[Supa] Minutes of the 06/08/15 SUPA call

["Narasimha Rao Vadrevu" <[e-mail]>]

Hi,

Please find below the meeting notes from the SUPA call on 6/8. Please let me
know if anything important is missing, or I got it wrongly.

Thanks and Regards,

Narasimha

Agenda:

1.     Note Well, logistics, agenda bashing (chairs, 0 min)

2.     Updated charter (Juergen/John/Scott, 20 min)

 <https://github.com/IETF-SUPA/SUPA/wiki/Charter>
https://github.com/IETF-SUPA/SUPA/wiki/Charter

3.     Q&A Wiki and BoF wiki (all, 20 min)

https://github.com/IETF-SUPA/SUPA/wiki

http://trac.tools.ietf.org/bof/trac/#

4.     Document #1 (all, 20 min)

5.      one I-D explains things well, converged from

*
http://datatracker.ietf.org/doc/draft-karagiannis-supa-problem-statement/

*        http://datatracker.ietf.org/doc/draft-bi-supa-gap-analysis/

*        http://datatracker.ietf.org/doc/draft-zhou-supa-framework/

---------------------------------------------------------------

*        Tina note well

Charter discussion

Benoit will only be for the first 1/2 hour

Dan and Tina entered BOF request in the wiki and would like to know from
benoit in terms of if information provided is sufficient or anything missing

Benoit: meeting for BOF declaration is this Thursday. Need to update the
Wiki with a couple of items are yet to complete like add key words etc.
Benoit wants to forward to IESG/IAB members. Didn't read the new draft yet

Dan: Draft name is SUPA proposition. Need Benoit to provide his input by
today to finalize the draft to IAB

Benoit: Draft needs to be combined as a single draft to IESG. When (IAB)
people get the draft, they need to know the guidelines for SUPA. Want to
know if the current one on the WiKi is the combined

Dan: It is mostly but doesn't include all the problem statement and gap
analysis discussions. Union of drafts and include the important information

Benoit: Is the #1 under BOF the current? A couple of open questions in the
scope:

1.      How SUPA propose to work with other policy models doing in IETF

2.      Whether SUPA information model will be reused for an ACL

 

Juergen: ACL is optimized for speed which can be interpreted packet arrival
times, which need specific requirement different from SUPA space

Benoit: SUPA abstraction operation needs technology specific models for
usage, what does it mean?

John: ACL is a set of conditions and set of actions. An ACL can select
something and applies to a security operation and forward or open a port
etc. But SUPA can send something and the manager can transfer to a normal
ACL

Benoit: is SUPA focusing only on the northbound interface between manager
and controller and not reapplying the principles from the information model
to routing policies, firewall policies and qos policies?

John: why wouldn't we?

Benoit: Then why we are not going for a ?charter. Some policies are not in
the scope of supa, if SUPA is only on the north bound interface why SUPA is
used on an ACL

John: what I interpreted from Juergen's text is,  Packet filtering, routing
import/export are embedded within the configuration and take a very specific
form SUPA policy can result in a routing/firewall policy indirectly defines
the ACL  which are not embedded in the configuration file

Benoit: That I understand

John: Then, I or someone need to clarify the WiKi with this.

Benoit: what is missing from the charter is SUPA is only focusing on the
north bound controller

John: That text is in and out of charter, Juergen edited the last charter
and he needs to answer

Benoit:  in the end the charter and Q&A will be updated, you need to agree
as a group on what we want to do

Juergen: we do not need to mention where it is located, in a network element
or outside the controller. Need to specify the functionality.

Benoit: understand the point. Want to avoid the confusion on routing
policies. Need to avoid potential confusion on routing policies and network
elements etc. Need you to update correctly in the BOF

Juergen: Ok

Benoit:  One more feedback. Need use cases more as an applicable statement
and looking at one specific use case in detail on how it can be applied as
information and data model.

