Fragmentation and allocation of object-oriented databases for simple attributes and complex methods: a cost-based technique

The efficiency of the distributed systems is most often dominated by the network communications between the different sites. In distributed database systems, the communication between the different sites is almost essential to satisfy a query. In order to minimize the time required to respond to a query in such systems, a reliable fragmentation technique is required to place the data as close to its user as possible. The placement of these fragments to the proper sites is a complementary step to assure the right position of these fragments. The previous fragmentation techniques provided a way for fragmentation of the database that is followed by an allocation technique to allocate the produced fragments as separate steps. The key idea of this paper is that it introduces a novel technique that combines both the fragmentation and allocation processes into one composite process thus putting into consideration the allocation constraints while fragmenting the data. It also simplifies the effort required to get the best fragments and the best location of these fragments. This proposed technique relies on a cost-based model rather than an affinity model as being used in previous techniques. The case of simple methods and complex attributes database is considered in this paper. The results obtained by the technique proofed the effectiveness and usefulness of the proposed technique