This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Resource Management – Task Scheduling • Program execution requires resources, most of which are shared. One essential responsibility of an operating system is to schedule the use of shared resources. Resource scheduling may be preemptive or nonpreemptive. In nonpreemptive scheduling a task once allocated a resource can keep it until the task voluntarily relinquishes its control (usually when it is no longer needed). With preemptive scheduling tasks may be interrupted and the resource removed prior to completion. Typical processor scheduling is preemptive, in a round-robin fashion. • A processor scheduling problem is defined by three parameters: workload, processing system, and performance measure. Processor scheduling (especially multiprocessor scheduling) has been studied extensively in deterministic scheduling theory in which workload is defined by a fixed set of tasks with known arrival times and execution times, and a scheduling algorithms are applied to the workload on the given processing system, based on the specified performance measure(s). Common performance objectives for evaluation include: minimize mean response time, minimize mean wait time, maximize throughput (or equivalently, minimize schedule length), minimize the number of deadline violations, etc. Schedules may be represented using Gantt charts. • Some known results in deterministic scheduling for multiprocessor scheduling: s Shortest-job-first algorithm produces schedules with minimum mean response time. Shortest-job-first algorithm produces schedules with minimum mean response time....
View Full Document
- Spring '08
- Queueing theory, Computer multitasking, Scheduling algorithm