Editor's note: These minutes have not been edited.
IETF'37: PacketWay- WG Minutes
------------------------------
The PacketWay working group met at IETF'37 in San Jose on Wednesday,
December 11. There were about 17 attenders [see attached list].
Danny Cohen opened the meeting with a PacketWay overview and the meeting
agenda.
Robert George (of Mississippi State University, MSU) gave a presentation
of the work they have been doing on the implementation of PacketWay and
the Secure extensions to PacketWay. Currently MSU is supporting two
different underlying transport protocols: Myrinet and UDP. In the
future, MSU plans to extend this support to RACEway and the Avalon link.
MSU will demonstrate interoperability of PktWay between clusters using
Myrinet, RACEway, and Avalon links.
Nathan Doss (of Lockheed-Martin-Sanders) gave a presentation on the work
Sanders is doing on a second implementation of PacketWay. This work
involves Sanders's two-level high-performance multi-computer testbed.
It is intended that the two PacketWay implementations (MSU and Sanders)
be interconnected to demonstrate interoperability.
Thom McMahon (of MSU) gave a presentation of the preliminary work MSU is
doing in adding multicast capability to the base PacketWay
specification.
The group then discussed what steps we should be taking to move
PacketWay forward on the IETF standardization path. It was agreed that
Danny should reissue the PacketWay specification with just the core
features (Levels A through C, without Level-D, the dynamic route
discovery). This separation is motivated by the initial
implementations that cover only Levels A through C, without Level-D.
This document would then be submitted (in January?) to the IETF for
release at the "proposed standard" level. The next step would be to
receive IETF-wide scrutiny of the document and complete the
two-implementation interoperability demonstration. These efforts should
lead to "draft standard" status within the year.
Separately, dynamic-route-discovery, the Secure PacketWay, and the
PktWay-Multicast extensions should be refined and entered onto the
standardization path as independent additions to the core PktWay
specifications (similarly to the way IP is augmented by similar
protocols).
Recorded by: Fred Shirley
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
ATTENDANTS LIST
PktWay-WG, IETF'37, Dec-11-1996
Name Organization E-address
<01> Danny Cohen Myricom <cohen@myri.com>
<02> Craig Lund Mercury <clund@mc.com>
<03> Vince Laviano George Mason Univ. <vlaviano@cne.gmu.edu>
<04> Lava K. Lavu George Mason Univ. <llavu@gmu.edu>
<05> Fred Shirley Sanders <fshirley@sanders.com>
<06> Robert George MSU <robert@erc.msstate.edu>
<07> Thom McMahon MSU <thom@erc.msstate.edu>
<08> Glenn McGuire Cross Comm <gmcguire@crosscomm.com>
<09> Phil Irey NSWC <pirey@nswc.navy.mil>
<10> Mark Pullen GMU <mpullen@gmu.edu>
<11> Jim Barnes Baynetworks <barnes@xylogics.com>
<12> Guevthu Schmuelling Pyramid <gs@pyramid.com>
<13> Seshadri Srinivasan Pyramid <seshadri@pyramid.com>
<14> Scott Wasson Whittaker Xyplex <swasson@xyplex.com>
<15> Jon Postel USC/ISI <Postel@ISI.EDU>
<16> Nathan Doss Sanders <ndoss@sanders.com>
<17> Frank Kastenholtz FTP Software <kasten@ftp.com>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[end]