Web caching proxy servers are essential for improving web performance and scalability, and recent research has focused on
making proxy caching work for database-backed web sites. In this paper, we explore a new proxy caching framework that exploits
the query semantics of HTML forms. We identify two common classes of form-based queries from real-world database-backed web
sites, namely,
keyword-based queries and
function-embedded queries. Using typical examples of these queries, we study two representative caching schemes within our framework: (i) traditional
passive query caching, and (ii)
active query caching, in which the proxy cache can service a request by evaluating a query over the contents of the cache. Results from our experimental
implementation show that our form-based proxy is a general and flexible approach that efficiently enables active caching schemes
for database-backed web sites. Furthermore, handling query containment at the proxy yields significant performance advantages
over passive query caching, but extending the power of the active cache to do full semantic caching appears to be less generally
effective.
Keywords Web proxy caching - Database-backed web sites