Skip to main content

Minutes IETF113: alto
minutes-113-alto-00

Meeting Minutes Application-Layer Traffic Optimization (alto) WG
Title Minutes IETF113: alto
State Active
Other versions plain text
Last updated 2022-03-24

minutes-113-alto-00
      Agenda for the IETF#113 ALTO WG Session
      -------------------------------------
      https://datatracker.ietf.org/meeting/113/materials/agenda-113-alto

      Session:

         Wednesday, March 23,2022
         Session I (09:00-11:00 UTC,04:00-06:00 EST, 07:00-09:00
         PDT,17:00-19:00 Beijing Time)

      WG Chairs:

         Jan Seedorf        (ietf@j-f-s.de)
     Mohamed Boucadair  (mohamed.boucadair@orange.com)
         Qin Wu             (bill.wu@huawei.com)

      Minutes Taker: Richard Yang
                     Daniel King

      Available During the Session:

         ICS:          
         https://datatracker.ietf.org/meeting/113/sessions/alto.ics MeetEcho:  
            https://meetings.conf.meetecho.com/ietf113/?group=alto Audio Only: 
           http://mp3.conf.meetecho.com/ietf113/alto/1.m3u Jabber:       
         xmpp:alto@jabber.ietf.org?join

      Available During and After the Session:

         CodiMD:        https://codimd.ietf.org/notes-ietf-113-alto
         Slides:        https://datatracker.ietf.org/meeting/113/session/alto
         Drafts (PDF): 
         https://datatracker.ietf.org/meeting/113/agenda/alto-drafts.pdf Drafts
         (TGZ):  https://datatracker.ietf.org/meeting/113/agenda/alto-drafts.tgz

      Available After the Session:

         Recording:     http://www.meetecho.com/ietf113/recordings#ALTO
         Jabber Logs:   https://www.ietf.org/jabber/logs/alto
Introduction (Chairs) (5 minutes)
         Session Intro & WG Status

         - No Agenda Changes
Socializing ALTO
Luis: We could socialize ALTO activity across other working groups.
Richard (via Chat): All are welcome to atend the weekly informal meeting.
Med (via Chat):To volunteer to help building a tutorial, just drop an email to
the chairs and we will see how to organize ourselves. Richard (via Chat): Some
of us (jordi, I) have ALTO tutorials that we presented recently, A more formal
tutorial is a great idea and we will be happy to work with chairs. Med (via
Chat): making them easily available is also something we are looking for
Richard(via Chat):Good idea, hosting at https://github.com/ietf-wg-alto/

Chartered Items
ALTO OAM Support (20 minutes)
         https://tools.ietf.org/html/draft-zhang-alto-oam-yang-02
         Discussion Leaders: Jensen Zhang/Dhruv Dhody
Discussion Points

Q1
Qin: The proposal works for Q1, but we should check with YANG-Doctors
Richard: There could be some complexity, so we need to understand the impact.
Med (via Chat): The main advantage of relying upon an IANA registry is that new
values are supported for free without having to define an augment /etc. Dhruv
(via Chat): agreed to that Med: Otherwise, each time a new metric is supported
you need to define a new identity for it. We don’t need that as the
IANA-maintained module will be generated each time a new value is assigned.

Q2
Qin: Do we have a standard algorithm? Also we need to agree a set of basic
parameters. Jensen: (missed comment about algorithm). RE: parameters, yes

Q3
Martin: You are asking the right questions. Strive to reduce scope where
possible. If no strong support for specific features, then they should be
omitted. Richard: Good comments from Martin. An interesting issue is how to
configure ALTO clients. Qin: I agree with Richard, we need some configuration
for ALTO clients. Dhruv: If we develop RESTCONF/YANG mechanisms for ALTO, will
existing clients support the function. Med: A data model will allow for
consistent configuration/management features nindependently of which “channel”
is used to touch that part Jordi: Additional use case includes Proxy or Load
Balancers that could benefit from an ALTO client.

Chairs: Overall, we see the authors have made good progress and we can proceed
with an adoption call on the list.

ALTO over New Transport (20 minutes)
         https://tools.ietf.org/html/draft-schott-alto-new-transport-00
         Discussion Leaders: Roland Scott
Qin (via Chat): For alto new transport in page 25,is there any out of order
delivery issue for streaming management? Richard: We expect a lot of concurrent
streams, we were careful to ensure there is not out of order delivery. 1. PUSH
packets are sent in order, in a single stream. 2. Built-in mechanism to include
seq number to incremental updates. These are two methods to avoid out of order
delivery. Qin (via Chat):How reciever set functionality works with other new
functionalities introduced? So no new extensions are required. Richard? Yes. We
will use http/2 for stream management. Qin: For the alto functionalities
describe, are these all mandatory, or will some be optional Richard: Mandatory,
but you can turn off the incremental updates. [Missed the comment about
application] Med (via Chat): Can we claim that the design follows recos in
draft-ietf-httpbis-bcp56bis? Richard: Good question. We have not checked but
will. Thanks for the suggestion. Martin: What is the impact of http/3 for alto
stream managment. It is fine for initial versions of the proposal to support
http/2, but we should analyse use of http/3, its due diligence. Roland: Our
main focus was http/2. but I’m happy to investigate http/3.

Deployment experience Update
         ALTO Code bases and Deployment (20 minutes)
         https://github.com/ietf-wg-alto/wg-materials
         Discussion Leaders: Jordi Ros Giralt/Kai Gao/Sruthi Yellamraju
Martin: Impressive. Good to see multiple drafts incorporated. Prior to this
project, did Rucio have alto support? Kai/Jordi: [will need to check audio]

Non-Chartered items
Supporting Bottleneck Structure Graphs in ALTO: Use Cases and Requirements
(20minutes)
        https://tools.ietf.org/html/draft-giraltyellamraju-alto-bsg-requirements/
        Discussion Leader: Jordi Ros Giralt

Qin (via Chat): @Jordi, can alto provide sufficient information to build
bottleneck structure graph? is there data translation needed? in the scope of
your draft or not? Richard (via Chat): There are two interpretations to this
question: (1) build bottleneck structure on top of alto and (2) make bottleneck
structure a (new) service provided by alto. For (1), we need some
modifications; I think the focus here is (2). Jordi? Qin: Lets move this
discussion to the list.

Misc (10 minutes)
        ALTO activities related to compute-aware networking (Luis)

Qin: Using alto could be good for compute aware networking. Please take this
discussion to the list.

Wrap-up (5 minutes) Chairs