User priority based efficient CPU scheduler algorithm for real time systems

Abstract

Operating systems at present face a large workload and are restricted by limited processing power. This may lead to lower than expected performance in practice. To counter this the operating system can apply an organizing algorithm to execute the required processes in strategically selected order. This activity is known as scheduling. Scheduling allows the computer to operate in an efficient manner and achieve the targets set for it. On this way schedulers drastically improve CPU performance. Due to its importance, quite a few scheduler algorithms have popped up over the years and research into scheduling remains a hot topic in computing. In this thesis attempts have been made to take the currently available scheduling algorithms and mold them in a planned manner into a procedure where the whole is greater than the sum of the components. As such a hybrid scheduler, named the ‘User Priority Based Efficient CPU Scheduler Algorithm For Real Time Systems’ is proposed. The suggested scheduler is subsequently designed and implemented in a simulation environment. The performance metrics of this complex algorithm are then measured. These values are at that point compared with corresponding values found for the traditional algorithms to establish standards of performance of the novel scheduler.