to "release to pivot", one of the LNP call
routing scenarios that requires the donor remain in the signaling loop, and
therefore disqualified by the FCC.


-------------------------------------------------------------
Mark D. Foster                  | EMAIL: mark.foster@neustar.com
CTO                             | NeuStar, Inc.
TEL: +1 202 533 2800/2702       | FAX: +1 202 533 2975
1120 Vermont Ave NW, Suite 550  | Washington, DC 20005


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Wed Dec 15 21:29:32 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 VAA21706
	for <iptel-archive@odin.ietf.org>; Wed, 15 Dec 1999 21:29:32 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 0E3B5530B; Wed, 15 Dec 1999 21:16:53 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id CDD7652DF; Wed, 15 Dec 1999 21:13:57 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-Id: <199912151529.JAA08677@b04a24.exu.ericsson.se>
Subject: Re: URL Generalization for Portability (was TRIP)
To: dean.willis@wcom.com (Dean Willis)
Date: Wed, 15 Dec 1999 09:29:14 -0600 (CST)
Cc: Adam.Roach@ericsson.com, iptel@lists.research.bell-labs.com,
        sip@lists.research.bell-labs.com
In-Reply-To: <001b01bf46c1$23940980$d19623a6@directlink.net> from "Dean Willis" at Dec 14, 99 11:56:32 pm
From: "Adam B. Roach" <Adam.Roach@ericsson.com>
X-Mailer: ELM [version 2.4 PL24]
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

Dean Willis writes:

>I was talking with some guys today about scope and context, and I'm inclined
>to think we need the nature of address and numbering plan indicator fields
>out of some ISUP variants. Scope appears to be global, local, or private,
>with some need to represent local and private within some context like area
>code or private numbering plan.
>
>So, for example
>
>tel:+19727294162 is a global number with unspecified context,
>tel:9727294162 is a local number
>tel:7724162;phone-context=mcivnet34 is a private number with context in the
>MCI private numbering plan #34
>tel:4512345687724162;phone-context=mciswtrunk is a private number, in
>numbering plan 45, relative to switch 123 and trunk group 45678
>. . . and whatever other sick private variants one might come up with

If we proceed with the idea of scoping numbers "like ISUP," could we 
align it with ISUP a little more closely? It would make interoperation
*much* easier.

ISUP defines four scopes; the list of them follows, along with
suggestions for representing them.

1) Subscriber number (significant within an area/city code)
   tel:5555555;phone-context=local

2) Unknown (usually a private numbering plan)
   tel:7724162;phone-context=mcivnet34
   or
   tel:4512345687724162;phone-context=mciswtrunk
   or
   tel:34251;phone-context=unknown
   or
   tel:23456;phone-context=private
   or
   ...

3) National number (sigificant within a country)
   tel:2145555555;phone-context=national

4) International number (globally significant)
   tel:12145555555;phone-context=global
   or
   tel:+12145555555


--
Adam Roach, Ericsson Inc. |  Ph: +1 972 583 7594 | 1010 E. Arapaho, MS L-04
adam.roach@ericsson.com   | Fax: +1 972 669 0154 | Richardson, TX 75081 USA

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Dec 20 02:10:58 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 CAA25862
	for <iptel-archive@odin.ietf.org>; Mon, 20 Dec 1999 02:10:58 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7B61252DC; Mon, 20 Dec 1999 01:56:44 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 83FE452DB; Mon, 20 Dec 1999 01:56:39 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <19991218055255.63239.qmail@hotmail.com>
X-Originating-IP: [24.113.248.38]
From: "Kam Lee" <kamleesys@hotmail.com>
To: iptel@lists.research.bell-labs.com
Subject: nesting of switch statements
Date: Fri, 17 Dec 1999 21:52:55 PST
Mime-Version: 1.0
Content-Type: text/plain; format=flowed
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk

Does CPL allow the nesting of string-switch'es and time-switch'es?

