SYSTEM MODELING AND EVALUATION ON FACTORS INFLUENCING POWER AND PERFORMANCE MANAGEMENT OF CLOUD LOAD BALANCING ALGORITHMS

Cloud is an on-demand IT resource provisioning technology uses server virtualization and load balancing as the underlying techniques. Power and performance management are the major concern of cloud to achieve Total Cost Ownership (TCO) in terms of user acceptance and societal importance. In this concern, there is a need to investigate the power and performance influencing factors to design a novel cloud load balancing algorithms with respect to recent hardware and software advancements. Hence, the work studied these approaches to allocate only required amount of virtual servers for varying cloud workload. In this regard, the cloud system model is designed and evaluated for different scenarios like reactive system model, cloud workload and different scaling and sizing of Virtual Machine (VM) servers for various load balancing algorithms. The simulation results infer that the launching of an optimal number of virtual machines, the cost of VM setup time in the data centre, control considerations - dynamic regulation of frequency of controller invocation, adaptive algorithms instead of dynamic algorithms, and multi-core CPU architectures are to be considered while implementing cloud load balancing methods. Appropriate consideration of the above-mentioned parameters is required to make a powerful, flexible and cost-effective load balancing methods for power and performance management for cloud data centre.