Research On Fault Tolerance Scheduling Algorithm In Real-Time Operating System

The real-time systems are widely used in many fields which include fabrication, life of humanity and so on. More and more real-time systems are playing a very important role in every kinds of control system. So the high reliability of real-time systems is required. In order to guarantee the system's reliability, the task must be re-executed before its deadline once a fault occurred. Fault tolerant requests that the timeliness of time and accuracy of computing result of real-time systems should be guaranteed after faults occurred. As an important technology which guarantees the reliability of real-time systems, fault tolerant has already become the hot topic in the field of real-time systems. Fault tolerant real-time scheduling algorithm is an important fault tolerant method to promote the reliability of systems. In order to realize the fault tolerant of real-time systems, it is necessary to design fault tolerant real-time scheduling algorithm which is practical, simple and highly effective to realize the fault tolerant scheduling and enhance the reliability of systems.In this paper, the related concept and theory to real-time system s' are introduced first, which include the definition, structure and characteristic of real time systems. The real-time operating systems are introduced especially. Then the reliability technology of real-time systems is analyzed, which includes the type of faults and the method of promoting the reliability of systems. Fault tolerant is analyzed especially. Fault tolerant scheduling algorithm mainly stems from the real-time scheduling algorithm. In this paper, some classical scheduling algorithm for real-time task will be analyzed. And some fault tolerant real-time scheduling algorithms which base on the classical scheduling algorithm for real-time task will also be analyzed.In the time before, researchers usually describe faults by assuming a constant number of faults or a minimum time of fault inter-arrival. In practice, the occurrence of faults is stochastic. We model the occurrence of faults as a stochastic process with a Poisson distribution having a mean inter-arrival rate ofλin this paper. And a fault tolerant task scheduling algorithm which is a greedy algorithm is explored according to the fault model. The algorithm provides enough slack time in order to guarantee that the tasks can be re-executed once fault occurred. The simulation result shows that the lost ratio and rejected ratio of tasks can be reduced although the task's arrival and fault occurrence is very frequently. So this algorithm is effective to the fault tolerance in real-time operating system.