Department of Computer Engineering
S E M I N A R
Prediction of Stale Queries
Computer Engineering Department
We propose a new mechanism to predict stale queries in the result cache of a search engine. The novelty of our approach is in the use of timestamp values in making staleness predictions. We assume an incrementally updated index to which all modifications (addition, deletion, and update of documents) are continuously reflected by a document parser. Our approach does not involve blind decisions such as the TTL-based invalidation approach. We show that our approach incurs very little overhead on the system while its prediction accuracy is comparable to earlier works.
DATE: 06 December, 2010, Monday @ 17:00