Performance Analysis of Three Database Server Distribution Algorithms

Although the concept of the distributed database has been around for over 20 years, it has not dominated the computer landscape especially in business-related applications. This paper will explore the effectiveness of distributed database under a variety of conditions by conducting experiments using a number of different combinations of variables listed below. Specifically, the following questions will be researched: How does the workload intensity influence the need and performance of distributed database applications? How does the number of nodes the database is stored upon affect the data access time? How does the method used to assign a given query to a specific database node influence the access time? The first variable is workload intensity. It is expected as intensity increases the need to utilize some form of distributed database increases. The second factor is number of nodes upon which the database is distributed. One would expect that as the number of nodes increases, access time would be reduced. The third variable is the algorithm used to distribute the inquiries across multiple nodes. A symmetric algorithm, one that provides an equal chance of any given inquiry landing on any specific node, would be expected to offer the most promise. However, results indicate the load balanced method outperforms both the sequential and random selection methods