Rotational-position-aware real-time disk scheduling using a dynamic active subset (DAS)

Scheduling disk requests with service guarantees has to bring the demand to meet guarantees in line with the need to optimize disk utilization. This paper presents the design, implementation, and experimental evaluation of a disk-scheduling framework which optimizes the disk utilization under the constraints of both hard and statistical service guarantees. The framework is based on two principles: 1) upon each scheduling decision, the calculation of a subset of the outstanding disk requests such that all service guarantees can be enforced under worst-case assumptions; and 2) the scheduling of this subset based on the rotational position of requests in order to improve the disk utilization. Results are presented through an implementation of the scheduling framework in DROPS, the Dresden real-time operating system.