[e2e] Research testbeds: Report to NSF

[Karen R. Sollins]

The NSF does not send grant money to PIs in other countries, but they 
do encourage international collaboration in several ways.  First, 
they simply encourage it in their US PIs.  Second, the NSF has a 
separate International program, which provides some funds to American 
PIs who happen to have collaborators in other countries, specifically 
to help with the collaboration, for example with travel money for 
visits (in both directions); this is often done by a supplement to a 
grant from a regular program, made by the International Program. 
Third, a number of countries have the equivalent of the NSF and the 
NSF has agreements with some of these organizations, such that 
collaborators in the US and the other countries can dually submit the 
same proposal to both organizations, in order to apply for "joint" 
support of such a project.  These proposals go through the review 
process appropriate to each of the countries.  Proposals in both of 
the latter two categories are handled somewhat separately from normal 
NSF proposals, both for reviewing and funding.  The NSF's 
International program has a separate pool of money to help support 
them and participates in the reviewing process, so from the point of 
view of a program officer in a particular field (say networking), all 
this money is a supplement to normal programs.  To make things like 
this work, it is often important first to talk with the appropriate 
program directors at the NSF, especially if they are unaware of these 
sorts of programs, so that they can both learn and extend the 
effectiveness of their funds.

The other thing the NSF has done in the past (I don't know how long 
this will continue) is to run the Startap (& friends) and often 
provide small amounts of money to other countries to help support 
international connections into the US.  My understanding (I wasn't 
really involved at all in this) was that this was generally in 
support of big science projects (e.g. various flavors of physics).  I 
never saw any networking research here, but this was money that 
actually did go outside the borders of the US.

			Karen

At 9:30 AM +0000 1/14/03, Jon Crowcroft wrote:
>superb report - of course i agree with all of it except that it could
>be a bit more global/international (i udnerstand that the NSF Is a US
>agency but) - for example, the successes of the ARPANET, DARTNET and
>other testbeds included international collaboration (AND
>infrastructure) as does the Internet II program (even broader, though
>its more an application testbed outside the US, not strictly in the
>spirit of this report - it IS however a network of networks!)
>
>the recommednation on unlicensed testbed spectrum is especially
>welcome but last time i checked,
>the FCC doesnt have any jurisdiction outside the US where there's more
>wireless (last time i looked) but more daft regulation, so it might be
>nice if someone could find the right heads to bang together in the EU
>(and i guess asia/pacific, though as far as i can see, with docomo and
>other stugff, they have their act together there already) and i guess
>south america (scuse my compelte lack of knowledge of wireless there -
>i assume they use satellites a lot due to demographics and distance!)
>
>so great stuff, but how do we up the ante?
>
>In message <200301102250.WAA02640 at gra.isi.edu>, Bob Braden typed:
>
>  >>
>  >>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
>  >>
>  >>
>  >>
>  >>
>  >>
>
>  cheers
>
>    jon