Semantic Query Processing in the VENUS Environment

Abstract

Recent years have seen the publication of several methodologies for query processing in a Multidatabase (MDB) environment. Whilst different approaches are pertinent to specific CIS architectures, the issue of efficiency of processing has received relatively little attention. This is somewhat surprising considering the potentially large costs associated with MDB queries. In this paper, we consider the efficiency of processing queries in the VENUS (Esprit 6398) environment. The goal of this system is to allow a user to query multiple data sources by interacting with a conceptually single database, and by providing a query interface based on the use of visual formalisms. This system takes advantage of existing multidatabase query processing strategies, and deals with optimization issues by considering the exploitation of inter- and intra-schema semantics. We present a methodology which harnesses knowledge of the set-theoretic relationships between classes participating in a query, and demonstrate that significant saving can be made in MDB query processing. The methodology presented here is based on a powerful data model, the Graph Model, which we argue is suitable as the canonical model of a MDB. We also introduce a logic language to formally express interdependencies between classes belonging to different schemas. These assertions form the "backbone" of the MDB and, we may benefit in several ways from the possibility of reasoning about them, specifically by applying them in MDB query processing.