A middleware system which intelligently caches query results

This paper describes how caching was used to improve performance in the Accessible Business Rules framework (ABR) for IBM's Websphere. ABR is a middleware system which enables application writers to build applications where the time and situation-variable parts of their business logic are externally applied entities known as business rules. The cache significantly reduced the number of queries to remote databases by storing query results. A key problem we faced was how to keep the cache current after database updates. This was solved using data update propagation (DUP). Two enhancements we made to DUP were to employ an update strategy which considers the values of database updates in order to perform intelligent cache invalidations and to automatically compute dependencies using compile and run-time analysis. Our techniques can be applied to other caching environments besides ABR. We show how our cache invalidation strategies perform for applications with database updates having queries similar to those in the Set Query benchmark.