From owner-mboned@lists.uoregon.edu  Mon Apr  4 16:56:08 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07830
	for <mboned-archive@lists.ietf.org>; Mon, 4 Apr 2005 16:56:08 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34Kij7i015245;
	Mon, 4 Apr 2005 13:44:45 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j34Kij1L015244;
	Mon, 4 Apr 2005 13:44:45 -0700 (PDT)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34KihjO014987
	for <mboned@lists.uoregon.edu>; Mon, 4 Apr 2005 13:44:43 -0700 (PDT)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id j34KiQR06006;
	Mon, 4 Apr 2005 23:44:26 +0300
Date: Mon, 4 Apr 2005 23:44:26 +0300 (EEST)
From: Pekka Savola <pekkas@netcore.fi>
To: Beau Williamson <bwilliam@cisco.com>
cc: "Manfredi, Albert E" <albert.e.manfredi@boeing.com>,
        John Kristoff <jtk@northwestern.edu>, mboned@lists.uoregon.edu
Subject: RE: mboned: I-D   ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
In-Reply-To: <6.0.0.22.2.20050329010513.0359e930@edison.cisco.com>
Message-ID: <Pine.LNX.4.61.0504042327180.5105@netcore.fi>
References: <EF40C42ACAB7A649B2EAE70C19B6CD6E037C3918@xch-ne-02.ne.nos.boeing.com>
 <6.0.0.22.2.20050329010513.0359e930@edison.cisco.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk

On Tue, 29 Mar 2005, Beau Williamson wrote:
>> As far as I can tell from RFC 2365, the IANA has access to 256 addresses in 
>> the 239/8 range, and I'm not using those. Perhaps that was our 
>> misunderstanding?
>
> Yes, I believe you misunderstand Scope Relative addresses.  This are the top 
> 256 addresses of any arbitrary scope defined within the 239/8 address space. 
> There is only one well-defined scope and that is the Site-Local 239.255/16 
> space which is considered the smallest scope possible.  This was 
> (well)defined so that applications like MADCAP would have a well-known 
> starting address in the smallest scope in the organization.

Using this as a stepping stone to try to move this discussion 
forward..

My reading of RFC2365 is pretty clear that the sites/enterprises that 
use any other part of 239/8 rather than just 239.192.0.0/14 or
239.25[345].0.0/16 are basically violating the RFC.  IMHO, this is 
pretty clear in 6.2.1:

    The ranges 239.0.0.0/10, 239.64.0.0/10 and 239.128.0.0/10 are
    unassigned and available for expansion of this space.  These ranges
    should be left unassigned until the 239.192.0.0/14 space is no longer
    sufficient. This is to allow for the possibility that future
    revisions of this document may define additional scopes on a scale
    larger than organizations.

In particular "allow for the possibility ..." clearly states that the 
intent is NOT to use these addresses, so that in future we could carve 
out blocks from there if we wanted (exactly like a proposal was to do 
now).

However, in practice ... many sites have made their own plans 
regardless of this prohibition (instead of using 239.192.0.0/14 -- 
that would have allowed about 60 /24's worth of scoping!).  (We 
certainly have used 239.1.2.3 ourself for an app but that's it.) 
Whether or not it would be even possible to use the lower bits if we 
wanted to is a question mark.

What is a bit less clear is which set of those addresses (per scope) 
are to be assigned by IANA (called confusingly scope-relative 
addresses in some contexts).  This is what RFC 2365 section 9.1 
describes, and http://www.iana.org/assignments/multicast-addresses 
lists.  These are at the moment restricted to about 256 diffferent 
apps; this may or may not be sufficient.

But aside from that as it seems that taking bits out of 239/8 might be 
a non-starter.

Two alternatives:
  - use a non-239/8 space for these assignments (reason: just to make 
it easier to block them in MSDP)
  - just forget about it completely, pointing out that RFC 2365 already 
has a policy allowing obtaining scoped multicast addresses (however, 
this is limited to 8 bits per scope, so this might not be sufficient 
for long term)

Thoughts?

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Mon Apr  4 17:53:02 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA16064
	for <mboned-archive@lists.ietf.org>; Mon, 4 Apr 2005 17:53:01 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34LhqwF002169;
	Mon, 4 Apr 2005 14:43:52 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j34LhqgL002166;
	Mon, 4 Apr 2005 14:43:52 -0700 (PDT)
Received: from slb-smtpout-01.boeing.com (slb-smtpout-01.boeing.com [130.76.64.48])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34Lhp4S002050
	for <mboned@lists.uoregon.edu>; Mon, 4 Apr 2005 14:43:51 -0700 (PDT)
Received: from stl-av-01.boeing.com ([192.76.190.6])
	by slb-smtpout-01.boeing.com (8.9.2.MG.10092003/8.8.5-M2) with ESMTP id OAA06423;
	Mon, 4 Apr 2005 14:43:38 -0700 (PDT)
Received: from XCH-NE-05P.ne.nos.boeing.com (localhost [127.0.0.1])
	by stl-av-01.boeing.com (8.11.3/8.11.3/MBS-AV-LDAP-01) with ESMTP id j34Lhb703880;
	Mon, 4 Apr 2005 16:43:37 -0500 (CDT)
Received: from XCH-NE-02.ne.nos.boeing.com ([128.225.80.202]) by XCH-NE-05P.ne.nos.boeing.com with Microsoft SMTPSVC(5.0.2195.6713);
	 Mon, 4 Apr 2005 17:43:36 -0400
X-MimeOLE: Produced By Microsoft Exchange V6.0.6603.0
Content-Class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: mboned: I-D   ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
Date: Mon, 4 Apr 2005 17:43:36 -0400
Message-ID: <EF40C42ACAB7A649B2EAE70C19B6CD6E037C3938@xch-ne-02.ne.nos.boeing.com>
Thread-Topic: mboned: I-D   ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
thread-index: AcU5V9OYMX4jk7IJSq+MFG5R6w5alwABP5QA
From: "Manfredi, Albert E" <albert.e.manfredi@boeing.com>
To: "Pekka Savola" <pekkas@netcore.fi>
Cc: <mboned@lists.uoregon.edu>
X-OriginalArrivalTime: 04 Apr 2005 21:43:36.0780 (UTC) FILETIME=[5B62B8C0:01C5395F]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by darkwing.uoregon.edu id j34Lhp4S002146
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk
Content-Transfer-Encoding: 8bit

> From: Pekka Savola [mailto:pekkas@netcore.fi]

> My reading of RFC2365 is pretty clear that the sites/enterprises that 
> use any other part of 239/8 rather than just 239.192.0.0/14 or
> 239.25[345].0.0/16 are basically violating the RFC.  IMHO, this is 
> pretty clear in 6.2.1:
> 
>     The ranges 239.0.0.0/10, 239.64.0.0/10 and 239.128.0.0/10 are
>     unassigned and available for expansion of this space.  
> These ranges
>     should be left unassigned until the 239.192.0.0/14 space 
> is no longer
>     sufficient. This is to allow for the possibility that future
>     revisions of this document may define additional scopes on a scale
>     larger than organizations.
> 
> In particular "allow for the possibility ..." clearly states that the 
> intent is NOT to use these addresses, so that in future we 
> could carve 
> out blocks from there if we wanted (exactly like a proposal was to do 
> now).

For future expansion, yes, but not for future expansion controlled by the IANA!

I'd say that the intent, also reflected in RFC 3171, is that the 239/8 block always be available for assignments made by organizations, much like RFC 1918. IANA is given access to the upper 256 in each scope only. This concept seems supported also by RFC 3171, which states:

9. Administratively Scoped Address Block (239/8)

   Addresses in the Administratively Scoped Address block are for local
   use within a domain and are described in [RFC2365].

9.1. Assignment Guidelines

   Since addresses in this block are local to a domain, no IANA
   assignment policy is required.

9.1.1. Relative Offsets

   The relative offsets [RFC2365] are used to ensure that a service can
   be located independent of the extent of the enclosing scope (see RFC
   2770 for details).  Since there are only 256 such offsets, the IANA
   should only assign a relative offset to a protocol that provides an
   infrastructure supporting service.  Examples of such services include
   the Session Announcement Protocol [RFC2974].  Pursuant to section
   4.4.2 of RFC 2780 [RFC2780], assignments of Relative Offsets follow
   an Expert Review, IESG Approval or Standards Action process.  See
   [IANA] for the current set of assignments.

> Two alternatives:
>   - use a non-239/8 space for these assignments (reason: just to make 
> it easier to block them in MSDP)
>   - just forget about it completely, pointing out that RFC 
> 2365 already 
> has a policy allowing obtaining scoped multicast addresses (however, 
> this is limited to 8 bits per scope, so this might not be sufficient 
> for long term)
> 
> Thoughts?

I'd say both of your options should be viable. From the point of view of the 239/8 range, seems to me that it's hard to say whether the "relative offsets" will run out first, or whether multicast addresses available to individual organizations will run out first. Is there a compelling reason now to violate RFC 3171?

Bert

_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Mon Apr  4 18:11:50 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA18137
	for <mboned-archive@lists.ietf.org>; Mon, 4 Apr 2005 18:11:49 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34M2WqS001601;
	Mon, 4 Apr 2005 15:02:32 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j34M2WFn001600;
	Mon, 4 Apr 2005 15:02:32 -0700 (PDT)
Received: from multicasttech.com (IDENT:root@lennon.multicasttech.com [63.105.122.7])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34M2UIt001532
	for <mboned@lists.uoregon.edu>; Mon, 4 Apr 2005 15:02:30 -0700 (PDT)
Received: from [70.179.108.58] (account <marshall_eubanks@multicasttech.com>)
  by multicasttech.com (CommuniGate Pro WebUser 3.4.8)
  with HTTP id 2748777; Mon, 04 Apr 2005 18:02:21 -0400
From: "Marshall Eubanks" <tme@multicasttech.com>
Subject: Re: mboned: I-D  
  ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
To: Pekka Savola <pekkas@netcore.fi>, Beau Williamson 
 <bwilliam@cisco.com>
Cc: "Manfredi, Albert E" <albert.e.manfredi@boeing.com>,
        John
  Kristoff <jtk@northwestern.edu>, mboned@lists.uoregon.edu
X-Mailer: CommuniGate Pro Web Mailer v.3.4.8
Date: Mon, 04 Apr 2005 18:02:21 -0400
Message-ID: <web-2748777@multicasttech.com>
In-Reply-To: <Pine.LNX.4.61.0504042327180.5105@netcore.fi>
MIME-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk
Content-Transfer-Encoding: 8bit

On Mon, 4 Apr 2005 23:44:26 +0300 (EEST)
 Pekka Savola <pekkas@netcore.fi> wrote:
> On Tue, 29 Mar 2005, Beau Williamson wrote:
> >> As far as I can tell from RFC 2365, the IANA has access to 256 addresses in 
> >> the 239/8 range, and I'm not using those. Perhaps that was our 
> >> misunderstanding?
> >
> > Yes, I believe you misunderstand Scope Relative addresses.  This are the top 
> > 256 addresses of any arbitrary scope defined within the 239/8 address space. 
> > There is only one well-defined scope and that is the Site-Local 239.255/16 
> > space which is considered the smallest scope possible.  This was 
> > (well)defined so that applications like MADCAP would have a well-known 
> > starting address in the smallest scope in the organization.
> 
> Using this as a stepping stone to try to move this discussion 
> forward..
> 
> My reading of RFC2365 is pretty clear that the sites/enterprises that 
> use any other part of 239/8 rather than just 239.192.0.0/14 or
> 239.25[345].0.0/16 are basically violating the RFC.  IMHO, this is 
> pretty clear in 6.2.1:
> 
>     The ranges 239.0.0.0/10, 239.64.0.0/10 and 239.128.0.0/10 are
>     unassigned and available for expansion of this space.  These ranges
>     should be left unassigned until the 239.192.0.0/14 space is no longer
>     sufficient. This is to allow for the possibility that future
>     revisions of this document may define additional scopes on a scale
>     larger than organizations.
> 
> In particular "allow for the possibility ..." clearly states that the 
> intent is NOT to use these addresses, so that in future we could carve 
> out blocks from there if we wanted (exactly like a proposal was to do 
> now).
> 
> However, in practice ... many sites have made their own plans 
> regardless of this prohibition (instead of using 239.192.0.0/14 -- 
> that would have allowed about 60 /24's worth of scoping!).  (We 
> certainly have used 239.1.2.3 ourself for an app but that's it.) 
> Whether or not it would be even possible to use the lower bits if we 
> wanted to is a question mark.
> 
> What is a bit less clear is which set of those addresses (per scope) 
> are to be assigned by IANA (called confusingly scope-relative 
> addresses in some contexts).  This is what RFC 2365 section 9.1 
> describes, and http://www.iana.org/assignments/multicast-addresses 
> lists.  These are at the moment restricted to about 256 diffferent 
> apps; this may or may not be sufficient.
> 
> But aside from that as it seems that taking bits out of 239/8 might be 
> a non-starter.
> 
> Two alternatives:
>   - use a non-239/8 space for these assignments (reason: just to make 
> it easier to block them in MSDP)
>   - just forget about it completely, pointing out that RFC 2365 already 
> has a policy allowing obtaining scoped multicast addresses (however, 
> this is limited to 8 bits per scope, so this might not be sufficient 
> for long term)
> 
> Thoughts?
> 

Carve out a /24 with potential for expansion for a first come, first serve
"swamp" for application developers. I would suggest 239.64.0.0/24

If enterprise customers complain, point 
them to the RFC2365, and tell them that they shouldn't using reserved space.

