A real-time self-adjusting scheduling algorithm for tandem precedence structure in cluster computing

This paper proposes an efficient algorithm to schedule a set of real-time tandem tasks on cluster computing environments. It utilizes the available processing power on each processor to accommodate as many tasks as possible on different cluster nodes. It also minimizes the communications among tasks by grouping together as many tasks as possible and allocating them to a single node. Analytical analysis and simulation results show that the algorithm produced a better admission rate compared to treating each application as one unit.