- From: Tuexen Michael <Michael.Tuexen AT icn.siemens.de>
- To: "'discussion AT sctp.de'" <discussion AT sctp.de>
- Cc: Wells Kris-KWELLS1 <Kris.Wells AT motorola.com>, Christian Ursula-UCHRIST1 <UCHRIST1 AT motorola.com>, "'Phillip Conrad'" <conrad AT acm.org>, "'Guo Marc-YGUO1'" <YGUO1 AT motorola.com>
- Subject: RE: [SCTP-Discussion] SCTP RFC v15
- Date: Tue May 7 08:42:00 2002
- List-id: SCTP Discussions <discussion.sctp.de>
Dear all,
please see my comments in-line.
Best regards
Michael
Michael Tuexen Tel.: +49 89 722 47210
Siemens AG Fax: +49 89 722 48212
ICN WN CC SE 7 E-mail:
Michael.Tuexen AT icn.siemens.de
>
-----Original Message-----
>
From: Phillip Conrad
>
[SMTP:conrad AT acm.org]
>
Sent: Tuesday, May 07, 2002 5:38 AM
>
To:
>
discussion AT sctp.de
>
Cc: Wells Kris-KWELLS1; Christian Ursula-UCHRIST1
>
Subject: Re: [SCTP-Discussion] SCTP RFC v15
>
>
At 03:26 PM 5/6/2002 -0500, Guo Marc-YGUO1 wrote:
>
>Hi, all
>
>
>
>We heard that SCTP RFC is now in version 15 and is not compatible with RFC
>
>version 13. Can anybody tell me if it is true?
>
>
I think there may be some confusion in terminology here. Perhaps you are
>
referring to "versions" of the Siemens/University of Essen software, rather
>
than "versions" of an "RFC".
>
>
There is no such concept as a "version" of an RFC. SCTP is specified in
>
RFC2960. Once the IETF issues an RFC number, that text is fixed for all
>
time as unchangable and unmodifiable.
>
>
Now, the text can be "obsoleted" by a NEW RFC with a different
>
number. Eventually, that will happen with SCTP.
>
>
For the moment, the "SCTP implementors guide" contains various
>
updates/corrections/clarification/upgrades to the SCTP protocol, expressed
>
as "changes" in the text of RFC2960. That document (the SCTP implementors
>
guide) is currently NOT an RFC, but is actually an Internet
>
Draft. Internet Drafts, DO have version numbers, starting with version
>
00. The SCTP Implementors Guide is currently at version 04 (see
>
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-04.txt).
>
Eventually these changes are likely to be folded in with the current text
>
of RFC2960 into a new RFC for SCTP, but it may be some time before that
>
happens.
>
>
For more information on the process by which RFCs, Internet Drafts, and
>
Internet Standards develop, go to www.ietf.org Especially helpful may be:
>
http://www.ietf.org/overview.html and http://www.ietf.org/tao.html in
>
clearing up some confusion about terminology here.
>
[Tuexen Michael] I totally agree with Phil.
>
>Which version of SCTP RFC or which drafts of SCTP that Siemens is
>
>implementing?
>
>
I don't speak for that group, but I would imagine that each release tries
>
to comply with RFC2960, as modified by the latest version of the
>
implementor's guide, and the latest version of the checksum draft. I would
>
check the announcements that accompany each release for exact details...
>
you can find these announcements in the web archive of this mailing list,
>
at:
>
>
http://www.sctp.de/pipermail/discussion/
[Tuexen Michael] The announcements are send to
http://www.sctp.de/pipermail/announce
[Tuexen Michael] and cced to the discussion list. So if you are
only interested in the announcement, there is a list for you.
>
Here are the announcements for version pre16 and version pre17
>
>
version pre16
>
http://www.sctp.de/pipermail/discussion/2002-February/000637.html
>
version pre17
>
http://www.sctp.de/pipermail/discussion/2002-February/000688.html
>
>
>
>Are all released versions of Siemens SCTP compatible with each
>
>other?
>
>
I don't speak for Siemens or or the Univ. of Essen, but given that the
>
checksum definition has changed at least once, and the latest version
>
probably implements the new checksum, I think it is unlikely... and more to
>
the point, not even desirable or possible... that all released versions are
>
compatible.
[Tuexen Michael] Well the newer support both checksums, so you can
choose (at runtime). Of course two instances using different checksum
algorithms can not interwork, but if you choose the same they should.
On the network they should be able to interwork (we do NOT test this),
from the API some extensions are necessary to provide new
functionality.
>
>Is Seimens planning v18 and beyond? When v18? How different from v17?
>
>
That's really for the Siemens and the Univ. of Essen folks to say .. but in
>
response to your next question...
[Tuexen Michael] We do have a version pre18 in testing/debugging.
We will release it, when the bugs are found and corrected. This
should be soon.
>
>Are future releases backward compatible with previous releases?
>
>
Well, at least one version is not... v16 introduced the new checksum
>
calculation,
>
(see http://www.sctp.de/pipermail/discussion/2002-February/000637.html )
>
so I would imagine it is impossible for that version to be compatible with
>
the
>
previous version.
>
>
However, if you look at the announcement for version pre17:
>
http://www.sctp.de/pipermail/discussion/2002-February/000688.html
>
one could infer that the version is reasonably backwards compatible
>
with version pre16, except in the areas where bugs were fixed.
[Tuexen Michael] They are compatible. Also in future versions you
can always enable the old algorithm. When the checksum ID is an RFC we
will only change the default algorithm from Adler32 to CRC32. If you
want,
you can change it back.
But it is not a goal, that older version can interwork. They should,
but
if they don't then it is a bug in the old version. And these won't be
fixed.
It is important for us that developer can change the version without
to much
effort. This means that we try top keep changes to the API as small
as possible.
>
I hope this is helpful in addressing your question, but again, please
>
understand that I do not speak for the development team at Siemens or the
>
University of Essen in any way. I just hang out with them sometimes.
>
[Tuexen Michael] I think it was helpful!
>
>Thanks.
>
>Marc Guo
>
>_______________________________________________
>
>discussion mailing list
>
>discussion AT sctp.de
>
>http://www.sctp.de/mailman/listinfo/discussion
>
>
_______________________________________________
>
discussion mailing list
>
discussion AT sctp.de
>
http://www.sctp.de/mailman/listinfo/discussion
Archive powered by MHonArc 2.6.19+.