Current hot items are
- IEPM - Bandwidth to the World project.
This is a new proposal to evolve from the PingER lightweight network
measurement activities into more intense network and application
monitoring for links that utilize high
performance networks (e.g. ESnet, Internet2 and Academic &
Research network) and applications such as Grid enabled bulk data replication.
- SC 2001 Bansdwidth Challenge This year the IEPM group
will be participating the the
Challenge. The goal will be to achieve high throughput to over 20
sites worldwide from the SLAC/FNAL booth at SC2001 in Denver in November 2001.
At the same time we hope to explore the impact of the QBone Scavenger Service
(QBSS) on traffic management on a high speed congested link. For more on this
Bandwidth to the World proposal.
- Bulk Throughput. High performance transport
of data over the network is critical to the success of HENP and other
fiels' applications such as file replication for data grids. As part
of the Particle Physics Data Grid efforts and for the BaBar experiment,
we have been conducting measurements of both network (TCP using iperf)
and application (file transfer using bbcp) throughputs from SLAC to
various sites. This is to understand what is achievable, how to go about
achieving it (large windows, multiple parallel streams etc.), what the impact
is on other users of the network or the hosts etc. For more on this start at:
Bulk throughput measurements.
- QBSS. Differentiated Services (DiffServ) are cited in
terms of increasing performance by labelling packets with "expedited
forwarding (forward me first)" and "assured forwarding (drop me last)".
However, IEPM is leading the way in making use of the QBone Scavenger
Service (qbss) being developed by the Qbone group in Internet2.
This form of DiffServ is used to provide less than best effort.
Huge data transfer will be trickled across the network and will back down
whenever other applications need the bandwidth.
Go to the qbsss webpage.
- WEB100. The modern high capacity networks used
by research groups have raised expectations amongst scientist. Often
these expectations are not reached. One of the reasons for failure
is the TCP/IP stack on the end-node that manages the networking for
the applications is not tuned to make full use of the resources. Work
at SLAC and elsewhere has attempted to understand how to tune the
parameters to improved performance. Now the IEPM group is working
with the web100 project to
understand this fully and optimise TCP performance.
Go to the web100 webpage.
- HENP Networking Group. IEPM is a the first
monitoring group to become involved with the HENP Networking
Group, within the applications area of Internet2. This group
will develop new understanding of what is required for
the HENP community and work closely with the Internet2 End-to-end
- Grid Forum. IEPM has become involved in
monitoring for the grid and is a founder member of the Global
Grid Forum Network Performance Working Group. Activities are
under way to implement the Grid Monitoring Architecture (GMA)
and integrate pingER monitoring with other tools to create
a monitoring framework for the particle physics data grid
Other achievements are detailed in the monthly report, see
Recent IEPM Status
Reports and Meeting Notes.
For a more detailed site map of the iepm webpages, see the
Revised September 30, 2001
Comments to email@example.com