An Expert System Controlling an Adaptable Distributed Data Base System

The task of concurrency control is to synchronize the user transactions on this data base and with this to avoid conflicts that lead to faulty or inconsistent data. Concurrency control has been subject to active research for quite some time and several concurrency control algorithms have been proposed and implemented. The relative performance of these methods varies depending on the system environment (work load, user performance, reliability requirements etc.). No method showed a superior performance in all or a majority of applications. A adaptable concurrency control mechanism is needed that provides the optimal behavior in terms of response time and throughput in all or at least many possible applications. This mechanism is able to change the concurrency control algorithm depending on the system environment. The target of the research that is described in this paper is the selection of the concurrency control algorithm that is most suitable for achieving a high performance in the given system environment. An expert system is proposed for this function that operates on-line, real-time and without human assistance. A prototype software has been implemented in PROLOG.