Data allocation design in computer networks: LAN versus MAN versus WAN

Abstract

Using off-the-shelf distributed database management systems, distributed database developments have significantly proliferated. The data allocation design is an essential factor affecting the efficiency and effectiveness of a disturbed database in meeting geographically dispersed database processing demands. Data allocations in local area networks (LANs), metropolitan area networks (MANs), and wide area networks (WANs) are subject to different requirements and considerations. In particular, the primary incentives for distributed database implementations in a LAN are load balancing and distributed computing, while the development of distributed databases in a WAN is mainly motivated by the notion of locality of reference. MANs benefit from both balanced nodal processing loads and maximized local access. Furthermore, the performance levels and cost structures in LANs, MANs and WANs are different. Moreover, LANs, MANs and WANs exhibit distinct network control and medium contention schemes. This paper analyzes the unique considerations and requirements for data allocation in each type of network and further examines the impact of network environments on the structures and the effectiveness of optimal data allocation models. © 1992 J.C. Baltzer A.G. Scientific Publishing Company.