Work related to Collaborations
The XIWT/IPERF WG met in Reston, VA, and Warren & Les attended some of it via
conference call. We discussed rate limiting, self pinging and also decided to
collect data on all 10 pings.
The group also had a phone conference call. The NIST folks have been developing some analysis code
using the S-plus package. We hope to get a trial version of S-plus to see what they've done and
decide if it's worth getting a full licence and using what they've started.
Monitoring Site in France
Warren emailed Rene Bilhaut about a PingER monitoring site in France. On his
suggestion contacted Gilles Farrach of IN2P3 in Lyon. Initially they seemed concerned
about BaBar connectivity
but have since lost interest and say there is no point adding more ICMP traffic to the French
network as CERN already monitors most of the BaBar collaborators in France and the French regional
center at IN2P3 in Lyon is routed to from SLAC via a 6Mbps link from CERN.
Les corresponded with Hans Werner Braun of NLANR and we jointly agreed to join the
AMP project by placing an AMP machine at SLAC.
The machine has now arrived at SLAC, has an IP address and name
(halimede) and is being installed. Arranged to attend a
NLANR sponsored meeting
on network monitoring at SDSC at the end of June and make a presentation.
Warren talked to Tracie Monk about installing a CAIDA/Skitter machine at SLAC.
Worked with RIPE-IDM project co-ordinator to set up their machine at SLAC.
Problem with set up, only works with up to 40m cable for the GPS antenna.
Our wireless person says he
can make something work. Has requested specs on Antenna because intrigued
by length limitation.
The NIMI machine was moved on to the video subnet.
Warren learned how to reconfigure networking
in FreeBSD and NIMI is back up.
Corresponded with Matt Zekauskas of Surveyor to get some surveyor data to enable a
more careful check on the correlation of Surveyor and ping data.
ESSC University Committee
Les briefed the ESSC university committee on the state of PingER and possible
futures. In response to action items from the meeting we created a graph of Gbytes/hour
technically possible between ESnet sites and other groups of sites that we
monitor. We also worked with and met with Walter Toki (head of BaBar's documentation committee) and
M. I. Williams of BaBar/Queen Mary College to create a page with
network monitoring for the BaBar collaboration.
Chris Witzig of RHIC to see if he is interested in a PingER RHIC monitoring
page similar to the BaBar page above.
Assistance for DoE
Reviewed SBIR (DoE Small Business Innovation Research) proposal
on real-time network monitoring and intrusion detection.
Reviewed 7 NGI network infrastructure proposals and attended a meeting in Washington
to review the Networking NGI proposals.
Tools & Analysis
Warren talked with Robin Tasker about the U.K. upgrade, and he helped
clarify networking in the U.K. for the BaBar paper we are working on. Warren also
tweaked a subroutine in mon-lib.pl to order nodes based
on domain rather than the fully qualified node name
following request from Robin. Warren looked at web response time using the unix
time function and found it rose and fell in a similar manner to the rises
and drops in ping round trip time seen by pingER. However, pingER observes
high packet loss, which would make the link unusable if it were true for
all types of packets. We considered whether smurf filters had been installed, but Robin
could not confirm it.
Warren contacted Pat Pomatto of HEPNRC about amending HEPNRC web pages. The
changes have been made.
Warren worked with BNL to find out why reverse traceroute server doesn't work.
They are upgrading their systems, and things are still flakey. We also requested
Colorado State to
set up a reverse traceroute server and sent information to Walter Toki.
Warren created two new pinger groups; Abilene and FUSION-CIT (the latter requested by
Marty Greenwald of MIT/ESSC) and a
page for CIT group.
We have conducted more beta tests of the Poisson Ping, and Warren installed
and tested the amendments from XIWT.
We asked the HEP/PingER contacts to install reverse ping. Most of them have.
Bug fix median value of monthly medians.
The web server had some problems which meant all the reportting didn't run.
Warren had to manually make the reports for previous day.
Some failures of gathering and processing data, had to do it manually.
Trouble shoot a problem with pinging jlabs. They had a security issue and Warren
worked with the folks there to monitor a node that wasn't effected by the
Worked with John MacAllister of Oxford to get traceping at SLAC to
monitor all the beacon sites, ran into a problem with insufficient resources to
run the program. Ordered more memory for the machine. Documented
the Current Interest in
Ping data analysis
high statistics ping data. Wrote programs to assist in the reduction and analysis
of the data. Found some interesting behavior
(regions in round trip time
which have very different behavior), researched and documented results in High statistics ping results. Also
ran into an anomaly observed when PCs ping PCs and documented findings in
PingER Measurement Pathologies.
reported this anomaly to the IETF/IPPM mailing list.
We hired a student to work on NIKHEF ping with TOS and look at the effect of TOS
over the test VoIP network. We ordered a machine (Epaphus) and
arranged to have it configured.
The student is currently part-time. Warren worked with him to provide an introduction
to SLAC, and he has started working on perl and C.
We also have a volunteer. Warren has been working with her to bring her up to speed
with networking and hope she will do some analysis looking for evidence of
rate limiting. In addition we have interviewed and hired another summer student,
and are looking for another to work on the web site. Finally we have a
student visitor returning
from France for the summer to work on monitoring the border network at SLAC.
Test VoIP Network.
Set up ssh account at esnet, set up pinger to ping nereus.
lots of pinging. set up sudo, set up NIKHEF ping on doris.
the NIKHEF ping has resolution to 1 µsec compared to 1 msec for the regular ping and
also allows one to set the TOS bits. Discovered, reported and got
a fix from the author for a bug in the latest NIKHEF ping.
Set up ssh account with RSA authentication for Becca.
Obtained IPv4 network details for IPv6 machine, Pontus. Worked on setting up
IPv6 under Linux/RedHat. Calculate IPv6 address based on RFC1897.
Set up aliases for non-PingER Mon work machines, and created
for WAN-MON nodes (Inside SLAC firewall only).
Work Related to Publicity and Information
Worked on paper for BaBar. Graphs etc.
Co-ord with XIWT/IPERF chair regarding printing of brochure.
Talk to techpubs and other experts, reviewed use of digital camera and scanner.
Decided to postpone brochure due to restriction imposed.
New web site
Talked to Bebo about setting up a virtual web server WWW-IEPM.
Requested and eventually got AFS space for new files.
Current Web site
Les created pages on PingER deployment,
chart of Gbytes/hour to the Throughput
vs. Loss page in response to a request from Walter Toki of the ESSC (see above).
Back to Top
Revised 26 May 1999