Status Report for SLAC/DOE IEPM Project
For April & May 2000
Work completed or in progress since the last report.
PingERWarren installed new versions of timeping at the ANL, BNL, JLAB, Stanford and CERN2 monitoring sites. Warren worked with Connie Logg to adapt some of her LAN performance tools to begin gathering data for Wide-Area Application monitoring. Warren installed IRRD to provide AS numbers with traceroute. WE are working with security to get it enabled for public access.
Les updated the new Beacons list to take advantage of the new timeping directives, in order to reduce the load on some sites.
Java Analysis StudioKaushik made some final changes to his program before he departed and Warren spoke to Tony Johnson about putting it into production but some changes need to be made.
VoIPWarren attended the I2 VoIP working group meeting.
New sites and Site assistanceData is now available from the JLAB monitoring site. Installation of PingER is underway at CALTECH. The Stanford monitoring site has moved from Patabwa to Vicky.
Rate LimitingMit & Les are working with Stefan Savage of UW on some sting pathologies. Mit is comparing sting with ping for losses, and also to analyzing tail drop behavior. We are putting together some web pages on the findings, see ICMP Rate Limiting.
Reverse Traceroute ServerIn response to SLAC security requests, Les added several features to traceroute.pl to limit number of simultaneous processes, to limit information on the internal site, to send email to people if the number of simultaneous processes is exceeded, to throttle the email if the condition persists, to provide help if such a condition occurs, to monitor and produce reports on the usage to see if it being attacked (see http://www-iepm.slac.stanford.edu/about/webstats/traceroute.html for the reports). Reviewed the reports to satisfy ourselves that no denial of service attacks were happening. Extra functionality was also added to incorporate the NIKHEF traceroute so we can see ASs, also added ttls of returned packets, and provide more help for people who do not understand an IP address and enter things like email addresses or URLs or IP addresses with invalid characters in them. The new version is available via Traceroute Server for HENP and ESnet. The SLAC server is now receiving up to 1600 hits/day.
Internet TroubleshootingWe worked on pin-pointing and by-passing a problem between U Colorado and SLAC, and Stanford and SLAC, for more see see: Big vs. Small Packets and Stanford to SLAC file transfer problems. The latter is SLAC internal use only at the request of security since it contains information on the internal SLAC network architecture.
Got account at Colorado to help in future debugging and Les rewrote pingroute.pl in perl (was in REXX) and ported to Stanford campus, Colorado (Digital OpenVMS), AIX and Linux. Also customized traceroute server for Colorado.
Les spent sometime assisting and working with the UK networkers understanding some QoS problems they were seeing, see: UK to US QoS Testing.
Back to Top
XIWT/IPERFWarren and Les met with Chuck Brownstein of CNRI and the SLAC technology transfer folks to go over the SLAC sub-contract for the CNRI/DARPA proposal.
IEEEThe paper on The Pinger Project: Active Internet Performance Monitoring for the HENP Community by Warren Matthews and Warren and Les Cottrell was published in the IEEE Communications Magazine May 2000 Vol 38 No. 5 issue.
ESCCWarren and Les attended the ESCC meeting at Pleasanton California. Warren presented a review of IPv6 deployment at SLAC and Les presented an Update on the Internet Monitoring.
I2Warren presented a talk on Network Performance Monitoring related to HENP research to the NLANR/I2 Joint Techs meeting.
StanfordWarren and Les met with computer music people from campus to look at a novel way of enunciating Internet RTT, jitter and loss.
Revised 27 May, 2000.
Comments to firstname.lastname@example.org