Real-time fault-tolerant scheduling algorithm with rearrangement in cloud systems

摘要

Cloud computing has aroused great interest from both academia and industry in the past decade. Due to Service Level Agreement (SLA), the cloud service providers have to guarantee the reliability of their systems. However, unpredictable hardware or software failures are usually inevitable. For real-time tasks, system failure may cause them to miss their deadlines. Fault tolerance technique can efficiently avoid such cases. Primary-backup model is widely used to realize fault tolerance by duplicating a task into two copies - a primary copy and a backup copy. The redundancy introduced by backup copies incurs extra overhead for cloud systems. To address this issue, in this paper we propose a real-time fault-tolerant scheduling algorithm with rearrangement (RFTR) in cloud systems. Different from traditional scheduling algorithms, RFTR dynamically rearranges the execution orders of tasks. Simulation experiments are conducted on CloudSim platform. Compared with the existing scheduling algorithm FESTAL, RFTR shows excellent performance in both task schedulability and resource conservation.