Customer Churn Aware Resource Allocation and Virtual Machine Placement in Cloud

Abstract

Cloud computing industry is growing by leaps and bounds since businesses are increasingly moving its computing infrastructure to datacenters to take the advantage of economy of scale that cloud computing model offers. Like any other industry, cloud service sellers also face tough competitions from rival cloud service providers. Therefore, cloud service industry is also prone to customer churn where even a small customer churn rate has the potential to threaten a business with wastage of resources, loss of jobs and declining revenue. This paper proposes a novel profit maximizing customer retention framework that integrates user experience with resource allocation and placement on physical machines (PM). Our framework addresses the customer churn problem by identification of dissatisfied users and then allocation of additional resources to improve the level of satisfaction through a viable retention action policy. Furthermore, virtual machine (VM) placement problem accounting for both operation and interference overheads is formulated as integer programming problem to place the allocated resources in the form of VMs on PMs. Simulation results using both real and synthetic data show the effectiveness of proposed work.