Status Report for SLAC/DOE IEPM Project
For November & December 2000
Work completed or in progress since the last report.
QoSThe project collaborators at Daresbury Lab created a mailing list. Warren installed the pingER software on loggy. loggy had some hardware problems that have been temporarily fixed by the Stanford administrators.
JNFlowThe JNFlow analysis code is broken, probably due to the upgrade of netflow to version 8. Warren is working with the developer to fix the problem.
NSWarren installed the ns (network simulator) package, and the modifications made by the developers for SLAC investigations. With help from George Riley at GA Tech started using the network simulator ns to try and understand the impact of packet reordering. So far Les has compared the simulation with observed data between SLAC and Daresbury Lab England, IN2P3, Lyon, France, LBNL, Berkeley, CA and Caltech, Pasadena, CA. and the agreement is encouraging, see http://www-iepm.slac.stanford.edu/monitoring/bulk/ns.html Made some simple modifications to the simulator and the configuration script. Wrote perl script to simplify running it. Started to put together a pre-proposal for validating the ns-2 simulator versus real bulk data throughput measurements.
IPv6Warren revamped the IPv6 network to fit into the new SLAC LAN design. Warren submitted a Field Work Proposal for further IPv6 deployment and development.
Last 120 daysWarren added a report for the last 120 days performance for all metrics. See the web page.
Performance mapsWarren installed the Cichlid package in order to automate the generation of performance maps.
Missing dataThis month there is a few days missing data from the UMD, ANL and Monash monitoring sites. Missing data from Transpac, Caltech, DoE/MICS, and University of Wisconsin was recovered.
NetPredictLes installed the new netpredict softwrae, and then worked with the NT folks to get SNMP installed on his workstation. Netpredict has some nice features which could be very useful for helping to tune/understand TCP performance.
TracerouteLes has been working with IDmap people at UMich to assist with mapping the Internet, by providing a traecroute server.
ComparisonsPut together a comparison of PingER, skping and skitter, see http://www.slac.stanford.edu/comp/net/wan-mon/iepm-cf.html and http://www-iepm.slac.stanford.edu/about/faq.txt. Put together a web page comparing passive and active Internet monitoring, see http://www.slac.stanford.edu/comp/net/wan-mon/passive-vs-active.html.
Choice of remote hostsLes looked at beacon hosts not responding to PingER to see if pings are blocked, then working to get alternative hosts. Found 2 such hosts out of 75 (since Jun00). Also found 5 other hosts out of about 250 hosts monitored from SLAC that stopped responding to pings in the last 6 months. Starting to update the Beacon list. Also removed several non Beacon sites that are not responding and sent email to contacts to try and get alternate hosts.
For the 248 hosts monitored by ping from SLAC, Les analyzed the fractions that responded to 3 common ports (80=web, 25=SMTP & 21=FTP control). 115 did not respond to port 80, 107 did not respond to port 25 and 124 did not respond to port 21. In addition, 77 did not respond to any of the 3 ports, 56 responded to only one port, 56 to only two ports and 50 to all 3 ports. Reviewed site being monitored in Venezuela and corresponded with person there to understand performance.
HEPAt the request of Charley Young of BaBar, Les looked at connectivity between SLAC and BINP, Novosibirsk, see http://www.slac.stanford.edu/grp/scs/net/case/nsk/. In response to request from Walter Toki put together a web page on network performance to IN2P3, see http://www.slac.stanford.edu/grp/scs/net/case/in2p3/. In response to a question from Joe Izen of UT Dallas, Les looked at the throughput between SLAC and UT Dallas and documented the findings, see http://www.slac.stanford.edu/grp/scs/net/case/utdallas/.
DoESeveral email exchanges with DoE/MICS folks on projects past & future. Communicated with Vern Paxson on possible collaboration. Talked to Deb Agarwal of LBL, Nagy Rao & Bill Wing of ORNL, and Matt Mathis of PSC. Talked to the PI of the Network Weather Service. Talked to DoE folks about the solicitation for network monitoring proposals. Read the paperwork from DoE for the solicitation etc.
New HiresOffers to for a full-time and a part-time post has been sent out. We hope to have the new hires on board in January.
PAM2001The abstract submitted for the PAM2001 conference in April 2001 has been accepted.
XIWTLes attended XIWT/IPEX meeting at HP. See http://www.slac.stanford.edu/grp/scs/trip/xiwt-nov00.html. More work on the IPEX proposal with Bob Kahn of CNRI and SLAC's tech transfer office to get the CRADA into place.
IUPAPLes put together some slides for a talk for Burt Richter to give to the IUPAP concerning the performance of networks to developing countries where there are physicists. Did some research on world population and % people on line. Also put together a web page on Internet NREN performance at the end of 2000, see http://www-iepm.slac.stanford.edu/monitoring/end-2000/.
Computer MusicWe met with the campus folks. Then worked with Tony Johnson to put together a Java applet for doing audio pings with real time graphs running on my laptop. Looking for suitable hosts to ping (i.e. interesting ones with high loss, RTT or variability).
Back to Top
Revised 28 December, 2000.
Comments to email@example.com