Mobile Based OLAP Using Parallel Processing and Multithreading

This work is committed to discovering the influence of parallel mechanism on I/O heavy applications such as OLAP operations on remote data set running on a resource constrained client. In this research, we have chosen an Android based smartphone as our client. We have implemented a mobile phone based client-centric OLAP application utilizing the potentiality of parallel processing mechanism of Android. We are dividing user request and associated caching tasks into smaller units and distributing them to multiple threads to execute concurrently. While serving user's requested query data, we are downloading and caching (hash based pre-emptive caching) extra sub-cube data in the background by employing multiple threads running in parallel. This yields better performance for succeeding cross-tab or drill-down actions. We have captured test results by varying number of threads and amount of data size associated with each thread. Result shows that parallel approach has striking performance improvement over sequential approach.