Heuristic Approaches to Delivering Cloud Resources at Minimal Cost in IT Infrastructure Management

This paper presents a survey of various heuristic algorithms for load balancing in cloud resource provisioning. The main goal is to reduce the task execution time of available virtual machines and thereby achieve lower cloud resources cost. The specific load-balancing problem is formulated in the context presented. Using the Proxmox virtual infrastructure, the efficiency of four selected heuristic algorithms was investigated under different circumstances in a production lab environment consisting of five virtual machines. The experimental results show that both the greedy strategy approach and the combined greedy and probabilistic strategy approach are potentially effective solutions for achieving load balancing in cloud resource provisioning at minimum cost.