If no one bites, give the space back.

Regards
Marshall

P.S. I stand by my offer to assist the first entity in need
get e-GLOP space through a registry.

> -- 
> Pekka Savola                 "You each name yourselves king, yet the
> Netcore Oy                    kingdom bleeds."
> Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
> _______________________________________________________________
> user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
> web archive:  http://darkwing.uoregon.edu/~llynch/mboned/

_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Mon Apr  4 18:32:37 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA20071
	for <mboned-archive@lists.ietf.org>; Mon, 4 Apr 2005 18:32:36 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34MNcRB006400;
	Mon, 4 Apr 2005 15:23:39 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j34MNcaa006399;
	Mon, 4 Apr 2005 15:23:38 -0700 (PDT)
Received: from sj-iport-5.cisco.com (sj-iport-5.cisco.com [171.68.10.87])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34MNbRx006289
	for <mboned@lists.uoregon.edu>; Mon, 4 Apr 2005 15:23:38 -0700 (PDT)
Received: from sj-core-1.cisco.com (171.71.177.237)
  by sj-iport-5.cisco.com with ESMTP; 04 Apr 2005 15:23:34 -0700
Received: from edison.cisco.com (edison.cisco.com [171.71.180.109])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id j34MNPDr018909;
	Mon, 4 Apr 2005 15:23:25 -0700 (PDT)
Received: from bwilliam-t30.cisco.com (bwilliam-t30.cisco.com [64.101.135.161]) by edison.cisco.com (8.8.6 (PHNE_14041)/CISCO.SERVER.1.2) with ESMTP id PAA16987; Mon, 4 Apr 2005 15:23:29 -0700 (PDT)
Message-Id: <6.0.0.22.2.20050404155101.03364360@edison.cisco.com>
X-Sender: bwilliam@edison.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 6.0.0.22
Date: Mon, 04 Apr 2005 17:23:36 -0500
To: Pekka Savola <pekkas@netcore.fi>
From: Beau Williamson <bwilliam@cisco.com>
Subject: RE: mboned: I-D  
  ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
Cc: "Manfredi, Albert E" <albert.e.manfredi@boeing.com>,
        John Kristoff <jtk@northwestern.edu>, mboned@lists.uoregon.edu
In-Reply-To: <Pine.LNX.4.61.0504042327180.5105@netcore.fi>
References: <EF40C42ACAB7A649B2EAE70C19B6CD6E037C3918@xch-ne-02.ne.nos.boeing.com>
 <6.0.0.22.2.20050329010513.0359e930@edison.cisco.com>
 <Pine.LNX.4.61.0504042327180.5105@netcore.fi>
Mime-Version: 1.0
Content-Type: multipart/alternative;
	boundary="=====================_23427516==.ALT"
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk

--=====================_23427516==.ALT
Content-Type: text/plain; charset="us-ascii"; format=flowed

At 4/4/2005 03:44 PM, Pekka Savola wrote:

>Using this as a stepping stone to try to move this discussion forward..

Great. :)


>My reading of RFC2365 is pretty clear that the sites/enterprises that use 
>any other part of 239/8 rather than just 239.192.0.0/14 or
>239.25[345].0.0/16 are basically violating the RFC.  IMHO, this is pretty 
>clear in 6.2.1:
>
>    The ranges 239.0.0.0/10, 239.64.0.0/10 and 239.128.0.0/10 are
>    unassigned and available for expansion of this space.  These ranges
>    should be left unassigned until the 239.192.0.0/14 space is no longer
>    sufficient. This is to allow for the possibility that future
>    revisions of this document may define additional scopes on a scale
>    larger than organizations.

I don't think it's quite that clear.  See the following map that outlines 
the space defined in RFC 2365 (including Scope Relative ranges for each 
scope) for the remaining comments and discussions.

    239.255.255.255  +-------------------------+
                     |       Local Scope       |
      239.255.255.0  |      Scope Relative     |
    239.255.254.255  +-------------------------+
                     |                         |
                     |       Local Scope       |
        239.255.0.0  |                         |
    239.254.255.255  +-------------------------+
                     |                         |
                     |       Local Scope       |
                     |     Expansion Range     |
        239.253.0.0  |                         |
    239.252.255.255  +-------------------------+
                     |                         |
                     |                         |
                     |                         |
                     |        UNASSIGNED       |
                     |                         |
                     |                         |
        239.196.0.0  |                         |
    239.195.255.255  +-------------------------+
                     |     Org-Local Scope     |
      239.195.255.0  |      Scope Relative     |
    239.194.255.255  +-------------------------+
                     |                         |
                     |                         |
                     |                         |
                     |     Org-Local Scope     |
                     |                         |
                     |                         |
        239.192.0.0  |                         |
    239.191.255.255  +-------------------------+
                     |                         |
                     |                         |
                     |                         |
                     |     Org-Local Scope     |
                     |     Expansion Range     |
                     |                         |
                     |                         |
                     |                         |
          239.0.0.0  |                         |
                     +-------------------------+

>In particular "allow for the possibility ..." clearly states that the 
>intent is NOT to use these addresses, so that in future we could carve out 
>blocks from there if we wanted (exactly like a proposal was to do now).

One the other hand, it clearly states that the Org-Local Expansion space 
can be used if the Org-Local space fills up.  This tells me that the intent 
is that Enterprises can make use of the entire 239.0.0.0 - 239.195.255.255 
space for their private Org-Local space.  If they want to carve up this 
space into scopes of other sizes, then they are free to do so.

The same statement is applicable for the Local Scope in that 239.254/16 and 
239.253/16 is the expansion space for the Local Scope should you run out of 
room for the Local Scope.  This tells me that the intent is Enterprises can 
make use of the entire 239.253.0.0 - 239.255.255.255 space for their 
private Local space.

The end result is that the only space not implicitly defined in RFC 2365 is 
the range of addresses between 239.196.0.0 and 239.251.255.255.


>However, in practice ... many sites have made their own plans regardless 
>of this prohibition (instead of using 239.192.0.0/14 -- that would have 
>allowed about 60 /24's worth of scoping!).  (We certainly have used 
>239.1.2.3 ourself for an app but that's it.) Whether or not it would be 
>even possible to use the lower bits if we wanted to is a question mark.

I think the only thing that can clearly be stated about RFC 2365 is that it 
is quite unclear as to how this space can be carved up for private use 
especially when scopes between the size of Local (smallest possible) and 
Org-Local (largest possible within the Org or Enterprise).

I think the primary goal that Dave Meyer had when this was written (and 
we've had some conversations on this) was to provide at least two 
well-known scope ranges (Local and Org-Local) that applications could 
always COUNT on existing within the Enterprise network.  (I'm sure Dave 
will correct me if I misspeak.)


>What is a bit less clear is which set of those addresses (per scope) are 
>to be assigned by IANA (called confusingly scope-relative addresses in 
>some contexts).  This is what RFC 2365 section 9.1 describes, and 
>http://www.iana.org/assignments/multicast-addresses lists.  These are at 
>the moment restricted to about 256 diffferent apps; this may or may not be 
>sufficient.

I personally don't think its confusing at all. I think it is quite clear in 
section 9 that one of the primary tenets of RFC 2365 that should be 
observed is that the two Scope Relative ranges of the Org-Local Scope and 
the Local Scope should be left alone for Scope Relative addressing in these 
two scopes.

This gives applications two blocks of well-known address, one Local and the 
other Org-Local.  Thus MZAP always knows that it can count on 
239.255.255.254 being the multicast address for Local Scope MZAP and 
239.195.255.254 being the multicast address for Org-Local Scope MZAP.


>But aside from that as it seems that taking bits out of 239/8 might be a 
>non-starter.

I agrree. Other than the assigning addresses out of the UNASSIGNED section 
(see address map), all other addresses already assigned.  However, in 
reality, I agree that assigning any addresses out of the 239/8 space (which 
is pretty well considered to be "private" address space) is a non-starter.


>Two alternatives:
>  - use a non-239/8 space for these assignments (reason: just to make it 
> easier to block them in MSDP)

Well I don't think that's really what we want.  The reason I say this is 
that if a developer hard codes this address into an application, it screws 
up the Enterprise admin who has decided that this application should ONLY 
work within his/her Enterprise network in scope X whose configured 
boundaries are multicast ranges 239.mumble - 239.fratz.  Having hard coded 
addresses in the application requires an increasing number of exceptions to 
be added to the scope range ACLs. This results in multicast boundary ACL's 
that get out of hand.

No, IMHO, the proper way is to give the Enterprise operator control of what 
multicast address an application is to use based on the desired 
scope.  That means no hard coded addresses assigned to applications with 
very few exceptions.

>  - just forget about it completely, pointing out that RFC 2365 already 
> has a policy allowing obtaining scoped multicast addresses (however, this 
> is limited to 8 bits per scope, so this might not be sufficient for long term)

But it still doesn't solve the problem because the application doesn't know 
ahead of time what the scopes are that the Enterprise admin has defined and 
within which one he/she wants the app to run in.  Thus, the application 
still can't use hard coded addresses and some other 
"run-time"/"deploy-time" address assignment/discovery method must be employed.

So regardless of what we do, we STILL need this light-weight address 
discovery protocol so we can encourage app developers to do the right thing 
and NOT use hard coded addresses.

Beau 
--=====================_23427516==.ALT
Content-Type: text/html; charset="us-ascii"

