Simplifying Statistical Modeling for Infrastructure-as-a-Service Cloud Computing

Abstract

Cloud computing is a multitenant system where multiple users share the resources provided by the service provider. It is important to adjudge the performance of the cloud servers to ensure the successful deployment and execution of software. The inherent nature of various data centers and the request rate of users make it challenging to evaluate the performance of the cloud centers mathematically. The current work proposes a novel queuing-based analytical model based on a general expansion method to assess the performance of cloud servers. The proposed work's main advantage is that the model can be applied irrespective of the type of distribution function assumed for inter-arrival times and service time. The queuing characteristic of both the single server system and multi-server are taken into consideration.The model determines the key performance indicators such as response time, waiting time, and other performance metrics. The model critically analyzes the relationship between the number of the server, service rate probabilistic distribution function of the servers, response time, waiting time, and the number of the buffers. It is observed that the system's performance having a general distribution service rate outperforms the system having an exponential service rate.