Juergen:  will change to use cases

Benoit: Use case is a debate, is it applicable and spend a lot of time apply
to one use case and not need to apply too much time on one use case use case
is a generic policy and deliver an ECA

Dan: Thanks for the feedback Benoit and complete the input after reading the
document

Benoit: Has one or two persons lined up reading the Q&A. They won't read the
docs but BOF wiki and Q&A wiki.  Please spend some time on those. These
matter if BOf is approved

Juergen: Is the Q&A more pertinent

Benoit:  No. Just giving guideline about what IAB/IESG will look into

John: Proposition draft explains a lot of everything that is on the wiki. Is
it possible to update that document this late?

Benoit: No problem. 

Tina: What key words are missing from the wiki

Benoit: Q&A wiki need to be updated with the above points 

John. I would update the Q&A wiki, but it is more import to move the wiki
content into proposition draft

Juergen: I just made a change based on Benoit's feedback and wiki needs to
be updated by whoever controls it.

Tina: wiki is open to everybody with a GitHub account

Juergen: Ok, I will make the changes then

BOF Wiki and SUPA GitHub wiki

Tina: One question, draft content has SUPA mapping, is it the draft
proposition was converged from this, do we use any content from this, do we
keep it or delete 

John: proposition draft will not incorporate content that uses mapping
because that is talking about specific mapping which we haven't agreed upon

Tina: Once you update delete it then

John:  Draft WANG is the YANG model but doesn't define how an information
model is mapped.

Dan: I think that is the answer, we are not interested in solving the
generic information model and that is not the purpose of this group

John: if the answer is draft WANG, why did we write the information model?
Draft Wang is part of the solution. Here is the information model and here
is the YANG draft to implement the structure? Information model has 3 sets,
generic, requisition to ECA. Draft yang has something that looks like ECA
but it needs a structure. If it is a yang, then the model needs objects. We
need YANG models and how we built the YANG models. I didn't phrase it
properly.

Tina: Please update the wiki as well

Dan: We have a proposal from Jun Bi off line, can we share the proposal

Juergen:  Charter text is important now and we need to come up with 000
draft and data can be looked into later

Max from the chat box: we need may be good use cases mentioned by Benoit
instead of solution implementation

Tina: Information model draft will be written by Ben Yi and John will help 

Juergen: Information model needs to be more stable than Data model

Dan: GPIM need to be more specific

Juergen: For the time being we need to concentrate on proposition draft

John: Agree with Juergen and I Don't think the proposition draft can cover
everything from the Q&A wiki

Dan: Section 6 is a conclusion

Juergen: section should use content from 1 through 5.

John is going to update the Q&A wiki since he wrote most of the stuff.
Juergen is going to update the BOF wiki with the latest charter

Dan: Need to finalize by tomorrow. Max needs to work on the prop draft. By
Wednesday end of day we have better Q&A wiki and proposition draft

John: easy to split the comments from Benoit. 

Tina: Juergen will send an email in his time zone 

Max: use cases should be much more planned than the one we have

Sue:  question to Max, did Juergen say that in the email or wiki where did u
get that input. Is someone working on rewriting the use cases?

Tina: Scott Bratner

Sue: Did Scott take up the pen? What's the timeline on the use case draft? I
can help with contingent on John. Is the work flow changed? Does it need BGP
model or not and based on that I can update

John: Benoit needs BGP

Sue: In that case I can work on that after the Wednesday.

Tina: Next couple of days is prop draft and do we need resources for use
cases?

Sue:  I can write the converged use cases and come back and write after she
is done with BGP end of the week.

Dan: Benoit says it is not the highest priority

Sue: It is higher priority than the BGP for the first couple of days.whats
Benoit's take? Will ask Benoit. Can Dan send the drafts? Tina sent them in
the chat. I will write the use cases into one

Max: Volunteered to help. Sue will take first pass and send it to Max

Tina:  It is getting close to Thursday meeting. Can Dan reemphasize what we
need to do before the IESG meeting?

Dan: Plan of record is-

*         Juergen is updating the charter wiki

*         John working on the Q&A wiki and wrap up

*         Everybody read the proposition draft and q&a wiki

We debate a lot and by tomorrow we close on the wiki and proposition and by
Wednesday we update and work with Benoit to work with folks from ISG.

Tina: We cross our fingers!

Dan: Continue the meetings even if the BOF is approved. We need to run a
successful second BOF and working group. Head the week in glory

Thanks everyone.