An execution/sleep scheduling policy for serving an additional job in priority queueing systems

Abstract

In a priority-based computer system, besides the regular jobs, an additional job (refereed to as job A ) is invoked infrequently but requires a significant amount of CPU time. To avoid CPU hogging, job A receives (up to) a fixed amount of CPU time whenever it is served. When the time expires, job A immediately relinquishes the CPU and puts itself to sleep for a period of time. By doing so, jobs with low priority may be processed in a timely manner. When the sleep time is over, job A is awakened and waits to resume service according to its priority. Then, the whole process is repeated until job A service is completed. In this paper, such an execution/sleep (ES) scheduling policy is analyzed for serving job A in a nonpreemptive priority queuing system. The Laplace Transforms are derived for: (i) the conditional response time of job A and (ii) the response time for jobs with priorities higher and lower than job A. This work is motivated by the ES policy in a switching system in which job A is invoked in response to the failure of signaling links. The proposed model is applicable to other real-time computer systems, and the modeling techniques can be applied or generalized to analyzing other scheduling policies in which timers are involved.