Status Report for SLAC/DOE IEPM Project
For May 2001
Work completed or in progress since the last report.
Web100Doug has created a web page documenting his work on Web100. He also installed web100alpha 2 patch, modified rh-slac.rh to work on redhat 7.0 for slac services, and evaluated web100 performance metrics for troubleshooting. They fall into 3 categories, 1) time tcp socket buffer is empty, 2) time cwin showing multiplicative decrease, 3) time rwin is empty. Doug modified avd to print out values to file for graphing and verification and evaluated the web100 ability to write to tcpstack for alpha1; does not work, causes stack faults.
Warren attended the web100 evaluators conference call and the web100 BOF at the Internet2 meeting in Nebraska.
JPlot and JDBCImran has been modifying and updating the Guthrie program used to manage the details of the large number of nodes involved in the pingER monitoring.
ChinaWe are still working with KEK, IHEP Beijing, Abilene to try and improve the routing from the US to CAS China. See Comparison of ihep.ac.cn with cas.ac.cn
Missing dataWorked with BNL to clear some Beacon site problems there. Now working with ICFA and others to try and get new Beacon sites for Brazil and Sweden.
The BNL monitoring site is still down. ILAN in Israel and ANSP in Brazil are also down.
New SitesThere is a new PingER monitoring site at the University of Melbourne in Australia.
Warren has been working with two sites in South Africa to bring up PingER. Installation is hindered at one site by a firewall and at the other because they only have NT4.
QoSWe have been exchanging email with Stanislav Shalunov of Internet 2 and NASA/GSFC and others to look into setting up some Qbone Scavenger Service tests. Made some measurements of throughput etc. between SLAC and GSFC. See http://www.slac.stanford.edu/grp/scs/net/case/qbss/ for details.
Warren met with Stanislav and George Uhl (GSFC) at the Internet2 meeting in Nebraska to discuss implementing QBSS for the GLAST collaboration. However, GSFC is a large site and without specific details from the physicists it is impossible to know the requirements. Warren will work out a scheme to mark at least some traffic with the required qbss bits set.
Passive monitoringLes and Connie Logg have been working to better understand the netflow data details (e.g. flow tie offs, word wrap, zero port numbers), and to define new ways of looking at the data. Also communicated with Nevil Brownlee, the Netramet author, about how it handles fragments, and with Cisco to find out exactly how some of their MIB variables are defined.
IPv6Warren attended the IPv6 workshop at the Internet2 Meeting in Nebraska. Most of the workshop involved basic configuration of routers and setting up DNS, both of which are already in place at SLAC. The group may seek funding from NSF to deploy and develop IPv6 in Internet2 and it is hoped SLAC can be part of this.
ProposalsLes worked with the Active Internet measurement for Esnet proposal team to put together a response to a set of questions posed by Thomas Ndousse of DoE. See http://www.slac.stanford.edu/grp/scs/net/talk/infra-project/thomas-response.htm
TrainingLes put together a syllabus for 5 lectures on the Internet that are to be presented at the International Nathiagali Summer College in Pakistan in June. Followed this up by writing the lectures.
Distribution listRedid the distribution list for the monthly status reports following a request by someone to be removed. The list now has nearly 50 subscribers (mostly volunteers).
Regular readers of the status report who are not currently subscribed to the notification list can join by sending email to firstname.lastname@example.org with 'subscribe pinger-distribution' in the body.
BaBar TalkWarren gave a talk at the BaBar Data Distribution meeting
Internet2Warren attended the Internet2 Tech Meeting in Lincoln, Nebraska.
Revised 30 May, 2001.
Comments to email@example.com