GENIOpticalWorkshop: event_OptWS_Menyuk_summary.txt

THE PROPOSAL:

NSF should fund a sizable procurement for cross-layer research that includes the following components:  (1) Involvement of applications, networking, and optical layers, in which the needs of the applications drive the optical and network layer developments.  (2) Interdisciplinary teams that involve application users, networking experts, and physical/optical layer experts.  (3) It should involve the development of real small-scale prototypes.  At the same time, it should involve simulations of both the network layer and physical layer behavior to both design the experiments and explore scalability to larger networks.

AND, VERY IMPORTANTLY:  Proposals should be jointly reviewed and the work should be jointly funded from both the CISE and ECCS directorates.


THE CONCERN:

If I were the director of NSF, I would ask the following question:  The telecom companies are in fact very concerned about the issues that David Clarke raised, and they are making progress towards resolving them.  David may not like their solutions, but there will be solutions.  This situation is very different from that of other large-scale NSF projects.  CIENA is not going to build a telescope in Peru or send ships to the Antarctic (at least not until global warming is much farther along).  They will be trying to solve the same issues that GENI is trying to solve.  So, why should NSF (and the American taxpayer) "pay for the cow when they can get the milk for free?"  Why should we compete with the telecommunications companies?

There is, I think, a very good answer to this objection, which Kristin Rauschenbach noted in her summary:  The goal is not to compete with the telecom companies, but rather to complement what they are doing and will be doing by focusing on research questions that are longer-term than the telecom industry in its current state is willing to tackle.  The goal is even-handed cooperation, not competition.  However, in order to be a serious partner, we (the academic community and NSF) must show that we are aware of the scope of the problems to be solved, which are cross-disciplinary and span the traditional research areas that are supported by CISE AND ECCS.  Our industrial colleagues are well aware that the bottlenecks for many applications are in determining what physical layer information is needed at the higher network layers to ensure quality of service and communicating that information to the upper layers (or, more properly, the management and control planes, which then communicate with the upper layers, and vice versa).  In brief, my opinion is that we --- and in particular NSF --- must get its act together and act cooperatively.  Otherwise, GENI will not succeed.