030.50.3 The view from the Gigabit networking World by David J. Farber <farber@central.cis.upenn.edu> In subsequent issues of the ISOC Newsletter, this column will cover specific research issues and accomplishments of both the NSF/DARPA/CNRI Gigabit Testbeds Initiatives and other gigabit network related research activities. This first contribution is timely since it is concurrent with the passage of the US HPCC bill.
The US testbeds, Aurora, Blance, CASA, Nector, and VISTANET are an ongoing activity now in the second year of a three year initial program. It aims are to provide a set of experimental testbed of different characteristics with the intent of allowing the research community to have the opportunity create and explore working prototype networks. This would in collaboration with other research activities
provide feedback to the hardware and software designers as to what the long term technical requirements would be for a future gigabit NREN.
This will provide an insight into the hardware/software alternatives that must be faced in the creation of the Gigabit NREN as envisioned in the HPCC bill. It is not the intention of the testbeds to have a bakeoff of technology. . We will have a handful of major switch designs, some ideas of what we have to do with protocols and operating systems -- for example the National Backplane ideas, the application level protocols, the speed up of TCP/IP, ATM, PTM etc. These approaches will not be examined in order to "pick one" but rather to get a better handle on the fundamental issues. At the end of the initiative we will be where the Arpanet was back in the mid 70's -- a working set of demos, some protocol ideas and lots of enthusiasm.
And finally and most importantly, I believe we will have a good enough feel for the benefits of gigabit networks to the science community, that we will be able to answer the question we should be asking -- is such a network a reasonable use of limited science resources?
