Modelling and Analysing QoS for Real-Time Interactive Applications on the Cloud

Abstract

The complexity of determining resource provisioning policies for applications in such complex environments introduces significant inefficiencies in the cloud. Novel approaches are needed to efficiently model and analyse Quality of Service for such applications, especially those with real-time constraints. This chapter investigates some of the techniques that can be used to explore these trade-offs and to find business models where value can be provided at all stages in the value-chain. For example, how can an application user identify a set of service providers that provide the best solution to their workflow in combination? How much risk is involved, e.g. what is the probability and impact of a failure of a service provider to deliver the QoS they promise? How can Service Level Agreements be specified that provide the flexibility to accommodate variability in service use, yet don’t result in unnecessarily high cost to the consumer due to the service provider having to use massive over-provisioning of resources to ensure they can meet times of peak-load?