An Improved Random Walk Algorithm for Resource Scheduling in Cloud Datacenter

TL;DR: In this article, the authors proposed an improved random walk algorithm which searches the global optimal scheme with simpler computing and compared the proposed algorithm with Round Rabin algorithm and Particle Swarm Optimization algorithm.

Abstract: Resource scheduling plays a crucial role in improving resource utilization rate and user service quality of cloud datacenter. An efficient resource scheduling algorithm enables the datacenter to achieve load balancing, becoming the core of enterprise development. However, at present, the scheduling algorithm of cloud datacenter is usually lack of dynamics, and the calculation is relatively complex. When searching for the optimal scheme, it is easy to fall into the local optimal value, resulting in a large amount of calculation, high energy consumption, low QoS (Quality of Service) and low resource utilization. In this paper, we focus on the prevalent problems of lacking of dynamics, the high makespan and energy consumption in cloud datacenter and design a dynamic load balancing schedule framework. In this framework, we propose an improved random walk algorithm which searches the global optimal scheme with simpler computing. We compare our proposed improved random walk algorithm with Round Rabin algorithm and Particle Swarm Optimization (PSO) algorithm. The experimental results prove that our proposed algorithm improves the utilization rate of resources. Particularly, the makespan of our proposed random walk algorithm is 7% lower than PSO's and the overall energy consumption of ours algorithm is about 15% lower than PSO's.