Cluster Based Real Time Scheduling for Distributed System

Real-time applications nowadays are implemented using a distributed architecture. Real-time task scheduling on such distributed system is a complex problem. The existing real-time task scheduling techniques for the distributed system are primarily based on partitioned and global scheduling. In this paper, we proposed cluster-based real-time tasks scheduling on a distributed system which is a hybrid scheduling approach where processing nodes group into clusters. The real-time tasks are not allowed to migrate among the processing nodes of the different cluster which reduce the migration and preemption overheads problem of global scheduling and improve system utilization which is the problem in partitioned scheduling approach. The performance of the proposed scheduler analyzes with different configurations of the cluster. The analysis of the result shows that task acceptance rate and system utilization increase as the number of clusters tuned to a smaller value. The migration, preemption, and scheduling overheads reduce with increasing the size cluster. The simulation result shows that the proposed scheduler increases the system utilization by approximately 20%. The migration and preemption overheads reduce by approximately 42% and 32 % as compared to benchmark global schedulers.