Integrating Lowest Priority Approach with Largest Point Scheme for Faster Feasibility Analysis

Recently many solutions have been proposed to lower the computational cost of feasibility analysis for real-time systems. The computational cost of feasibility tests can be lowered by strategies such as lowering the number of scheduling points needed during analysis, starting feasibility analysis from lowest priority, or starting schedulability tests for a task with larger scheduling point. All these techniques significantly reduce the computation time of feasibility analysis for fixed priority systems. The computation time of such tests can be further reduced by combining various solutions for efficient feasibility analysis of periodic task sets. In this work, we integrate both lowest priority first with largest points first solution to derive a faster feasibility analysis test for fixed priority system. Our experimental evaluations suggest that the proposed technique significantly lowers the computational cost of the test when system utilization is in the range of 80% or when the ratio between the task period of a lower priority task and the highest priority task is large.