thanks
kam
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com


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


From owner-iptel-outgoing@lists.research.bell-labs.com  Mon Dec 20 02:23:18 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 CAA26801
	for <iptel-archive@odin.ietf.org>; Mon, 20 Dec 1999 02:23:17 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id CE72D52DD; Mon, 20 Dec 1999 01:58:45 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D476152D5; Mon, 20 Dec 1999 01:58:39 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <385CF141.C88A05B7@ericsson.com>
Date: Sun, 19 Dec 1999 15:52:50 +0100
From: Ulf Andersson <ulf.andersson@ericsson.com>
Organization: Ericsson Inc
X-Mailer: Mozilla 4.04 [en] (Win95; I)
MIME-Version: 1.0
To: "tony@netspeak.com" <tony@netspeak.com>,
        "noreilly@netspeak.com" <noreilly@netspeak.com>,
        "artru@8x8.com" <artru@8x8.com>, "marc@8x8.com" <marc@8x8.com>,
        "isabelle@8x8.com" <isabelle@8x8.com>, "jfc@8x8.com" <jfc@8x8.com>,
        "scottsc@8x8.com" <scottsc@8x8.com>,
        "dpetrie@pingtel.com" <dpetrie@pingtel.com>,
        "patel@pingtel.com" <patel@pingtel.com>,
        "terry@catapult.com" <terry@catapult.com>,
        "richard@catapult.com" <richard@catapult.com>,
        "ronald@catapult.com" <ronald@catapult.com>,
        "imene@catapult.com" <imene@catapult.com>,
        "Ashok_Ganesan@Mitel.COM" <Ashok_Ganesan@Mitel.COM>,
        "Jerry_yin@mitel.com" <Jerry_yin@Mitel.COM>,
        "hgs@cs.columbia.edu" <hgs@cs.columbia.edu>,
        "lennox@cs.columbia.edu" <lennox@cs.columbia.edu>,
        "xiaotaow@cs.columbia.edu" <xiaotaow@cs.columbia.edu>,
        "orton@nortelnetworks.com" <orton@nortelnetworks.com>,
        "ingle@nortelnetworks.com" <ingle@nortelnetworks.com>,
        "billj@nortelnetworks.com" <billj@nortelnetworks.com>,
        "oste@nortelnetworks.com" <oste@nortelnetworks.com>,
        "williaml@nortelnetworks.com" <williaml@nortelnetworks.com>,
        "mohammad.vakil@wcom.com" <mohammad.vakil@wcom.com>,
        "robert.sparks@wcom.com" <robert.sparks@wcom.com>,
        "saravanan.balasubramanian@wcom.com" <saravanan.balasubramanian@wcom.com>,
        "steven.r.donovan@wcom.com" <steven.r.donovan@wcom.com>,
        "sachin.danait@wcom.com" <sachin.danait@wcom.com>,
        "buzz.mcdermott@wcom.com" <buzz.mcdermott@wcom.com>,
        "john.k.gallant@wcom.com" <john.k.gallant@wcom.com>,
        "paul.schultz@wcom.com" <paul.schultz@wcom.com>,
        "douglas_carson@agilent.com" <douglas_carson@agilent.com>,
        "joyce@broadsoft.com" <joyce@broadsoft.com>,
        "dave@broadsoft.com" <dave@broadsoft.com>,
        "jamie@broadsoft.com" <jamie@broadsoft.com>,
        "alex@broadsoft.com" <alex@broadsoft.com>,
        "rkrishna@vtel.com" <rkrishna@vtel.com>,
        "vmath@vtel.com" <vmath@vtel.com>,
        "wkwok@telogy.com" <wkwok@telogy.com>,
        "mwinslow@radcomusa.com" <mwinslow@radcomusa.com>,
        "dnevo@radcomusa.com" <dnevo@radcomusa.com>,
        "eb@indigo-software.com" <eb@indigo-software.com>,
        "nzi@indigo-software.com" <nzi@indigo-software.com>,
        "cz@indigo-software.com" <cz@indigo-software.com>,
        "cteoh@nuera.com" <cteoh@nuera.com>, "mluu@nuera.com" <mluu@nuera.com>,
        "memami@nuera.com" <memami@nuera.com>,
        "jryner@nuera.com" <jryner@nuera.com>,
        "adam.roach@ericsson.com" <adam.roach@ericsson.com>,
        "sean.olson@ericsson.com" <sean.olson@ericsson.com>,
        "mark.peck@ericsson.com" <mark.peck@ericsson.com>,
        "cjessen@nortelnetworks.com" <cjessen@nortelnetworks.com>,
        "psollee@nortelnetworks.com" <psollee@nortelnetworks.com>,
        "kelvin.porter@wcom.com" <kelvin.porter@wcom.com>,
        "mark.x.johnson@wcom.com" <mark.x.johnson@wcom.com>,
        "ajay.deo@wcom.com" <ajay.deo@wcom.com>,
        "tina.guo@wcom.com" <tina.guo@wcom.com>,
        "vpatel@dynamicsoft.com" <vpatel@dynamicsoft.com>,
        Jonathan Rosenberg <jdrosen@dynamicsoft.com>,
        "trang@dynamicsoft.com" <trang@dynamicsoft.com>,
        "manojb@cisco.com" <manojb@cisco.com>,
        "sudiptom@cisco.com" <sudiptom@cisco.com>,
        "clindner@cisco.com" <clindner@cisco.com>,
        "rhicks@delta-info.com" <rhicks@delta-info.com>,
        "prandall@delta-info.com" <prandall@delta-info.com>,
        "nurban@delta-info.com" <nurban@delta-info.com>,
        "jon2@andrew.cmu.edu" <jon2@andrew.cmu.edu>,
        "mrappard@andrew.cmu.edu" <mrappard@andrew.cmu.edu>,
        "ak@hplb.hpl.hp.com" <ak@hplb.hpl.hp.com>,
        "jose@tct.hut.fi" <jose@tct.hut.fi>,
        "queca@tct.hut.fi" <queca@tct.hut.fi>,
        "Jerry_Mahler@mw.3com.com" <Jerry_Mahler@mw.3com.com>,
        "Rick_Dean@mw.3com.com" <Rick_Dean@mw.3com.com>,
        "jii@oz.com" <jii@oz.com>, "oli@oz.com" <oli@oz.com>,
        "etremblay@mediatrix.com" <etremblay@mediatrix.com>,
        "jfwaddell@mediatrix.com" <jfwaddell@mediatrix.com>,
        "hans@erix.ericsson.se" <hans@erix.ericsson.se>,
        "mml@erix.ericsson.se" <mml@erix.ericsson.se>,
        "calle@erix.ericsson.se" <calle@erix.ericsson.se>,
        "Ravandhu_hariram@3com.com" <Ravandhu_hariram@3com.com>,
        "anoop_tripathi@3com.com" <anoop_tripathi@3com.com>,
        "clark@facetcorp.com" <clark@facetcorp.com>,
        "larry@facetcorp.com" <larry@facetcorp.com>,
        "johnm@facetcorp.com" <johnm@facetcorp.com>,
        "jim@facetcorp.com" <jim@facetcorp.com>,
        "brad@facetcorp.com" <brad@facetcorp.com>,
        "alex@ipcell.com" <alex@ipcell.com>,
        "sankar@ipcell.com" <sankar@ipcell.com>,
        "ldang@vovida.com" <ldang@vovida.com>,
        "cjennings@vovida.com" <cjennings@vovida.com>,
        "skumar@vovida.com" <skumar@vovida.com>,
        sip@lists.research.bell-labs.com, ensc-tia@sfu.ca,
        sip-implementors@cs.columbia.edu, confctrl@isi.edu,
        iptel@lists.research.bell-labs.com, eriietf@kk.ericsson.se
Cc: barbara.boyle@ericsson.com
Subject: Bakeoff Press Release
Content-Type: text/plain; charset=iso-8859-1
X-MIME-Autoconverted: from 8bit to quoted-printable by gwa.ericsson.com id IAA19690
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 CAA26801

Hi!

Attached is a copy of the 3rd SIP Bakeoff press release that went out earlier this week.
Please refer all questions regarding the press release to Barbara Boyle at 972-583-5481.
Thank you!

Regards,
Ulf


-- Attached Press Release --

ERICSSON

FOR IMMEDIATE RELEASE

LEADING EQUIPMENT VENDORS, SERVICE PROVIDERS AND RESEARCH
ORGANIZATIONS UNITE TO ADVANCE INTERNET TELEPHONY STANDARD

Third SIP Bake-Off brings 26 organizations together to test for interoperability using
Session Initiation Protocol (SIP)

Richardson, Texas — December 14, 1999 — Engineering groups representing leading vendors,
service providers and research organizations met last week to test the interoperability of
their
Internet telephony products.

The technology "bake-off" was the industry’s third event organized to test the
interoperability of
software and hardware devices using the Session Initiation Protocol (SIP), which is used
to initiate,
manage and terminate interactive sessions, including voice, across a wide area network.
SIP became
an official standard of the Internet Engineering Task Force (IETF) in March 1999 and has
gained
significant support among telco equipment vendors, carriers, CLECs, cable operators and
web
based voice chat providers.

The Third SIP Bake-Off, sponsored by Ericsson, was the largest event to date to test for
industry
interoperability using SIP, with 31 teams from 26 companies and research organizations
taking part.
Participants tested interoperability of basic call setup and of more advanced features,
including
security, location and mobility features, call routing services and call center
applications. Participants
paired up in rounds, and in each round attempted to execute a suite of tests to determine
the level of
interoperability.

