|
Status Report for SLAC/DOE IEPM ProjectFor July 2001
Work completed or in progress since the
last report.
|
PingER
Les is working with Frank Nagy of FNAL to understand and provide, where possible, replacements for some hosts/sites (about 6) that are no longer reachable.IPEX
Warren and Les attended a telephone conference call to discuss the initial deployment and configuration for the IPEX project co-ordinated by the XIWT. Fred Murphy of SLAC's business services is close to finalizing the agreement with CNRI for SLAC being a subcontractor on the DARPA funded IPEX proposal.INCITE
The INCITE project has been funded by DoE. SLAC is a collaborator on this project and is funded at the level of about 1 FTE. Warren and Les talked with Wu-chun Feng of Los Alamos National Lab and Rich Baraniuk of Rice University about setting up pingER at their sites and integrating the tools from the INCITE project. Les put together a statement of work and timelines for the SLAC part of the proposal and sent to Thomas Ndousse. Exchanging email with PSC folks concerning adding NIMI nodes at LANL and Rice.Web100
Doung Chang attended the Web100 Web100 workshop in Boulder Colorado. He gave a status report at the SLAC network group. He maintains a web page on activitiesQBSS and file transfer throughput
Warren worked with Paola Grosso to implement and test setting the 'qbone scavenger service' bits in the ip header using the bbftp file transfer tool. Les worked with Andy Hanushevsky, the author of the bbcp copy program to help debug it, recommand and add new features including support for setting QBSS bits, /dev/null, to deploy it at IN2P3 and Daresbury, and to make throughput measurements with it. As part of this Les communicated with people at Daresbury and IN2P3 to understand the configurations of host and networks at those sites. We made measurements with tcpdump to further understand how bbcp etc. was working. Paola and Gary are looking to set up a test network at SLAC where we can impose a bottleneck and use QBSS to see how it works. Les did some work on measuring the cpu utilization of iperf. Les worked with a student on automating bbcp measurements.Les has been in communication with Stanislav Shalunov of the QBSS Design Team and Les will make a presentation on applications that can use QBSS at the I2 meeting in October.
NTON
NTON is down.Grid
Warren has been working with the global grid forum's performance working group and the newly created network performance working group.Traceping
Warren and Imran started to implement the 'traceping' tool in java. Java was chosen to allow users to run on a variety of platforms without having to install new programs. Java should also ease the task of deploying the tool to nimi. Les and Teresa found the source to the original traceping (in VMS/DCL). Les sent email to the original traceping author, John MacAllister informing him of our plans.Traceroute
Les responded to about 2 a week for false intrusion alarms raised by firewalls which appeared to come from www.slac.stanford.edu. They wre false positives in response to reverse traceroutes. Les updated and added information in the web page explaining how traceroute works and may appear as a scan.Missing data
The Pinger site in Korea has been down for the entire month. It seems some changes were made and the local administrator cannot reverse it.Problem assistance
We assisted BNL to understand why they were seeing poor throughput performance. It resolved to be due to a Cisco PIX firewall resetting the window scaling bits.
DoE
Several members of the team from DoE/MICS visited SLAC and Warren presented a review of the research they are funding.ESCC
Warren attended the ESnet Co-ordinating Committee (ESCC) meeting at Brookhaven National Lab (BNL). He presented a talk about SLAC projects titled Advanced Network Collaborations at SLAC.IP Perf
Warren submitted an abstract for the IP Perf conference to be held in November.ACM Sigcom Performance Workshop
Warren worked with Giuseppe Paleologo, a statistician formerly at Stanford now with IBM, to write and extended abstract on comparing the data from the Surveyor and RIPE-TT network performance monitoring projects. The Abstract is online in ps and pdf format. The workshop will be held in November.GridForum
Warren attended the GGF meeting in Washington, DC. He presented a review of network monitoring to the performance working group and attended a bof regarding the creation of a network working group.CHEP
Connie and Les worked on the paper on Passive Monitoring for CHEP01 and submitted it. Warren & Les submitted a paper on high performance throughput and IPv6. Andy hanushevsky and Les submitted a paper on the bbcp copy program. Warren submitted a paper titled 'IPv6 in ESNet' with Bob Fink (ESnet), Susan Hicks (ORNL) and Vyto Grigaliunas (FNAL).SC2001
Les drafted the text and outline for a panel for the SC2001 SLAC/FNAL booth in Denver in November.Other Publicity
Les gave a lecture on "Characterizing the Internet" to summer students on the Stanford campus.Les gave 6 lectures on the Internet as a faculty member at the International Nathiaglai Summer College in Pakistan.
Les also talked to people in Psakistan about setting up a PingER monitoring host in Pakistan, and to people from Myanmar and Bangladesh about monitroing hosts in those countries.