DOMAIN-BASED DISTRIBUTED MEDIATION SYSTEM FOR LARGE-SCALE DATA INTEGRATION

ABSTRACT. The increasing use of computers and the development of communication infrastructures resulted in abundance of information on the networks and extremely need for querying and integrating data from large number of data sources, especially for scientific applications. Distributed mediation system performs this task fragmentally, where each mediator in the system architecture responsible for querying subset of data sources. To make such type of systems success in large scale, we must deal with two issues: logical distribution of the data sources and mediators in the system architecture, and logical interaction between the mediators. To handle these issues, this paper proposes a domain based distributed mediation system called Multi-Mediators System for Large scale Data Integration (MMSLDI). MMSLDI enables the reducing of the general query response time by eliminating unnecessary visits to the data sources that do not contribute to the answer of the query. Moreover, it satisfies the requirements of adaptation to various kinds of domains, decentralized control, and automation of its processes