[e2e] Research testbeds: Report to NSF

[Bob Braden]

Folks,

This mailing list is intended to be primarily devoted to discussion
of network research topics.  This note falls into that category.

One of the enduring issues for network researchers is the need for
effective testbeds to conduct experimental work.  Historically, the
end-to-end Research Group played a significant role in designing and
coordinating the DARPA Research Network Testbed (DARTnet), which
resulted in the development of IP multicast, video teleconferencing,
and Internet packet scheduling (and in particular int-serv).  We are
convinced that there is a continuing need in this area and that this
need is not currently being filled.  There are testbeds that primarily
provide production-quality high-speed connectivity for developing
advanced distributed applications or grid computing, and there are
testbeds that are primarily for testing of advanced physical-layer
transmission technologies.  But we also need testbeds for conducting
fundamental research on network- and transport-layer issues, for
example.

Recently the US NSF (National Science Foundation) commissioned a small
group of academic network researchers to study the issues in network
research testbeds and to make recommendations to them.  The resulting
report is available at http://gaia.cs.umass.edu/testbed_workshop.
Some of you may find it interesting.

Bob Braden