A model for determining the optimal topology and database allocation in computer networks

The authors propose a comprehensive model to be used in the design of distributed database networks. The designer has a number of decisions to make, including the allocation of databases to nodes, the location and capacity of communication lines between the nodes, and the amount of computational power required at each node. The designer must consider storage and maintenance costs of the databases, the expected traffic from each node to each database, and the desired time of a response to queries and updates to the databases. The proposed model considers these factors and, in addition, routing of the queries and updates; both one- and two-hop paths are included. The model accounts for the costs of the communication lines and processors, the storage and maintenance costs of databases, and the communication costs of queries and updates. The traffic from each node to each database is specified by the designer. Given a maximum desired response time to a query or update, the design goal becomes the minimization of the total operating cost of the network.<>