Performance evaluation of concurrency control algorithms in distributed database systems
An important consideration in the design of database systems is the concurrency control, the database mechanism concerned with scheduling individual transactions in such a way that access into the database does not compromise its integrity. This concurrency control is exacerbated in distributed database environments for several reasons. The abundance of research papers dealing with this problem signifies its importance. This research takes a unique approach in that it examines the performance of the algorithms in typical network environments, giving quantitative in addition to qualitative results. Five database concurrency control algorithms, namely the Primary Copy (ALSB PA 76), the Majority Consensus (THOM RH 79), the Timestamp (ROTH JB 78), the Broadcast (ELLI CA 77), and the Logical Clock (KANE A 79), are evaluated in terms of response time, database convergence time, throughput, and channel utilization. These algorithms are evaluated in two network environments, a LAN and a RAN. Queueing theory is used as the primary tool in analyzing the performance of these algorithms.
- Research Organization:
- Polytechnic Inst. of Brooklyn, NY (USA)
- OSTI ID:
- 5255154
- Country of Publication:
- United States
- Language:
- English
Similar Records
Towards a general concurrency control algorithm for database systems
Consistent state detection and recovery for concurrent processing