Research on Parallel Real-time Scheduling Algorithm of Hybrid Parameter Tasks on Multi-core Platform

Nowadays, multi-core processors are becoming main stream in computer market due to their high performance, low cost and less power consumption characteristics. However, multi-core processors give rise to new problems when they are applied to real-time system. Little attention has been focused on tasks' DMR(deadline miss rate) by using previous multi-core real time scheduling algorithms under the overload condition. As a result, the domino effect will happen because of many tasks failing to meet their deadlines. Meanwhile, the system performance is leaded to drop sharply. In order to alleviate this problem, this paper proposes a new multi-core real time scheduling algorithm which extends the Pfair scheduling method using tasks' hybrid parameters. In addition, the paper will also discuss the tasks' allocation method which would decrease the switch cost. Experimental results for schedules demonstrate that our scheme enables the real time tasks to be scheduled more efficiently on multi-core platform by adopting hybrid parameter priority. Furthermore, the system performance has gained the robust characteristic, because more real time tasks can meet their deadline under the overload condition.