Providing quality of service for disks I/O sub-system

摘要

The disk scheduling algorithms try to service a request with lower priority and strict deadline only if serving this request is not going to violate the deadline constraints of a higher priority request. One of the algorithms uses a single queue to hold all the requests regardless of their priority. The single queue is reorganized after the arrival of a new request according to the order requests would be serviced. A second algorithm uses multiple queues. Each queue holds only requests with the same priority, and the requests in each priority queue are sorted according to their deadlines. The decision of servicing a request is made at scheduling time.