An analysis of resource sharing in a large scale information system

Date

1972

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Resource sharing in a large scale information system provides for concurrent DEMAND (interactive) and BATCH processing. This sharing is normally controlled by a scheduling algorithm that resides within the operating system. The sharing algorithm attempts to provide satisfactory service to each user. As a result of the values set for the parameters in the sharing algorithm, the impact of the sharing algorithm on the system performance is determined. The system performance is also contingent on the resources available to the operating system and the characteristics of the input environment. An analysis of an existing sharing algorithm is made by examining the UNIVAC 1108-EXEC 8 system in an effort to study the ability of this algorithm to control the DEMAND to BATCH balance. The hardware configuration utilized by this operating system was also evaluated., The vehicle used to perform this analysis was a discrete stochastic model of the UNIVAC system implemented in FORTRAN. The model is event oriented in that most of the important events in the actual system are logically and statistically represented. The effect of making the sharing algorithm adaptive to the current processing environment is considered with respect to response time for DEMAND users and the total number of tasks processed for BATCH users (throughput). The configuration is changed to determine its impact with respect to DEMAND response time and BATCH throughput. An evaluation of the simulation is presented to point out potential future work.

Description

Keywords

Citation