A Web-Based Database Management System Supporting Parallel Data Mining Service on PC Clusters

In this study, we develop a database management system which provides users with on-line web services to build and manage their databases and perform parallel data mining on Internet. We exploit two different programming toolkits, i.e., MPI and DSM to parallelize association rules on PC clusters in order to minimize response time. In addition, we have evaluated the performance of these two programming toolkits on parallelization of data association mining. Our experimental results show that both of MPI and DSM are effective for parallel data association mining while MPI provides a 30~45% performance improvement than DSM.