Load Sharing in Cluster Service Provision

Many of the factors that are driving the development and the adoption of grids, such as scale, dynamic and distributed characteristics and heterogeneity, are also pushing to the forefront service quality issues in grid computing. These include performance, reliability and security. Although grid middleware can address some of these issues, it has become imperative to ensure adequate service provision from compute servers. This paper is concerned with the presentation of a load sharing scheme which can contribute to service provision in grid computing, at cluster level. As the aim of the scheme is to enhance performance and reliability, a proactive, non-preemptive and distributed approach was adopted. This ensures that load information is gathered continuously before it is needed, and that a task is allocated to the most appropriate node for execution. Reliability is enhanced by the decentralised nature of the scheme and the symmetric roles that the nodes can