SCTP Discussions

Text archives Help


Re: [SCTP-Discussion] Re: SCTP implementation


Chronological Thread 
  • From: "La Monte Henry Piggy Yarroll" <piggy AT baqaqi.chi.il.us>
  • To: discussion AT sctp.de
  • Subject: Re: [SCTP-Discussion] Re: SCTP implementation
  • Date: Tue Mar 13 13:29:01 2001
  • List-id: SCTP Discussions <discussion.sctp.de>

Yiwen Jiang
<Yiwen.Jiang AT tic.siemens.ca>
writes:
> > "Immature" is more accurate than "not very bug free". :-). The
> > latest release (now on sctp.refcode.org) actually passes all the
> > functional and unit tests that we've written so far. If you are
>
> Cool... My next question would be, how complete are the functional and unit
> tests written so far, percentage wise?

I'd say 30%-50%, but progress should go faster on the remaining
components. We've beaten most of the tough kernel-specific problems.

> > interested in a kernel implementation, I invite you to help us steal
> > as much of Andreas' and Michael's fine code as possible. :-)
>
> Are you saying that their code implements more features than yours?

Precisely.

> I just started to work on this project, and I'm trying to determine which
> implementation we should lean towards, and help out... If we do use the
> kernel implementation, sure!

What we have now is just enough that an adventuresome developer could
write a working SCTP application with our code. Our direction is
toward implementing the UDP-style interface first. Naturally, we'd be
thrilled to have some developers start pounding on our
implementation.

If stability and OO methods are important to you, I strongly recommend
Andreas' and Michael's implementation. If you'd like to help shake
out the design of the UDP-style interface, try our implementation.

> > Sun is co-authoring the SCTP API draft with us and they have a good
> > start on a kernel implementation. They brought a very slick
> > single-homed version to the last bakeoff.
>
> How does one get involved in the bakeoff? and how much information
> are the vendors providing in the bakeoff? Was your kernel
> implementation part of the bakeoff?

Yes, our kernel implementation was at the second bakeoff (just
barely). We managed to exchange a little data with two other
implementations. The bakeoff was where we had our first working
kernel...

Bakeoffs are a way of validating the spec (RFC 2960). Vendors bring
their implementations and we all run a large number of
interoperability experiments, and then analyze the failures and
confusions.

The bakeoffs are generally of interest only to people with their own
implementations of SCTP. The next bakeoff is at the end of April in
Nice. Drop me a note at work
<piggy1 AT motorola.com>,
and I can dig up
registration details for you.

> > HP has a very nice implementation for HPUX, and Cisco's IOS
> > implementation performed well too.
> >
> > That's all the publicly announced kernel implementations. If you wish
> > to hear about more, watch the TSV list. Several vendors have
>
> What is the TSV list? How do I subscribe it?

Start here: http://www.ietf.org/html.charters/tsvwg-charter.html

> > contacted us about information for their kernel implementations,
> > and there are at least two other groups seriously considering open
> > source kernel implementations.
>
> Thank you very much for the info... I really appreciate it!




Archive powered by MHonArc 2.6.19+.

Top of page