Status Report for SLAC/DOE IEPM Project

For February & March 1999

Coordination/Publicity/Deployment

Les gave a talk at Stanford main campus as part of their series of seminars on networking (see http://www.slac.stanford.edu/grp/scs/net/talk/cs-mar99 ). Also Les has been preparing a talk he will present at a Cisco workshop on peering.

Les and Warren have been co-ordinating with Bill Lidinksy at HEPNRC (the PingER archive site) to resolve some issue on broken links and keeping the information up-to-date.

Les has been working to increase the number of sites running reverse traceroute servers. We are hoping to increase our understanding of changes in network performance by better understanding routing changes.

Warren has been working on a brochure and a pamphlet to give a non-technical overview of PingER and network monitoring in general.

Les worked with NetPredict to prepare their ESnet presentation. Also Les prepared some slides for Walter Toki to present at the HEPAP meeting.

Development/Analysis/Documentation

Les and Warren worked with Linh Khuu from IBM to help them set up PingER. Warren worked with Jeremy Hylton from the XIWT archive site at CNRI to resolve some issues and develop visualization and analysis techniques. Warren worked with Michael O'Conner at BNL and was provided with a login to BNL in order to get their monitoring site set up again. Warren reviewed the data that was gathered in his absence and retrieved much of the data that was missing due to temporary glitches.

Les worked on understanding systematic and random errors in measuring ping repsonse times compared to the wire time. He also documented the effects of ICMP traffic shaping (see http://www.slac.stanford.edu/comp/net/wan-mon/pathology.html ). Les also met with Duong Van Minh to discuss the use of sub-fractals and chaos theory to analyze Internet traffic.

Les worked with Becca Nitzan of ESnet to help gather information on VoIP and analyse it. Warren is working to set up a machine on the VoIP subnet and will install PingER.

Les and Warren, together with the folks at DESY and in Russia worked to try and understand the causes of some performance changes observed recently to Novosibirsk (see for example http://www.slac.stanford.edu/grp/scs/net/talk/peering-mar99/sld017.htm). Some of the problem is the shadow of earth where the satellite gets to from 17.30 GMT till 20.00 GMT. The Satellite Control Center swithces off the satellite transimter at this time. They save battery energy.

Revised 22 March 1999
URL: http://www-iepm.slac.stanford.edu/about/status/status-03-99.html
Comments to iepm-l@slac.stanford.edu