As Web services grow more popular, they are more frequently adopted in mission critical applications. Query result caching is a representative technique for scaling up the throughput of database processing, which is a major bottleneck point in the Web sites. It is located in front of the DBMS and services queries with significantly less overhead by reusing previous query results. However, like many other caching systems, it is crucial that cache consistency and coherency problems must be dealt with efficiently in order to be useful. In this paper, we propose a novel cache consistency mechanism that has the following characteristics. First, it imposes little overhead to the database server. Thus it provides the scalability to the number of query results and the number of cache nodes in a cluster. Second, it is transparently deployed in that it does not require any significant modifications or administration efforts to existing systems, i.e., Web application servers and database servers. Third, it ensures strong cache consistency by immediately invalidating the query results affected by any updates. We present the experimental results that verify the effectiveness and scalability of the proposed mechanism.
