Approach to Event Store Decision
6/97 STAR Computing Plan and Model Working Group:
- ODBMS and, specifically, Objectivity offers major potential benefits in terms of
- cost, given that manpower is by far the greatest component of the cost
- realizability of an effective solution by startup
- performance and capability to provide an efficient STAR data processing implementation
- compatibility with the wider STAR software development effort.
- We recommend that Objectivity be very quickly evaluated by a representative community of RHIC computing interests (RCF, the Grand Challenge, representatives of the experiments)
- Decision has to be RHIC-wide: hardware/OS impact, scale of effort, timescale
- Should it be adopted, its implementation should be aggressively pursued with an effort to leverage the work of BaBar, RD45 and other groups developing Objectivity-based event stores
7/97 RHIC Computing review recommendations:
- Decide rapidly on an event store
- Do not consider product cost in evaluating Objectivity; its price is only 2-3 FTE-years
- If Objectivity is chosen, deploy rapidly.
8/97 RHIC Event Store Task Force is convened
- Assemble requirements, evaluation criteria, advocated event store alternatives
- Measure event store alternatives against requirements and criteria
- Alternatives with advocates were Objectivity and ROOT
- Decide!