Minutes for ALTO at interim-2015-alto-1
minutes-interim-2015-alto-1-1

ALTO Virtual Interim Meeting
Tuesday, October 27, 2015 8:00AM US Central - 10:10 AM US Central
Vijay K. Gurbani and Jan Seedorf, ALTO chairs.

Attendees: 22-25
Note takers: May （Haoran）<whr_hr_r@....com>
             Richard Yang
Jabber scribe: Vijay K. Gurbani

Proceedings are archived at:
https://www.ietf.org/proceedings/interim/2015/10/27/alto/proceedings.html

Meeting started and agenda was bashed.  Lyle Bertz indicated that he 
will coalesce two talks into one timslot.  Vijay noted that there is one
document pending with the WG: draft-ietf-alto-deployments-12.  This
draft has been through WGLC and is gated on the shepherd (Vijay) to 
finsih the proto-writeup.  The draft should be out of the WG by Nov. 15,
2015 the latest.

W. Roome, ALTO Incremental Updates
draft-ietf-alto-incr-updates-sse-01
Wendy presented the changes since -00 and noted that it is past its
milestone deadline.  (NOTE TO CHAIRS: update milestone deadlines.)
There was a discussion on why must ordinal costs be integers (as 
specified in the base ALTO RFC).  The chairs urged this discussion to
be taken to the list so the WG could focus on the next steps with this
draft.  A few WG members (Hans) indicated that they are working on 
implementing this draft and have been involved in off-line interoperability
tests.  Big question under discussion for next steps was: should the
draft wait until HTTP/2 became widely deployed or should we take the
draft to WGLC now?  The general consensus of the group was to proceed
with the current version of the draft and not wait until HTTP/2 became
widely deployed.  At the appropriate time that wide deployment of 
HTTP/2 occurs, the decision to revise the (then) RFC corresponding to
this draft can be made.

W. Roome, ALTO Interoperability Test Specification
draft-roome-alto-interop-ietf93-01
Wendy indicated that the IETF 93 interoperability event was a resounding
success, no bugs were found in the ALTO RFC (RFC 7285).  All slips that
occurred during the interoperability were huam foibles.  There are on-line
test tools available from Alcatel-Lucent.  The big question was: should 
we make this an RFC?  The general discussion in the group was that yes,
the work in this draft is important should proceed towards an RFC track for
archival purposes.  The chair (Vijay) noted that there are ways to move 
work towards an RFC in an AD-sponsored independent stream if there is no
charter deliverable corresponding to the work item, and if the WG 
feels that this work deserves to move ahead as such, then the chairs will 
facilitate.  (NOTE TO CHAIRS: talk to AD to move this work towards RFC as 
an independent stream.)

L. Bertz, e_ALTO project
draft-bertz-alto-aggrimpl-00
Lyle discussed e_ALTO (Erlang ALTO) that he had brought to the IETF 
93 interoperatability event.  Talked further about how and where to aggregate
data coming from multiple sources to ALTO: the client can aggreagte or
third-party aggregation via another ALTO server.  Lyle has some relevant
experience from running aggregation in e_ALTO.  Richard Yang liked this
work, as did other folks attending the meeting.  The chair (Vijay) noted
that this draft may benefit from Piotr's inter-ALTO work and urged the
relevant authors to coordinate activities by next IETF.  The chair (Vijay)
summarized the work as being of interest to the working group in general,
but where the work ends up in the WG stream needs to be determined by the
working group, preferably before the next IETF in March 2016.

L. Bertz, Mobility Network Representation in ALTO
draft-brtz-alto-mobilitynets-00
Lyle went through the draft on how to provide an EPCS based mechanism
that applies to many different applications (CDNi, DMM) and that 
generalizing point of attachment for a mobile host (UE) as a PID allows
the network topology to change without impacting the costs.  The chair
(Vijay) summarized the general consensus of the attendees towards this
work as positive and for now the work should continue on the list as an
independent submission.

