An Overview Of Virtual Machine Consolidation Techniques: Perspectives From Bin Packing And Knapsack Algorithms

Abstract

Load balancing in cloud data centers is inevitable that ensures better resource utilization and improved system performance. Load balancing is not just evenly distributing the load across the available servers but to distribute the load efficiently among the servers.  Load balancing can be made efficient by distributing the load among minimum number of servers to save energy. The servers that are executing with minimum load and idle servers are considered to be power consuming since they are not delivering their fullest potential. Rather they can be freed of and moved to sleep state. Virtual Machine (VM) consolidation is one such efficient load balancing technique. Virtual machine consolidation involves the process of migrating virtual machines among the physical machines. This ensures that no server is overloaded or under loaded. This is done in order to reduce the number of active physical machines. The process of virtual machine consolidation has multiple objectives that include energy saving and improved performance. Various approaches to virtual machine consolidation are found in the literature, among which bin packing and knapsack algorithms are compared and analyzed in this paper. This paper provides an overview of few bin packing and knapsack algorithms based virtual machine consolidation techniques.