A Strategy for Maintaining Client-Based Web Warehouse

Currently there is a common problem of real-time data inconsistency between source transactional systems and data warehouse. ECA, a well-known algorithm, is quite effective in addressing the view maintenance anomaly by making use of compensation queries. However, ECA does not maintain information about the data warehouse, and has to send compensating queries to the source system resulting in performance degradation. This paper proposes a new strategy for maintaining data warehouse in a Web environment without sending update queries to the source system. To maintain consistency in realtime, automatically installed web client program create modify queries for each update query to source system. Modify queries associated with each update query are processed at the web warehouse level without connecting to the source system. Hence the problem of a single modify query having to access multiple source systems disappears and the system performance improves. Through simulation, the performance of the proposed strategy is compared with previous solutions.