A comparison of two view materialization approaches for disease surveillance system

The effectiveness of a disease surveillance system such as RODS depends heavily on the performance of the underlying data management components. Given that such system's core functionality is to support decision making and data mining, the response time of complex queries involving aggregation requires special attention. Traditionally, materialized views have been advocated to address this issue. The question that this paper examines is which approach implementing materialized views is more suitable in a disease surveillance environment. Our comparison focuses on the two most common approaches, namely, the Cache Tables approach and the Data Warehousing approach. Our evaluation shows that the choice of an approach is not limited by the cost-performance but by the cost-flexibility of usage as well.