Scheduling Mixed Tasks with Deadlines in Grids Using Bin Packing

This paper addresses the problem of scheduling independent tasks with different priorities and deadline constraints in computational grids. The problem is first formulated as a bin packing problem. We propose a heuristic algorithm named Residual Capacity Maximization Scheduling (RCMS), which integrates the ideas of a classical bin packing algorithm (Best Fit) and a mixed integer quadratic programming modeling approach. RCMS is highly scalable as it does not need to know the global state of the grid. Simulation results based on a real-world trace demonstrate that with respect to the total number of schedulable tasks meeting deadlines, RCMS outperforms existing approaches by 8%. With respect to the number of the highest priority tasks meeting deadlines, RCMS outperforms them by over 20% on average. Moreover, RCMS is a fully distributed algorithm with a low complexity.