Distributed top-k full-text content dissemination

The World Wide Web (WWW) is experiencing an explosive growth of live content. To timely disseminate end users with fresh content that is relevant to users' interests is a very useful but challenging task. In this paper we design a distributed top-k document dissemination scheme: by inputting query terms and a number k as subscription conditions, end users subscribe to the top-k most relevant documents. When the documents and subscription conditions are distributed, the difficulties are how to forward the personalized top-k documents to the needed users with low forwarding cost. To this end, we propose document forwarding and filtering algorithms on a set of dedicate servers. Experiments based on real query logs and document datasets show the efficiency of our proposed algorithms.