Managing Cache Consistency to Scale Dynamic Web Systems

Abstract

Data caching is a technique that can be used by web servers to speed up the response time of client requests. Dynamic websites are becoming more popular, but they pose a problem –- it is difficult to cache dynamic content, as each user may receive a different version of a webpage. Caching fragments of content in a distributed way solves this problem, but poses a maintainability challenge: cached fragments may depend on other cached fragments, or on underlying information in a database. When the underlying information is updated, care must be taken to ensure cached information is also invalidated. If new code is added that updates the database, the cache can very easily become inconsistent with the underlying data. The deploy-time dependency analysis method solves this maintainability problem by analyzing web application source code at deploy-time, and statically writing cache dependency information into the deployed application. This allows for the significant performance gains distributed object caching can allow, without any of the maintainability problems that such caching creates.