The highlight of the event was a complex call scenario, involving four clients and three
servers, each
from a different vendor. In this scenario, a caller initiating a call was immediately
challenged to
authenticate himself to a server. After successfully authenticating, the call was routed
through three
servers, causing three separate phones to ring at once. After one of the ringing phones
was picked up,
the proxies exchanged messages causing the other phones to stop ringing. At that point,
the two
parties were connected and were able to talk to each other and then disconnect.

Any problems that arose during the tests were discussed among the teams to determine
whether they
were caused by troubles with implementation or by an ambiguity in the specification. All
ambiguities
were recorded so they could be clarified in the next version of the SIP specification.

These kinds of interoperability tests are a key component of the IETF standards process.
Requests for
Comments (RFCs), or official standards from the IETF, progress through three phases of
maturity as
the protocols are rolled out and deployed. SIP is currently at the first phase, proposed
standard RFC,
and interoperability of every feature must be demonstrated in order to advance to the next
stage, draft
standard RFC.

Ericsson is the first vendor to sponsor a SIP Bake-Off. Previous SIP Bake-Off events have
been
sponsored by Pulver.com in August and Columbia University in April 1999.


Organizations participating in the Third SIP Bake-Off included: 3Com Corporation
(NASD:COMS),
8x8 (NASD:EGHT), Agilent, BroadSoft, Catapult Communications, Cisco Systems (NASD:CSCO),
Columbia University, Delta, dynamicsoft, E*Club (Carnegie Mellon University),
Ericsson Inc. (NASD:ERICY), Helsinki University of Technology, Hewlett-Packard Labs,
Indigo,
IPCell Technologies, MCI WorldCom (NASD:WCOM), Mediatrix, Mitel Corp (NYSE:MLT).,
Nortel Networks (NYSE:NT), Nuera Communications, OZ.COM, Pingtel Corp.,
Radcom Equipment, Inc., RaveTel, Telogy Networks, and Vovida Networks, Inc.

"When 26 vendors, service providers and research organizations come together to test their
equipment
for interoperability, it really attests to the industry's commitment to meeting the needs
of customers,"
said Mats Nilsson, corporate director of technical standards strategy at Ericsson.
"Ericsson is proud to
be the first vendor sponsor of an event that provides a truly open, unbiased network to
test for
interoperability and that furthers the development of SIP as one of the key components in
the
new telecoms world."

"This third round of SIP testing gives all leading IP telephony vendors a great
opportunity to test more
advanced features. We welcome participating in the SIP Bake-Offs in order to ensure
product
interoperability -- from intelligent phones to media gateways and back-end servers -- and
to be a part
of introducing SIP as an effective protocol for implementing voice over IP," said Andrew
Bezaitis,
director, business development and marketing, Internet Communications Group, 3Com
Corporation.
"As the IP Telephony market continues its rapid maturation, SIP-based solutions stand to
provide
significant value in enabling enhanced services.”

"Participation in industry interoperability events such as the SIP bakeoff allows us to
validate the
operation of tools such as call trace and protocol analysis for the carrier networks of
the future.
Agilent network monitoring solutions help operators reduce costs, increase quality of
service and
deliver industry leading services over any network." said Tom White, senior vice president
and
general manger for Agilent Technologies' Communications Solutions Group.

"Having been designed with interoperability in mind, SIP is key to ensuring that enhanced
services run
smoothly across packet networks," said Scott Hoffpauir, vice president of engineering at
BroadSoft.
"BroadSoft is dedicated to providing service delivery and creation platforms that create a
basis for
new value propositions.  The SIP Bakeoff is an ideal forum for ensuring interoperability
among
vendors who are forming the basis of the next generation network."

"The camaraderie between participants was outstanding at the Third SIP Bake-Off. This
interoperability event enabled us to learn the testing needs of our customers first-hand
and put our
multiple testing capabilities to the test by interacting with numerous vendors
simultaneously," said
Richard Mott, Catapult Communications regional manager.

"Interoperability events have proven to be effective, and the SIP Bake-Off was no
exception," said
Neil Urban, development manager at Delta Information Systems.  "This latest SIP Bakeoff
provided
an excellent opportunity to interact with a wide array of multi-vendor products, ensuring
the reliability
and robustness of our protocol analysis equipment."

"This bakeoff demonstrated that multi-vendor interoperability for real-world call
scenarios is now a
reality," said Jonathan Rosenberg, chief scientist at dynamicsoft and co-author of the SIP
specification.
"This further validates that SIP is ready for deployment."

"It was a great experience to test our implementation with others from different companies
and see
that, after all....it works!!" said Inmaculada Espigares from Helsinki University of
Technology. "We
had never seen so many people so happy after hearing a telephone RING!"

“We see SIP as the key for opening the long-awaited Internet telephony era. Its natural
integration
with web and email services makes it the ideal lever to make good on the promises of
Internet
telephony. Adopting this technology will enable ISPs and operators to offer resolute new
services
to their customers,” said Emmanuel Bertrand, CEO of Indigo Software.

"Mitel is fully supportive of all activities directed towards standardization in the IP
telephony market.
Interoperability is of course key to this, and the SIP bakeoffs are proving to be
excellent testing
ground for the technology as the market develops," said Trefor Davis, IP telephony product
line
manager at Mitel.

"This event went a long way to resolving outstanding issues in implementing the IETF RFC
2543
and in establishing interoperability among multiple vendors' SIP implementations. It
hastened the
day when SIP will be a pervasive protocol in voice over IP network deployments," said
Neville O'Reilly, vice president, Business Development, NetSpeak.

"SIP is a great standard to interoperate between existing telephony service and new IP
based
services" says Kjartan Pierre Emilsson, CTO of OZ.COM.  "By using SIP in the iPulseTM
product that we co-develop with Ericsson, we effectively provide a true convergence
platform,
on which third parties can easily add new communication services."

"The tremendous momentum behind SIP is clearly evident from both the attendance and
enthusiasm at this bake-off, " said Jay Batson, CEO of Pingtel. "SIP's inherent
functionality,
scalability and opportunities for endless innovation using an intelligent endpoint are
making SIP
the preferred protocol for both enterprise and service provider VoIP networks."

“RADCOM is committed to providing innovative, effective solutions to interoperability and
other convergence problems," said Anthony S. Wiley, president of RADCOM Equipment Inc.
"We are therefore pleased to have had the opportunity to prove our decodes and provide
performance analyses on the SIP-based, Internet telephony devices represented at this
exciting
event."

