Status Report for SLAC/DOE IEPM Project
For May & June 1998
Beacon SitesMost of the monitoring sites are now pinging the set of beacon sites that were defined previously. There has been some discussion on whether to alias pinged sites (perhaps as ping or ping0).
ContactsWe have finally re-established communications with Drew Baden at UMD, hopefully now we can get the UMD data and pinger tools (eg reverse traceroute server) set up. UMD is an important monitoring site due to its connectivity. We are also creating a list of contacts for all the monitoring sites and beacon sites.
SurveyorThe GPS Antenna has been delivered and fitted. The surveyor itself has also arrived and will be set up soon.
New SiteThe Niels Bohr Institute (NBI) has become a PingER Monitoring Site and are pinging the beacon sites.
Back to Top
Report GeneratorWe are developing new tables with hourly and daily ticks, in order for the statisticians at ORNL to analyze the trends in the data in more detail.
PingERA new release of the timeping code is currently in the final stages if testing and will be released soon. New features include a poisson distribution to the sampling rather than the half hourly sampling currently implemented.
We have also added the Inter-Quartile-Range (IQR) and the standard deviation to the values recorded. Network research indicates the load on the network is proportional to the spread of the round trip times, and we shall be looking to confirm this.
Network MapsWe are working on a new tool that uses the reverse traceroute servers at the monitoring sites, and logs the routes taken to a specific site. The program combines the individual routes and displays a network map which, combined with the ping data, may give an indication of network congestion points.
PingTableSeveral minor changes have been made to table.pl and pingtable.pl namely more options, better layout and documentation and bug fixes.
Passive MeasurementsWe have been working with the Internet Performance Working Group of the Cross-Industry Working Team (XIWT), investigating methods to determine network performance without probing the network. To this end we have been customizing iweb log files and comparing the apparent performace based on time taken to serve a web page with the metrics with gather from ping. Further we are looking at amending the web server to give a more accurate value for this time taken.
Revised 22 June 1998
Comments to firstname.lastname@example.org