Status Report for SLAC/DOE IEPM Project
For October 2002
Work completed or in progress since the last report.
MAGGIEWarren attended a conference call with Vern Paxson and Andrew Adams regarding the progress of the MAGGIE proposal. Warren worked with Teri Church to complete a budget projection for the project. Funding for IEPM/PingER has been extended. The MAGGIE proposal will most likely be submitted next year.
GlobusWarren worked with SLAC Computer Services people to reconfigure the Globus MDS system to serve network performance data. Globus does not lend itself to sharing files, so Warren brought up a instance of the toolkit seperate from the SCS and BaBar installations.
IEPM-BWWarren discussed setting up IEPM-BW target sites with North Carolina State University. Connie reworked the extraction/analysis parts of the IEPM-BW toolkit. She also compared the data from the multiple monitoring sites, see 60 Days Comparisons for Monitoring Hosts to provide input to the iGrid2002 paper. The IEPM-BW measurement toolkit is now deployed at 9 measurement sites: SLAC, FNAL, U Michigan, APAN Japan, NIKHEF Netherlands, Internet2, Manchester England, University College England and INFN Milan Italy. Currently most of the sites are monitoring only one or 2 hosts, while they get experience and evaluate the toolkit. The exceptions are SLAC (of course) monitoring about 37 remote hosts, NIKHEF monitoring 35 remote hosts, INFN/Milan in production monitoring 4 remote hosts and APAN is in production monitoring 3 hosts. FNAL is also moving into production to provide monitoring for CDF, D0, SDSS, Atlas and CMS. Jerrod finished off the scatterplot comparisons of iperf/bbcp and bbftp (see Correlation plots.
Web100Warren worked with Connie to install the latest web100 release on Antonia. Warren also implemented Sylvain Ravot's modifications. Warren installed the net100 (web100 with extra features) patches on Hercules. Les successfully followed up by testing the net100 GUI utility (gutil). Les is also trying the netw00 HSTCP (Sally Floyd) modifications in Net100. There appear to be some problems with the latest version of iperf (June 2002 1.6.3) and net100 (iperf seems to stop transmitting after a few tens of seconds), but the exact symptoms/configuration need further evaluation
SC2002Warren built 8 machines for SC2002. He installed Red Hat Linux Gigabit card. He configured NFS, NTP, IPCHAINS and web100.
While in Europe, Les had discussions with Paul Fernes of Level 3, Harvey Newman of Caltech, Olivier Martin of CERN and Linda Winkler of StarLight. Out of this came a plan to get an OC192/POS connection from the Level 3 gateway in Sunnyvale (near SLAC) to StarLight, connect it up at StarLight at 10Gbits/s to SC2002 at Baltimore and CERN at 2.5 Gbits/s. Level 3 will also provide colocation space at Sunnyvale. Caltech will provide servers (disk and cpu) to go at Sunnyvale. there will also be servers at SC022, CERN & StarLight. Later Cisco came on as a partner to loan a GSR OC192/POS capable router to go at Sunnyvale. This has evolved into the "Extreme bandwidth": SC2002 Bandwidth Challenge.
We are also working on setting up a QBSS demo for SC2002, and a high performance monitoring tools demo, see SC2002: Bandwidth to the World).
Jiri, updated his avaiable bandwidth estimation (packet pair) monitoring tool for Sc2002. Heng, Connie & Les defined and created a Java tool for monitoring the switch port utilizations at SC2002. This will be used with the QBSS demo.
AdminWarren worked with the administrator at JLAB to upgrade and reinstall PingER on the JLAB monitoring machine.
Warren worked with unix-admin to upgrade up the IPERF Server that was broken because the SLAC supported iperf binary was not available for the platform.
Digital DivideLes gave a presentation on Monitoring Internet Connectivity to Research and Educational Institutions at the ICTP meeting in Trieste on "Developing nations access to digital publishing meeting." Following the meeting there is considerable interest in extendinPingER monitoring to more developing nations, in particular Africa. Les is working with the people at ICTP/Trieste to get relevant sites/hosts and contacts for such nations.
GGFAs part of the GGF NMWG working group, Les worked on the GGF recommendation A Hierarchy of Network Measurements for Grid Applications and Services. Later in the month he attended the GGF6 in Chicago, see the trip report.
Internet2Warren attended the Internet2 members meeting in L.A. He spoke at the PIPEFitters meeting and the PIPES BOF. He also had many productive conversations with participants.
HEPLes visted Padova and CERN to share information and ideas with collaborators there. For more see the trip report. While in Trieste, Les learnt that the Italian Nuclear & HE Physics Newtork (INFN) has recently awarded a grant to the INFN/Trieste group to procure, configure and deploy about 14 embedded Linux boxes running the SLAC developed IEPM-BW high performance monitoring toolkit at INFN sites. Six of the boxes will have GE interfaces and be 1 unit high, the remaining will have three 100Mbit/s interfaces and be 1 unit high but 0.5 unit wide. For reliability there will be no hard disk rather they will use flash card memory. The data will be uploaded from the measurement hosts to an analysis/archive host. They currently have 4 hosts running, 2 are embedded and 2 general purpose PCs. Two are at Trieste and 2 at Milan. Les pointed out to Roberto Percacci (the PI) and Fabrizio Coccetti that he had contacts at the Padova, and Rome INFN sites where it could be very advantageous to place the measurement engines. As a result of this the Trieste and Padova people have made contact and this looks to be a fruitful collaboration.
MeetingsResearchers from LBL visited SLAC and the IEPM group to discuss work and collaboration. We are also working with LBL and the SLAC security folks to get approval to install a "Self Configuring Network Monitor" at SLAC.
Warren attended a meeting about some aspect of the European Data Grid (EDG) extensions to Globus and how web services might be used.
iGrid2002Les with help from Connie and Jiri wrote and submitted a paper to iGrid2002 recounting experiences and work done.
IPEXWarren prepared a few slides for the IPEX co-ordinators to present at the PI meeting.
Warren worked with Les, Chuck and accounts to prepare a report on spending and activities related to the IPEX
NetPredict SBIRLes sent off to DoE, the final report on the NetPredict SBIR.
SupportWarren responded to an enquiry about using tcpdump.
Warren responded to an enquiry about the NIKHEF ping.
Created November 1, 2002. Last updated November 5, 2002.
Comments to firstname.lastname@example.org