A QoS Scheduling Scheme with Availability Constraint in Distributed Systems

An efficient resource management mechanism is important in a heterogeneous system to discover available resources, to allocate an appropriate subset of resources to applications, and to map data or tasks onto selected resources. The key component, tasks scheduling, draws our attention. Make span is the principal concern many existing researches. But, other QoS requirements are also important in more and more realistic applications. Like Grid service it is expected that the service provider be reliable, robust, or highly available. In this study, an existing availability-aware scheduling model called SSAC is investigated first. Then, we proposed an optimization approach to increase the availability and to reduce the make span of tasks running in heterogeneous systems. Three quantitative conditions are used to guide the optimization. Our experimental results show that compared with three existing solutions (Minmin, SSAC and ACO), our algorithm significantly improves the availability and performance of multiclass tasks running in heterogeneous systems.