[Infowarrior] - NSF picks ARPANET creator to build new experimental network

[Richard Forno]

NSF picks ARPANET creator to build new experimental network

By John Timmer | Published: May 28, 2007 - 07:09AM CT

http://arstechnica.com/news.ars/post/20070528-nsf-picks-arpanet-creator-to-b
uild-new-experimental-network.html

The National Science Foundation has tapped BBN Technologies to build and
manage the GENI project, intended as a proving ground for next-generation
network technologies. For BBN, this is a bit of a return to its roots: the
company was instrumental in the construction of the US ARPANET project,
which gradually morphed into the Internet that we all know and love.

GENI will be a Major Research and Equipment Facility Construction Project,
which is the NSF's mechanism for funding big-ticket items. It's notable as
being the first such project that the NSF has targeted to the computer
science community. Planning for GENI has been going on for a number of
years, as the NSF has solicited input from the research community regarding
their needs and commissioned a number of study groups to do preliminary
planning. BBN will now have a few years to create a formal design proposal
and cost estimate, after which the NSF will decide whether it's worth
constructing.

The project isn't expected to follow ARPANET's trajectory to widespread
adoption‹this is not a "next generation internet." Instead, it is intended
to allow computer scientists and networking engineers to test how well their
ideas work under real network conditions. Dedicated fiber connections will
be used to link a number of major academic institutions, which will host
hardware such as remote sensors, compute clusters, and experimental wireless
hardware. To facilitate such experimental uses, all of the hardware used in
GENI is expected to be programmable, allowing the system to be reconfigured
and updated according to users' needs and its successes and failures.

Preliminary documents show plans for dedicated fiber connections among
research institutions that looks suspiciously like the existing high-speed
academic network managed by Internet2/Lambda Rail. Those projects have
already dedicated part of its bandwidth to tests of networking technology.
Given the apparent overlap in function, it's surprising that none of the
GENI documents mention them at all. It's possible that the next few years'
worth of architectural work will make it clear whether these academic
networks will share resources with GENI or perform parallel functions.

Although the GENI network itself won't be merged with the general-use
Internet, any successful technology it produces is likely to make its way
onto the public network eventually. To speed this process along the NSF is
encouraging commercial entities to take part in the project.