K. Gao, Routing State Abstraction
draft-gao-alto-routing-state-abstraction-01
Kai presented his work on how path vector as a solution to the problem of 
allowing applications that require multi-flow coordination.  The chair
(Vijay) noted that this is important work towards the charter deliverable
of a graph representation format in ALTO.  The next steps for the draft
was to continue its trajectory in the WG as an independent submission until
such time that one or more drafts are used to satisfy the charter deliverable
of a graph representation format.

J. Wang, ECS for flows: ALTO Extension
draft-wang-alto-ecs-flows-00
Jason made the case for ALTO to support multiple address types (besides
the IP address type it supports now) plus some flow constraints (by
naming protocols, limiting port numbers, or ToS bits).  There appears to
be support and interest for this work in the WG, especially because the
work in the draft may be beneficial to large enterprise networks.  The
next step for the draft was to continue on the independent stream for 
1-2 more cycles.

J. Zhang, Experiences of implementing ALTO in OpenDaylight
draft-zhang-alto-opendaylight-impl-00
Jensen presented his work in designing and implementing an ECS ALTO service
in OpenDaylight.  Unfortunately, time constrains prevented an indepth
look at lessons learned from this experience.  The chair (Vijay) urged
Jensen start a discussion on the working group mailing list to distill
lessons learned from his work.

R. Yang, ALTO Extension: General cost types for path vectors and others
draft-yang-alto-general-cost-type-00
draft-yang-alto-pv-topology-00
Richard presented his work on generalizing the cost-mode to a broader
set of data types and on "parameterizing" the cost map service.  The
chair (Vijay) noted that this work ties in with the charter deliverable
of a graph representation format, for which the WG does not yet have
any candidate drafts.  In this context, it is urged that the WG
participants pay close attention to the set of drafts that discuss
path vector and other similar extensions to determine whether we can
coalesce around a set of drafts to address the open charter item.
Richard was asked to continue moving the work ahead as an independent
submission until the time that the WG decides on candidate draft(s).

P. Wydrych, Inter-ALTO problem statement and next steps
draft-dulinski-alto-inter-problem-statement-02
Piotr is mostly interested in determining next steps for this draft.
The draft has not changed since IETF 93.  The chair (Vijay) noted
that the work has been around for a few IETF cycles, and in fact,
has some synergies with the work the Lyle Bertz presented on
e_ALTO and aggregation.  Piotr and Lyle were urged to consult and 
determine how to move their respective works ahead, possibly as a
unified draft (?).

S. Randriamasy, Multi-cost ALTO updates
draft-ietf-alto-multi-cost-01
Sabine went through the changes in the draft (see slides).  She has
received updates on Section 1 (Introduction) from Richard Yang and
Wang Xin that need to be included in a revised version.  She would
like to get WG feedback and prepare for WGLC.  The chair (Vijay)
noted that the draft has been discussed at many IETFs, even before
it became a working group item and as such, is fairly mature.  He
requested the authors to revise the draft to include all known
comments so far in preparation for a WGLC, preferably before March
2016 IETF.
 
H. Seidel, Experiences with ALTO map calculation from network data
draft-seidel-alto-map-calculation-00
Hans presented his work on the challenges in creating ALTO maps
from network data.  The general consensus of the attendees appeared
to favour the work as an interesting exercise.  The chair (Vijay)
asked whether the work had any bearings on SSE.  Hans indicated
that he has been implementing SSE.

S. Randriamasy, ALTO Cost Calendar updates                     
draft-randriamasy-alto-cost-calendar-05
Sabine went through the updates to the draft.  The chairs (Vijay)
noted that this draft has been presented in many IETFs but is not
a WG item.  Clearly, there appears to be interest in this draft,
especially for scheduling large-scale data transfers.  The decision
on whether or not to adopt as a WG item will be made on the list
before the March 2016 IETF.

End of meeting.