EVALUATION OF DISTANCE-BASED PRIORITY SCHEME FOR TASK STREAMS WITH. DEADLINE CONSTRAINTS

Abstract

A real-time application is usually comprised of a set of cooperating tasks, which are often invoked at regular intervals and have deadlines by which they must complete their execution. When activities have timing constraints, scheduling them to meet their timing constraints is a major problem. This dissertation addresses this problem. The dissertation evaluates and analyses the performance of new priority- based scheme, called" Distance-Based Priority (DBP) scheme ", for scheduling real-time streams with (ni,k)-firm deadlines on a single server. A stream with (m,k)-firm deadline experiences dynamic failure if fewer than in out of any k consecutive customers meet their deadlines. The scheme assigns higher priorities to customers from streams that are closer to dynamic failure so as to improve their chances of meeting their deadlines. The effectiveness of this approach is evaluated through simulation under various system patterns. The scheme is compared to other conventional schemes such as Single Priority scheme, Fixed Priority scheme and Imprecise Computation model. The dissertation then proposes an improved version of the DBP scheme and it is found through simulation that the proposed scheme gives still better performance than DBP scheme.