<html>
<body>
At 4/4/2005 03:44 PM, Pekka Savola wrote:<br><br>
<blockquote type=cite class=cite cite>Using this as a stepping stone to
try to move this discussion forward..</blockquote><br>
Great. :)<br><br>
<br>
<blockquote type=cite class=cite cite>My reading of RFC2365 is pretty
clear that the sites/enterprises that use any other part of 239/8 rather
than just 239.192.0.0/14 or<br>
239.25[345].0.0/16 are basically violating the RFC.&nbsp; IMHO, this is
pretty clear in 6.2.1:<br><br>
&nbsp;&nbsp; The ranges 239.0.0.0/10, 239.64.0.0/10 and 239.128.0.0/10
are<br>
&nbsp;&nbsp; unassigned and available for expansion of this space.&nbsp;
These ranges<br>
&nbsp;&nbsp; should be left unassigned until the 239.192.0.0/14 space is
no longer<br>
&nbsp;&nbsp; sufficient. This is to allow for the possibility that
future<br>
&nbsp;&nbsp; revisions of this document may define additional scopes on a
scale<br>
&nbsp;&nbsp; larger than organizations.</blockquote><br>
I don't think it's quite <b>that </b>clear.&nbsp; See the following map
that outlines the space defined in RFC 2365 (including Scope Relative
ranges for each scope) for the remaining comments and discussions.&nbsp;
<br><br>
&nbsp;&nbsp; 239.255.255.255&nbsp; +-------------------------+<br>
&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; Local
Scope&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp; 239.255.255.0&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Scope Relative&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp; 239.255.254.255&nbsp; +-------------------------+<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&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; Local
Scope&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 239.255.0.0&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;
|<br>
&nbsp;&nbsp; 239.254.255.255&nbsp; +-------------------------+<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&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; Local
Scope&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp; Expansion Range&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 239.253.0.0&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;
|<br>
&nbsp;&nbsp; 239.252.255.255&nbsp; +-------------------------+<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&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;
UNASSIGNED&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 239.196.0.0&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;
|<br>
&nbsp;&nbsp; 239.195.255.255&nbsp; +-------------------------+<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp; Org-Local Scope&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp; 239.195.255.0&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Scope Relative&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp; 239.194.255.255&nbsp; +-------------------------+<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp; Org-Local Scope&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 239.192.0.0&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;
|<br>
&nbsp;&nbsp; 239.191.255.255&nbsp; +-------------------------+<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp; Org-Local Scope&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp; Expansion Range&nbsp;&nbsp;&nbsp;&nbsp; |<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 239.0.0.0&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;
|<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
+-------------------------+<br><br>
<blockquote type=cite class=cite cite>In particular &quot;allow for the
possibility ...&quot; clearly states that the intent is NOT to use these
addresses, so that in future we could carve out blocks from there if we
wanted (exactly like a proposal was to do now).</blockquote><br>
One the other hand, it clearly states that the Org-Local Expansion space
can be used if the Org-Local space fills up.&nbsp; This tells me that the
intent is that Enterprises can <b>make use of the entire 239.0.0.0 -
239.195.255.255 space for their private Org-Local space</b>.&nbsp; If
they want to carve up this space into scopes of other sizes, then they
are free to do so.<br><br>
The same statement is applicable for the Local Scope in that 239.254/16
and 239.253/16 is the expansion space for the Local Scope should you run
out of room for the Local Scope.&nbsp; This tells me that the intent is
Enterprises can <b>make use of the entire 239.253.0.0 - 239.255.255.255
space for their private Local space</b>.<br><br>
The end result is that the only space not implicitly defined in RFC 2365
is the range of addresses between 239.196.0.0 and 239.251.255.255.&nbsp;
<br><br>
<br>
<blockquote type=cite class=cite cite>However, in practice ... many sites
have made their own plans regardless of this prohibition (instead of
using 239.192.0.0/14 -- that would have allowed about 60 /24's worth of
scoping!).&nbsp; (We certainly have used 239.1.2.3 ourself for an app but
that's it.) Whether or not it would be even possible to use the lower
bits if we wanted to is a question mark.</blockquote><br>
I think the only thing that can clearly be stated about RFC 2365 is that
it is quite unclear as to how this space can be carved up for private use
especially when scopes between the size of Local (smallest possible) and
Org-Local (largest possible within the Org or Enterprise).&nbsp;
<br><br>
I think the primary goal that Dave Meyer had when this was written (and
we've had some conversations on this) was to provide at least two
well-known scope ranges (Local and Org-Local) that applications could
always COUNT on existing within the Enterprise network.&nbsp; (I'm sure
Dave will correct me if I misspeak.)<br><br>
<br>
<blockquote type=cite class=cite cite>What is a bit less clear is which
set of those addresses (per scope) are to be assigned by IANA (called
confusingly scope-relative addresses in some contexts).&nbsp; This is
what RFC 2365 section 9.1 describes, and
<a href="http://www.iana.org/assignments/multicast-addresses" eudora="autourl">http://www.iana.org/assignments/multicast-addresses</a>
lists.&nbsp; These are at the moment restricted to about 256 diffferent apps; this may or may not be sufficient.</blockquote><br>
I personally don't think its confusing at all. I think it is quite clear in section 9 that one of the <b>primary </b>tenets of RFC 2365 that should be observed is that the two Scope Relative ranges of the Org-Local Scope and the Local Scope should be left alone for Scope Relative addressing in these two scopes.&nbsp; <br><br>
This gives applications two blocks of well-known address, one Local and the other Org-Local.&nbsp; Thus MZAP <b>always </b>knows that it can count on 239.255.255.254 being the multicast address for Local Scope MZAP and 239.195.255.254 being the multicast address for Org-Local Scope MZAP.<br><br>
<br>
<blockquote type=cite class=cite cite>But aside from that as it seems that taking bits out of 239/8 might be a non-starter.</blockquote><br>
I agrree. Other than the assigning addresses out of the UNASSIGNED section (see address map), all other addresses already assigned.&nbsp; However, in reality, I agree that assigning any addresses out of the 239/8 space (which is pretty well considered to be &quot;private&quot; address space) is a non-starter.<br><br>
<br>
<blockquote type=cite class=cite cite>Two alternatives:<br>
&nbsp;- use a non-239/8 space for these assignments (reason: just to make it easier to block them in MSDP)</blockquote><br>
Well I don't think that's really what we want.&nbsp; The reason I say this is that if a developer hard codes this address into an application, it screws up the Enterprise admin who has decided that this application should ONLY work within his/her Enterprise network in scope X whose configured boundaries are multicast ranges 239.mumble - 239.fratz.&nbsp; Having hard coded addresses in the application requires an increasing number of exceptions to be added to the scope range ACLs. This results in multicast boundary ACL's that get out of hand.&nbsp; <br><br>
No, IMHO, the proper way is to give the Enterprise operator control of what multicast address an application is to use based on the desired scope.&nbsp; That means no hard coded addresses assigned to applications with very few exceptions.&nbsp; <br><br>
<blockquote type=cite class=cite cite>&nbsp;- just forget about it completely, pointing out that RFC 2365 already has a policy allowing obtaining scoped multicast addresses (however, this is limited to 8 bits per scope, so this might not be sufficient for long term)</blockquote><br>
But it still doesn't solve the problem because the application doesn't know ahead of time what the scopes are that the Enterprise admin has defined and within which one he/she wants the app to run in.&nbsp; Thus, the application still can't use hard coded addresses and some other &quot;run-time&quot;/&quot;deploy-time&quot; address assignment/discovery method must be employed.<br><br>
So regardless of what we do, we STILL need this light-weight address discovery protocol so we can encourage app developers to do the right thing and NOT use hard coded addresses.<br><br>
Beau</body>
</html>

--=====================_23427516==.ALT--
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Mon Apr  4 18:38:32 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA20571
	for <mboned-archive@lists.ietf.org>; Mon, 4 Apr 2005 18:38:31 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34MTrHp019004;
	Mon, 4 Apr 2005 15:29:53 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j34MTrDq019000;
	Mon, 4 Apr 2005 15:29:53 -0700 (PDT)
Received: from sj-iport-5.cisco.com (sj-iport-5.cisco.com [171.68.10.87])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34MTqjL018850
	for <mboned@lists.uoregon.edu>; Mon, 4 Apr 2005 15:29:53 -0700 (PDT)
Received: from sj-core-1.cisco.com (171.71.177.237)
  by sj-iport-5.cisco.com with ESMTP; 04 Apr 2005 15:29:48 -0700
Received: from edison.cisco.com (edison.cisco.com [171.71.180.109])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id j34MTbDr026453;
	Mon, 4 Apr 2005 15:29:37 -0700 (PDT)
Received: from bwilliam-t30.cisco.com (bwilliam-t30.cisco.com [64.101.135.161]) by edison.cisco.com (8.8.6 (PHNE_14041)/CISCO.SERVER.1.2) with ESMTP id PAA24536; Mon, 4 Apr 2005 15:29:36 -0700 (PDT)
Message-Id: <6.0.0.22.2.20050404172449.034aa548@edison.cisco.com>
X-Sender: bwilliam@edison.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 6.0.0.22
Date: Mon, 04 Apr 2005 17:29:43 -0500
To: "Marshall Eubanks" <tme@multicasttech.com>,
        Pekka Savola <pekkas@netcore.fi>
From: Beau Williamson <bwilliam@cisco.com>
Subject: Re: mboned: I-D  
  ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
Cc: "Manfredi, Albert E" <albert.e.manfredi@boeing.com>,
        John Kristoff <jtk@northwestern.edu>, mboned@lists.uoregon.edu
In-Reply-To: <web-2748777@multicasttech.com>
References: <Pine.LNX.4.61.0504042327180.5105@netcore.fi>
 <web-2748777@multicasttech.com>
Mime-Version: 1.0
Content-Type: multipart/alternative;
	boundary="=====================_23794424==.ALT"
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk

--=====================_23794424==.ALT
Content-Type: text/plain; charset="us-ascii"; format=flowed

At 4/4/2005 05:02 PM, Marshall Eubanks wrote:


>Carve out a /24 with potential for expansion for a first come, first serve
>"swamp" for application developers. I would suggest 239.64.0.0/24

Sorry, but that's just a non-starter for two reasons:

1) That range has already been reserved for Enterprises to "Expand" the 
Org-Local scope "should the normal address space of 239.192/14 be 
insufficient".

2) It's still inside of the 239/8 PRIVATE address space.  Would you 
consider allocating Unicast addresses from net 10???  That's crazy. 
Allocating addresses from 239/8 makes about as much sense.


>If enterprise customers complain, point
>them to the RFC2365, and tell them that they shouldn't using reserved space.

They are NOT using reserved space. They are using the EXPANSION space for 
the Org-local space.

Beau 
--=====================_23794424==.ALT
Content-Type: text/html; charset="us-ascii"

<html>
<body>
At 4/4/2005 05:02 PM, Marshall Eubanks wrote:<br><br>
<br>
<blockquote type=cite class=cite cite>Carve out a /24 with potential for
expansion for a first come, first serve<br>
&quot;swamp&quot; for application developers. I would suggest
239.64.0.0/24</blockquote><br>
Sorry, but that's just a non-starter for two reasons:<br><br>
1) That range has already been reserved for Enterprises to
&quot;Expand&quot; the Org-Local scope &quot;should the normal address
space of 239.192/14 be insufficient&quot;. <br><br>
2) It's still inside of the 239/8 <b><i>PRIVATE </i></b>address
space.&nbsp; Would you consider allocating Unicast addresses from net
10???&nbsp; That's crazy. Allocating addresses from 239/8 makes about as
much sense.<br><br>
<br>
<blockquote type=cite class=cite cite>If enterprise customers complain,
point <br>
them to the RFC2365, and tell them that they shouldn't using reserved
space.</blockquote><br>
They are <b>NOT </b>using reserved space. They are using the EXPANSION
space for the Org-local space.<br><br>
Beau</body>
</html>

--=====================_23794424==.ALT--
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Mon Apr  4 18:44:26 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA21122
	for <mboned-archive@lists.ietf.org>; Mon, 4 Apr 2005 18:44:26 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34MZfaG029580;
	Mon, 4 Apr 2005 15:35:41 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j34MZfka029579;
	Mon, 4 Apr 2005 15:35:41 -0700 (PDT)
Received: from sj-iport-5.cisco.com (sj-iport-5.cisco.com [171.68.10.87])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j34MZewq029517
	for <mboned@lists.uoregon.edu>; Mon, 4 Apr 2005 15:35:40 -0700 (PDT)
Received: from sj-core-1.cisco.com (171.71.177.237)
  by sj-iport-5.cisco.com with ESMTP; 04 Apr 2005 15:35:35 -0700
Received: from edison.cisco.com (edison.cisco.com [171.71.180.109])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id j34MZSDr002768;
	Mon, 4 Apr 2005 15:35:29 -0700 (PDT)
Received: from bwilliam-t30.cisco.com (bwilliam-t30.cisco.com [64.101.135.161]) by edison.cisco.com (8.8.6 (PHNE_14041)/CISCO.SERVER.1.2) with ESMTP id PAA02092; Mon, 4 Apr 2005 15:35:28 -0700 (PDT)
Message-Id: <6.0.0.22.2.20050404173405.035cb850@edison.cisco.com>
X-Sender: bwilliam@edison.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 6.0.0.22
Date: Mon, 04 Apr 2005 17:35:35 -0500
To: "Manfredi, Albert E" <albert.e.manfredi@boeing.com>,
        "Pekka Savola" <pekkas@netcore.fi>
From: Beau Williamson <bwilliam@cisco.com>
Subject: RE: mboned: I-D  
  ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
Cc: <mboned@lists.uoregon.edu>
In-Reply-To: <EF40C42ACAB7A649B2EAE70C19B6CD6E037C3938@xch-ne-02.ne.nos.
 boeing.com>
References: <EF40C42ACAB7A649B2EAE70C19B6CD6E037C3938@xch-ne-02.ne.nos.boeing.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk

At 4/4/2005 04:43 PM, Manfredi, Albert E wrote:

> > In particular "allow for the possibility ..." clearly states that the
> > intent is NOT to use these addresses, so that in future we
> > could carve
> > out blocks from there if we wanted (exactly like a proposal was to do
> > now).
>
>For future expansion, yes, but not for future expansion controlled by the 
>IANA!
>
>I'd say that the intent, also reflected in RFC 3171, is that the 239/8 
>block always be available for assignments made by organizations, much like 
>RFC 1918. IANA is given access to the upper 256 in each scope only. This 
>concept seems supported also by RFC 3171, which states:
>
>9. Administratively Scoped Address Block (239/8)
>
>    Addresses in the Administratively Scoped Address block are for local
>    use within a domain and are described in [RFC2365].
>
>9.1. Assignment Guidelines
>
>    Since addresses in this block are local to a domain, no IANA
>    assignment policy is required.
>
>9.1.1. Relative Offsets
>
>    The relative offsets [RFC2365] are used to ensure that a service can
>    be located independent of the extent of the enclosing scope (see RFC
>    2770 for details).  Since there are only 256 such offsets, the IANA
>    should only assign a relative offset to a protocol that provides an
>    infrastructure supporting service.  Examples of such services include
>    the Session Announcement Protocol [RFC2974].  Pursuant to section
>    4.4.2 of RFC 2780 [RFC2780], assignments of Relative Offsets follow
>    an Expert Review, IESG Approval or Standards Action process.  See
>    [IANA] for the current set of assignments.

Yes, exactly.  Thanks Bert, for pulling up the references that make this 
point.

