Showing papers on "Concurrency control published in 1899"

Approaches to concurrency control in distributed data base systems

Abstract: Whenever multiple users or programs access a data base concurrently, the problem of concurrency control arises The problem is to synchronize concurrent interactions so that each reads consistent data from the data base, writes consistent data, and is ultimately processed to completion In a distributed data base this problem is exacerbated because a concurrency control mechanism at one site cannot instantaneously know about interactions at other sites No fewer than 30 papers on this topic have appeared to date Our purpose is to survey this literature, concentrating on three approaches—locking, majority consensus, and SDD-1 protocols—which together subsume the bulk of the literature **

An analysis of the roll-back and blocking operations of three concurrency control mechanisms

Abstract: Transactions in database systems are run concurrently to achieve optimal resource utilization. Concurrent execution of transactions is managed by concurrency control mechanisms for maintaining the database consistency. These mechanisms use activities like transaction roll-backs and transaction blockings for serializing the concurrent execution, and they have significant effect on the performance of database systems; however, their relationship with throughput, workload, and other aspects of the system is unclear. Further it is not clear how the read: write ratio affects the performance. This paper attempts to show the effect of roll-back, blocking, and read: write ratio on the performance of database systems under several different types of workloads. We have used detailed and realistic simulation models to conduct our investigation; and, unlike other performance studies, we have avoided simplifying assumptions as far as possible to include most of the attributes of real database systems. In this study we show that neither a roll-back nor a blocking scheme is consistently better for all types of workloads; they are rather workload sensitive. We also show that it is not the write-only transactions but the read-only transactions that need special treatment for efficient processing. We report that transaction wait-time does not have significant effect on the throughput and the effect of read: write ratio is very short lived. We have introduced a new term Domain of Efficiency (DoE) to explain the behavior of these mechanisms.