"Standards-based IP communication is the future of converged networks," said Ben Jenkins,
CEO
of RaveTel.  "By participating in this kind of forum, we are ensuring that our products
will work
seamlessly with those of other vendors."

"Telogy Networks welcomes participation in interoperability events such as this Third SIP
Bake-off,
which we view as a tool for accelerating the rapid deployment of new standards," said Ed
Morgan,
executive director, Research and Development for Telogy Networks, A Texas Instruments
Company.
"Telogy plans to offer SIP as a protocol option for VoIP applications such as gateways, IP
phones,
cable modems, and other new products."

"We gladly participate in industry interoperability testing such as this so companies
using our stacks
don't have to replicate the test effort," said Luan Dang, senior vice president, CTO and
co-founder
of Vovida Networks.  "Vovida continues its support of the SIP, MGCP and H.323 protocols in
our
products and is committed to supporting events such as these that further interoperability
and open
source adoption of VoIP standards."

More information about the Third SIP Bake-Off can be found at http://www.sipbakeoff.org.
More information about SIP can be found at http://www.cs.columbia.edu/~hgs/sip.

For additional information, contact the press relations department of any of the
participating
organizations:


Ericsson is the leading provider in the new telecoms world, with communications solutions
that combine
telecom and datacom technologies with freedom of mobility for the user. With more than
100,000
employees in140 countries, Ericsson simplifies communications for its customers -network
operators,
service providers, enterprises and consumers – the world over.

With over 300 million customer connections worldwide, 3Com Corporation connects more
people
and organizations to information and each other in more innovative, simple and reliable
ways than any
other networking company. 3Com delivers e-Networking solutions through information access
products and network systems to enterprises, small businesses, consumers, carriers and
network
service providers. For further information, visit 3Com's World Wide Web site at
www.3com.com
or the press site at www.3com.com/pressbox.

Agilent Technologies, Inc. is a diversified technology company, resulting from
Hewlett-Packard
Company's plan to strategically re-align itself into two fully independent companies. With
42,000
employees serving customers in more than 120 countries, Agilent Technologies, as a
subsidiary of
HP, is a global leader in designing and manufacturing test, measurement and monitoring
instruments,
systems and solutions, and semiconductor and optical components. The company serves
markets
that include communications,  electronics, life sciences and healthcare. The businesses
comprising
Agilent had net revenues of more than $8.3 billion in fiscal year 1999. Information about
Agilent
Technologies can be found on the Web at http://www.agilent.com.

BroadSoft is the leading provider of service delivery systems and service creation tools
for
next-generation communications networks.  The company's flagship product, BroadWorks,
allows
service providers to offer innovative new communications services such as IP Centrex and
other
enhanced services, on an open, hardware-independent platform.  Founded in 1998, BroadSoft
is
located outside of Washington, DC and is a privately held, venture backed company.
(www.broadsoft.com)

Catapult Communications delivers comprehensive telecom test systems that enable testing of
both
established networks and emerging technologies such as Voice over IP and 3rd Generation
Cellular.
Offering multi-protocol, multi-user capabilities in a single solution, Catapult helps
manufacturers of
Gateways, Gatekeepers, MCUs and SIP terminal equipment ensure interoperability in today's
converging markets.

Delta's Network Systems group, specializing in the development of conferencing and VoIP
multi-protocol  test equipment, has provided audio, video and data protocol analyzers
since 1992.
Delta is committed to providing cost effective protocol analyzers for new standards as
they develop.
We also recognize that no standard product can meet the needs of all users, so we work
with our
customers to provide customized solutions. Delta also manufactures application specific
audio and
video compression products and systems. Visit us at www.delta-info.com/products/vt320.

dynamicsoft Inc. delivers connection management and end-to-end server solutions to service

providers and communications equipment manufacturers, enabling the scalable, interoperable
and
immediate deployment of applications and services for converged voice/data networks. Visit

dynamicsoft at  www.dynamicsoft.com.  dynamicsoft is funded byComVentures (www.comven.com)

and VantagePoint Venture Partners (www.vpvp.com), two world class venture capital firms
investing
in technologies.

Telecommunications Technology Department at Helsinki University is interested in the
technology,
engineering, information and enterprise aspects of the functional areas of communications
services and
switching such as signaling in telecommunications networks, routing in circuit and packet
switched
networks including Internet routing protocols, interworking of networks and services, and
telecommunications product development. Recent research has been concentrated on IP
switching
and Voice over IP. For further information http://keskus.hut.fi/engl.shtml or contact with
Jose
Costa Requena at http://keskus.hut.fi/tutkimus/ipana/pers/jose.shtml.

Indigo Software, headquartered in Brussels, Belgium, is leading the way in
Internet-speaking
telephony solutions. Designing and providing cutting-edge solutions for ISPs and operators
worldwide,
Indigo Software is the European innovator in Java-based software for the worldwide
Internet telephony
market. Supported by major industry players, Indigo software is a young, forward-looking,
venture-backed company.

Mitel (NYSE: MLT; TSE: MLT) is a growing global provider of converging voice and data
systems
and applications, and specialty semiconductors for the communications industry.  The
company has
revenues of more than CDN $1.3 billion and employs approximately 6,000 people worldwide.

A pioneer in Voice over IP (VoIP) communications, NetSpeak Corporation (NASDAQ: NSPK)
is a leading developer and marketer of advanced telephony solutions for IP networks.
NetSpeak
solutions are powered by products that have been strategically designed to meet the
rapidly evolving
business needs of its customers.  Through a protocol-independent software architecture,
NetSpeak
products deliver efficient management of network resources and enable cost-effective
introduction
of new VoIP applications.

OZ.COMTM is a pioneer in the advancement and development of scalable, interactive
real-time
communications products and solutions. The company’s mission is to bring
"people above networksTM" and develop technology that integrates the Internet, traditional
wireline
and wireless communications services. With corporate headquarters in Boston and European
offices
in Reykjavik, Iceland, and Stockholm, Sweden, OZ.COM has over 80 employees and can be
found on the World Wide Web at http://www.oz.com/.

Pingtel is fostering a voice services and applications renaissance with the world's first
intelligent
voice-over-IP telephone for service providers and enterprises. As the delivery platform
for
next-generation IP PBX systems and virtual IP PBX services, it not only makes today's
voice
features much easier to use, but enables a whole new collection of valuable, personalized
voice
services and applications. A venture-backed start-up, Pingtel is located near Boston, in
Woburn,
MA, and can also be found at www.pingtel.com.

