Status Report for SLAC/DOE IEPM Project
For June 2002
Work completed or in progress since the last report.
IEPM-BWWork continued bringing up the IEPM-BW target at FermiLab. Warren worked on understanding the FermiLab Kerberos environment and setting it up at SLAC. The kerberized ssh and the kerberized tool for running cron jobs work. Warren has requested help from unix-admin to properly configure these tools for the SLAC environment.
Warren worked on plugging GridFTP and UDPMON into the IEPM-BW framework.
Warren continued to work with Yee-Ting Li (UCL) and Richard Hughes-Jones (University of Manchester) to understand the European Datagrid (EDG) extensions.
Richard Hughes-Jones visited SLAC for 1.5 weeks. We shared information on IEPM-BW and UDPmon. We then installed the IEPM-BW monitoring code at Manchester and documented the procedures. Follwing this Connie & Jerrod worked on automating the installation in prepration for installing at FNAL. As part of the porting Les & Connie rewrote many scripts to parameterize the site/host dependencies and to clean up/improve their functions.
Based on our experiences with netFlow, Les submitted a couple of suggestions to the IETF ipfix working group.
MAGGIEThere was an encouraging response to the initial drafts of the Maggie ( Measurement and Analysis for the Global Grid and InternetEnd-to-end Performance ) proposal. The initial deadline has been postponed and the proposal will be submitted in August for consideration for FY03.
Warren and Les attended a conference call with Shawn Mckee (UMich) and Eric Boyd, Andrew Adamson and Matt Zekauskas (all Internet2) to review the Atlas and Internet2 monitoring proposal and how they meshed and with the MAGGIE proposal, in particular what areas were complimentary. Warren followed up with Russ Hobby (Internet2) and the end-to-end performance initiative and Bob Grossman (UIC) about his distributed data storage work.
IPEXWarren is working on an interface to the IPEX passive monitoring data. He is configuring the system to authenticate users using X509 certificates and is waiting for the Telcordia firewall to permit https traffic. Hung Mei worked on re-analyzing the IPEX passive data following a disk crash at Telcordia.
GlobusWarren is working on putting PingER and IEPM-BW data into MDS.
The GridFTP monitoring has been extended to include the University of Michigan. Warren is woking with the Grid team at KEK to bring up a gatekeeper that will trust DoE certificates.
IPv6Warren has received several more enquiries about IPv6 activities at SLAC.
PingERWarren debugged a problem preventing the traceroute server at ANL from functioning.
Les updated connectivity.pl and mon-lib.pl to incoporate modifications from the Eu DataGrid and NASA. the mods provided for an extra color (blue) and a new application type (vnd-ms-excel). ,p> Les reviewed a paper form Italy where they had analyzed PingER data.
Disk performanceAjay has been analyzing read & write disk performance at the remote sites. This includes the effects of caching, the file system etc. See Disk Throughputs
Web100Heng Mei is starting to look at the Web100 data with a view to seeing what can be usefully used with IEPM-BW. Les discussed using Web100 with Heng and Ajay. Ajay is working on adding web100 & IPv6 to the formal distribution for iperf, and also using the web100 to provide less network intensive estimates of throughput.
Missing Data and remote Site problemsWarren worked with Jerrod to retrieve occassional glitches of missing data.
Les responded to questions on port utilization from U Florida; worked with NASA folks following a disk crash, to get ssh working again, the main problems were due to making ssh from OpenSSH and SSH.com interwork. It required reformatting of the OpenSSH key and installing part of OpenSSH at NASA (where they are using SSH.com).
MiscThe usual paperwork, installations and administration. Les removed the icfa-scic-ntf mailing list that was mainly beings used for Spam. He also moved some of the members of icfa-scic-ntf to the pinger-dstribition email list (mainly used for enailing this report).
Warren worked with a user having problems accessing SLAC and IEPM-BW data.
Warren received email from the University of Plymouth (UK), about installing PingER. Warren worked with an administrator at a company to install and configure the tools.
Les initiated a long email discussion on jumbo frames. It would be good to measure the effect of jumbo frames, however, it is hard to set things up.
Travel & TalksWarren arranged his trip to the Global Grid Forum (GGF). He will give a talk at the Network Monitoring Working Group (NMWG) meeting.
Warren will also attend the Internet2/Web100/ESCC meetings in Boulder on the way back from GGF.
Jiri and Les attended and each gave a talk at the CAIDA organized Bandwidth Estimation Workshop in San Diego:
- An experience,results and remarks to BW estimation and Cross Traffic measurement tools talk given by Jiri Navratil at the CAIDA Bandwith Estimation workshop,June 5-7,2002.
- Results from a new infrastructure for application and throughput performance measurement talk presented by Les Cottrell at the CAIDA banwdith Estimation workshop, SDSC June 6 '02.
Working GroupsLes attended 4 GGF NMWG phone meetings, wrote the section on delay, and reviewed and commented on the rest of the paper. This will be presented at the GGF in Edinburgh in July.
Les organized the logistics (conference room, VRVS, drinks and donuts) for the ICFA/SCIC meeting at SLAC, Jun 1 '02.
Les travelled to FNAL and participated in the Sc2002 face to face meeting. One of the SLAC/FNAL booth panels will present network monitoring and took Rough meeting notes. He also attended 2 voice meetings on the topic. ,p> While at FNAL les met with the FNAL PingER folks, see the Rough meeting notes
Les worked with Tony Johnson to get a real-time Java applet display of ping response times to various regions of the world. See Ping World demo.
Les has been working with Cees Delaat and Athony from NIKHEF to put in a proposal for a demo at iGrid2002 in Amsterdam in September. It was accepted.
DoELes prepared and sent the annual IEPM report to DoE. See http://www-iepm.slac.stanford.edu/about/status/iepm-yr-may02.doc
Les reviewed 3 proposals from young networkers to DoE and submitted the reviews.
Les worked with LANL and Rice to put together the first annual SciDAC INCITE report to doE.
Created June 24, 2002, last updated June 25, 2002.
Comments to firstname.lastname@example.org