Attendees:
Kevin Barron,
Steve Francis,
Nathan Freitas,
Jim Frew,
Ben Humphrey,
Elise Meyer,
Larry Murdock,
Stephen Pope,
Kevin Schmidt, and
Jason Simpson, Alexandria Project
Unable to Attend:
Vince Sefcik (who will probably not be able to attend until October 1997)
General Business
Kevin Barron distributed a copy of "MPOA Ties It All Together," and also mentioned that
the Data Communications Tech Tutorials was a helpful resource for ATM information.
All URLs for good resource areas are welcome, and Elise will put them up on the
web page.
The group was pleased about the possibility of OC-12 connections to the vBNS
as noted in the minutes instead of the originally planned OC-3.
Elise distributed pp 9-10 and A26-A28 from the CalREN-2 Proposal (PDF), which had
the High Performance Attributes and Application Descriptions for four of the
CalREN-2 research group projects.
Research Group Reports
Our first task is to determine what each research groups requirements are for
bandwidth and quality of service, and what their infrastructure is like. We
started to fill in the Research Group Requirements Worksheet for the High
Energy Physics, Electronic Art Technology Lab, Alexandria, and ICESS projects.
In the course of discussing those four research projects we got into some
lively discussions in the following areas:
- Steve Francis updated us on the latest UCNET plan, which is to have
the UCNET traffic use the CalREN-2 inter-campus network as soon as
it is stable. This means that if any of the research groups have no
Quality of Service requirements and have their bandwidths needs being
met by their current connection to the campus FDDI backbone, they don't
need anything else. He also mentioned that there would be an "undernet"
connection between the ATM switch and the existing campus router commgw,
which would be used as an alternate path out of campus if the GSR had
problems.
- We discussed the difference between multi-homing on a host by host
basis vs. multi-homing a network. Steve Francis created a diagram
to explain this along with a possible campus network topology.
- We discussed whether non-CalREN-2 UCSB traffic would go to a vBNS site
via the vBNS network, or via the vBNS site's alternate ISP. Steve Francis
explained that all UCSB traffic bound for a vBNS site would go via the
vBNS network with the following diagram.
Stephen Pope gave a brief description of the Next Generation Networked Multimedia
(NGNM) project. They are building systems that use 64 channels of data. Their
worlds require 40Mb/s. They are concerned with transport latency (the time
delay in sending data across a network) and latency jitter (latency variance
over time). They are developing adaptive programs for latency jitter. More
details about this project are available from their web page. Stephen noted that
they wanted to be able to do direct atm between UCSB and their other CalREN-2
destination sites.
Action Items for Next Meeting: August 1, 1997
Prior to the meeting, the remaining
research groups are to mail Elise their requirements information that she will
incorporate with the data from today's meeting.
Back to Main CalREN-2 Page
EMM
