Cloud computing has gradually spread throughout the global computing environment; speed, reliability, quality of service, security, and standardization of services are the main issues in its ecosystem. To evaluate the performance of the infrastructure as a service (IaaS) approach in the cloud environment, response delay and request rejection are considered. The authors estimate the performance of the fault-prone IaaS with a faulty virtual machine and atmosphere, through proposing a stochastic queueing model and evaluating its closed results using data from a campus cloud.

With emphasis on the elasticity of the IaaS environment and its appropriateness for the fluctuating workload, the authors simulate their proposal based on the performance analysis of a queueing network system. In the proposed stochastic model, the inputs are the arriving requests; third-party owners of hardware, software, storage, and so on are the resources; and the defined admission buffers are mounted on the OpenStack environment. Virtual-machine-manager would be responsible for the admission management and fulfillment of expected delay estimation. Speed calculation and fault scenarios are provided, as the estimation of the performance at different heavy-loaded scenarios is the main objective of the trial.

A nice stochastic analysis based on the Markov process is composed, and its performance estimation is evaluated and verified via experimental results. Next, according to the achieved results, an optimum responsiveness determination problem is formalized: a nonlinear, discrete programming scenario. The solution and implementation algorithm is provided and its analytic graphs and results are demonstrated.

The paper is a coherent and classical scientific artifact. It has a smooth stream in which a passion and diligence to solve the problem elegantly are apparent.