ABSTRACT

Various concurrency control algorithms have been proposed for use in distributed database systems. But, the number of algorithms available for the distributed concurrency control, come into one of three basic classes: locking algorithms, Timestamp algorithms and optimistic (or certification) algorithms. In this paper we are presenting a Distributed Transaction Processing Model and an approach for concurrency control in distributed database systems. The analysis of our approach is a decomposition of the concurrency control problem into two major sub-problems: read-write and write-write synchronization. We describe a series of synchronization techniques for solving each sub-problem and will show how to combine these techniques into algorithms for solving the entire concurrency control problem. Such algorithms are called "concurrency control methods". Our approach concentrates on the structure and correctness of concurrency control methods and also the performance of such methods up to some extent.