Resolving data conflicts with multiple versions and precedence relationships in real-time databases

Abstract

We propose a multiversion concurrency control algorithm for real-time databases. Based on a priority cognitive algorithm commonly used in real-time processing, the proposed algorithm resolves data conflicts among transactions in favor of the urgent transactions, i.e., those with higher priority. Our algorithm increases the degree of transaction concurrency while minimizing the transactions that miss their deadlines. This is achieved by using precedence relationships and multiple versions. A precedence relationship, which implies a serialization order among conflicting transactions, is utilized to resolve data conflicts and avoid deadlocks, The algorithm dynamically determines the abort of lower priority transactions depending on the likelihood of roll-backs and their states of execution. This leads to saving system resources by reducing unnecessary aborts of the lower priority transactions. We present some experimental results of our algorithm, comparing other realtime concurrency control algorithms. It is briefly shown that the proposed algorithm has superior performance to several algorithms for real-time databases. (C) 1997 Elsevier Science B.V.