Client-side caching for client-server OLAP systems

Recently, client-side data caching has become popular for building highly interactive web applications. This research is about client-side data caching in a client-server OLAP system. Contrary to the traditional OLAP systems, the system we study here is client-centric which contains a light-weight OLAP engine and a data cache, such that queries posed by the user may be processed on the client without a round-trip to the server. It has been shown that this system works better than the traditional ones in some data visualization scenarios. In this study, we focus on the client-side data caching in this system. We implemented a new data structure for storing multidimensional data on the client-side, which occupies a small memory footprint and decreases the load time significantly with a minimal client-side processing overhead. Also, we investigate the efficiency of client-side caching when the processing of the query needs extra data that are not already client-resident. Our results show that in some cases it could be cost-effective to download more than %80 of the required data from the server, rather than to download the entire sub-cube from the server