Beau Williamson 
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Mon Apr  4 22:38:06 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA08440
	for <mboned-archive@lists.ietf.org>; Mon, 4 Apr 2005 22:38:05 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j352S3YB016357;
	Mon, 4 Apr 2005 19:28:03 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j352S3lo016356;
	Mon, 4 Apr 2005 19:28:03 -0700 (PDT)
Received: from aharp.ittns.northwestern.edu (aharp.ittns.northwestern.edu [129.105.153.69])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j352S288016298
	for <mboned@lists.uoregon.edu>; Mon, 4 Apr 2005 19:28:03 -0700 (PDT)
Received: from dsl017-022-068.chi1.dsl.speakeasy.net (aharp [127.0.0.1])
	by aharp.ittns.northwestern.edu (smtpd) with SMTP id A4FB4136C82
	for <mboned@lists.uoregon.edu>; Mon,  4 Apr 2005 21:27:57 -0500 (CDT)
Date: Mon, 4 Apr 2005 21:27:58 -0500
From: John Kristoff <jtk@northwestern.edu>
To: mboned@lists.uoregon.edu
Subject: Re: mboned: I-D   ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
In-Reply-To: <Pine.LNX.4.61.0504042327180.5105@netcore.fi>
References: <EF40C42ACAB7A649B2EAE70C19B6CD6E037C3918@xch-ne-02.ne.nos.boeing.com>
	<6.0.0.22.2.20050329010513.0359e930@edison.cisco.com>
	<Pine.LNX.4.61.0504042327180.5105@netcore.fi>
X-Mailer: Sylpheed-Claws 0.9.12 (GTK+ 1.2.10; i686-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Message-Id: <20050405022757.A4FB4136C82@aharp.ittns.northwestern.edu>
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk
Content-Transfer-Encoding: 7bit

On Mon, 4 Apr 2005 23:44:26 +0300 (EEST)
Pekka Savola <pekkas@netcore.fi> wrote:

> Thoughts?

After the initial volley being lead by your note I think you have
more good feedback and material with which to update your draft with.
I haven't really seen a whole lot just change right now, just some
good comments and clarifications.  If you want some help with the
text let me know and I can update what I sent last time.

John
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Mon Apr  4 23:08:34 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA10969
	for <mboned-archive@lists.ietf.org>; Mon, 4 Apr 2005 23:08:33 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j352xYUU007405;
	Mon, 4 Apr 2005 19:59:34 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j352xYqA007404;
	Mon, 4 Apr 2005 19:59:34 -0700 (PDT)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j352xW7s007348
	for <mboned@lists.uoregon.edu>; Mon, 4 Apr 2005 19:59:33 -0700 (PDT)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id j352xMO12468;
	Tue, 5 Apr 2005 05:59:23 +0300
Date: Tue, 5 Apr 2005 05:59:22 +0300 (EEST)
From: Pekka Savola <pekkas@netcore.fi>
To: "Manfredi, Albert E" <albert.e.manfredi@boeing.com>
cc: mboned@lists.uoregon.edu
Subject: RE: mboned: I-D   ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
In-Reply-To: <EF40C42ACAB7A649B2EAE70C19B6CD6E037C3938@xch-ne-02.ne.nos.boeing.com>
Message-ID: <Pine.LNX.4.61.0504050553550.12289@netcore.fi>
References: <EF40C42ACAB7A649B2EAE70C19B6CD6E037C3938@xch-ne-02.ne.nos.boeing.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk

On Mon, 4 Apr 2005, Manfredi, Albert E wrote:
> For future expansion, yes, but not for future expansion controlled 
> by the IANA!

Not quite.

IMHO, it's clear that IANA does not have control over the "future 
expansion".  Rather, it is reversed _to the IETF_ for future expansion 
to.

That's why http://www.iana.org/assignments/multicast-addresses lists 
these blocks as:

  239.000.000.000-239.063.255.255 Reserved                    [IANA]

In all other contexts, when the address space is "reserved", it's not 
supposed to be used -- not by the users, nor assigned by IANA; 
until we IETF folks decide that the space is no longer reserved.

All these blocks have (IMHO) been reserved for the _IETF_ to decide 
what to do with them.

Too bad this is confusing, and to a degree slightly contradictory 
between 2365 and 3171..

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Fri Apr  8 12:15:20 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA08364
	for <mboned-archive@lists.ietf.org>; Fri, 8 Apr 2005 12:15:19 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j38G7Ahr020593;
	Fri, 8 Apr 2005 09:07:11 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j38G7A7b020591;
	Fri, 8 Apr 2005 09:07:10 -0700 (PDT)
Received: from m106.maoz.com (m106.maoz.com [205.167.76.9])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j38G7AO4020558
	for <mboned@lists.uoregon.edu>; Fri, 8 Apr 2005 09:07:10 -0700 (PDT)
Received: from m106.maoz.com (localhost.localdomain [127.0.0.1])
	by m106.maoz.com (8.13.2/8.13.2) with ESMTP id j38G54r2017159;
	Fri, 8 Apr 2005 09:05:04 -0700
Received: (from dmm@localhost)
	by m106.maoz.com (8.13.2/8.12.11/Submit) id j38G54YH017158;
	Fri, 8 Apr 2005 09:05:04 -0700
X-Authentication-Warning: m106.maoz.com: dmm set sender to dmm@1-4-5.net using -f
Date: Fri, 8 Apr 2005 09:05:04 -0700
From: David Meyer <dmm@1-4-5.net>
To: mboned@lists.uoregon.edu
Cc: dthaler@microsoft.com, mohitt@microsoft.com, amitag@microsoft.com,
        lorenzo@cisco.com, pusateri@juniper.net
Subject: mboned: WGLC for draft-ietf-mboned-auto-multicast-04.txt
Message-ID: <20050408160504.GA17082@1-4-5.net>
Mime-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="r5Pyd7+fXNt84Ff3"
Content-Disposition: inline
User-Agent: Mutt/1.4.1i
X-public-key: http://www.1-4-5.net/~dmm/public-key.asc
X-gpg-fingerprint: 2409 8B50 B389 A307 BA5C 2A16 3918 03D6 A099 D8A7
X-philosophy: "I find your lack of faith disturbing." -- Darth Vader, Star Wars Episode IV.
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk


--r5Pyd7+fXNt84Ff3
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable


	Folks

	This note starts the WG Last Call for comments on
	draft-ietf-mboned-auto-multicast-04.txt, "Automatic IP
	Multicast Without Explicit Tunnels (AMT)". It can be
	found on =20
=09
	 ftp://ftp.ietf.org/internet-drafts/draft-ietf-mboned-auto-multicast-04.txt

	Please review the document carefully, and send your
	feedback to the list.  Please also indicate whether or
	not you believe that this document is ready to go to the
	IESG.=20

	This Last Call will end on 22 Apr 2005 at 1400 PDT (UTC/GMT-7).

	Thanks,

	Dave









--r5Pyd7+fXNt84Ff3
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFCVquwORgD1qCZ2KcRAjIOAKCT2VT2ySuMdW0KQQ1AptNghzwp7wCdEQ3p
8SdaKIEBC9gPUo7LrkSFJUM=
=2XNL
-----END PGP SIGNATURE-----

--r5Pyd7+fXNt84Ff3--
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Fri Apr  8 12:23:17 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA09333
	for <mboned-archive@lists.ietf.org>; Fri, 8 Apr 2005 12:23:16 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j38GHuge008503;
	Fri, 8 Apr 2005 09:17:56 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j38GHuxn008492;
	Fri, 8 Apr 2005 09:17:56 -0700 (PDT)
Received: from m106.maoz.com (m106.maoz.com [205.167.76.9])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j38GHt4M008457
	for <mboned@lists.uoregon.edu>; Fri, 8 Apr 2005 09:17:55 -0700 (PDT)
Received: from m106.maoz.com (localhost.localdomain [127.0.0.1])
	by m106.maoz.com (8.13.2/8.13.2) with ESMTP id j38GHr3Q017605;
	Fri, 8 Apr 2005 09:17:53 -0700
Received: (from dmm@localhost)
	by m106.maoz.com (8.13.2/8.12.11/Submit) id j38GHpCT017604;
	Fri, 8 Apr 2005 09:17:51 -0700
X-Authentication-Warning: m106.maoz.com: dmm set sender to dmm@1-4-5.net using -f
Date: Fri, 8 Apr 2005 09:17:51 -0700
From: David Meyer <dmm@1-4-5.net>
To: mboned@lists.uoregon.edu
Cc: iana@iana.org
Subject: mboned: Re: I-D ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
Message-ID: <20050408161751.GA17531@1-4-5.net>
Mime-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="tKW2IUtsqtDRztdT"
Content-Disposition: inline
User-Agent: Mutt/1.4.1i
X-public-key: http://www.1-4-5.net/~dmm/public-key.asc
X-gpg-fingerprint: 2409 8B50 B389 A307 BA5C 2A16 3918 03D6 A099 D8A7
X-philosophy: "I find your lack of faith disturbing." -- Darth Vader, Star Wars Episode IV.
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk


--tKW2IUtsqtDRztdT
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

	Folks,

	All of this discussion, in addition to the accelerating
	land-rush on IPv4 multicast addresses, makes me think
	that we need new thinking in this entire area. I've been
	speaking with our area director a bit and we've been
	kicking around the idea of holding a "multicast
	addressing BOF" in Paris. This BOF would have the
	objective of generating new thinking (as opposed to
	exploring the possible need for a new working group).=20

	What I would like to do is:

	(i).	Get the WG's feeling about the idea of having a
		BOF on this topic, and

	(ii).	Solicit topics from those that think a BOF is a
		good idea. Topics would include (but are not
		limited to) formulations of the problem statement,
		and innovative solutions to the addressing problem.=20

	Again, the objective of such a BOF would be to solidify a
	problem statement, and generate some new thinking around
	the whole IPv4 multicast address allocation issue.

	Comments?

	Dave

--tKW2IUtsqtDRztdT
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFCVq6vORgD1qCZ2KcRAsbQAJ9amvcVWSQ0aj8rolinPdES3euzcACeLzkz
BQOPX5v1np/jTrVktFZMWS8=
=Vj0d
-----END PGP SIGNATURE-----

--tKW2IUtsqtDRztdT--
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Sun Apr 10 12:14:57 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA29772
	for <mboned-archive@lists.ietf.org>; Sun, 10 Apr 2005 12:14:57 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3AG4dlV017220;
	Sun, 10 Apr 2005 09:04:39 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j3AG4d1P017219;
	Sun, 10 Apr 2005 09:04:39 -0700 (PDT)
Received: from sj-iport-4.cisco.com (sj-iport-4.cisco.com [171.68.10.86])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3AG4cPY017185
	for <mboned@lists.uoregon.edu>; Sun, 10 Apr 2005 09:04:38 -0700 (PDT)
Received: from sj-core-3.cisco.com (171.68.223.137)
  by sj-iport-4.cisco.com with ESMTP; 10 Apr 2005 09:04:33 -0700
Received: from xbh-sjc-211.amer.cisco.com (xbh-sjc-211.cisco.com [171.70.151.144])
	by sj-core-3.cisco.com (8.12.10/8.12.6) with ESMTP id j3AG3ugu017418;
	Sun, 10 Apr 2005 09:04:30 -0700 (PDT)
Received: from xmb-sjc-22c.amer.cisco.com ([128.107.191.47]) by xbh-sjc-211.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	 Sun, 10 Apr 2005 09:04:20 -0700
Received: from [192.168.1.8] ([10.21.146.36]) by xmb-sjc-22c.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	 Sun, 10 Apr 2005 09:04:19 -0700
In-Reply-To: <20050408160504.GA17082@1-4-5.net>
References: <20050408160504.GA17082@1-4-5.net>
Mime-Version: 1.0 (Apple Message framework v619.2)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <4ee003b57c0ebfa8470d21d6e1e262b1@cisco.com>
Content-Transfer-Encoding: 7bit
Cc: mohitt@microsoft.com, John Zwiebel <jzwiebel@cisco.com>,
        mboned@lists.uoregon.edu, lorenzo@cisco.com, dthaler@microsoft.com,
        pusateri@juniper.net, amitag@microsoft.com
From: John Zwiebel <jzwiebel@cisco.com>
Subject: Re: mboned: WGLC for draft-ietf-mboned-auto-multicast-04.txt
Date: Sun, 10 Apr 2005 09:04:32 -0700
To: David Meyer <dmm@1-4-5.net>
X-Mailer: Apple Mail (2.619.2)
X-OriginalArrivalTime: 10 Apr 2005 16:04:19.0754 (UTC) FILETIME=[F42190A0:01C53DE6]
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk
Content-Transfer-Encoding: 7bit


On Apr 8, 2005, at 9:05 AM, David Meyer wrote:

>
> 	This Last Call will end on 22 Apr 2005 at 1400 PDT (UTC/GMT-7).
>
> 	Thanks,
>

I went through it for the first time.
There were a couple of places that I got lost in the context switch
of which system is the relay and which is the gateway -- not that
I didn't understand where they were in the network, but there was
confusion about which system was doing what.

