Static Task Cluster Size Determination in Homogeneous Distributed Systems

In a distributed system which consists of unknown number of machines, many task clustering heuristics have been proposed in order to satisfy requirements such as determining the number of machines and minimizing response time for scheduling Directed Acyclic Graph(DAG). However, those heuristics are not aware of the actual number of existing machines, because their objective is only to minimize response time. As a result, the number of machines determined by an existing task clustering may exceed the number of actually existing machines. Therefore, conventional approaches adopt merging each cluster for reducing the number of clusters at the expense of increase of response time. In this paper, we present a static cluster size determination method and requirements for a task clustering in order to suppress response time while reducing the number of clusters. Our experimental evaluation by simulation shows the effectiveness of a task clustering which adopts the requirements.