A Quantitative Risk Analysis Framework for Evaluating and Monitoring Operational Reliability of Cloud Computing

Cloud computing offers the advantage of on-demand, reliable and cost efficient computing solutions without the capital investment and management resources to build and maintain in-house data centers and network infrastructures. Scalability of cloud solutions enable consumers to upgrade or downsize their services as needed. In a cloud environment, hardware and software resources are distributed among one or more locations, from within a single datacenter to multiple data centers located around the globe. In a cloud environment, the consumers may share both physical and logical resources of the cloud provider, increasing the threats of insider attacks or data leakages along with the issues of different legal jurisdictions while implementing accountability and compliance. Potential security and reliability concerns need to be thoroughly analyzed prior to adoption of cloud based services and contingency plans need to be outlined. Risk factors in cloud environments also need continuous monitoring by providers, consumers and other stakeholders involved in order to reduce risk impacts while ensuring the highest level of performance from the cloud.;In this research, we present a quantitative framework to systematically analyze the risks of cloud computing, applying probability distributions and Monte Carlo (MC) simulation. Potential risk factors associated with reliability and security concerns are first identified and then analyzed based on available information. Risk is defined as the product of the probability that an event would occur and the consequences it might have, both of which can vary. Based on the nature of uncertainties and available data, probability distributions are assigned to the input variables. Monte Carlo simulation is performed to construct a probability distribution of outcomes of the risks. The reverse cumulative distribution function (CDF) indicates the probability of exceeding any given risk cost in the graph. Two case studies are presented in order to demonstrate how this framework can be implemented in evaluating and monitoring operational reliability of cloud computing.