How caching improves efficiency and result completeness for querying linked data

Abstract

Link traversal based query execution is a novel query approach which enables applications that exploit the Web of Data to its full potential. This approach makes use of the characteristics of Linked Data: During query execution it traverses data links to discover data that may contribute to query results. Once retrieved from the Web, the data can be cached and reused for subsequent queries. We expect such a reuse to be beneficial for two reasons: First, it may improve query performance because it reduces the need to retrieve data multiple times; second, it may provide for additional query results, calculated based on cached data that would not be discoverable by a link traversal based execution alone. However, no systematic analysis exist that justifies the application of caching strategies based on these assumptions. In this paper we evaluate the potential of caching to improve efficiency and result completeness in link traversal based query execution systems. We conceptually analyze the potential benefit of keeping and reusing retrieved data. Furthermore, we verify the theoretical impact of caching by conducting a comprehensive experiment that is based on a real-world application scenario.