W3C home > Mailing lists > Public > ietf-discuss@w3.org > December 2000

Minutes

From: Hilarie Orman <HORMAN@novell.com>
Date: Mon, 18 Dec 2000 20:13:22 -0800
Message-Id: <5.0.2.1.2.20001218201209.00b01220@FMSMSX63.fm.intel.com>
To: ietf-openproxy@imc.org
Cc: discuss@apps.ietf.org
Minutes of the OPES BOF, IETF49, San Diego, December 12, 2000
Chairs: Michael Condry, Hilarie Orman
Area Director: Patrick Fahlstrom

"R" indicates a response given a the meeting; "A" indicates
an answer from the chairs after IETF concluded.

Q:  What is the overlap with the MIDCOM and can we set up
a liaison with that group?
A: (we've talked to Melinda Shore and might share a some policy
definitions and identity representations with MIDCOM; also, the
MIDCOM charter will be narrowed from what was proposed before
the BOF).

Q:  (from Micah Beck ) Feels that the service being provided have very little
tie to the content (e.g. ad insertion).  He would like to see more examples
involving programming using and interpretive language.

R:  (Hilarie)  Until an implementation framework is defined, this is
out of scope.

Q:  I would like to see other services as examples.
R:  See the drafts and the presentation.

Q: (Patrick Fahlstrom)
    Three comments:
    * The IETF likes to see more focus in the charters.
    * Feels that we do not agree on our deliverables.
    * Charter should have a list of deliverables, or the charter
           should be to identify the list of deliverables.
R: More details on deliverables are in the roadmap but can be moved
to the charter.

After the Examples Slides:

Q:  (Pei Cao) With the example list this broad ("all over the map"), I am
unconvinced that all these services will be supported.
R:  These examples are used to create the baseline requirements
so we can implement the common functionality.  The
requirements/architecture draft explains the framework in more detail.

Q:  What protocols will we be talking about in this working group?
R:  ICAP, configuration and distribution.

Q:  The internet is a layered system with each layer adding
value.  I hope that ICAP and the rest of the framework will show value.
R:  ICAP is well defined and where we will start.

Q:  (Patrick Fahlstrom):  I am concerned that we are saying what problems
we are solving and not what we are doing.  (I interpret this as we are
concentrating on the example problems that will be solved with the framework,
rather than the framework itself.)

R:  This is only a perception from the presentation, and that is
do to a lack of time.

The ICAP slide
Q:  How much easier will ICAP protocol make [ICAP] application writing?
R:  Departing from HTTP made writing the application much
easier, and is an easy adaptation of Apache (mostly requiring the removal
of 300-500 lines of code). There where some serious issues raised
when trying to meet the HTTP standard that were done away with.

Q:  Will ICAP be open source?
A:  That is up to the various implementors.  There is commitment at this time.

CDN Peering from Mark:

CDN Peering is what it takes to transport information across
administrative boundaries.  There is little relationship with OPES.

Q:  Will OPES be over top of  CDNP?
R:  More along the lines of beside it.  OPES could run over a
CDNP, and we will need some common ground - naming, in particular.

Q:  (Balaji Pitchaikani) Would they share accounting?
R:  Probably in terms of what intermediaries should be trusted?
The answer could come from the accounting system.
Q:  (same person) Would like to see some documentation or participate in
creating a document.
RR:  We need to define the when, why and where the OPES
extensions will happen and define the protocol issues to provide the core
functionality first.

Q: (Barron Housel) Will adaptation, interpretation,
etc depend upon user awareness?  e.g. identifying display devices.

R:  We are ironing out just what will be happening from the
user.  For instance, we may look into using avatars.

Q:  Authentification and identification - how are we discovering this?  ICAP?
R:  ICAP can be used, but we are concerned with maintaining and
sending identity rather then discovering it.

Q:  (Scott Brim) Just a concern - there are 5 or 6 groups that are close to
each other, in this space.  Howe do we divide and separate the concerns?
Perhaps a generalization of groups and a refocus.
R:  Here we are being told to narrow the focus and not make it more
general.
Q:  (same person) Again a mention that there is a lot in common with MIDCOM,
which has a Thursday morning BOF.





Received on Monday, 18 December 2000 23:17:18 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:38:00 UTC