Byzantine fault-tolerance in federated cloud computing

Abstract

Cloud computing has emerged as popular paradigm that enables the establishment of large scale, flexible computing infrastructures that can offer significant cost savings for both businesses and consumers by allowing compute resources to be scaled dynamically to deal with current or anticipated usage [1]. This concept has been further strengthened with the emergence of federated computing Clouds that allow users to scale applications across multiple domains to meet Quality of Service targets [2]. However, the challenge of building dependable and robust Clouds remains a critical research problem that has not yet been clearly understood [3], and yet is vital for establishing user confidence in Clouds. This is particularly true when considering Byzantine faults that are arbitrary in nature. This paper analyses the application of Byzantine fault-tolerance to federated Clouds in detail, and presents experimentation performed to analyse the effectiveness of Byzantine fault-tolerance in federated Clouds. We have developed a Cloud framework called FT-FC that allows us to very quickly create diversity-based Byzantine fault-tolerant systems and apply them to federated Clouds, and have produced initial results to demonstrate the feasibility and potential of this approach. We have furthermore identified a number of research problems and challenges that need to be addressed in order to progress this area further. Our current experimental results, although very initial, are highly encouraging figures, and demonstrate the effectiveness of the FT-FC framework.