Transport protocols and bulk file transfer

[stanislav shalunov <>]

                     Bulk Transport Protocol Call
                      Friday, February 25, 2005

Attendees:

   1. Lisong Xu
   2. Yunhong Gu
   3. Injong Rhee
   4. Larry Dunn
   5. Stanislav Shalunov
   6. Matt Zekauskas
   7. T. Charles Yun
   8. Steven Senger
   9. Susan Evett (scribe)

Call started at 12:04PM US/Eastern.

Agenda
   1. Review of previous action items.
   2. Editor for Protocol Specification -- volunteer needed.
   3. Support of Windows (and implications) -- position.

1. Action Items Overview

   A. Stanislav to develop a charter and side materials to request
      Internet2 working group status.  [ETA for charter: 2005-02-25.]
      [Done.]  -- Stanislav sent a draft around to the group and
      forwarded comments from Matt.  Stanislav asked the group for
      comments on the charter; group largely felt it was fine.
      Stanislav noted that milestones only run until December; Matt
      suggested adding a milestone in December to "Update Milestones."
      Group agreed -- Stanislav to add this milestone to the charter.
      He will wait a few days for comments and will submit the charter
      to the Internet2 WG review panel.

   B. Stanislav to revise the appendix, based on Larry's comments, and
      verify questions he has over equations in the document (see
      email).  -- [Done.]  Removed and changed body of the text to
      indicate the removal of the appendix.  Larry to look through the
      document to ensure all references to the appendix were removed.

   C. Group to discuss whether weekly, bi-weekly, monthly, or
      'as-needed' calls would best suit the needs of the group and the
      work to complete.  Group agreed that weekly calls were too
      often; Stanislav suggested biweekly calls instead.  Steven
      suggested that biweekly would be often enough to be sure that
      the group stayed 'on task'.  Larry suggested holding off on
      scheduling only 'as needed' calls until tasks have been
      sufficiently distributed.  Group agreed on bi-weekly calls for
      the near future.

   D. Steven to modify the API to incorporate a function call within
      the socket structure.  Steven has been working on it and will
      submit a draft by the middle of next week so that the group can
      generate feedback for discussion on the mailing list and he can
      generate another draft by the next call (03/12).  Stanislav
      suggested that Steven post questions and concerns to the list so
      that the group can provide input along the way.  Steven agreed
      to post information on the mailing list in the future.

   E. Group to discuss name / value of the tool within the API (tied
      to project name).  [On hold.]

2. Protocol Specification -- Editor

   No one stepped forward; on hold.

3. Support for Windows (and Implications)

   Stanislav reported that he does not know enough about how Windows
   works with advanced applications or implications of using Windows.
   Steven reported that his clients run on Windows (which makes it a
   requirement); Gu reported that UDP works on Windows.  Steven
   reported annoyances but not obstacles and Gu felt it would not be
   difficult.  Stanislav asked if it was possible to read the TSC
   register on Windows; Gu said that it does, at least on Windows XP.
   However, to reach the highest rate with UDP on Windows, you need
   1024 bytes/packet, else the performance falls off.  Steven found
   this interesting and wanted to do some experimenting with this.

   Larry suggested contacting the TCP stack guys at Microsoft or
   perhaps talk to the man from Microsoft who presented at one of Russ
   Hobby's panel; Matt reported that he has some contacts in this
   group, also.  Gu and Steven will collect their 'gotchas' while Matt
   and Stanislav will pursue a contact at Microsoft.

   Stanislav asked what other areas should the group be working on;
   Larry suggested porting (in the future), a log to report how things
   were working over time (also in the future) -- Stanislav called
   this the 'experience report'.  He asked about application needs
   statement -- would this be useful?  Larry thought this was in the
   Application archive -- he felt that there had always been a
   difficulty for Internet2 developers and users to be in contact.
   Stanislav noted that defining the needs would not be trivial.
   Larry wanted to, eventually, document the linkage between this work
   and NSF funding (potential).

Re: IRTF

   Injong has not heard from Aaron Falk (presentation at PFLDnet)
   regarding setting up a Working Group.  Stanislav reported a
   discussion with Thomas Ndousse at Joint Techs in re DoE and NSF
   funding for TCP research.

ACTION ITEMS FROM THIS CALL:

1. Stanislav to add a milestone to the Charter for December 2005 to
   ``update milestones'' and submit the charter for review by the WG
   panel.  [ETA: 2005-03-04.]

2. Larry to review current draft to ensure all references to the
   deleted appendix are removed.

3. Group agreed to meet biweekly in the future; because of a conflict
   on March 11, the group agreed to meet on March 4 and biweekly after
   that date.

4. Steven will submit a draft by the middle of next week for group
   feedback (via the mailing list); he will incorporate changes and
   generate another draft by the next call.  [ETA: 2005-03-04.]

5. Group to discuss name / value of the tool within the API (tied to
   project name).  [Continued.  On hold.]

6. Gu will collect his 'gotchas' on programming with Windows.

7. Steven will collect his 'gotchas' on programming with Windows.

8. Matt will pursue a contact at Microsoft to help with Windows
   porting.

9. Stanislav will pursue a contact at Microsoft to help with Windows
   porting.

Next call is at the same time on 2005-03-04.
Call ended at 12:47PM US/Eastern.

-- 
Stanislav Shalunov              http://www.internet2.edu/~shalunov/

I never let school stand in the way of my education.  -- Mark Twain