Status Report for SLAC/DOE IEPM Project
For November & December 1997
Warren spent two days in San Diego visiting NLANR/CAIDA discussing various projects and getting an overview of the work they do there.
Met with the chair of the ESSC to brief him on current monitoring activities and concerns with connectivity to the University of Cincinnati.
Communicated with TRIUMF networkers to understand their new traceroute tools. Added ARM to the WAN monitoring table of sites running PingER. Worked with person from GTE to help them install the PingER tools.
Sent Email to the ESnet/HENP collection sites encouraging them to run MRTG (a router monitoring tool) and to make the information available for their external links. Charley Granieri installed the MRTG tools at SLAC. This would facilitate getting an indication of how much traffic each site is exchanging with offsite.
Minor changes to link to the new SAS plot reports.
Using the new SAS more uniform tabular format we were able to remove a large section of code that was previously needed to accommodate the varying formats.
Grouping the data
The grouping has been integrated with the HEPNRC graph_ping code. This allows users to select predefined groups of sites (e.g. the set of sites involved in a large HEP experiment) to be plotted. The next step is to install this at HEPNRC.
After talking with the author of the original MapNet Warren has rewritten some of the code, and sketched out how to code the remaining issues. The latest version at http://www.slac.stanford.edu/~warrenm/work/java/newjava/mapping.html supports coloring of links & sites by various metric measurements. The only collection site incorporated so far is SLAC.
Long Term Analysis
Warren has written a Perl script to provide for each month a statistical summary of each month's data. The summary includes mean, median, percentiles, inter-quartile distance, MTBF, for each site.
Warren also extracted the by-month data from es.net to show-by-site data in tab separated values format suitable for viewing in Excel. Unfortunately as we learnt by discussing with Chin Guok and Jim Leighton of ESnet, the ESnet data has many problems over the last year due to the resignation of a key person.
Warren has started work on visualizing route information following the work at TRIUMF & NLANR A rough version is running, bu the code is not stable.
Warren has been looking at improved ways to distribute the code. We have located a suitable package and spent some time evaluating. It would require significant effort to incorporate but would give our tools a very professional looking setup. It is unclear whether the effort would justify the limited distribution that we expect.Back to Top
Revised 22 December 1997
Comments to email@example.com