A Three Level Search Engine Index Based in Query Log Distribution

Queries to a search engine follow a power-law distribution, which is far from uniform. Hence, it is natural to adapt a search engine index to the query distribution. In this paper we present a three level memory organization for a search engine inverted file index that includes main and secondary memory, as well as precomputed answers, such that the use of main memory and the answer time are significantly improved. We include experimental results as well as an analytical model.