Improving Performance of Object Oriented Databases, BaBar Case Studies

The BABAR experiment at SLAC is designed to perform high-precision studies of the decays of the B-meson, that are produced in e+e- collisions at the PEP-II accelerator. Since it was brought on-line in May 1999, it is continuously increasing luminosity approaching designed goal of recording 30 hadronic events per second, or 3*108 events per year. In order to promptly process and store persistently all generated events, a powerful processing software and hardware is essential. Current estimates of the number of computing nodes needed by first level reconstruction range from 200 to 350. Providing correctly configured hardware and software for physicists ’ analysis become equally important and even more complex task. The part of the system responsible for concurrency issues in this environment is a database component. The paper will describe performance studies, chosen hardware and software configuration of the underlying Objectivity/DB database system used by BABAR Project