RADCOM specializes in innovative, packet telephony solutions for providers of integrated
voice,
data & Internet products and services.   Its line of integrated WAN/FastLAN analyzers,
which
supports over 350 voice, data and cellular protocols, provide an ideal solution for
testing VoATM,
VoFR as well as VoIP. RADCOM's test solution suites are used in the development and
manufacturing of network equipment; the installation of networks; and the ongoing
maintenance of
operational networks to facilitate real-time identification, diagnosis, isolation and
resolution of
network problems. RADCOM's sales network includes over 50 distributors in 35 countries
worldwide
and over 14 manufacturer's representatives across North America. (www.radcom-inc.com)

RaveTel is one of the first providers of presence-aware communications applications for
converged
networks.  The company is a newly-formed spin-off from the E*Club project at
Carnegie-Mellon
University in Pittsburgh, PA. For more information, visit RaveTel at
http://sip.ini.cmu.edu.

Telogy Networks, a Texas Instruments Company, is the leading provider of embedded
communications
software to global equipment manufacturers. Telogy's Golden GatewayR software enables its
customers
to develop connected products that send real-time voice, fax and data over packet networks
(such as
Internet/Intranet, Frame Relay, and ATM). As one of the few embedded software companies
with both
microprocessor and DSP expertise, Telogy Networks offers its customers truly comprehensive
product
solutions. Telogy's products are in use by virtually every communications equipment
manufacturer,
including Cisco Systems, 3COM, Nortel Networks, Motorola and many others throughout the
world.
Company news and product information are available at http://www.telogy.com.

Vovida Networks, a founding member of the International Softswitch Consortium, is a
Linux-based
communications infrastructure company and the global pioneer of open source telephony.
Vovida is a
global provider of communications software products that scale from silicon to systems and
enable
advanced telephony features, applications and services on packet-switched networks.  In
addition,
Linux-based, open source implementations of MGCP, RTP, H.323 Annex F and SIP are freely
available on our web site: www.vovida.com.

     ###

For more information, please contact:
Barbara Boyle
Marketing Communications Manager
972-583-5481





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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 21 00:44:06 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 AAA24879
	for <iptel-archive@odin.ietf.org>; Tue, 21 Dec 1999 00:44:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 547CF52DE; Tue, 21 Dec 1999 00:37:05 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 7855152DB; Tue, 21 Dec 1999 00:37:03 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <385F12AD.3D1B4DC0@dynamicsoft.com>
Date: Tue, 21 Dec 1999 00:39:57 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Kam Lee <kamleesys@hotmail.com>
Cc: iptel@lists.research.bell-labs.com
Subject: Re: nesting of switch statements
References: <19991218055255.63239.qmail@hotmail.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

Sure. You can do something like:

<string-switch field="from">

   <string is="sip:jdrosen@dynamicsoft.com">

        <string-switch field="to">
            
            <string is="sip:lennox@cs.columbia.edu">

               <proxy/>
 
            </string>

        </string-switch>

   </string>

</string-switch>

-Jonathan R.

Kam Lee wrote:
> 
> Does CPL allow the nesting of string-switch'es and time-switch'es?
> 
> thanks
> kam
> ______________________________________________________
> Get Your Private, Free Email at http://www.hotmail.com
> 
> ---------
> This message came from the IETF IPTEL Working Group Mailing List.

-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Sat Dec 25 12:54:05 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 MAA29498
	for <iptel-archive@odin.ietf.org>; Sat, 25 Dec 1999 12:54:05 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 52D0F52E2; Sat, 25 Dec 1999 12:35:06 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id E392E52E6; Sat, 25 Dec 1999 12:34:58 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <3863E69D.3BC6922@dynamicsoft.com>
Date: Fri, 24 Dec 1999 16:33:17 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: James Yu <james.yu@neustar.com>
Cc: John Mainwaring <crm312a@nortelnetworks.com>,
        Mark Foster <mark.foster@neustar.com>,
        iptel@lists.research.bell-labs.com, sip@lists.research.bell-labs.com
Subject: Re: TRIP
References: <ED88182BFF78D211A4D800A0C9E9435C3BE93C@dc02.npac.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



James Yu wrote:
> 
> For overlap signaling, a GW or LS that receives a call with partial address
> information may not know that the address info. is complete.  If it launches
> a query (DNS or NPDB), the entity that has the "number portability" info.
> won't be able to find the info. because the address info. is incomplete.  It
> seems that either this LS or GW may wait a little while to see whether there
> are subsequent address info. before launching the query, or it can launch a
> query (DNS or NPDB-like), receives an error (hope that it is specific such
> as incomplete address instead of something ambiguous), and then wait for
> additional address info. before timing out.

I think Dean had sent out some pseudo-code some time back on how trip
and enum and overlap interact, and this was more or less what the code
suggested. I believe he thought it was a bad thing, and so do I.

Overlap dialing seems fundamentally in conflict with LNP or enum queries
at the local outbound proxy.

To do LNP, the local outbound proxy needs to have the entire string of
digits before it can perform the query. But, overlap seems to imply that
it won't ever know the entire string of digits, as it doesn't have
knowledge of those numbering plans. It just forwards the request to some
other proxy or gateway that does.

Sooo, it would seem that a proxy should only use enum or LNP for numbers
it knows to be complete. This means that for overlap dialing, the LNP
query will happen close to the destination (possibly even in the PSTN),
not unlike the N-1 case we've been discussing. However, for en-bloc, it
can happen anywhere. This does seem to imply that a proxy needs to know
whether the number in the request URI of an incoming request is en-bloc
or overlap. It could make this determination if it knew the structure of
the numbering plans, but thats unlikely. Might we need a parameter
somewhere (ugh)? I really hate to crowd SIP with things to support old
legacy technologies like overlap, so if we can avoid this, I'd be very
happy.



-Jonathan R.
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 28 10:19:40 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 KAA00972
	for <iptel-archive@odin.ietf.org>; Tue, 28 Dec 1999 10:19:40 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 211CB52D5; Tue, 28 Dec 1999 10:08:00 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id A89E652BB; Tue, 28 Dec 1999 10:07:55 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <ED88182BFF78D211A4D800A0C9E9435C3BE9A6@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>
Cc: John Mainwaring <crm312a@nortelnetworks.com>,
        Mark Foster <mark.foster@neustar.com>,
        iptel@lists.research.bell-labs.com, sip@lists.research.bell-labs.com
Subject: RE: TRIP
Date: Tue, 28 Dec 1999 08:58:09 -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

Overlap dialing probably happens often for international calls where the
switch in the originating country don't know much about the numbering plan
of the destination country.  The criteria to begin processing the received
digits at the originating switch may be based on the minimum number of
digits and the expiration of the inter-digit interval as have been indicated
by several people.   For a voice call to a destination Switched Circuit
Network (SCN) that involves IP,  the call can be originated from an IP phone
(e.g., SIP client, H.323 terminal) or from an SCN phone.  The IP phones may
be designed to allow en-bloc signaling.  But overlap signaling probably
cannot be avoided for the SCN-IP-SCN case.  The originating switch may not
know whether subsequent digits will be dialed.  So even if a mechanism to
carry the "overlap vs. en-bloc" signaling exists, the originating switch may
need to set it to "overlap signaling" in case it is not sure whether all
digits have been collected (e.g.., international calls).

