Evaluation of customer's losses and value-at-risk under cloud outages

Cloud services may suffer from outages, which interrupt customers' activities and result in economic losses. A linear model is proposed here to translate outages into economic losses, accounting for both the frequency and the duration of outages. Two tools, the exceedance loss probability and the Value-at-Risk, are employed to assess the maximum economic exposure of a customer. The use of both tools is demonstrated by applying them to a dataset concerning five major cloud providers. Those tools can be employed to drive the choice of a cloud provider towards the least risky one.