Co-scheduling of Disk Head Time in Cluster-Based Storage

Disk time slicing is a promising technique for storage performance insulation. To work with cluster-based storage, however, time slices associated with striped data must be co-scheduled on the corresponding servers. This paper describes algorithms for determining global time slice schedules and mechanisms for coordinating the independent server activities. Experiments with a prototype show that, combined, they can provide performance insulation for workloads sharing a storage cluster -- each workload realizes a configured minimum efficiency within its time slices regardless of the activities of the other workloads.