It seems that certain proxy servers may need to be designated as the check
point when overlap signaling may be involved based on the From and To
information and others.  Hope that we can get some ideas/solutions at the
1/25-27 ITU IP & Telecoms Interworking Workshop.

James

> -----Original Message-----
> From: Jonathan Rosenberg [mailto:jdrosen@dynamicsoft.com]
> Sent: Friday, December 24, 1999 4:33 PM
> To: James Yu
> Cc: John Mainwaring; Mark Foster; iptel@lists.research.bell-labs.com;
> sip@lists.research.bell-labs.com
> Subject: Re: TRIP
> 
> 
> 
> 
> James Yu wrote:
> > 
> > For overlap signaling, a GW or LS that receives a call with 
> partial address
> > information may not know that the address info. is 
> complete.  If it launches
> > a query (DNS or NPDB), the entity that has the "number 
> portability" info.
> > won't be able to find the info. because the address info. 
> is incomplete.  It
> > seems that either this LS or GW may wait a little while to 
> see whether there
> > are subsequent address info. before launching the query, or 
> it can launch a
> > query (DNS or NPDB-like), receives an error (hope that it 
> is specific such
> > as incomplete address instead of something ambiguous), and 
> then wait for
> > additional address info. before timing out.
> 
> I think Dean had sent out some pseudo-code some time back on how trip
> and enum and overlap interact, and this was more or less what the code
> suggested. I believe he thought it was a bad thing, and so do I.
> 
> Overlap dialing seems fundamentally in conflict with LNP or 
> enum queries
> at the local outbound proxy.
> 
> To do LNP, the local outbound proxy needs to have the entire string of
> digits before it can perform the query. But, overlap seems to 
> imply that
> it won't ever know the entire string of digits, as it doesn't have
> knowledge of those numbering plans. It just forwards the 
> request to some
> other proxy or gateway that does.
> 
> Sooo, it would seem that a proxy should only use enum or LNP 
> for numbers
> it knows to be complete. This means that for overlap dialing, the LNP
> query will happen close to the destination (possibly even in 
> the PSTN),
> not unlike the N-1 case we've been discussing. However, for 
> en-bloc, it
> can happen anywhere. This does seem to imply that a proxy 
> needs to know
> whether the number in the request URI of an incoming request 
> is en-bloc
> or overlap. It could make this determination if it knew the 
> structure of
> the numbering plans, but thats unlikely. Might we need a parameter
> somewhere (ugh)? I really hate to crowd SIP with things to support old
> legacy technologies like overlap, so if we can avoid this, I'd be very
> happy.
> 
> 
> 
> -Jonathan R.
> -- 
> Jonathan D. Rosenberg                       200 Executive Drive
> Chief Scientist                             Suite 120 
> dynamicsoft                                 West Orange, NJ 07052
> jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
> http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
> http://www.dynamicsoft.com
> 

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 28 16:10:24 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 QAA05378
	for <iptel-archive@odin.ietf.org>; Tue, 28 Dec 1999 16:10:24 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id BC6AA52C4; Tue, 28 Dec 1999 16:05:44 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id 34E1252BB; Tue, 28 Dec 1999 16:05:44 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
From: "Vance Shipley" <vances@motivity.ca>
To: "James Yu" <james.yu@neustar.com>,
        "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>
Cc: "John Mainwaring" <crm312a@nortelnetworks.com>,
        "Mark Foster" <mark.foster@neustar.com>,
        <iptel@lists.research.bell-labs.com>,
        <sip@lists.research.bell-labs.com>
Subject: RE: TRIP
Date: Tue, 28 Dec 1999 15:58:44 -0500
Message-ID: <NBBBIOHCCJPPDNIBBIKCEEHBCCAA.vances@motivity.ca>
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 9.0.2416 (9.0.2910.0)
In-Reply-To: <ED88182BFF78D211A4D800A0C9E9435C3BE9A6@dc02.npac.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300
Importance: Normal
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7bit


James Yu writes:
}
}  Overlap dialing probably happens often for international calls where the
}  switch in the originating country don't know much about the numbering plan
}  of the destination country.  The criteria to begin processing the received

Actually no.  The ITU recommendation Q.767 "Application of the ISDN user part
of CCITT Signalling System No. 7 for international ISDN interconnections" is
used today on the International Plane between countries public telephone
networks.
This recommendation defines how ISUP signaling is to be applied toward the
application of international gateway interconnections to other countries
gateways.

Q.767 does not allow for overlap signaling.  If the local serving office switch
for the originating subscriber uses overlap signaling to speed up call routing
then an intermediate switch, or the gateway switch itself, will gather up all
the digits and send the call forward enbloc.

Overlap signaling is not used in ANSI ISUP at all.  Overlap signaling was used
in the days of in band signaling to speed up the delays caused by the relatively
long duration of the tones used.  One area where it was almost always used was
in fact international calls.  In the case of FGD signaling in North America when
an international call was presented the country code was sent first using
<KP>CC<ST> so that the switch could begin route selection and grab an outgoing
route while the rest of the digits where pulsed in.  This becomes unnecessary
when ISUP signaling is used.

In my mind overlap signaling should never be generated but may have to be
supported
when received.  The receiving switch must buffer the received digits until it
decides
that all have been received and then forward the call enbloc.  There are two
methods
the switch may use to recognize when all digits have been received.  The best is
when
the switch has an accurate view of the dialing plan(s).  It would know that
country
code XX, city code YYY requires exactly 7 digits in the subscriber number and
when it
received all 12 digits it would forward the call enbloc.  If it never received
enough
digits it would eventually time out and release the call with a cause code
indicating
"address incomplete".  The second method simply applies an inter digit timer and
maximum
number of digits.  Digits are buffered until either the maximum expected are
received,
the inter digit delay timer expires or an end-of-dial digit is recognized (e.g.
<ST>
or '#').  These methods are generally both used where some dial plans are
recognized
(e.g. North American Plan) and the rest use the time out.

