QoS-driven service configuration in computational grids

Abstract

Distributed applications on a computational grid infrastructure require the provision of quality of service (QoS) to ensure application performance. However current grid solutions do not support QoS management, and are limited in performance optimization. In this paper we propose to use a platform managed QoS-aware service configuration approach that enables application developers to separate functional specification and QoS requirements from implementation decisions that depend on the deployment environment. The middleware platform is responsible for achieving application QoS goals at deployment time. This task of the middleware platform is called service planning. In order to show the benefit of the middleware framework, we present an analytical model of service planning for finding QoS-tradeoff points in making configuration decisions. This model is by itself an improvement of a common solution technique used for QoS management. We apply the model to a video-based object tracking grid application.