Analysis of Scheduling Algorithms for Computer Clusters
Igor Grudenić, Nikola Bogunović · 2008
Abstract – Most scheduling problems are hard to solve optimally for the reasonable size of the input. These problems are even harder to tackle for distributed computing environments. Since distributed computing is a rapidly evolving trend in data processing, there is increased interest in making this approach efficient. Organization of heterogeneous resources in a constantly changing environment makes cluster scheduling a challenging task. In this paper we present different scheduling algorithms that are most frequently employed on computer clusters. We recognize different metrics and system architectures, and analyze various approaches to job and resource matching in different environments.