Transparent data access in a multivendor, distributed data-acquisition and data processing system

Location‐independent data access methods are being developed at Lawrence Livermore National Laboratory to manage and use data acquired by a multivendor distributed laboratory computer network. An interprocess communication system (IPCS) is used to connect tasks which produce, manage, or use experimental data. These tasks may run on any of the computers in the network. A central database and database server maintain supervisory records of experimental activity and locations of data produced by the experiment. Data may be located in several subsidiary local databases, file systems, or tape archives. Data access requests, if they cannot be satisfied locally, are routed via the IPCS to database servers, which locate and retrieve (or store) the requested data. We describe the design details and give examples of the current implementation.