To enhance the performance in processing the database operations, the main memory database system(MMDS) has been introduced and studied by many researchers. However, parallel execution of one database operation primitive has not been studied in muti-processor main memory resident database system environment. In this thesis, query processing in main memory resident database is studied for general multi-processor system with large shared memory.
First, a parallel processing strategy for the queries is proposed to decrease the average response time. Next, an efficient parallel processing algorithm for join operations is proposed. Finally, the performance of parallel hash based equijoin algorithms is evaluated for the multi-processor MMDS.