I guess my main concern is the gateway definition doesn't explain
how to handle PIM joins received by the gateway from a multicast
island that the gateway is suppose to be serving.  And the restriction
on the source/group address that a gateway can use seems to make
it impossible for a source within a multicast island to send SSM,
perhaps I missed the sentence that says, "only a gateway can source
multicast traffic to other gateways and/or relays to be forwarded to
the rest of the multicast enabled network."?

FWIW: perhaps its because I've been around forever, but the MBONE is,
IMHO, dead.  The MBONE used DVMRP tunnels to extend multicast as an
overlayed logical network across the internet.  Now that most providers
do multicast natively, there is no need for the MBONE.  Using a term
that has a certain meaning in a new context happens all the time,
but there are some terms (IMHO, MBONE being one) that have been
so strongly defined that trying to morph them into something else is
a mistake.  Of course, I should be willing to provide an alternative,
but I don't have one.

For example, when IPv6 is deployed almost everywhere, are we still
going to connect to the 6-bone?  and what is the internet now?  the 
4-bone?
If we're going to communicate with one another, we need to be
explicit in our definitions and terms.

Following is other notes that includes comments on missing commas etc.
Mostly though its a request for some clarity on certain portions where
I found myself thinking of at least two ways to interpret what I was
reading.

Intro:
    2. Allowing isolated sites/hosts to transmit the SSM flavor of
       multicast.
 > after reading this entire thing, I see how an isolated host that is
also configured as an AMT gateway can transmit SSM, but I do not see
how a host within an isolated site can transmit SSM.

Page 5: extra paren
  This requires the relay to keep state for each
    gateway which has joined a particular group or (source, group) pair).


Page 5:
  Since an AMT gateway may be a host,
    and hosts typically do not implement routing protocols, gateways will
    use IGMP/MLD as described in Section 5 below.

John says: I think that there are plenty of FreeBSD/Linux systems 
available to
	do PIM.  Is this draft primarily to be implemented on a host (eg a
	Windows PC)

Para 4.1.1: what happens if a relay has too many tunnels?  How does the
	AMT gateway find a new relay?


5.5.  AMT Membership Update

    "An AMT Membership Update is sent from the originator to the
    respondent containing the original IGMP/MLD Membership/Listener
    Report or Leave/Done received over the AMT pseudo-interface.  It
    echoes the Response MAC received in the AMT Membership Query so the
    respondent can verify return routability to the originator."

John says: the IGMP/MLD report... isn't this -generated- by the AMT 
Gateway or Relay?
	And isn't it really a report sent because the Gateway/Relay wants to
	receive multicast traffic and so should be equivalent to sending a PIM
	join (on a router) or sending a IGMP/MLD report (from a host)?  Are you
	suggesting that there are separate entities on the gateway?  The
	switch from talking about the gateway to talking about the interface
	on the gateway was totally missed.


Para 6.4: I'd like some kind of a note that says something like
	"as opposed to forward 	multicast traffic for a source inside
	the domain it serves".  I'm getting confused between a host that has
	no native multicast support and acts as its own AMT gateway and
	a router that is providing AMT gateway service for a multicast
	domain.

Para 6.5: "ditto"  This only makes sense to me if this is a host acting
	as its own gateway.  Explicitly stating this fact will clear up
	confusion.  And explicitly stating that "e" is any value the host
	wants to append will also help.  (Since if I'm not guessing right...)


6.6.  Receiving IGMPv3/MLDv2 Reports at the Gateway

    When an AMT Request is received by the gateway, it follows the same
    3-way handshake procedure a relay would follow if it received the AMT
    Request.

John says: HUH?  Let's just say I read this 3 times, then 
cut-and-pasted it before
	I figured out what you meant.  Put the period before "follow" and 
delete
	the rest.

inside of 6.6 it says:
    If S is not the AMT Gateway Anycast Address, the packet is dropped.
    If G does not contain the low bits of the global unicast address (as
    described above), the packet is also dropped.

So now I'm really, really confused because the beginning of this draft
suggested that we are dealing with a Gateway that could support an
entire multicast island (domain).  Suddenly it appears that we have
a bunch of restrictions on the group addresses that can be used
for sources inside of that island and I totally missed that there would
have to be some kind of NAT performed (or address translation) by
the AMT gateway to achieve this support.  If it was there, it needs to
be emphasized.

"7.3...
  When data arrives for that group, "

 From where?  One assumes that it comes in on any interface other than 
the
AMT interface, but its helpful to be explicit.

Para 7.4: again, how do you deal with the multicast island?


8.  IANA Considerations

    The IANA should allocate a prefix dedicated to the public AMT Relays
    to the native

 > It isn't -just- the relays.  Or is it...

    Furthermore, to support sourcing SSM traffic from AMT gateways, the
    IANA should allocate a subnet prefix dedicated to the AMT link.

John says: Let's see, I understood this was suppose to emulate a NBMA 
network?
     ie, this is a single subnet on which both gateways and relays 
reside.
     and /32 routes are advertised.  Where did I miss this?
     OIC now... FWIW I think the terms "link-layer" and "NBMA" are used
     in a misleading context because it sounds as if these are suppose
     to emulate being on the same subnet.  This is not true.  There is
     a subnet for all the AMT relays and a subnet for all the AMT 
gateways.
     Or did I get something else wrong?


I don't think PIMSM should refer to RFC2362.  The newer draft should
be used since there have been significant changes.

_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Sun Apr 10 22:40:54 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA13065
	for <mboned-archive@lists.ietf.org>; Sun, 10 Apr 2005 22:40:54 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3B2WlV0021585;
	Sun, 10 Apr 2005 19:32:47 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j3B2Wl8s021584;
	Sun, 10 Apr 2005 19:32:47 -0700 (PDT)
Received: from students2.iit.edu (students2.iit.edu [216.47.143.102])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3B2WkPt021573
	for <mboned@lists.uoregon.edu>; Sun, 10 Apr 2005 19:32:46 -0700 (PDT)
Received: from iit.edu (localhost [127.0.0.1])
 by students2.iit.edu (iPlanet Messaging Server 5.2 Patch 2 (built Jul 14
 2004)) with ESMTP id <0IER00JFPG31Q3@students2.iit.edu> for
 mboned@lists.uoregon.edu; Sun, 10 Apr 2005 21:25:49 -0500 (CDT)
Received: from [216.47.130.232] by students2.iit.edu (mshttpd); Sun,
 10 Apr 2005 20:25:49 -0600
Date: Sun, 10 Apr 2005 20:25:49 -0600
From: Venketesan <venkthi1@iit.edu>
Subject: mboned: Re:WGLC for draft-ietf-mboned-auto-multicast-04.txt
To: John Zwiebel <jzwiebel@cisco.com>
Cc: David Meyer <dmm@1-4-5.net>, mohitt@microsoft.com,
        mboned@lists.uoregon.edu, lorenzo@cisco.com, dthaler@microsoft.com,
        pusateri@juniper.net, amitag@microsoft.com
Message-id: <6042996604418c.604418c6042996@iit.edu>
MIME-version: 1.0
X-Mailer: iPlanet Messenger Express 5.2 Patch 2 (built Jul 14 2004)
Content-type: text/plain; charset=us-ascii
Content-language: en
Content-transfer-encoding: 7BIT
Content-disposition: inline
X-Accept-Language: en
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk
Content-Transfer-Encoding: 7BIT

Hi,

I am not clear about the following points about AMT:
1. I agree with john on the point that AMT may be having restrictions 
on the usage of the multicast addresses in SSM mode (Also AMT does not 
support sending to ASM groups). 

2. Abstraction of a "source" for SSM is an AMT gateway. It seems in a 
domain, even if there are more than one source they are treated as a 
single S in <S,G> by receivers external to the domain i.e. all the 
sources within an AMT gateway are represented by the same AMT GW.

Consider an example, say the unicast address of the AMT gateway is 
a.b.c.d and the anycast address prefix allotted to it is x.y/16, then 
the usable SSM address range by the hosts within the AMT gateway's 
domain is 232.c.d/24. The S in <S,G> is already fixed as x.y.a.b, 
hence the address space is <x.y.a.b, 232.c.d.(0-255)> with a size of 
2*8.     

3.How does AMT support receiving ASM traffic? it is said that the 
gateways announce the groups they are interested in
to a relay and the relay maintains a list of gateway_ips-groups. So it 
sends multicast traffic it "receives" to all these
gateways downstream. Now how does the AMT relay receive the traffic? 
does it forward/flood some kind of a join to other 
relay? Is DVMRP running between relays in the MBONE and is it scalable?


4. Does AMT in ASM mode force all traffic to go through the relays in 
the MBONE thus bypassing/overiding the connectivity 
between the RPs of the domains formed by MSDP/MBGP. I mean how does 
AMT incorporate MSDP and MBGP. My understanding is this is the 
reason "AMT does not support sending to non-SSM groups" as stated in 
the draft.

5.This is a question about the progressive deplotyment capability of 
AMT
Suppose an initial connectivity is as shown below
------------
AMT gateway |---------------
------------		    |	
			------------
			|AMT relay  |--------MBONE
			------------	
------------		    |	
AMT gateway |----------------
------------	
subsequently another AMT gateway is added higher in the domain 
(towards MBONE) of the existing gateways as shown below:

------------
AMT gateway |---------------
------------		    |---------------------	
			------------	        --|---------
		        |AMT gateway |-------- | AMT Relay |----- MBONE
			------------	        ----------- 
------------		    |---------------------|	
AMT gateway |----------------
------------	
Now although a new AMT gateway has been added the nearest relay 
remains the same for the
existing gateways. How to ensure that all the traffic is sent from the 
relay 
the new gateway that duplicates it to the old gateways, instead of 
sending two copies to the existing gateways.
meaning u no more need the connectivity between the original gateways 
and the relay

regards,
venkat

----- Original Message -----
From: John Zwiebel <jzwiebel@cisco.com>
Date: Sunday, April 10, 2005 10:04 am
Subject: Re: mboned: WGLC for draft-ietf-mboned-auto-multicast-04.txt

