Optimistic Concurrency Control in Firm Real-Time Databases

Concurrency control algorithms for real-time database systems satisfy not only consistency requirements but also meet transaction-timing constraints. Two Phase Locking (2PL) is used often in traditional database systems. However, it has some inherent problems such as the possibility of deadlocks as well as long and unpredictable blocking times. Optimistic concurrency control protocols are non-blocking and deadlock free, but they have the problems of late conflict detection and transaction restarts. Other Concurrency Control techniques, such as Dynamic Adjustment of Serialization Order (DASO) have been found to be better at reducing number of transaction restarts. In this paper, we propose a new optimistic concurrency control algorithm based on DASO using firm deadline in order to effectively reduce number of unnecessary restarts. Since firm real time transaction imparts no value to the system once its deadline expires, therefore in our algorithm, we adjust the timestamp intervals of all conflicting active transactions only after the validating transaction is guaranteed to meet its deadline during the validation phase. A simulator is designed to verify the effectiveness of the proposed method. The simulation results show that the proposed method can significantly reduce the number of unnecessary restarts and thereby improve the miss ratio, commit ratio.