Re: Common Hardware Platform for D&P?

Paul, All;

I have been thinking along the same lines.  Many of the experiments that may want to use GENI will be focused on end equipment or end-to-end protocols, apps, etc,  and therefore not require the same granular flexibility that network-based experiments will require.  I suggest that there may be a partition (slice?) of GENI substrate that is standardized and fairly stable for this sort of experimentation. Concurrently, the ability to partition and slice at a very granular level some pool of resources independent from the ‘standardized’ platform is required.  I tend to think of this model in terms of virtual private networks (VPNs). 

To Paul’s point, ATCA is certainly one solution for unifying the substrate resources into a common chasis with benefits of common power supply, hardware interconnection bus (can be bypassed using port to port front panel connections) and a common low-level h/w management capability (resource configuration & inventory).  The openness of ATCA would seem to simplify the development and introduction of new H/W and S/W into the substrate, with the potential downside that the ATCA specification could be overly constraining to some H/W development not yet mature enough to fit into the necessary form factor.

-Stu

Please respond to this suggestion, plus, additionally, provide perspectives from other communities (e.g. wireless, networking, …) on how such a platform could be developed and used in this first cycle of D&P.  An important aspect of such a D&P proposal would be the true ‘cross layer’ R&D to be carried out on the hardware.