> 
> On Apr 8, 2005, at 9:05 AM, David Meyer wrote:
> 
> >
> > 	This Last Call will end on 22 Apr 2005 at 1400 PDT (UTC/GMT-7).
> >
> > 	Thanks,
> >
> 
> I went through it for the first time.
> There were a couple of places that I got lost in the context switch
> of which system is the relay and which is the gateway -- not that
> I didn't understand where they were in the network, but there was
> confusion about which system was doing what.
> 
> I guess my main concern is the gateway definition doesn't explain
> how to handle PIM joins received by the gateway from a multicast
> island that the gateway is suppose to be serving.  And the 
restriction
> on the source/group address that a gateway can use seems to make
> it impossible for a source within a multicast island to send SSM,
> perhaps I missed the sentence that says, "only a gateway can source
> multicast traffic to other gateways and/or relays to be forwarded to
> the rest of the multicast enabled network."?
> 
> FWIW: perhaps its because I've been around forever, but the MBONE is,
> IMHO, dead.  The MBONE used DVMRP tunnels to extend multicast as an
> overlayed logical network across the internet.  Now that most 
> providersdo multicast natively, there is no need for the MBONE.  
> Using a term
> that has a certain meaning in a new context happens all the time,
> but there are some terms (IMHO, MBONE being one) that have been
> so strongly defined that trying to morph them into something else is
> a mistake.  Of course, I should be willing to provide an alternative,
> but I don't have one.
> 
> For example, when IPv6 is deployed almost everywhere, are we still
> going to connect to the 6-bone?  and what is the internet now?  
> the 
> 4-bone?
> If we're going to communicate with one another, we need to be
> explicit in our definitions and terms.
> 
> Following is other notes that includes comments on missing commas 
etc.
> Mostly though its a request for some clarity on certain portions 
where
> I found myself thinking of at least two ways to interpret what I was
> reading.
> 
> Intro:
>    2. Allowing isolated sites/hosts to transmit the SSM flavor of
>       multicast.
> > after reading this entire thing, I see how an isolated host 
> that is
> also configured as an AMT gateway can transmit SSM, but I do not see
> how a host within an isolated site can transmit SSM.
> 
> Page 5: extra paren
>  This requires the relay to keep state for each
>    gateway which has joined a particular group or (source, group) 
> pair).
> 
> Page 5:
>  Since an AMT gateway may be a host,
>    and hosts typically do not implement routing protocols, 
> gateways will
>    use IGMP/MLD as described in Section 5 below.
> 
> John says: I think that there are plenty of FreeBSD/Linux systems 
> available to
> 	do PIM.  Is this draft primarily to be implemented on a host 
(eg a
> 	Windows PC)
> 
> Para 4.1.1: what happens if a relay has too many tunnels?  How 
> does the
> 	AMT gateway find a new relay?
> 
> 
> 5.5.  AMT Membership Update
> 
>    "An AMT Membership Update is sent from the originator to the
>    respondent containing the original IGMP/MLD Membership/Listener
>    Report or Leave/Done received over the AMT pseudo-interface.  It
>    echoes the Response MAC received in the AMT Membership Query 
> so the
>    respondent can verify return routability to the originator."
> 
> John says: the IGMP/MLD report... isn't this -generated- by the 
> AMT 
> Gateway or Relay?
> 	And isn't it really a report sent because the Gateway/Relay 
wants to
> 	receive multicast traffic and so should be equivalent to 
sending 
> a PIM
> 	join (on a router) or sending a IGMP/MLD report (from a 
host)?  
> Are you
> 	suggesting that there are separate entities on the gateway?  
The
> 	switch from talking about the gateway to talking about the 
interface
> 	on the gateway was totally missed.
> 
> 
> Para 6.4: I'd like some kind of a note that says something like
> 	"as opposed to forward 	multicast traffic for a source inside
> 	the domain it serves".  I'm getting confused between a host 
that has
> 	no native multicast support and acts as its own AMT gateway and
> 	a router that is providing AMT gateway service for a multicast
> 	domain.
> 
> Para 6.5: "ditto"  This only makes sense to me if this is a host 
> acting	as its own gateway.  Explicitly stating this fact will 
> clear up
> 	confusion.  And explicitly stating that "e" is any value the 
host
> 	wants to append will also help.  (Since if I'm not guessing 
right...)
> 
> 
> 6.6.  Receiving IGMPv3/MLDv2 Reports at the Gateway
> 
>    When an AMT Request is received by the gateway, it follows the 
> same    3-way handshake procedure a relay would follow if it 
> received the AMT
>    Request.
> 
> John says: HUH?  Let's just say I read this 3 times, then 
> cut-and-pasted it before
> 	I figured out what you meant.  Put the period before "follow" 
and 
> delete
> 	the rest.
> 
> inside of 6.6 it says:
>    If S is not the AMT Gateway Anycast Address, the packet is 
> dropped.    If G does not contain the low bits of the global 
> unicast address (as
>    described above), the packet is also dropped.
> 
> So now I'm really, really confused because the beginning of this 
draft
> suggested that we are dealing with a Gateway that could support an
> entire multicast island (domain).  Suddenly it appears that we have
> a bunch of restrictions on the group addresses that can be used
> for sources inside of that island and I totally missed that there 
> wouldhave to be some kind of NAT performed (or address 
> translation) by
> the AMT gateway to achieve this support.  If it was there, it 
> needs to
> be emphasized.
> 
> "7.3...
>  When data arrives for that group, "
> 
> From where?  One assumes that it comes in on any interface other 
> than 
> the
> AMT interface, but its helpful to be explicit.
> 
> Para 7.4: again, how do you deal with the multicast island?
> 
> 
> 8.  IANA Considerations
> 
>    The IANA should allocate a prefix dedicated to the public AMT 
> Relays    to the native
> 
> > It isn't -just- the relays.  Or is it...
> 
>    Furthermore, to support sourcing SSM traffic from AMT 
> gateways, the
>    IANA should allocate a subnet prefix dedicated to the AMT link.
> 
> John says: Let's see, I understood this was suppose to emulate a 
> NBMA 
> network?
>     ie, this is a single subnet on which both gateways and relays 
> reside.
>     and /32 routes are advertised.  Where did I miss this?
>     OIC now... FWIW I think the terms "link-layer" and "NBMA" are 
> used     in a misleading context because it sounds as if these are 
> suppose     to emulate being on the same subnet.  This is not 
> true.  There is
>     a subnet for all the AMT relays and a subnet for all the AMT 
> gateways.
>     Or did I get something else wrong?
> 
> 
> I don't think PIMSM should refer to RFC2362.  The newer draft should
> be used since there have been significant changes.
> 
> _______________________________________________________________
> user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
> web archive:  http://darkwing.uoregon.edu/~llynch/mboned/
> 
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Wed Apr 13 06:53:46 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA14209
	for <mboned-archive@lists.ietf.org>; Wed, 13 Apr 2005 06:53:46 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3DAgtJg012092;
	Wed, 13 Apr 2005 03:42:55 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j3DAgtsm012090;
	Wed, 13 Apr 2005 03:42:55 -0700 (PDT)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3DAgrdA012051
	for <mboned@lists.uoregon.edu>; Wed, 13 Apr 2005 03:42:54 -0700 (PDT)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id j3DAgYK03555;
	Wed, 13 Apr 2005 13:42:35 +0300
Date: Wed, 13 Apr 2005 13:42:34 +0300 (EEST)
From: Pekka Savola <pekkas@netcore.fi>
To: David Meyer <dmm@1-4-5.net>
cc: mboned@lists.uoregon.edu, iana@iana.org
Subject: Re: mboned: Re: I-D ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
In-Reply-To: <20050408161751.GA17531@1-4-5.net>
Message-ID: <Pine.LNX.4.61.0504131337270.3445@netcore.fi>
References: <20050408161751.GA17531@1-4-5.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk

On Fri, 8 Apr 2005, David Meyer wrote:
> 	What I would like to do is:
>
> 	(i).	Get the WG's feeling about the idea of having a
> 		BOF on this topic, and

I think this would make sense.

> 	(ii).	Solicit topics from those that think a BOF is a
> 		good idea. Topics would include (but are not
> 		limited to) formulations of the problem statement,
> 		and innovative solutions to the addressing problem.

It would probably be good if draft-ietf-mboned-addrarch-01.txt was in 
the RFC editor's queue, so folks would have one stable background 
document. I guess it could also be argued the other way around, 
though..

First off, the people should understand which high-level problems 
we're talking about (draft-ietf-mboned-addrdisc-problems-00.txt added 
three different ones, addressing just one of them).

It would probably be a good idea to also include some 
concrete statistics / interpretation of the IANA requests.. because I 
think basically what we're doing depends on what people are actually 
requesting and what they would be satisfied with.


-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Wed Apr 13 08:01:01 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA18660
	for <mboned-archive@lists.ietf.org>; Wed, 13 Apr 2005 08:01:01 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3DBi0XW011676;
	Wed, 13 Apr 2005 04:44:00 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j3DBi0kS011673;
	Wed, 13 Apr 2005 04:44:00 -0700 (PDT)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3DBhvqo011643
	for <mboned@lists.uoregon.edu>; Wed, 13 Apr 2005 04:43:58 -0700 (PDT)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id j3DBhmj04754;
	Wed, 13 Apr 2005 14:43:48 +0300
Date: Wed, 13 Apr 2005 14:43:48 +0300 (EEST)
From: Pekka Savola <pekkas@netcore.fi>
To: David Meyer <dmm@1-4-5.net>
cc: mboned@lists.uoregon.edu, dthaler@microsoft.com, mohitt@microsoft.com,
        amitag@microsoft.com, lorenzo@cisco.com, pusateri@juniper.net
Subject: Re: mboned: WGLC for draft-ietf-mboned-auto-multicast-04.txt
In-Reply-To: <20050408160504.GA17082@1-4-5.net>
Message-ID: <Pine.LNX.4.61.0504131348460.3592@netcore.fi>
References: <20050408160504.GA17082@1-4-5.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk

On Fri, 8 Apr 2005, David Meyer wrote:
> 	 ftp://ftp.ietf.org/internet-drafts/draft-ietf-mboned-auto-multicast-04.txt
>
> 	Please review the document carefully, and send your
> 	feedback to the list.  Please also indicate whether or
> 	not you believe that this document is ready to go to the
> 	IESG.
>
> 	This Last Call will end on 22 Apr 2005 at 1400 PDT (UTC/GMT-7).

I believe this is very important work (which also means it the 
time-to-market is important, so we should be finished with this soon) 
in making multicast available to everyone, but I think this still 
needs about 1-2 revisions (below).

Dave M., what is the intended category?  Experimental or Proposed 
Standard?  This requires extensive work for PS, but I could be 
convinced that if we go for Experimental first the document might not 
need to be as good at this point.  The most important thing, in the 
end, is creating a sufficiently stable spec and making the IANA prefix 
assignments so this can be deployed.

My comments to -03 (sent on the list on November 2004) have not been 
responded or integrated; I guess they were lost.  They are below.

Also, many comments to -02 weren't discussed or adopted either; I've 
included these at the end as well (removing the obsolete ones).

new issues
----------

1) MAC calculation also includes the source port in the calculation. 
However, note that the relays must not do so for the packets received 
from gateways because if there is NAT in the path, the source address 
and port will be changed and MAC does not verify.  Maybe the spec 
should be clearer which how to create and verify the MAC (from 
different nodes' perspectives)?

2) the document should be clearer about how this interoperates with 
NATs, i.e., in which cases the gateway can be behind a NAT.  If the 
traffic between gateways is sent directly (this is a MAY), the sender 
cannot be behind a NAT, at least, because otherwise NAT translation 
state does not exist and the IGMP join is discarded by the NAT.  So, 
as NAT traversal requires much more than just using UDP and tuning 
IGMP/MLD intervals, for example the following requires clarification:

    All multicast packets (data and control) are encapsulated in unicast
    packets.  To work across NAT's, the encapsulation is done over UDP
    using a well-known port number [TBD IANA].

substantial
-----------

==> the document does not sufficiently early spell out how it works (or
rather does not work under certain modes) with NATs.  Easy fix would be
s/isolated/isolated non-NAT/ below:

    2. Allowing isolated sites/hosts to transmit the SSM flavor of
       multicast.

...

    Each gateway has an AMT pseudo-interface that serves as a default
    multicast route.  Requests to join a multicast session are sent to
    this interface and encapsulated to a particular relay reachable
    across the unicast-only infrastructure.

==> it's required to have some discussion here about locally-scoped groups,
or global groups which should not be propagated inter-domain (like service
discovery etc.) because the relay will be in a different administrative
domain.  Should the gateway recognize those and not unnecessarily
send those upstream?  Should the relay be configured to ignore them?

[...]
    Finally, to support sourcing traffic to SSM groups by a gateway with
    a global unicast address, the AMT Subnet Prefix is treated as the
    subnet prefix of the AMT pseudo-interface, and an anycast address is
    added on the interface.  This anycast address is formed by
    concatenating the AMT Subnet Prefix followed by the high bits of the
    gateway's global unicast address.  For example, if IANA assigns the
    IPv4 prefix x.y/16 as the AMT Subnet Prefix, and the gateway has
    global unicast address a.b.c.d, then the AMT Gateway's Anycast
    Address will be x.y.a.b.  Note that multiple gateways might end up
    with the same anycast address assigned to their pseudo-interfaces.

==> here and in many places as well you discuss how v4 prefix tricks
are to be used, but you should also be discussing how the
manipulations/assignments would be done for v6.

6.3.  Responding to Relay Changes

    When a gateway determines that its current relay is unreachable
    (e.g., upon receipt of a ICMP Unreachable message [ICMP] for the
    relay's unicast address), it may need to repeat relay address
    discovery.  However, care should be taken not to abandon the current
    relay too quickly due to transient network conditions. Some
    implementations may find it difficult to send a discovery packet to
    the anycast relay address while the gateway has an address configured
    on the AMT pseudo-interface on the same anycast prefix. Therefore, it
    may be necessary to tear down the AMT pseudo-interface to rediscover
    a new relay.

==> why would it be difficult to do that on some implementations?  because
the anycast address is onlink with the interface?  But remember that the
prefix should not be configured on-link, as it's used only for the purposes
of discovering the local unicast address, thus should never be configured
on the interface, right ?

    The IANA should also allocate an Autonomous System ID which can be
    used as a pseudo-AS when advertising routes to the above prefix.
...
    An autonomous system number dedicated to a pseudo-AS for multicast is
    already in use today (AS 10888), and so it is expected that no
    additional AS number is required for this prefix.

==> aren't these in conflict ?

==> there is an important issue which needs to be discussed.  Who will
maintain the prefix/AS records in RADB/RIPE/.. etc. whois databases?

    Within the native MBGP infrastructure, there is a risk that a rogue
    router or a rogue AS could introduce a bogus route to the AMT Subnet
    Prefix, and thus divert joins and cause RPF failures of multicast
    traffic.  Again, network managers have to guarantee the integrity of
    the MBGP routing to the AMT subnet prefix in much the same way that
    they guarantee the integrity of all other routes in the M-RIB.

==> this seems insufficient, because it's saying "just do it like you do for
unicast" -- but these prefixes are ANYCAST, also with their own AS number
which anyone can add to their own AS-macro, and even if you did everything
you can using unicast, you could not (completely) protect against
illegitimate hijacking.  This anycast speciality should be discussed in a
paragraph of its own.

    Gateways and relays will accept and decapsulate multicast traffic
    from any source from which regular unicast traffic is accepted. If
    this is for any reason felt to be a security risk, then additional
    source address based packet filtering MUST be applied:

==> this is totally bogus way to specify this, it's saying like "if the
implementor thinks it makes sense, do X", effectively creating a loophole
where you can be compliant with the spec but not do the checks.  Just say
that implementations MUST do these.










semi-editorial
--------------

    day) to re-resolve the unicast address of a close relay.  The gateway
    also initializes a timer used to send periodic membership reports to

==> s/initalizes/SHOULD initialize/ ?



editorial
---------

==> ToC should be just before the Introduction.  I suggest using tools where
this is trivial to achieve ;-)

The protocol introduced in this
    specification is deployed in a few strategically-placed network nodes
    and in user-installable software modules [...]

