Using SiteRank for Decentralized Computation of Web Document Ranking

The PageRank algorithm demonstrates the significance of the computation of document ranking of general importance or authority in Web information retrieval. However, doing a PageRank computation for the whole Web graph is both time-consuming and costly. State of the art Web crawler based search engines also suffer from the latency in retrieving a complete Web graph for the computation of PageRank. We look into the problem of computing PageRank in a decentralized and timely fashion by making use of SiteRank and aggregating rankings from multiple sites. A SiteRank is basically the ranking generated by applying the classical PageRank algorithm to the graph of Web sites, i.e., the Web graph at the granularity of Web sites instead of Web pages. Our empirical results show that SiteRank also follows a power-law distribution. Our experimental results demonstrate that the decomposition of global Web document ranking computation by making use of SiteRank is a very promising approach for computing global document rankings in a decentralized Web search system. In particular, by sharing SiteRank among member servers, such a search system also obtains a new means to fight link spamming.