Transport protocols and bulk file transfer

[Susan Evett <>]

All,

Attached are the notes from the 2/12 meeting in Minneapolis in .doc and 
.txt formats. The group agreed to reconvene on 3/27 from 2-3 pm CST. 
Call-in info is:

Starting Time: Mar 27, 2007 at 2:00 PM US/Central
Duration: 1 hour

To join the call, do one of the following:
a. Dial +1-734-615-7474 (NON-US/CA USERS MUST USE THIS NUMBER - SYSTEM 
WILL NOT CALL OUT), +1-866-411-0013 (toll free US/Canada Only) or 
+1-800-392-6130 (alternate toll free US/Canada Only) and enter access 
code 0125223, or
b. To have the system call you, click here: 
https://edial.internet2.edu/call/0125223

Talk to you all then,
--Susan

Attachment: 2007021207-jts.doc
Description: MS-Word document

BULK TRANSPORT WG Meeting
Monday, February 12, 2007

Attendees:
Steven Senger
Chet Rusczyk
Martin Swany
Glen Turner 
Ken Lindhal
Susan Evett (scribe)

Meeting started at 12:45 p.m. CST. 

Phoebus is considering using VFER for the rate-based transport protocol. 
Martin is 
considering introducing Phoebus as a bulk-transport platform to this group to 
get 
feedback (Phoebus is an offshoot of his black cloud proposal and logistical 
session layer 
work). Martin suggested plugging Phoebus into the requirement the group has 
already 
determined to determine where it is a match and/or there are holes; 
specifically, group 
should check against VFER/plug in; compare to API. Martin asked the group to 
imagine 
Phoebus as a service in the new Internet2 Network.

Glen Turner (AARNet) said that the EU e-VLBI community is looking to develop 
their 
own bulk transport software because they are having difficulties w/the one 
developed by 
Haystack – he said that there were problems when routing difficulties arise. 
Most all of 
the applications would need to sit in the kernel space; Chet asked for more 
detailed 
feedback from Glen’s group. (Chet and Glen took discussion of detailed 
comments 
offline, after the conclusion of the WG session.) Glen reiterated that 
anything that pushes 
across the implementation experience would be useful.

Martin expressed concern that the design space document is missing concerns 
about end-
to-end packet flow; Steve felt that the group was not wed to the VFER 
implementation. 
He worked on a draft implementation and API; he’s had problems getting it to 
run on 
their applications. Group agreed to review Martin’s logistical session layer 
(LSL) paper 
and Phoebus materials against the design space specifications. Group agreed 
to a March 
27, 2007 (03/27/07) meeting 2 pm CST; Susan will setup a call and send around 
info.

Meeting adjourned at 12:30 p.m. CST.

ACTION ITEMS:
1.      Steve handed off the API to Susan; she will post to 

 and 
post on the website w/a last call set for March 9, 2007.
2.      Steve is to check with to insure all changes to API requested by Ivan 
have been 
received by last call date.
3.      Steve, Martin, and Chet will compare Martin’s LSL and Phoebus-related 
materials against a) the design space document and b) their own application 
requirements. Areas of convergence and non-convergence should be specified.
4.      Susan will create a Relevant Work section on the Transport homepage 
(transport.internet2.edu) with links to Martin’s work and VFER; other 
relevant 
material that the group may need to access will be linked from here as needed.

Call ended at 1:35 pm CST. Next call is 03/27/07 at 2 pm CST.