Re: [Sip] Toward the Evolution of SIP and Related Working Groups

["Mary Barnes" <[e-mail]>]

I have a few comments below [MB].  And one more mega comment/proposal
here.  Mobile IP is an area that has a similar history as SIP in terms
of the demand and explosion of work. They considered our SIP/SIPPING
model (at that time we had few other spinoffs, as I recall) when they
split their WGs. They have a group for IPv6 work(mext), one for
IPv4(mipv4) and one for performance, signaling and
optimizations(mipshop). So, from you list of numbers, I could also see a
3 WG split for SIP/SIPPING (discounting all the current spinoffs) using
your #s (and assuming SIP WG is done):
A) #2+#3: SIP core protocol and extensions to fold into next protocol
version
B) #4:  Policies and security related aspects of SIP Protocol and then
put the burden of screening new work such as SPIT, identity, etc. in
this group until it's decided that the workload demands a separate WG.  
C) #5: Basically, SIPPING slightly reduced from current charter  -
filtering  (including OPS related items), howtos, discussing p-headers,
etc. - the primary difference here is that this group would not screen
basic SIP extension proposals or any of the policy/security, etc.
related work. Right now we don't do the security, but some of the other
stuff has been debatable.  

In my mind, the primary thing this would accomplish would be separating
out the policy and security work to ensure there's the right focus and
attention. I also think this makes scheduling only slightly easier in
that it might be easier to ensure security folks could attend. 

However, I don't think this really helps that much with overall workload
- it's just moving some things around. For example,  if A above does all
the SIP extension screening, that workload may make A be just as crazy
as SIP is now OR if in the end the security stuff is causing the
craziness, then group B becomes the crazy one, until it sheds more work
to new WGs and of course C stays fairly sane, which is what I think it
is now ;)  

In general, I agree that SIP should follow the model that SIPPING did a
couple years ago and NOT accept new work items until currently chartered
items are done (or at least down to a manageable handful (5-6)) rather
than the current 17 (use the tools page rather than WG charter to see
the delta), so putting the WG on a path to closure is an excellent idea.


Mary.

-----Original Message-----
From: [e-mail] [mailto:[e-mail]] On Behalf Of
Dean Willis
Sent: Monday, June 16, 2008 6:23 PM
To: [e-mail]
Subject: [Sip] Toward the Evolution of SIP and Related Working Groups


The SIP working group spun off from the MMUSIC working group when we
realized that the SIP work was reasonably separable from the other work
going on in MMUSIC.

The SIP working group first met officially at IETF 46 in November 1999,
although we did have an earlier interim meeting.

This means that the fall 1999 meeting will be our 10th anniversary. In
addition to having a big party, I'd also like to plan for this being the
last meeting of the SIP working group as it is now defined. Ten years is
long enough. It's time we restructured.

There will probably still be people who want to change, revise, extend,
or otherwise manipulate the family of SIP specifications after 2009.

So what should we do? How should we organize to get those needs met?   
Rest assured, we'll be having further discussions on this topic in the
coming months. But I'd like to start the discussion now.

By November 2009, I fully expect SIP to have completed the majority of
currently chartered items. Between now and then, we'll probably also
take on and complete a bit of new work (although I plan to be very
resistant to new work). I also expect that there are a few items on our
charter now that we're not going to finish. It may well be that some of
those items need a different organizational structure to progress than
what we have. Or maybe they don't need doing.

Further, I've come to believe that most working groups should be
shorter-lived and tightly constrained by their charters to produce a
single concise set of deliverables. There's still a need for a longer-
lived thing that maintains architectural oversight. I've come to believe
that this is the "area", not the working group.

Let's start the recap by looking at the current charter of the SIP WG.  
I've eliminated the WGLC milestones to shorten the list, as each has a
following IESG milestone. Note that there's a Feb 2008 milestone for
"Revise Charter". You can consider this topic the first installment on
that milestone, even though it is a bit late.

1) Jul 2007 Location Conveyance with SIP to IESG (PS)

2) Aug 2007 Connection reuse mechanism to IESG (PS)

3) Sep 2007 Session Policies to IESG as PS

4) Sep 2007 Example security flows to IESG (Informational)

5) Oct 2007 New resource priority namespaces for DISA to IESG (PS)

6) Nov 2007 Requirements for media keying to IESG (Informational)

7 )Dec 2007 Extensions to SIP UA Profile Delivery Change Notification
Event Package for XCAP to IESG (PS)

8) Dec 2007 Roadmap for SIP to IESG (Informational)

9) Dec 2007 Using SAML for SIP to IESG (PS)

10) Dec 2007 Extension for use in etags in conditional notification to
IESG (PS)

