Task scheduling algorithms to optimize the scheduling length of the applications in distributed environment

Abstract

The Heterogeneous Distributed Computing System (HDCS) is a well-orchestrated suite of high-performance processing cores with diverse computational capabilities and network bandwidth to solve computationally intensive applications. In such environments, there are many applications to be processed. These applications are distributed across the collection of heterogeneous computing resources to achieve a certain performance objective. A crucial problem for HDCS is in finding a suitable strategy or scheduling of a set of tasks to be executed. Such a strategy is capable of allocation of tasks that have different types of communication and computational cost to the suitable resources in different types of environments, including distributed, cluster, parallel, grid, cloud, and fog systems. Due to the arbitrary size of the tasks, precedence constraints between the tasks, and the bounded number of available resources for execution, finding an optimal scheduling solution for HDCS is proven as an NP-complete problem. Therefore, many heuristic scheduling strategies are proposed to provide a sub-optimal solution to minimize the scheduling length of the applications. However, it is still an open problem in HDCS due to the lack of standard benchmarks for evaluating the scheduling algorithms. In this thesis, effective heuristic static list-based scheduling strategies are proposed for HDCS to maximize the resource utilization rate, minimizing the scheduling length of the applications and the overall computational cost of the processors. newline