==> s/is deployed/can be deployed/ ? :-)

    3. Allowing isolated sites/hosts to receive general multicast (ISM
       [RFC1112]).

==> why not use "ASM" instead of "ISM" ?

    a AMT Relay Discovery message to the AMT Relay Anycast Address, and

==> s/a AMT/an AMT/


=======================================================
comments to -02 which apparently haven't been addressed
=======================================================

substantial
-----------

    2. The AMT Subnet Prefix is advertised for RPF reachability in the
    M-RIB by relays and gateways.

    3. Relays or gateways that receive a join for a source/group pair
    use information encoded in the address pair to rebuild the address

==> only relays need to do these, AFAICS. (should be clearer)

    The solution above will scale to an arbitrary number of relays, as
    long at the number of relays requiring multicast traffic from a
    given AMT site remains reasonable enough to not overly burden the
    site's gateway.

==> umm, e.g. with 10 or 15 recipients, with reasonably large packets,
slower links (such as those used by typical AMT deployers) would already be
highly congested -- and what's worse, all the transmissions would always be
simultaneous.

Luckily enough this is not a *huge* problem, as the sites are only likely
recipients, not senders.  But the problem is just changed the other way
around, to the relays, which might need to encapsulate, simultaneously, to
even 100 or 1000 sites.  This could be pretty high as well, especially if
the encapsulation is done in the slow router CPU..

    Finally, to support sourcing traffic to SSM groups by a gateway with
    a global unicast address, the AMT Subnet Prefix is treated as the

==> this wording, the same later in section 6.3, seems to be ignoring the
possibility that gateway is not necessarily the host sending/receiving.

==> further, it might be important to say a few words on the multicast
address assignment procedures which you're relying on here.  As it is, the
apps and hosts must know somehow that they use the correct group addresses...

6.3. Responding to Relay Changes

    When a gateway determines that its current relay is unreachable
    (e.g., upon receipt of a ICMP Unreachable message for the relay's
    unicast address), it immediately repeats the unicast address
    resolution step by sending a UDP encapsulated ICMP Echo Request to
    the AMT Relay Anycast Address, and storing the source address of the
    UDP encapsulated ICMP Echo Response as the new unicast address to
    use as a "link-layer" destination.

==> s/a ICMP/an ICMP/
==> there seem to be some remnants of ICMP here, which should be AMT Relay
Discovery process, I guess.
==> If there are existing joins (or transmissions) when the relay is
changed, should there be some text to suggest that the join should be
resent, or the transmission stopped until a new join is received (or the
like)

6.7. Sending data to SSM groups

[...]

    If the group address is an SSM group, then the packet is unicast
    encapsulated to each remote node from which the gateway has received
    an IGMPv3 report for the packet's (source, group) pair.

==> this seems to create an implicit requirement for "MSNIP--"
functionality: the AMT code must know whether there have been any joins to
the group, before sending out anything.  Regular multicast implementation would
just toss all the packets to the relay...

9. Security Considerations

==> there are at least [one] issue which need description:
  1) SSM (S,G) joins from native multicast to AMT gateway -- describe how
these are secured.





editorial
---------

    to exchange multicast traffic with the native multicast
    infrastructure (MBone) and does not require any manual

==> please eradicate any references to "MBone".  Such thing doesn't really
exist anymore :-)

This discovery should be done periodically (e.g., once a
    day) to re-resolve the unicast address of a close relay.

==> should this last "should" be upper-cased?

  The gateway MUST also advertise itself as the default
    route for multicast in the M-RIB (or it must be the default unicast
    router if unicast and multicast topologies are congruent).

==> would it make sense to clarify that this advertisement is only needed in
the site's internal topology?

   If S is not the AMT Gateway Anycast Address of the gateway, the
    packet is dropped.

==> s/dropped/silently discarded/ (everywhere) ?

[References]
==> as MBGP is mentioned so many times, a suitable reference should be added
here, I think.


-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Wed Apr 13 10:28:22 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA00986
	for <mboned-archive@lists.ietf.org>; Wed, 13 Apr 2005 10:28:22 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3DE3ZbN025285;
	Wed, 13 Apr 2005 07:03:35 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j3DE3Zk7025284;
	Wed, 13 Apr 2005 07:03:35 -0700 (PDT)
Received: from m106.maoz.com (m106.maoz.com [205.167.76.9])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3DE3YUJ025273
	for <mboned@lists.uoregon.edu>; Wed, 13 Apr 2005 07:03:34 -0700 (PDT)
Received: from m106.maoz.com (localhost.localdomain [127.0.0.1])
	by m106.maoz.com (8.13.2/8.13.2) with ESMTP id j3DE3WQm001371;
	Wed, 13 Apr 2005 07:03:32 -0700
Received: (from dmm@localhost)
	by m106.maoz.com (8.13.2/8.12.11/Submit) id j3DE3U99001370;
	Wed, 13 Apr 2005 07:03:30 -0700
X-Authentication-Warning: m106.maoz.com: dmm set sender to dmm@1-4-5.net using -f
Date: Wed, 13 Apr 2005 07:03:30 -0700
From: David Meyer <dmm@1-4-5.net>
To: Pekka Savola <pekkas@netcore.fi>
Cc: mboned@lists.uoregon.edu, iana@iana.org
Subject: Re: mboned: Re: I-D ACTION:draft-ietf-mboned-addrdisc-problems-00.txt
Message-ID: <20050413140330.GA1345@1-4-5.net>
References: <20050408161751.GA17531@1-4-5.net> <Pine.LNX.4.61.0504131337270.3445@netcore.fi>
Mime-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="MGYHOYXEY6WxJCY8"
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.61.0504131337270.3445@netcore.fi>
User-Agent: Mutt/1.4.1i
X-public-key: http://www.1-4-5.net/~dmm/public-key.asc
X-gpg-fingerprint: 2409 8B50 B389 A307 BA5C 2A16 3918 03D6 A099 D8A7
X-philosophy: "I find your lack of faith disturbing." -- Darth Vader, Star Wars Episode IV.
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk


--MGYHOYXEY6WxJCY8
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Apr 13, 2005 at 01:42:34PM +0300, Pekka Savola wrote:
> On Fri, 8 Apr 2005, David Meyer wrote:
> >	What I would like to do is:
> >
> >	(i).	Get the WG's feeling about the idea of having a
> >		BOF on this topic, and
>=20
> I think this would make sense.
>=20
> >	(ii).	Solicit topics from those that think a BOF is a
> >		good idea. Topics would include (but are not
> >		limited to) formulations of the problem statement,
> >		and innovative solutions to the addressing problem.
>=20
> It would probably be good if draft-ietf-mboned-addrarch-01.txt was in=20
> the RFC editor's queue, so folks would have one stable background=20
> document. I guess it could also be argued the other way around,=20
> though..
>=20
> First off, the people should understand which high-level problems=20
> we're talking about (draft-ietf-mboned-addrdisc-problems-00.txt added=20
> three different ones, addressing just one of them).
>=20
> It would probably be a good idea to also include some=20
> concrete statistics / interpretation of the IANA requests.. because I=20
> think basically what we're doing depends on what people are actually=20
> requesting and what they would be satisfied with.

	Not sure what you're asking for in this last
	question. http://www.iana.org/assignments/multicast-addresses
	tells (most of) the story.

	Dave

--MGYHOYXEY6WxJCY8
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFCXSayORgD1qCZ2KcRAjs3AJ4zXl2rigArDQSlFT0dzppQ2QuSOgCfT1XL
w8Ae8dXSKIQ0+ImSfAKmYB0=
=yHBD
-----END PGP SIGNATURE-----

--MGYHOYXEY6WxJCY8--
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Wed Apr 13 10:38:56 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA01888
	for <mboned-archive@lists.ietf.org>; Wed, 13 Apr 2005 10:38:51 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3DEJeET007433;
	Wed, 13 Apr 2005 07:19:40 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j3DEJePN007432;
	Wed, 13 Apr 2005 07:19:40 -0700 (PDT)
Received: from m106.maoz.com (m106.maoz.com [205.167.76.9])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3DEJdap007423
	for <mboned@lists.uoregon.edu>; Wed, 13 Apr 2005 07:19:39 -0700 (PDT)
Received: from m106.maoz.com (localhost.localdomain [127.0.0.1])
	by m106.maoz.com (8.13.2/8.13.2) with ESMTP id j3DEIPF1002024;
	Wed, 13 Apr 2005 07:18:25 -0700
Received: (from dmm@localhost)
	by m106.maoz.com (8.13.2/8.12.11/Submit) id j3DEIPFL002023;
	Wed, 13 Apr 2005 07:18:25 -0700
X-Authentication-Warning: m106.maoz.com: dmm set sender to dmm@1-4-5.net using -f
Date: Wed, 13 Apr 2005 07:18:25 -0700
From: David Meyer <dmm@1-4-5.net>
To: Pekka Savola <pekkas@netcore.fi>
Cc: mboned@lists.uoregon.edu, dthaler@microsoft.com, mohitt@microsoft.com,
        amitag@microsoft.com, lorenzo@cisco.com, pusateri@juniper.net
Subject: Re: mboned: WGLC for draft-ietf-mboned-auto-multicast-04.txt
Message-ID: <20050413141825.GA2013@1-4-5.net>
References: <20050408160504.GA17082@1-4-5.net> <Pine.LNX.4.61.0504131348460.3592@netcore.fi>
Mime-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="dDRMvlgZJXvWKvBx"
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.61.0504131348460.3592@netcore.fi>
User-Agent: Mutt/1.4.1i
X-public-key: http://www.1-4-5.net/~dmm/public-key.asc
X-gpg-fingerprint: 2409 8B50 B389 A307 BA5C 2A16 3918 03D6 A099 D8A7
X-philosophy: "I find your lack of faith disturbing." -- Darth Vader, Star Wars Episode IV.
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk


--dDRMvlgZJXvWKvBx
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Apr 13, 2005 at 02:43:48PM +0300, Pekka Savola wrote:
> On Fri, 8 Apr 2005, David Meyer wrote:
> >	 ftp://ftp.ietf.org/internet-drafts/draft-ietf-mboned-auto-multicast-04=
.txt
> >
> >	Please review the document carefully, and send your
> >	feedback to the list.  Please also indicate whether or
> >	not you believe that this document is ready to go to the
> >	IESG.
> >
> >	This Last Call will end on 22 Apr 2005 at 1400 PDT (UTC/GMT-7).
>=20
> I believe this is very important work (which also means it the=20
> time-to-market is important, so we should be finished with this soon)=20
> in making multicast available to everyone, but I think this still=20
> needs about 1-2 revisions (below).
>=20
> Dave M., what is the intended category?  Experimental or Proposed=20
> Standard?  This requires extensive work for PS, but I could be=20

	Experimental.


