Make it fresh, make it quick: searching a network of personal webservers

Mayank Bawa and Roberto J. Bayardo, Jr. and Sridhar Rajagopalan and Eugene J. Shekita
WWW '03: Proceedings of the 12th international conference on World Wide Web

Overview

YouSearch? is distributed search application for searching personal webservers. Every peer runs periodically an indexing process. This process starts searching for new and modified files since the last run. Subsequently, a number of keywords are obtained for these files. With the bloom filter the peer can be tested for documents that contain certain keywords. These bloom filters are then passed on to a centralized component, the registrar. The registrar aggregates all the bloom filters it receives from peers.

A search is performed as follows. The peer performing a search passes the query on to the registrar. Using the bloom filters the registrars determines which peers have content matching the request. The registrar sends the peerlist to the requesting peer. The peer contacts all the peers and request the URLs for matching documents.