Message undeliverable at this time Original text follows -----------------------------------------
From ITU-SG16@mailbag.jf.intel.com
X-Envelope-From: ITU-SG16@mailbag.jf.intel.com Received: (from mail@localhost) by woody.multilink.com (8.6.12/8.7.3) id QAA09586; Fri, 10 Oct 1997 16:11:29 -0400 Received: from mailbag.jf.intel.com(134.134.248.7) by woody via smap (V1.3) id sma009584; Fri Oct 10 16:11:11 1997 Received: from mailbag.jf.intel.com (mailbag.jf.intel.com [134.134.248.7]) by mailbag.jf.intel.com (8.8.6/8.8.5) with ESMTP id OAA13919; Fri, 10 Oct 1997 14:03:18 -0700 (PDT) Received: from MAILBAG.INTEL.COM by MAILBAG.INTEL.COM (LISTSERV-TCP/IP release 1.8c) with spool id 142404 for ITU-SG16@MAILBAG.INTEL.COM; Fri, 10 Oct 1997 14:03:15 -0700 Received: from axp.cmpu.net (axp.cmpu.net [204.96.11.17]) by mailbag.jf.intel.com (8.8.6/8.8.5) with ESMTP id OAA13897 for ITU-SG16@mailbag.jf.intel.com; Fri, 10 Oct 1997 14:03:13 -0700 (PDT) Received: from smithmicro.com (C1123-111.dialup.autonet.net [206.88.153.111]) by axp.cmpu.net (8.8.6/8.8.5) with ESMTP id PAA19922 for ITU-SG16@mailbag.jf.intel.com; Fri, 10 Oct 1997 15:59:52 -0500 X-Mailer: Mozilla 4.03 [en] (Win95; U) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Message-ID: 343E9787.2C343992@smithmicro.com Date: Fri, 10 Oct 1997 14:00:55 -0700 Reply-To: Mailing list for parties associated with ITU-T Study Group 16 ITU-SG16@mailbag.jf.intel.com Sender: Mailing list for parties associated with ITU-T Study Group 16 ITU-SG16@mailbag.jf.intel.com From: Paul Long plong@SMITHMICRO.COM Organization: Smith Micro Software, Inc. Subject: Clarification of H.245 MTSE To: ITU-SG16@mailbag.jf.intel.com Content-Transfer-Encoding: quoted-printable X-MIME-Autoconverted: from 8bit to quoted-printable by mailbag.jf.intel.com id OAA13919
(I posted this to the sg16.lbc reflector, but Tom Geary suggested that I = also post it here. Sorry for the cross-posting. I believe this only affects H.324, so = all you H.323, et al., people can continue with whatever you were doing. :-) )
Here is the text for a proposal I submitted back in Portland and that was= supposed to have been discussed in Sunriver. Since it wasn't, we need to discuss this on t= he reflector before the next meeting in Geneva. If you have any problem at all with th= is proposed "fix" to the MTSE, please speak up now.
This contribution proposes backward-compatible modifications to the H.245= Multiplex Table Signaling Entity (MTSE) in order to fix an apparent oversight. These modi= fications allow an implementation to use the MultiplexEntrySend message as it was intende= d=97-with multiple multiplex entries.
1. Introduction
There is a flaw in the H.245 MTSE SDL that precludes sending the Multiple= xEntrySend message with multiple entries if there are any entries in common between = it and any of the preceeding MultiplexEntrySend messages. When multiple entries are sent in= a MultiplexEntrySend message, they all share the same value in the sequence= Number field, yet each out-going MTSE maintains its own, per-entry sequence number in out_S= Q. Therefore, only those entries whose MTSE coincidentally have the same value for out_= SQ can be combined in a subsequent MultiplexEntrySend message. The trouble this cau= ses during implementation has resulted in some vendors deciding to not send multiple= entries in a MultiplexEntrySend message.
2. Solutions
There are two solutions to this problem.
2.1 Avoid multiple entries
Never send multiple entries in a MultiplexEntrySend message after the fir= st one. The first MultiplexEntrySend message may or may not contain multiple entries. The d= ownside is that one is not using the MultiplexEntrySend message as it was intended to be = used, thus wasting a small amount of bandwidth due to duplicated overhead for each M= ultiplexEntrySend message containing a single entry and the associated time loss, e.g., dur= ing session startup, as the sender waits for responses to each request.
2.2 Use a common sequence-number variable
Have all of the out-going MTSEs share a single sequence-number variable f= or determining what sequence number to use for out-going MultiplexEntrySend messages. Ea= ch MTSE continues to use its existing, MTSE-specific sequence-number variable for recording= the sequence number used in the MultiplexEntrySend message. This is so that responses = can be identified as belonging to the same dialogue as the MultiplexEntrySend request.
This is backwards compatible with the MTSE defined in previous versions o= f H.245 because there is no requirement in the in-coming MTSE that the values of MultiplexEntrySend.sequenceNumber be contiguous from one message to anoth= er for a given MTSE instance. In effect, sequenceNumber is used as a somewhat unique han= dle for a brief transaction, not as something that must increase by one each time. The re= sult is that this allows all out-going MultiplexEntrySend messages to contain multiple entr= ies.
2. Text Modifications
Here are the required text modifications for the second, common-sequence-= number-variable solution.
Replace the out_SQ definition in 8.7.3.2/H.245 with these two variables a= nd their descriptions.
out_SQ This state variable is used to indicate the most recently sent MultiplexE= ntrySend message for this MTSE instance. It is mapped to the MultiplexEntrySend message se= quenceNumber field before transmission of a MultiplexEntrySend message.
global_out_SQ This state variable is used to indicate the most recently sent MultiplexE= ntrySend message for any MTSE instance. It is incremented by one and assigned to out_SQ be= fore out_SQ is mapped to the MultiplexEntrySend message sequenceNumber field. Arithmetic= performed on global_out_SQ is module 256.
Replace the contents of the task symbol in Figure 28(i)/H.245 containing = "out_SQ :=3D out_SQ + 1" with this:
global_out_SQ :=3D global_out_SQ + 1 out_SQ :=3D global_out_SQ
3. Proposal
That the text modifications described above be applied to H.245 version 3= to allow all MultiplexEntrySend messages to contain multiple multiplex entries.
[end]
-- Paul Long___________________________http://www.cmpu.net/public/plong Smith Micro Software, Inc.__________http://www.smithmicro.com/