Distributed Computing and Databases for High Energy Physics

We are proposing to develop a fault-tolerant distributed computing and database system for use in High Energy Physics, but applicable to other disciplines. This system will allow us to solve the challenging problem of data processing, analysis and simulation by exploiting developments in distributed systems, fault tolerance, high performance networks and database technology. We will extend the data storage paradigm so that data are directly accessible by eld as well as by event. The system will be scalable to at least hundreds of processors. We will incorporate Asynchronous Transfer Mode network technology and develop the system for wide-area use, particularly over the National Research and Education Network.