Evaluation techniques for storage hierarchies

IBM Systems Journal, Volume 9, Issue 2, pp. 78-117 (1970).

The ACM Fellows Program recognizes and honors outstanding ACM members for their achievements in computer science and information technology and for their significant contributions to the mission of the ACM. ACM writes about Donald Slutz's election as an ACM Fellow, noting that “The paper by Mattson, Gecsei, Slutz and Traiger advances the art of performance evaluation of computing systems by introducing a method of computing a large number of different performance measures in a single pass through a computer program.”

This very highly cited paper discusses storage systems with a hierarchy of storage devices that have a wide range of cost and performance characteristics. In such storage hierarchies, the logical address space is often partitioned into pages that represent the blocks of information being moved between devices in the hierarchy. Naturally, many factors can affect the cost and performance of a storage hierarchy. Technologists must consider the capacity and characteristics of each storage device, the physical structure of the hierarchy, the way in which information is moved by the hierarchy management facility, the expected pattern of storage, the software required to move logical information, and the cost per bit of each device. These kinds of factors make it quite difficult to design an “optimal” hierarchy. To address these challenges, this paper discusses a novel, efficient method for determining the overall performance of storage hierarchies.

See Donald Slutz's ACM Fellows Award