Real-time scheduling in distributed systems

Abstract

Real-time computer systems are expected to greatly benefit from multiprocessor technology and employing multiprocessor systems for real-time applications has shown to be difficult. A major obstacle is that scheduling algorithms for real-time multiprocessor systems are significantly more complex than for uniprocessor systems. In multiprocessor systems, the scheduling algorithm must not only specify the order of tasks, but also determine the specific processor to be used. Any practical scheduling algorithm for assigning real-time tasks to a multiprocessor system presents trade-off between its computational complexity and its performance. Finding a minimal schedule for a given set of real-time tasks in a multiprocessor system has shown to be NP-hard. In this paper, we show several scheduling algorithms used in such multiprocessor systems and their performance with experimental results.