Minimizing Multiple-Priority Inversion Protocol in Hard Real Time System

Abstract

With the interesting of researchers to develop the scheduling of multiprocessor and distributed system in hard real time embedded systems, we present a new locking protocol called Minimizing multiple-Priority Inversion Protocol (MPIP), in this protocol, we proposed a novel mechanism to turn from preemptive to non-preemptive state to minimize multiple-priority inversion caused by low priority task on local processor. We focus on distributed locking semaphore and we showed its ability to lock the shared memory, thus, working in Multi-processor environment. We assumed all shared resources as global to avoid resource nested. With properties of our protocol, we define a new upper bound blocking time. We submit our proposal to experimental evaluation to compare it with MPCP protocol in term of hard real time schedulability.