	Dave

> convinced that if we go for Experimental first the document might not=20
> need to be as good at this point.  The most important thing, in the=20
> end, is creating a sufficiently stable spec and making the IANA prefix=20
> assignments so this can be deployed.
>=20
> My comments to -03 (sent on the list on November 2004) have not been=20
> responded or integrated; I guess they were lost.  They are below.
>=20
> Also, many comments to -02 weren't discussed or adopted either; I've=20
> included these at the end as well (removing the obsolete ones).
>=20
> new issues
> ----------
>=20
> 1) MAC calculation also includes the source port in the calculation.=20
> However, note that the relays must not do so for the packets received=20
> from gateways because if there is NAT in the path, the source address=20
> and port will be changed and MAC does not verify.  Maybe the spec=20
> should be clearer which how to create and verify the MAC (from=20
> different nodes' perspectives)?
>=20
> 2) the document should be clearer about how this interoperates with=20
> NATs, i.e., in which cases the gateway can be behind a NAT.  If the=20
> traffic between gateways is sent directly (this is a MAY), the sender=20
> cannot be behind a NAT, at least, because otherwise NAT translation=20
> state does not exist and the IGMP join is discarded by the NAT.  So,=20
> as NAT traversal requires much more than just using UDP and tuning=20
> IGMP/MLD intervals, for example the following requires clarification:
>=20
>    All multicast packets (data and control) are encapsulated in unicast
>    packets.  To work across NAT's, the encapsulation is done over UDP
>    using a well-known port number [TBD IANA].
>=20
> substantial
> -----------
>=20
> =3D=3D> the document does not sufficiently early spell out how it works (=
or
> rather does not work under certain modes) with NATs.  Easy fix would be
> s/isolated/isolated non-NAT/ below:
>=20
>    2. Allowing isolated sites/hosts to transmit the SSM flavor of
>       multicast.
>=20
> ...
>=20
>    Each gateway has an AMT pseudo-interface that serves as a default
>    multicast route.  Requests to join a multicast session are sent to
>    this interface and encapsulated to a particular relay reachable
>    across the unicast-only infrastructure.
>=20
> =3D=3D> it's required to have some discussion here about locally-scoped g=
roups,
> or global groups which should not be propagated inter-domain (like service
> discovery etc.) because the relay will be in a different administrative
> domain.  Should the gateway recognize those and not unnecessarily
> send those upstream?  Should the relay be configured to ignore them?
>=20
> [...]
>    Finally, to support sourcing traffic to SSM groups by a gateway with
>    a global unicast address, the AMT Subnet Prefix is treated as the
>    subnet prefix of the AMT pseudo-interface, and an anycast address is
>    added on the interface.  This anycast address is formed by
>    concatenating the AMT Subnet Prefix followed by the high bits of the
>    gateway's global unicast address.  For example, if IANA assigns the
>    IPv4 prefix x.y/16 as the AMT Subnet Prefix, and the gateway has
>    global unicast address a.b.c.d, then the AMT Gateway's Anycast
>    Address will be x.y.a.b.  Note that multiple gateways might end up
>    with the same anycast address assigned to their pseudo-interfaces.
>=20
> =3D=3D> here and in many places as well you discuss how v4 prefix tricks
> are to be used, but you should also be discussing how the
> manipulations/assignments would be done for v6.
>=20
> 6.3.  Responding to Relay Changes
>=20
>    When a gateway determines that its current relay is unreachable
>    (e.g., upon receipt of a ICMP Unreachable message [ICMP] for the
>    relay's unicast address), it may need to repeat relay address
>    discovery.  However, care should be taken not to abandon the current
>    relay too quickly due to transient network conditions. Some
>    implementations may find it difficult to send a discovery packet to
>    the anycast relay address while the gateway has an address configured
>    on the AMT pseudo-interface on the same anycast prefix. Therefore, it
>    may be necessary to tear down the AMT pseudo-interface to rediscover
>    a new relay.
>=20
> =3D=3D> why would it be difficult to do that on some implementations?  be=
cause
> the anycast address is onlink with the interface?  But remember that the
> prefix should not be configured on-link, as it's used only for the purpos=
es
> of discovering the local unicast address, thus should never be configured
> on the interface, right ?
>=20
>    The IANA should also allocate an Autonomous System ID which can be
>    used as a pseudo-AS when advertising routes to the above prefix.
> ...
>    An autonomous system number dedicated to a pseudo-AS for multicast is
>    already in use today (AS 10888), and so it is expected that no
>    additional AS number is required for this prefix.
>=20
> =3D=3D> aren't these in conflict ?
>=20
> =3D=3D> there is an important issue which needs to be discussed.  Who will
> maintain the prefix/AS records in RADB/RIPE/.. etc. whois databases?
>=20
>    Within the native MBGP infrastructure, there is a risk that a rogue
>    router or a rogue AS could introduce a bogus route to the AMT Subnet
>    Prefix, and thus divert joins and cause RPF failures of multicast
>    traffic.  Again, network managers have to guarantee the integrity of
>    the MBGP routing to the AMT subnet prefix in much the same way that
>    they guarantee the integrity of all other routes in the M-RIB.
>=20
> =3D=3D> this seems insufficient, because it's saying "just do it like you=
 do for
> unicast" -- but these prefixes are ANYCAST, also with their own AS number
> which anyone can add to their own AS-macro, and even if you did everything
> you can using unicast, you could not (completely) protect against
> illegitimate hijacking.  This anycast speciality should be discussed in a
> paragraph of its own.
>=20
>    Gateways and relays will accept and decapsulate multicast traffic
>    from any source from which regular unicast traffic is accepted. If
>    this is for any reason felt to be a security risk, then additional
>    source address based packet filtering MUST be applied:
>=20
> =3D=3D> this is totally bogus way to specify this, it's saying like "if t=
he
> implementor thinks it makes sense, do X", effectively creating a loophole
> where you can be compliant with the spec but not do the checks.  Just say
> that implementations MUST do these.
>=20
>=20
>=20
>=20
>=20
>=20
>=20
>=20
>=20
>=20
> semi-editorial
> --------------
>=20
>    day) to re-resolve the unicast address of a close relay.  The gateway
>    also initializes a timer used to send periodic membership reports to
>=20
> =3D=3D> s/initalizes/SHOULD initialize/ ?
>=20
>=20
>=20
> editorial
> ---------
>=20
> =3D=3D> ToC should be just before the Introduction.  I suggest using tool=
s where
> this is trivial to achieve ;-)
>=20
> The protocol introduced in this
>    specification is deployed in a few strategically-placed network nodes
>    and in user-installable software modules [...]
>=20
> =3D=3D> s/is deployed/can be deployed/ ? :-)
>=20
>    3. Allowing isolated sites/hosts to receive general multicast (ISM
>       [RFC1112]).
>=20
> =3D=3D> why not use "ASM" instead of "ISM" ?
>=20
>    a AMT Relay Discovery message to the AMT Relay Anycast Address, and
>=20
> =3D=3D> s/a AMT/an AMT/
>=20
>=20
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D
> comments to -02 which apparently haven't been addressed
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D
>=20
> substantial
> -----------
>=20
>    2. The AMT Subnet Prefix is advertised for RPF reachability in the
>    M-RIB by relays and gateways.
>=20
>    3. Relays or gateways that receive a join for a source/group pair
>    use information encoded in the address pair to rebuild the address
>=20
> =3D=3D> only relays need to do these, AFAICS. (should be clearer)
>=20
>    The solution above will scale to an arbitrary number of relays, as
>    long at the number of relays requiring multicast traffic from a
>    given AMT site remains reasonable enough to not overly burden the
>    site's gateway.
>=20
> =3D=3D> umm, e.g. with 10 or 15 recipients, with reasonably large packets,
> slower links (such as those used by typical AMT deployers) would already =
be
> highly congested -- and what's worse, all the transmissions would always =
be
> simultaneous.
>=20
> Luckily enough this is not a *huge* problem, as the sites are only likely
> recipients, not senders.  But the problem is just changed the other way
> around, to the relays, which might need to encapsulate, simultaneously, to
> even 100 or 1000 sites.  This could be pretty high as well, especially if
> the encapsulation is done in the slow router CPU..
>=20
>    Finally, to support sourcing traffic to SSM groups by a gateway with
>    a global unicast address, the AMT Subnet Prefix is treated as the
>=20
> =3D=3D> this wording, the same later in section 6.3, seems to be ignoring=
 the
> possibility that gateway is not necessarily the host sending/receiving.
>=20
> =3D=3D> further, it might be important to say a few words on the multicast
> address assignment procedures which you're relying on here.  As it is, the
> apps and hosts must know somehow that they use the correct group=20
> addresses...
>=20
> 6.3. Responding to Relay Changes
>=20
>    When a gateway determines that its current relay is unreachable
>    (e.g., upon receipt of a ICMP Unreachable message for the relay's
>    unicast address), it immediately repeats the unicast address
>    resolution step by sending a UDP encapsulated ICMP Echo Request to
>    the AMT Relay Anycast Address, and storing the source address of the
>    UDP encapsulated ICMP Echo Response as the new unicast address to
>    use as a "link-layer" destination.
>=20
> =3D=3D> s/a ICMP/an ICMP/
> =3D=3D> there seem to be some remnants of ICMP here, which should be AMT =
Relay
> Discovery process, I guess.
> =3D=3D> If there are existing joins (or transmissions) when the relay is
> changed, should there be some text to suggest that the join should be
> resent, or the transmission stopped until a new join is received (or the
> like)
>=20
> 6.7. Sending data to SSM groups
>=20
> [...]
>=20
>    If the group address is an SSM group, then the packet is unicast
>    encapsulated to each remote node from which the gateway has received
>    an IGMPv3 report for the packet's (source, group) pair.
>=20
> =3D=3D> this seems to create an implicit requirement for "MSNIP--"
> functionality: the AMT code must know whether there have been any joins to
> the group, before sending out anything.  Regular multicast implementation=
=20
> would
> just toss all the packets to the relay...
>=20
> 9. Security Considerations
>=20
> =3D=3D> there are at least [one] issue which need description:
>  1) SSM (S,G) joins from native multicast to AMT gateway -- describe how
> these are secured.
>=20
>=20
>=20
>=20
>=20
> editorial
> ---------
>=20
>    to exchange multicast traffic with the native multicast
>    infrastructure (MBone) and does not require any manual
>=20
> =3D=3D> please eradicate any references to "MBone".  Such thing doesn't r=
eally
> exist anymore :-)
>=20
> This discovery should be done periodically (e.g., once a
>    day) to re-resolve the unicast address of a close relay.
>=20
> =3D=3D> should this last "should" be upper-cased?
>=20
>  The gateway MUST also advertise itself as the default
>    route for multicast in the M-RIB (or it must be the default unicast
>    router if unicast and multicast topologies are congruent).
>=20
> =3D=3D> would it make sense to clarify that this advertisement is only ne=
eded in
> the site's internal topology?
>=20
>   If S is not the AMT Gateway Anycast Address of the gateway, the
>    packet is dropped.
>=20
> =3D=3D> s/dropped/silently discarded/ (everywhere) ?
>=20
> [References]
> =3D=3D> as MBGP is mentioned so many times, a suitable reference should b=
e added
> here, I think.
>=20
>=20
> --=20
> Pekka Savola                 "You each name yourselves king, yet the
> Netcore Oy                    kingdom bleeds."
> Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings

--dDRMvlgZJXvWKvBx
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFCXSoxORgD1qCZ2KcRAlzjAJ9kDgJTeEXmLdOuP6vkysn0HfcqFACeLqij
BE9cJQHfXLAUKAKzrQxU9FE=
=huSV
-----END PGP SIGNATURE-----

--dDRMvlgZJXvWKvBx--
_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


From owner-mboned@lists.uoregon.edu  Fri Apr 15 15:44:36 2005
Received: from darkwing.uoregon.edu (root@darkwing.uoregon.edu [128.223.142.13])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA10280
	for <mboned-archive@lists.ietf.org>; Fri, 15 Apr 2005 15:44:36 -0400 (EDT)
Received: from darkwing.uoregon.edu (majordom@localhost [127.0.0.1])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3FJZotl018137;
	Fri, 15 Apr 2005 12:35:50 -0700 (PDT)
Received: (from majordom@localhost)
	by darkwing.uoregon.edu (8.13.4/8.13.4/Submit) id j3FJZoIF018136;
	Fri, 15 Apr 2005 12:35:50 -0700 (PDT)
Received: from ietf.org (odin.ietf.org [132.151.1.176])
	by darkwing.uoregon.edu (8.13.4/8.13.4) with ESMTP id j3FJZmKT018113
	for <mboned@lists.uoregon.edu>; Fri, 15 Apr 2005 12:35:49 -0700 (PDT)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA08650;
	Fri, 15 Apr 2005 15:35:46 -0400 (EDT)
Message-Id: <200504151935.PAA08650@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: i-d-announce@ietf.org
Cc: mboned@lists.uoregon.edu
From: Internet-Drafts@ietf.org
Subject: mboned: I-D ACTION:draft-ietf-mboned-maccnt-req-00.txt
Date: Fri, 15 Apr 2005 15:35:46 -0400
Sender: owner-mboned@lists.uoregon.edu
Precedence: bulk

--NextPart

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

	Title		: Accounting, Authentication and Authorization 
			  Issues in Well Managed IP Multicasting Services
	Author(s)	: H. He, et al.
	Filename	: draft-ietf-mboned-maccnt-req-00.txt
	Pages		: 17
	Date		: 2005-4-15
	
This Internet Draft (I-D) describes problems in the area of 
     accounting and access control for multicasting.  General 
     requirements for accounting capabilities including quality-of-
     service (QoS) related issues are listed.  This I-D assumes that 
     these capabilities can be realized by functions implemented at 
     edges of a network based on IGMP or MLD.  By such functions, 
     information obtained from edge routers would be logged in a 
     dedicated database.  Finally, cases for Content Delivery Services 
     (CDS) are described as application examples which could benefit 
     from multicasting accounting and access control capabilities as 
     described in the I-D.  It is proposed that this I-D be used as a 
     starting point for further discussion on these issues.

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

To remove yourself from the I-D Announcement list, send a message to 
i-d-announce-request@ietf.org with the word unsubscribe in the body of the message.  
You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce 
to change your subscription settings.


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-mboned-maccnt-req-00.txt".

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


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

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

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

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

Content-Type: text/plain
Content-ID:	<2005-4-15152542.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-mboned-maccnt-req-00.txt

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

Content-Type: text/plain
Content-ID:	<2005-4-15152542.I-D@ietf.org>

--OtherAccess--

--NextPart--


_______________________________________________________________
user interface: http://darkwing.uoregon.edu/~llynch/mboned.html
web archive:  http://darkwing.uoregon.edu/~llynch/mboned/


