|
Status Report for SLAC/DOE IEPM Project
For June & July 1999
|
|
Work related to Collaborations
Visitors
Les met with Hungbin Hu from Wuhan university to go over the SLAC LAN,
monitoring and the IEPM effort.
HENP
Les attended an ICFA Standing Committee on Internet Connectivity (SCIC) video
meeting and made presentation on Status of SCIC Working Group
on Monitoring.
Followed up by talking to Jim Leighton of ESnet about the KEK ESnet
link, and providing information on how the .de<=>.ca performance
compares with.de<=>.edu performance, & how the German Russian connectivity is
performing. Had email discussion with Michael Ernst concerning DFN
and ICMP rate limiting.
Les sent email to the chairman of the ESSC suggesting a talk on ESnet monitoring
at the Santa Fe meeting at the end of August.
HEPNRC
Les arranged a phone meeting with the FNAL/HEPNRC folks to
re-involve them in the PingER
project. Had voice meeting with Bill Lidinsky, Michael, and Shiqi Hi
from HEPNRC and Warren to
get everyone up to speed on projects, to parse them out in terms of priority
and who is interested and decide on some actions. As part of preparation
for the meeting, Les put together some rough notes from the
ICFA/SCIC meeting
earlier in the month.
Les also put together some
rough notes of the
meeting with HEPNRC.
Warren has sent several emails to the new HEPNRC employee (Michael)
regarding the present
state and future plans for the data acquisition code. Warren has sent the
most recent
poisson code hoping he will take over it's development.
XIWT
Reviewed an early draft of a white paper from the XIWT. Started preparing
a web page providing a comparison between PingER/Surveyor/AMP/RIPE/skitter.
Corresponded with some of the project leaders in order to check information.
See
Comparison of some Internet Active End-to-end Performance Measurement
projects.
RIPE
Corresponded with RIPE and Waikato folks to compare their projects with
other active
active end-to-end performance measuring projects.
The machine for the RIPE-IDM monitoring arrived and has
been installed. We are waiting for the GPS aerial to be set up before the
machine is fully operational.
Surveyor
Warren obtained and processed the June Surveyor data.
NGI Testbed
Les attended the NGI Testbed workshop in Berkeley. As part of the
preparation he set up a PingER web page for the
Particle Physics Data
Grid project,
and one for
the
NGI testbeds.
Les also put together and presented a talk on the
IEPM/PingER project and
PPDG.
Other sites
Warren worked with Larry Gerhardstein and Cullen Tollbom to get PingER
installed at PNL. The software is installed and operational, but there
is not currently a hole in their firewall through which SLAC can retrieve
the data. If they make it available we will process it.
Tools & Analysis
PingER and Ping Validation
Rebecca and Les worked on extending the help for PingER. Following a
suggestion from Mike Sullenberger of Cisco, Les
worked with Rebecca to find out what the maximum segment size
is between SLAC and the beacon sites. Got confirmation from Vern Paxson.
Warren has been working with Boris Bertucci (a visiting French student) to
use tcpdump to passively measure and understand the SLAC DMZ traffic.
Warren installed a modified version of Pinger from Jeremy Hylton
(CNRI/XIWT). We
now store all the results from the ping probes, so we can do further analysis.
The pinger configuration details are now stored in an Oracle database, this
is the first step in converting much of the data and analysis work to
java and perl database connectivity.
As part of the database move, our student Yvonne Gbalazeh has written a
managemnt program tha twe have dubbed "Guthrie" to manage the database
access. The latest code has been made available on
our
ftp site.
Warren also wrote a perl program that updates the list of
remote
sites,
using the details stored in the database.
Tringer
Warren installed a program from Jeremy that logs traceroute results.
The data is made
available
but needs processing!
Sites & groups
Les worked with Rebecca to come up with a list of all the AMP/Pinger/Surveyor
and RIPE sites and see how they match up with each other and with various
HENP (Atlas, BaBar, Belle, CMS, D0, Zeus)/ESnet/DOE funded/Internet 2/vBNS
groupings (e.g. how many of the BaBar sites have Surveyor or AMPs) etc.
Warren also wrote a perl program that updates the list of
remote
sites,
using the details stored in the database.
VoIP
We are trying to use a modified TTCP to add load to the line,
but it wouldn't make a
connection. We used TCPDUMP to discover the code was creating packets with
the wrong IP address. Kishan found the part of the code that needed
modifying and we returned the changes to the author at NASA. We are now
testing with NASA, and Kishan is further developing the code to use UDP
packets for our tests on the test link.
Network Anomalies
There were some odd problems with several remote sites. Bizarre looping
effects were noticed when tracerouting to a set controlled by IBM and at JHU.
We suspect some sort of shaping is going on. Also Olemiss appears to have
implemented smurf filtering. Unfortunately all Warren's enquiries
attempting to get
confimration of his suspicions haven't been answered.
Missing data
We lost a significant amount of data from the RAL monitoring site.
No solution was found but a new monitoring site has been established
at RAL and we are retrieving data from that.
Related to that. The SLAC webserver receives a significant pounding when
pingER data is being called. I worked with systems to set up enough anonymous
ftp space for the data and moved it from the webserver area. All the data is
still available, but should impact SLAC resources less harshly.
IPv6
We have received a 7200 router from ESnet and Warren worked with Tony
Zele to upgrade
the IOS on a 2600 to the IPv6 aware version.
Students
Interviewed 3 RA student candidates to work on IEPM. Met with Bjorn Frogner
about students and discussed their (NetPredict's) work. Made offers to 2 of the
students.
Work Related to Publicity and Information
Papers
We have been invited to submit paper to IEEE communications magazine
on network traffic measurement and experiments. Warren has begun to put a
paper together. Also Warren made a first stab at an abstract for CHEP.
Documentation
Updated traceroute server web page to point to some archives of traceroutes.
Added network weather service and timeit to the
Interesting web sites for Internet Monitoring page.
Miscellaneous
By comparing SLAC phone CDR records with router utilization,
came up with an estimate
of the ratio of voice traffic between offsite & SLAC with IP traffic
between offsite & SLAC the ratio varies from 7:1 to 22:1 (data is higher).
Back to Top
Revised 24 July 1999
URL:
http://www-iepm.slac.stanford.edu/about/status/status-07-99.html
Comments to
iepm-l@slac.stanford.edu