Abstract
Background: Mobile distributed real time database is an emerging field of research. Transaction execute in these system should commit before elapsing the deadline. Assigning priority to arrived and executing transaction is the main function of transaction scheduler. Perform correct order of execution enhance the system performance in time constraint environment. In recent years a number of heuristic policies in DRTDBS are developed to enhance the quality estimation of scheduler. However due to various wireless limitations a very little work is investigated in the MDRTDBS. In this paper we have develop two heuristic policies and perform simulation using pessimistic and optimistic concurrency control algorithm to enhance the system performance.
Objective: Two heuristic policies based on write only locks and number of same data item existence has been introduced. These heuristic approaches enhance the performance of MDRTDBS scheduler.
Methods: Develop two heuristic policies based on data contention which cope with CC policies like DHP-2PL and FBOCC. These policies count the number of write only locks and existence of number of same data item. Ultimately these policies enhance the system performance.
Results: Investigated the performance using simulator which is designed using C language. Based on these simulation results using pessimistic and optimistic CCs different graphs are plotted which shows the performance of system for these developed heuristic approaches. During simulation the low and high workload of transaction arrival is taken to fair investigate of these heuristic approaches.
Conclusion: Simulation results show that the system performance is improved using these heuristic approaches. However, NWL_Modified heuristic approach outperform in MDRTDBS than earlier heuristic approaches.
Keywords: Transaction, concurrency control, mobile distributed real time database system, heuristic, scheduling, priority.
Graphical Abstract