A client on a client-centric OLAP system is capable of performing OLAP operations on data downloaded from the OLAP server. A previous research study has shown that this concept can work very well for some common data visualization scenarios. As a proof-of-concept, a research prototype was built where the client was a web-browser on a desktop machine. In this research, our focus has shifted to mobile devices, such as a tablet, which is more resource-constrained. Our objective is to find out the pros and cons of a web-based client vis-a-vis an app-based client, i.e., one that runs on the native operating system. Two research prototypes have been built, one for each type of client, to run in a client-centric OLAP system. We compare their architectures, software implementation aspects and performance on 'real life' data. We pay special attention on how software running on mobile devices handles unusually high volume of data. We also compare the performance of app-based clients running in a client-centric and a server-centric system respectively.
展开▼