|
|
|
|
|
PHASE 0: Install, test and operate a replica of the Objectivity/DB implementation from an existing CERN RD45/DEC TurboLaser project. The CPU power of the Caltech facility is considerably greater than that used in the RD45/DEC Joint Project. This allows us to make scaling tests in a much larger environment. It will also establish the vendor independence of the implementation.
PHASE 1: Install the CMS software environment including the large CMSIM simulation program and the new OO-based programs. Run production simulations to effectively utilize the available spare CPU cycles on the machine. Test the system with the new programs. This allows us to demonstrate that large scale systems such as this can be used as general purpose facilities for High Energy Physics, as well as efficient Object Database engines.
PHASE 2: Marry the existing, operational HPSS at Caltech to the Objectivity/DB database instance and evaluate the support for the various caching levels of the ODBMS. The marrying of the two systems allows us to gain direct experience with multiple HPSS clients on a very large SMP-architecture computer system.
PHASE 3: Evaluate the prototype LHC Computing Centre. This Phase follows naturally from success in the previous phases. The full details of this Phase will be defined during the Review of Phase 2 in the light of the available capacity and bandwidth, the status of the RD45 Project, and support from CERN. Since the Caltech facility will be used in networking projects such as Internet II and next generation Internet, it is expected to have excellent high bandwidth WAN connectivity by this time. The full system will become a model for a Regional Centre, running LHC application software that accesses the federated OO database. We should investigate how well this model works in terms of caching Objects locally and across the WAN. This is of crucial interest in scientific research because it tests the model of truly distributed data and access in the OO paradigm. One of the main questions will be: is the network reliable and fast enough?
|
|
|
|
|
|
|