	-Vance




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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 28 17:45:07 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 RAA06350
	for <iptel-archive@odin.ietf.org>; Tue, 28 Dec 1999 17:45:06 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id CE1D852D6; Tue, 28 Dec 1999 17:33:57 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id D513352DC; Tue, 28 Dec 1999 17:33:56 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <ED88182BFF78D211A4D800A0C9E9435C3BE9AB@dc02.npac.com>
From: James Yu <james.yu@neustar.com>
To: "'Vance Shipley'" <vances@motivity.ca>
Cc: John Mainwaring <crm312a@nortelnetworks.com>,
        Mark Foster <mark.foster@neustar.com>,
        "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>,
        iptel@lists.research.bell-labs.com, sip@lists.research.bell-labs.com
Subject: RE: TRIP
Date: Tue, 28 Dec 1999 16:24:34 -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

Vance,

Thanks for the explanation.  

The key questions now are whether overlap signaling (e.g., Feature Group-D
over Multi-Frequency facility) is still used somewhere and if used where the
digit collection/enbloc signaling is performed in the existing SCN.  

If overlap signaling is still used (someone please confirm this and indicate
where it is still used) and a call is to be routed from the originating SCN
to the Internet, I guess that the originating GW (SCN-Internet)probably has
to perform the digit collection/enbloc signaling function so that overlap
signaling does not propagate further into the Internet and/or the
destination GW/SCN.  Otherwise, some entities in the Internet must perform
that function.

James

> -----Original Message-----
> From: Vance Shipley [mailto:vances@motivity.ca]
> Sent: Tuesday, December 28, 1999 3:59 PM
> To: James Yu; 'Jonathan Rosenberg'
> Cc: John Mainwaring; Mark Foster; iptel@lists.research.bell-labs.com;
> sip@lists.research.bell-labs.com
> Subject: RE: TRIP
> 
> 
> 
> James Yu writes:
> }
> }  Overlap dialing probably happens often for international 
> calls where the
> }  switch in the originating country don't know much about 
> the numbering plan
> }  of the destination country.  The criteria to begin 
> processing the received
> 
> Actually no.  The ITU recommendation Q.767 "Application of 
> the ISDN user part
> of CCITT Signalling System No. 7 for international ISDN 
> interconnections" is
> used today on the International Plane between countries 
> public telephone
> networks.
> This recommendation defines how ISUP signaling is to be 
> applied toward the
> application of international gateway interconnections to 
> other countries
> gateways.
> 
> Q.767 does not allow for overlap signaling.  If the local 
> serving office switch
> for the originating subscriber uses overlap signaling to 
> speed up call routing
> then an intermediate switch, or the gateway switch itself, 
> will gather up all
> the digits and send the call forward enbloc.
> 
> Overlap signaling is not used in ANSI ISUP at all.  Overlap 
> signaling was used
> in the days of in band signaling to speed up the delays 
> caused by the relatively
> long duration of the tones used.  One area where it was 
> almost always used was
> in fact international calls.  In the case of FGD signaling in 
> North America when
> an international call was presented the country code was sent 
> first using
> <KP>CC<ST> so that the switch could begin route selection and 
> grab an outgoing
> route while the rest of the digits where pulsed in.  This 
> becomes unnecessary
> when ISUP signaling is used.
> 
> In my mind overlap signaling should never be generated but 
> may have to be
> supported
> when received.  The receiving switch must buffer the received 
> digits until it
> decides
> that all have been received and then forward the call enbloc. 
>  There are two
> methods
> the switch may use to recognize when all digits have been 
> received.  The best is
> when
> the switch has an accurate view of the dialing plan(s).  It 
> would know that
> country
> code XX, city code YYY requires exactly 7 digits in the 
> subscriber number and
> when it
> received all 12 digits it would forward the call enbloc.  If 
> it never received
> enough
> digits it would eventually time out and release the call with 
> a cause code
> indicating
> "address incomplete".  The second method simply applies an 
> inter digit timer and
> maximum
> number of digits.  Digits are buffered until either the 
> maximum expected are
> received,
> the inter digit delay timer expires or an end-of-dial digit 
> is recognized (e.g.
> <ST>
> or '#').  These methods are generally both used where some 
> dial plans are
> recognized
> (e.g. North American Plan) and the rest use the time out.
> 
> 	-Vance
> 
> 
> 

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Tue Dec 28 18:04:29 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 SAA06535
	for <iptel-archive@odin.ietf.org>; Tue, 28 Dec 1999 18:04:29 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 1E92052DC; Tue, 28 Dec 1999 17:54:02 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id F318E52DE; Tue, 28 Dec 1999 17:54:00 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Message-ID: <386940D2.B5C36E05@dynamicsoft.com>
Date: Tue, 28 Dec 1999 17:59:30 -0500
From: Jonathan Rosenberg <jdrosen@dynamicsoft.com>
Organization: dynamicsoft
X-Mailer: Mozilla 4.7 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: James Yu <james.yu@neustar.com>
Cc: "'Vance Shipley'" <vances@motivity.ca>,
        John Mainwaring <crm312a@nortelnetworks.com>,
        Mark Foster <mark.foster@neustar.com>,
        iptel@lists.research.bell-labs.com, sip@lists.research.bell-labs.com
Subject: Re: TRIP
References: <ED88182BFF78D211A4D800A0C9E9435C3BE9AB@dc02.npac.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

We've discussed this on the list. Several folks pointed out that
European systems support overlap, and it was decided that it was not
reasonable to mandate the originating gateway to do digit collection to
convert this to en-bloc, although it could be optionally done.

-Jonathan R.

James Yu wrote:
> 
> Vance,
> 
> Thanks for the explanation.
> 
> The key questions now are whether overlap signaling (e.g., Feature Group-D
> over Multi-Frequency facility) is still used somewhere and if used where the
> digit collection/enbloc signaling is performed in the existing SCN.
> 
> If overlap signaling is still used (someone please confirm this and indicate
> where it is still used) and a call is to be routed from the originating SCN
> to the Internet, I guess that the originating GW (SCN-Internet)probably has
> to perform the digit collection/enbloc signaling function so that overlap
> signaling does not propagate further into the Internet and/or the
> destination GW/SCN.  Otherwise, some entities in the Internet must perform
> that function.
> 
> James
> 
-- 
Jonathan D. Rosenberg                       200 Executive Drive
Chief Scientist                             Suite 120 
dynamicsoft                                 West Orange, NJ 07052
jdrosen@dynamicsoft.com                     FAX:   (732) 741-4778
http://www.cs.columbia.edu/~jdrosen         PHONE: (732) 741-7244
http://www.dynamicsoft.com

