Design and analysis of a grouped sweeping scheme for multimedia storage management

Abstract

This paper presents an efficient grouped sweeping scheme (GSS) for disk scheduling to support multimedia applications. The GSS scheme provides a formulation for balancing the two conflicting requirements: on the one hand we want to reduce the amount of seek delay, while on the other hand the worst case bound on the time between two consecutive services to the same request stream needs to be minimized. The GSS approach leads to a general formulation covering a family of disk scheduling schemes. Two previously proposed approaches, a fixed order round-robin scheme and a more sophisticated elevator type scheme, are special cases of the GSS formulation. Our analysis shows that by properly setting the grouping factor the optimized GSS scheme can lead to substantially lower buffer requirement, particularly when compared with that of the round-robin scheme. We develop an efficient procedure to determine the optimum value of the grouping factor. We also incorporate the effect of disk arrays in the GSS formulation and discuss how GSS can be used in dynamic settings such as to provide interactive service in an integrated media environment.