The instrumentation of a parallel, distributed database operation, retrieve-common, for merging two large sets of records

The Naval Postgraduate School's Laboratory for Database Systems Research is the site of the multi-backend database supercomputer (MBDS). Originally, MBDS supported a prototype primary operation (retrieve-common) which merged two sets of records in a distributed, parallel database environment. This thesis presents the testing, and modification of that prototyped primary operation. First, the design rationale of the MBDS is reviewed. Specifically, this review examines the reasons for a database-oriented supercomputer, the MBDS primary processes, and the methodology of distributing a database within loosely coupled and highly parallel database stores. Then, this study explains the methodology involved in developing theories on the cause of retrieve-common's defects and bottlenecks. Finally, in validating our theories, this study relates the process of discovering and correcting these discrepancies.