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


From owner-iptel-outgoing@lists.research.bell-labs.com  Thu Dec 30 06:44:44 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 GAA23095
	for <iptel-archive@odin.ietf.org>; Thu, 30 Dec 1999 06:44:43 -0500 (EST)
Received: by lists.research.bell-labs.com (Postfix)
	id 7D9EE52DB; Thu, 30 Dec 1999 06:33:55 -0500 (EST)
Delivered-To: iptel-outgoing-local@paperless.dnrc.bell-labs.com
Received: by lists.research.bell-labs.com (Postfix, from userid 20006)
	id BA25652DD; Thu, 30 Dec 1999 06:33:54 -0500 (EST)
Delivered-To: iptel-local@paperless.dnrc.bell-labs.com
Content-return: allowed
Date: Thu, 30 Dec 1999 11:31:11 +0000
From: "Drage, Keith" <Keith.Drage@siemenscomms.co.uk>
Subject: RE: TRIP
To: "'Vance Shipley'" <vances@motivity.ca>, James Yu <james.yu@neustar.com>,
        "'Jonathan Rosenberg'" <jdrosen@dynamicsoft.com>
Cc: John Mainwaring <crm312a@nortelnetworks.com>,
        Mark Foster <mark.foster@neustar.com>,
        iptel@lists.research.bell-labs.com, sip@lists.research.bell-labs.com
Message-id: <5AC7D3F55147D311AA1700A0C99583ED2C3607@beex03.sgcs.co.uk>
MIME-version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-type: text/plain
Content-transfer-encoding: 7BIT
Comments: ***********************************************
Comments: *    Our domain is now siemenscomms.co.uk     *
Comments: * Please update your address book accordingly *
Comments: ***********************************************
Sender: owner-iptel@lists.research.bell-labs.com
Precedence: bulk
Content-Transfer-Encoding: 7BIT

I don't know which part of Q.767 you are looking at, but my understanding
says that the mode of operation used across the international interface is
subject to bilateral agreement, and therefore neither method of address
sending is preferred. See statements throughout the document of the form:

"Note - The use of en-bloc or overlap signalling is based on bi-lateral
agreements"

I would also note that while Q.767 is still extensively used, particularly
for international operation, I would expect network providers to be
specifying base Q.761 through Q.764 nowadays for international operation,
noting that these documents do specify the behaviour of international
gateway exchanges. Current ETSI specifications (ISUP version 3) for
international operation are based on the 1997 version of the ITU-T
Recommendations, and will shortly move to the recently approved 1999
versions. Again these documents do not mandate either en-bloc or overlap
sending.

I have no direct understanding of the amount of use overlap dialling gets
across the international interface. I know operators try to collect as many
digits as possible before presenting the call to the international network
in the first place, if only because as soon as you send the Initial Address
Message (IAM), you have a circuit allocated that could potentially be used
elsewhere in generating revenue. This is done using timeouts as well as
knowledge of the numbering plan. This however does not necessarily mean that
they eliminate overlap dialling across the international interface, even if
they prefer not to use it.

The main justification for keeping overlap dialling has always been for
those countries with open numbering plans (primarily Germany and Austria).
Here, private networks using DDI are allowed to specify their own length of
the DDI number (within the constraints of the maximum number of digits
specified by E.164), and therefore even the receiving public network does
not necessarily know the length of any number unless it is an end subscriber
on their own system. I see no signs of these two countries wanting to remove
this capability, and therefore it will remain fundamentally part of E.164
(in that is unlikely that there will be a rule fixing the length of
subscriber numbers within a particular NDC). This probably also partly
explains the different views of these two countries on the best way to
handle number portability within their networks.

In my opinion, extensive use of timeouts is particularly unfriendly to the
calling user. To be effective, the timeouts have to be short, and occur just
when the user is referring back to his list for the remaining digits of the
long number he is dialling. The only real solution to avoiding overlap
dialling is where all terminals on the calling network are forced to predial
their numbers before hitting the send key. So far only the mobile networks
seem to have managed this.

Note that this is not a defence of overlap, merely a statement that it is a
fact of life in telecommunication signalling.

Keith Drage

> ---------------------------------------------------------------
> Keith Drage (keith.drage@siemenscomms.co.uk)
> Siemens Communications Limited,
> Technology Drive, Beeston, Nottingham NG9 1LA, UK
> Tel: +44 115 943 4991   Fax: +44 115 943 4992
> ---------------------------------------------------------------
> Internet communications are not secure and therefore Siemens
> Communications Limited does not accept legal responsibility for the
> contents of this message. Any views or opinions presented are solely
> those of the author and do not necessarily represent those of Siemens
> Communications Limited unless otherwise specifically stated.
> 


-----Original Message-----
From: Vance Shipley [mailto:vances@motivity.ca]
Sent: Tuesday, December 28, 1999 8:59 PM
To: James Yu; 'Jonathan Rosenberg'
Cc: John Mainwaring; Mark Foster; iptel@lists.research.bell-labs.com;
sip@lists.research.bell-labs.com
Subject: RE: TRIP



James Yu writes:
}
}  Overlap dialing probably happens often for international calls where the
}  switch in the originating country don't know much about the numbering
plan
}  of the destination country.  The criteria to begin processing the
received

Actually no.  The ITU recommendation Q.767 "Application of the ISDN user
part
of CCITT Signalling System No. 7 for international ISDN interconnections" is
used today on the International Plane between countries public telephone
networks.
This recommendation defines how ISUP signaling is to be applied toward the
application of international gateway interconnections to other countries
gateways.

Q.767 does not allow for overlap signaling.  If the local serving office
switch
for the originating subscriber uses overlap signaling to speed up call
routing
then an intermediate switch, or the gateway switch itself, will gather up
all
the digits and send the call forward enbloc.

Overlap signaling is not used in ANSI ISUP at all.  Overlap signaling was
used
in the days of in band signaling to speed up the delays caused by the
relatively
long duration of the tones used.  One area where it was almost always used
was
in fact international calls.  In the case of FGD signaling in North America
when
an international call was presented the country code was sent first using
<KP>CC<ST> so that the switch could begin route selection and grab an
outgoing
route while the rest of the digits where pulsed in.  This becomes
unnecessary
when ISUP signaling is used.

In my mind overlap signaling should never be generated but may have to be
supported
when received.  The receiving switch must buffer the received digits until
it
decides
that all have been received and then forward the call enbloc.  There are two
methods
the switch may use to recognize when all digits have been received.  The
best is
when
the switch has an accurate view of the dialing plan(s).  It would know that
country
code XX, city code YYY requires exactly 7 digits in the subscriber number
and
when it
received all 12 digits it would forward the call enbloc.  If it never
received
enough
digits it would eventually time out and release the call with a cause code
indicating
"address incomplete".  The second method simply applies an inter digit timer
and
maximum
number of digits.  Digits are buffered until either the maximum expected are
received,
the inter digit delay timer expires or an end-of-dial digit is recognized
(e.g.
<ST>
or '#').  These methods are generally both used where some dial plans are
recognized
(e.g. North American Plan) and the rest use the time out.

	-Vance




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

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


