Priority assignment heuristic to cohorts executing in parallel

  • Authors:

  • Udai Shanker;M. Misra;A. K. Sarje

  • Affiliations:

  • Department of Electronics and Computer Engineering, Indian Institute of Technology, Roorkee, Uttaranchal, India;Department of Electronics and Computer Engineering, Indian Institute of Technology, Roorkee, Uttaranchal, India;Department of Electronics and Computer Engineering, Indian Institute of Technology, Roorkee, Uttaranchal, India

  • Venue:
  • ICCOMP'05 Proceedings of the 9th WSEAS International Conference on Computers
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Transaction scheduling plays an important role in deciding the performance of a distributed real time database system (DRTDBS). It has been demonstrated that the priority based scheduling enhances the performance of a DRTDBS. Generally, the priority of a transaction is determined on the basis of its deadline. Previous work on priority assignment focuses either on centralized database or on distributed databases where subtasks (cohorts) of a transaction are executed sequentially. Several priority assignment heuristics have been proposed both for centralized and distributed databases. However the heuristics proposed for DRTDBS where subtasks of a transaction are executed sequentially may not suit well when the subtasks of a transaction are executed in parallel. In this paper, we consider parallel execution of cohorts and propose a heuristic to determine their priorities. The heuristic proposed in this paper improves the system performance by favoring cohorts demanding lesser number of locks. This minimizes the number of blocked/aborted cohorts. In our scheme a cohort's temporary priority is also calculated every time when a data contention occurs to favor near completion cohorts. This reduces a large amount of wasted work. We compare the proposed scheme with EDF priority assignment policy. Simulation results show that the proposed scheme combined with intermediate (temporary) priority assignment policy gives better performance than EDF based scheme.