Abstract
Search engines and large scale IR systems need to cache query results for efficiency and scalability purposes. In this study, we propose to explicitly incorporate the query costs in the static caching policy. To this end, a query's cost is represented by its execution time, which involves CPU time to decompress the postings and compute the query-document similarities to obtain the final top-N answers. Simulation results using a large Web crawl data and a real query log reveal that the proposed strategy improves overall system performance in terms of the total query execution time. © Springer-Verlag Berlin Heidelberg 2009.
Cite
CITATION STYLE
Altingovde, I. S., Ozcan, R., & Ulusoy, Ö. (2009). A cost-aware strategy for query result caching in web search engines. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5478 LNCS, pp. 628–636). https://doi.org/10.1007/978-3-642-00958-7_59
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
