Report on individual activity to the PPDG for the quarter ending September, 2001
 
Les Cottrell, SLAC

Project: Network throughput performance

We have extended measurements of bulk network throughput between SLAC 
and major BaBar and collaborator sites using iperf, to include PPDG and 
major HENP sites. We are now monitoring these sites on a regular basis by iperf
for 10 seconds each hour. These measurements are reported in 
http://www.slac.stanford.edu/comp/net/iperftests-html/International-Iperf-Tests.html. 
As part of understanding how best to make these measurements we have
also made spot measurements to understand how to achieve high performance 
and the impact of measurement duration, large windows and multiple parallel streams 
(see http://www-iepm.slac.stanford.edu/monitoring/bulk/window-vs-streams.html). 
From this we can set realistic expectations of what is achievable and how to go about 
achieving it. Our current work involves making the automated measurements
more robust, improving resistance to denial of service attacks, and
providing better reports and better ways to view them.

We have extended the measurements to the application level, 
and are using the SLAC written bbcp file copy program to make measurements
of file transfer performance between sites. We have worked with the author of bbcp
)Andy hanushevsky) to extensively test it, identify bugs and define new
features. The goals of this are to 
see what extra constraints are imposed by the application on top of the
network layer (e.g. security, disk access etc.), and how close one can 
get to network (iperf) performance. The early results are available at 
http://www-iepm.slac.stanford.edu/monitoring/bulk/bbcp.html. 
We presented details of this work at CHEP01 in Beijing (see 
http://www.slac.stanford.edu/grp/scs/net/talk/chep01-throughput_files/frame.htm).

We are looking into how to use QBone Scavenger Service (QBSS) 
to reduce the impact of the high throughput on other
users. We reported on this at the Virtual Internet 2 meeting (see
http://www.slac.stanford.edu/grp/scs/net/talk/qbss-i2-oct01_files/frame.htm).
As part of this we have set up 2 QBSS testbeds. More details can be found at
http://www-iepm.slac.stanford.edu/monitoring/qbss/measure.html where we
show how well QBSS manages traffic and how well interactive traffic
works in the presence of heavy QBSS marked bulk throughput traffic. 

We have also put together a proposal for the SC2001 Bandwidth Challenge.
See http://www-iepm.slac.stanford.edu/monitoring/bulk/sc2001/ for
more details. This proposal has been accepted. 
It includes over 20 collaborating sites (including all the PPDG sites) to which
we will be sending large amounts of bulk throughput from the SLAC/FNAL
booth at SC2001. We also hope to demonstrate the effectiveness of QBSS
for very high speed links (2Gbits/sec). Following SC2001 we expect to use
some of the infrastructure (accounts, privacy keys, installed software)
put together for this demonstration in order to provide long term monitoring
of bulk throughput between PPDG and some other key sites.