Minutes for ALTO at IETF-92
minutes-92-alto-1

IETF 92, Dallas, TX (USA)
Thursday, March 26, 2015 1520-1720 US Central Daylight Time ALTO Working Group
meeting ~20 people in room Chairs: Jan Seedorf, Vijay K. Gurbani and Enrico
Marocco (not present) Notetaker: Piotr Wydrych Jabber scribe: Vijay Gurbani

Administrivia                        Chairs            04'
Who has running code? 3: Richard Yang (Yale), Lyle Bertz (Spring), ?
Broad support for making an interop
Next IETF to early? Sabine Randriamasy not sure RY (Richard Yang): slightly
behind milestones, how should we push on? JS (Jan Seedorf): we'll come back to
this question

ALTO deployment considerations       M.Scharf          10'
draft-ietf-alto-deployments-11
Moving to WGLC? Yes - clear consenus

Extended e2e properties for ALTO     H.Song            10'
draft-deng-alto-p2p-ext-05
JS: should we hold information that changes frequently (like 2G/3G/4G)?
VG (Vijay Gurbani): How to prevent revealing private information?
JS: Max bandwidth / best technology for a given endpoint may be more stable.
JS: Who has read a document? 2 hands raised.
RY: Document touches dynamics and privacy.
JS: This topic is returning. According to IESG, ALTO is not for TE and avoiding
congestions, information changing within seconds are out of scope. Daily - in
scope. RY: What is too dynamic, what is not? JS (as indiv.): 3G/4G is too
dynamic. JS: Rich Woundy was always concerned about disclosing business
information. RY: Privacy is a serious issue. SS (Sergey Slovetskiy): What
defines what is in scope / out of scope? JS: ALTO Prolem Statement SS:
Fast-changing properties may be important for CDNs.

Unified approach to properties       W.Roome           10'
draft-roome-alto-pid-properties-03
JS: There are another services than map service. How does it relate?
WR (Wendy Roome): Not yet sure, but network map seems to be most apropriate.
VJ: If we want an unified approach, we need to work on it within the group.
WR: One approach is to define a PID property service. This is a second approach.
MS (Michael Scharf): We need to to define PID properties somehow.
JS: Does this overlay with previous presentation.
MS&VG: It is crucial to define a framework for defining endpoint and PID
properties. JS: There is IANA registry for registering new properties. VG: It
is not clear how should we define it. JS: We need to discuss it on list. Not
adopting any document as WG document. ZL (Zhu Li): We would be insterested in
using ALTO properties to define what is held in caches. LB (Lyle Bertz): We've
extended properties in our implementation many times. Good way to go. VG: Lot
of work before next IETF. We need an unified approach.

ALTO incremental updates             W.Roome           15'
draft-roome-alto-incr-update-sse-02
VG: How many people read the draft? 2 hands raised. We'll try to adopt the
draft as WG document within few weeks through the maillist. JS: It's an
important issue. There is no competing proposal.

Multi-cost ALTO                      S.Randriamasy     10'
draft-randriamasy-alto-multi-cost-10
JS: No competing approaches, been a while. Adopting?
Quiet hummm, but clear consensus. Adopting as WG draft.

ALTO cost calendar                   S.Randriamasy     10'
draft-randriamasy-alto-cost-calendar-03
VG: Who can discuss it now? RY (an author).
RY: The are some semantics issues. E.g., what does not mean when the calendar
says that some value is valid for 24h? Is it from request or from given point
in time? VG: We'll taking the document to the list.

ALTO use cases in NFV                Q.Fu              10'
draft-fu-alto-nfv-usecase-04
MS: This proves that we need a registry.
RY: The sequence of destinations is an important topic.

YANG model for base ALTO data        X.Shi             15'
draft-shi-alto-yang-model-02
WR: Is it for changing maps or server configuration (defining property types)?
XS (Xiao Shi): Changing maps data.

Topology/path-vector                 R.Yang            20'
draft-yang-alto-path-vector-00
draft-yang-alto-topology-06
draft-huang-alto-topocomp-framework-00
?: It is an important work on topology and reduction of topology.
MS: It's a cool stuff, not competing with other IETF works.
JS: Support. It's complex, people should have time to read through it.
Hummm. Support in room. To the list as well.

Inter-ALTO: A second attempt?        P.Wydrych         05'
No draft

Conclusion/Wrap-up                   Chairs            01'