Workload: 18 man/months
ProblemOur customer has developed Java integration for Palo OLAP Server. This server is a multidimensional in-memory OLAP Server. All the data it holds is organized in cubes, dimensions, elements and element attributes. The most rewarding aspect of the in-memory technology is the speed advantage. Compared to an old-style ROLAP system, in-memory technologies have the potential to be as much as 100 times faster than an OLAP system with a disk based relational database.
To browse data on Palo OLAP Server our customer decided to develop a Web-client.
We developed version 2.0 of Palo Web Client. Web Palo Client demo you can find here
The Palo Web Client provides browsing functionalities to view and edit Palo Server data. It allows to open cubes, change the dimensions’ assignment and change the selected element for the filter dimensions using the mouse. Nesting of dimensions and expanding/collapsing of dimension hierarchies is also supported. Cell values or consolidated values can be edited. It is possible to view multiple cubes simultaneously.
We developed the Palo Web Client using GWT technology. To browse data, user only needs a common browser. No additional software installation on the client computer is necessary.
Client application displays a table that represents a hypercube. In order to minimize data processing on the client site and reduce communication traffic, application downloads only actual data that is used to display hypercube. This functionality is implemented via AJAX capability of GWT. We implemented level-oriented loading of MOLAP-axes (in case of multiple elements, non-visible ones are not selected). Loading of data is optimized so that in case of minor changes in view (e.g. rotation of a single axis) only new data is selected. Besides, communication traffic is minimized when empty data is transmitted.