Status Report for SLAC/DOE IEPM Project
For November & December, 1999
Work completed or in progress since the last report.
PingER DevelpomentWarren worked on new directives send= size= type= to increase the configurability of the pinger program. He also worked on implementing the ULM format for future analysis. HEPNRC was contacted about beta testing.
Yvonne amended the Guthrie node-details manager program to write out all the countries being monitoring so this does not have to be maintained manually anymore.
Warren worked on some problems with table.pl. However we are moving to using database connectivity much more and the program will be completely re-written soon.
Warren added a new metric to be used to detect candidates for rate-limiting
SynackKishan worked on amending his synack code to provide sub-millisecond timimg
JAS (Java Analysis Studio)Kaushik, Les and Warren worked with Paul Raines and Tony Johnson to improve the stability of the java servlets. We now have a rudimentary version that produces simple bar charts and is callable from pingtable.pl
IPv6Warren continued to work with the administrators of remote sites involved in the IPv6 monitoring to trouble shoot connectivity problems.
Warren talked to contacts in England and on Stanford Campus about IPv6 testing, there are not many takers, but a new IPv6 machine is being set up at STAR-TAP and we will be given an account to run more monitoring and test applications such as ssh, mail and web.
Warren met with some members of the network administration group and the security group to discuss deployment of IPv6 on site. Warren will install a firewall to prevent opening a back door into the SLAC network.
Smart Bit BoxRajiv and a netcom engineer got the Smart Bit box working, and we tested it by monitoring the load seen by MRTG.
Evaluating new tools
JetmonWarren installed jetmon from UOREGON. It may be useful for real-time demos, e.g. for SC2000.
CichlidWarren looked at NLANR's cichlid tool.
stingWarren installed sting, a tool that can be used to detect forward and backwards, rather than just round trip, packet loss. Les tried it out on several links and it looks quite useful. In one case it indicates a possible poorly implemented TCP stack host in Macedonia. Les is communicating with the Macedonians to understand in more detail.
Rate LimitingAfter gaining permission from the site Mit ran some synack test to NorduNet to look for differences in the ping loss/RTT with TCP values. Perhaps sites that are rate limiting ICMP can be identified this way. It appears that packet loss to Nordunet since Oct 22 is dramatically decreased.
Les contacted people in Argentina, Macedonia and Poland about possible rate limiting.
New sitesWe have another new monitoring site, this one at Monash University in Melbourne, Australia. Two people from Monash visited SLAC and we discussed the monitoring among other things with them.
We installed a second monitoring site at CERN, most likely the original site will be phased out, but we are currently studying the differences in performance reported by each site.
Successfully got an Israeli IEPM/PingER monitoring site and worked with them to get Beacons, answer questions, & get added to monitor site list. Discussed rate limiting tests with the Israelis, but it looks like there is no rate limiting between SLAC & ILAN at the moment.
Got a host in Chile to monitor from someone in Argentina. Added hosts in Kazakhstan, Uzbekistan, Indonesia, Malaysia, Peru, Thailand, Singapore, Uruguay and Viet Nam to the monitored sites.
Missing Data and PathologiesThe usual glitches required following up, in particular we have not been able to collect data from ANSP.BR (Brazil) for over a week. We also stopped monitoring some hosts in Singapore and Pakistan since they were exhibiting high unreachability.
Back to Top
IEPM Web siteAdded a search capability to some pages on the site. For an example, see the top right of this page that you are now reading.
CHEP 2000Our abstracts for the CHEP conference hae been accepted. One will be presented as a talk and the other as a poster. We had a voice conference with Becca Nitzan of ESnet to look at what we need to do next for the VoIP paper.
Paper for PAMThe abstract we submitted for the PAM conference has been accepted.
IEEE PaperLes is working with Tom Chen from SMU to put together a paper on network monitoring.
SeminarsLes and Warren attended a seminar by Walter Willinger discussing the self-similar nature of Internet traffic, and the importance of creating models that provide insight into the nature of the network.
Les gave a seminar at SLAC on Quality of Service in National Research and Education Networks. As part of this Les analyzed the difference in performance of Internet 2, ESnet and XIWT sites.
BookWarren was contacted by a professor from Columbia regarding some previous talkes presented by IEPM, it seems we may be included in his new book.
Warren was contacted by an author working on a new book about IPv6 to discuss experiences with implementing the software.
XIWTWorking with XIWT to host the next meeting (1.5 days in February/March 2000) at SLAC. Les and Warren attended an XIWT voice conference. SLAC has joined CNRI as a sub in submitting the abstract of a proposal to DARPA for Network Modeling and Simulation (NMS). is available.
HEPNRCWe had a phone meeting with HEPNRC and resolved why HEPNRC have not been collecting data from several recent monitoring sites.
NetPredictWorked with Bjorn of NetPredict to put together a 2nd proposal to DARPA. Studied how SLAC could be involved in such a proposal. Generated a lot of paperwork required by DoE and SLAC in order to submit the proposal.
NATOCommunicated with a small startup interested in setting up satellite links to to Central and East Europe and FSU republics where we would provide them with measurement capabilities. Cleared things with SLAC's Technology Transfer Office.
Revised 21 December 1999
Comments to email@example.com