Atsuyuki Morishima(Institute of Information Sciences and Electronics, University of Tsukuba)
Hiroyuki Kitagawa(Institute of Information Sciences and Electronics, University of Tsukuba)
Integration of heterogeneous information sources has been one of the most important issues in recent advanced application environments. We are developing an information integration environment for the World Wide Web (or shortly, Web), relational databases, and structured document repositories. In this environment, manipulation of the information sources is performed through software modules called wrappers. In this paper, we present design and implementation of the Web wrapper. In general, Web page manipulation may cause very large data transfer cost if all the necessary pages are transferred to the Web wrapper. To alleviate this problem, we propose a query processing scheme which uses distributed agents, named navigators. The navigators cooperatively take part of the wrapper's functions at Web server sites, to reduce the cost of Web page transfer. This paper includes experimental results on data transfer costs in this scheme.