Meeting notes from ALTO WG meeting, IETF 86, Orlando, FL, USA.
Friday, March 15, 2013

About 40 attendees in room + 22 in Jabber session.

This report has been distilled from the detailed meeting notes taken
by Jan Seedorf and Haibin Song. Thanks, guys!

Chairs: Vijay Gurbani and Enrico Marocco.

Agenda
------

Meeting materials: 
  slides, 
  MP3, 
  Meetecho recording (HTML5) 

Enrico explains agenda and highlights that core protocol needs to be
done first before we can look at any extensions, that is why at this
IETF there was a side meeting on extensions and official ALTO agenda
time was only allocated to chartered items.

  Martin Stiemerling (as responsible AD): need of energy for the base
  specification, any re-chareting will not be approved before the core
  spec is finished, so please focus on this and help getting it done.

  Michael Scharf: when can we start looking at extensions in detail?
  Martin: when the core spec is on the way to the IESG, i.e. when it
  has passed the AD review.


ALTO protocol
-------------

Meeting materials: 
  slides, 
  MP3, 
  Meetecho recording (HTML5)

Richard Yang presents on core ALTO protocol and latest changes.

  Reinaldo Penno : not sure random string for constraints on cost type
  is a good idea, what about using well-defined company-ID? Martin:
  indeed a good point. Enrico: new name "cost-id" maybe misleading.
  Michael: why not call it cost type and name the old "cost type"
  differently, e.g. "cost metric"? Vijay: if nobody objects let's go
  with this proposal.

  Jan Seedorf: support for Sebastian Kiesel's proposal on
  restructuring the security section, will try to help.

  Martin: what is the problem with the current JSON schema? Richard:
  it may be ambigous in certain cases, that is why we want to revise
  it. Enrico: let's look closely at the wording regarding client
  authentication, i.e. let's not preclude other IETF client
  authentication mechansims.

  Reinaldo: on ALTO error codes vs. HTTP codes, a possible way, if
  HTTP error, just copy it to ALTO layer.


Third party server discovery
----------------------------

Meeting materials: 
  slides, 
  MP3, 
  Meetecho recording (HTML5)

Martin presents on third party ALTO server discovery. Sebastian has
developed a proof of concept, and it works.

  Michael: need input from DNS experts with real-world experience, but
  looks promising.

  Enrico (poll): how many people have read the draft? Five people in
  the room plus one on Jabber have read the draft, only positive
  feedback; WG adoption decision on 3rd-party draft postponed to the
  mailing list.


Anycast based server discovery
------------------------------

Meeting materials: 
  slides, 
  MP3, 
  Meetecho recording (HTML5)

Reinaldo presents on discovery based on well-known addresses.

  Martin: like the draft.

  Michael: there are a lot of open security issues regarding this
  approach. Enrico: other works/extensions should not depend on this
  approach.

  Enrico (poll): how many people have read the draft? 3 people have
  read the draft.

  Enrico (poll): how many people understand the approach? 8 people in
  the room understand the approach based on the presentation.

  Enrico (poll): how many people think it would be harmful to do
  discovery based on well-known IP? 0 people think it would be harmful
  to do discovery in the proposed way (i.e. with anycast addresses).

  Wolfgang Beck: what server would be returned?  Reinaldo: a server on
  your network.


Way forward on ALTO extensions
------------------------------

Meeting materials: 
  MP3, 

Vijay provides a short summary of the extensions discussed during the
side meeting on Wednesday.