An adaptive load sensing priority assignment protocol for distributed real-time database systems

Transaction processing in a distributed real time database system (DRTDBS) is coordinated by a concurrency control protocol (CCP). The performance of a CCP is affected by the load condition of a transaction processing system. For example, the performance of the Adaptive Speculative Locking (ASL) protocol degrades in high load conditions of the system. Priority protocols help a CCP by prioritizing transactions. The performance of the priority protocols is also affected by system load conditions, but they can be optimized by dynamically switching between priority protocols at run time when the system load changes. The objective of this research is to develop a protocol, Adaptive Priority Assignment protocol (APAP), which changes the priority protocol at run time to improve the performance of a CCP in a DRTDBS.;APAP is implemented in a DRTDBS, where ASL is used as the underlying CCP to validate APAP. The performance of APAP was tested under varying system load conditions with various combinations of the database system parameters. Under the scenarios tested, APAP performed better than other priority protocols and demonstrated that dynamic selection of priority protocols during run time is an effective way to improve the performance of a CCP in a DRTDBS.