Efficient cache answerability for XPath queries

The problem of cache answerability has traditionally been studied over conjunctive queries performed on top of a relational database system. However, with the proliferation of semistructured data and, in particular, of XML as the de facto standard for information interchange on the Internet, most of the assumptions and methods used for traditional systems - and cache answerability is no exception - need to be revisited from the point of view of the semistructured data and query model. In this paper, we present a formal framework for the efficient processing of XPath queries over XML documents in a cache environment that is based on the classic rewriting approach. Furthermore, we provide details on the implementation of our formal methods on top of HLCACHES, an LDAP-based distributed caching system for XML, and argue that our approach is more efficient than traditional query rewriting algorithms while, at the same time, supporting the full expressive power of XPath queries.