11) Feb 2008 Identify requirements for test matrix to move SIP to Draft
Standard

12) Feb 2008 Delivering request-URI and parameters to UAS via proxy to
IESG (PS)

13) Feb 2008 Revise charter

14) Feb 2008 Establishment of secure media sessions using DTLS-SRTP to
IESG (PS)

15) Apr 2008 MIME body handling in SIP to IESG (PS)

16) Apr 2008 Essential corrections to RFC 3261 (1st batch) to IESG (PS)

17) Jul 2008 Mechanisms for UA initiated privacy to IESG (BCP)

18) Aug 2008 Guidelines for double route recording to IESG (BCP)

19) Sep 2008 X.509 Certificates for TLS use in SIP to IESG (PS)

20) Sep 2008 X.509 extended key usage for SIP to IESG (PS)

21) Nov 2008 Termination of early dialog prior to final response to IESG
(PS)


Looking at this, we have a couple of discrete tasks that I think might
make the basis for different narrowly-scoped working groups. We've also
got a lot of stuff that we're late on delivering!

1) SIP WG (to finish by end of 2009) gets the bulk of these items.

2) SIP Draft Standard WG: Takes the milestone of "Identify requirements
for test matrix to move SIP to Draft Standard" and all of the follow-on
work needed to move SIP from Proposed to Draft on the standards track.
There's certainly enough work here for a dedicated working group, with
the last milestone being something like "Deliver SIP as a Draft
Standards Document to IESG". We may wish to consider whether we really
want to go this route or whether we should instead be thinking of a "SIP
3.0" effort. My personal though is that there's just too much cruft in
the RFC 3261 family to make Draft, but I could be proven wrong. I'd
rather see us put SIP 2.0 into maintenance mode at the end of 2009 and
move the bulk of our resources into developing a SIP 3.0 family designed
from the ground up to be Draft-Standard achievable.
[MB] I like this suggestion. [/MB]

3) RAI Essential Corrections WG: Does this justify its own working
group? Remember, it would be not just RFC 3261, but the entire family of
related RFCs that would need lock-step maintenance and corrections.  
Another approach might be to roll the corrections into the Draft
Standard version of SIP 2.0.
[MB] I think this should be roleed into 2) above. [/MB]


4) RAI Policies WG: We have these ongoing, not-quite-committed work
items like session policies, use of SAML, etc.  that we've never been
able to get fully wrapped up. I suspect they need their own WG. Or we
need to decide that these things just aren't worth doing and stop
feeling guilty about not finishing. As it is, I don't have a good
feeling about SIP finishing them.
[MB] I like this idea and would it also not cover any outstanding
ongoing security items or are you anticipating that to addressed
entirely by the currently chartered work items?  [/MB]

5) SIP Operations: I'd like to see SIPPING replaced with a "how to"  
group possibly under the OPS banner. They could deal with all the "how
do we do X with SIP" from other SDOs. When confronted by a need to
extend SIP to meet a given need, they could then BOF up a new WG to deal
with the required extension.
[MB] Are you proposing this to replace SIPPING WG?  If so, I don't see
it falling under the banner of OPS. I do see that there may potentially
be the need to initially discuss OPS related items, but as in the past,
I see those being completed in the appropriate OPS WGs (e.g.,
performance).  I have mixed views on how effective it would be to have a
BoF and new WG for new extensions. I don't think the model we currently
have is that ineffective in terms of first discussing requirements
before agreeing a solution.  I see a SIP Maintenace group perhaps to
handle these items, although in many cases, I think it would be far more
effective to roll these things into your item 2 - i.e., don't you see
SIP 3.0 as pulling in a lot of the headers that have been defined since
RFC 3261 has been published. [/MB]

We already have working groups looking at conferencing, peering,
telephony feature interop, provisioning, peer-to-peer, NAT traversal,

Other things that might need WGs: SPIT? Identity and Reputation?

[MB] I would agree these two topics might warrant separate WGs, although
some of that work overlaps with your WG 4, I believe, the division of
work may not be so clear.  [/MB]

I also suspect we need a couple of directorates that are not WGs to act
in an advisory capacity:

[MB] I think we already have these and more covered by our current RAI
area review team. These are specialized reviews and assignment of
reviewers is based on area of expertise and we've got security and event
package gurus already on the team. [/MB]

1) RAI Security Directorate: Focuses on security review and
architectural issues in RAI documents.

2) RAI Event Package Directorate: Reviews and approves new SIP event
packages




_______________________________________________
Sip mailing list  https://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol Use
[e-mail] for questions on current sip Use
[e-mail] for new developments on the application of sip
_______________________________________________
Sip mailing list  https://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [e-mail] for questions on current sip
Use [e-mail] for new developments on the application of sip