Modeling and performance analysis of cloudlet in Mobile Cloud Computing

Mobile Cloud Computing (MCC) is an emerging technology to overcome the resource constrained nature of mobile devices by offloading the resource-intensive operations to cloud-based data centers. The characteristics of MCC, including the mobility, the instability of 3G/WiFi connections, and the complexity of virtualization, make the prediction of performance difficult. To deal with these issues, an analytical performance model consisting of interacting stochastic sub-models is proposed. Also, the cyclic inter-dependency among these sub-models is resolved by the fixed-point iteration method. Moreover, closed-form solutions of the sub-models are presented whenever achievable. Specifically, this study models a type of MCC known as the Cloudlet in which the mobile devices, via a WiFi connection, receive services from a cloudlet (e.g., coffee shop) as an intermediary node. In this architecture, a dedicated Virtual Machine (VM) is provisioned on a Physical Machine (PM) while the PM can be located as a part of the cloudlet or a public cloud. The performance of such a class of MCC is affected by varied set of parameters, such as connection failure, and workload. The model illustrates the impact of the parameters on two important performance measures: request rejection probability and mean response delay. Using the SHARPE software package, the model is solved and numerical results presented. Furthermore, the analytical results are verified through discrete-event simulation. (C) 2016 Elsevier B.V. All rights reserved.