Abstract
Cloud Computing (CC) is often referred to in discussions on Green Cloud Computing (GCC) as a more energy efficient approach to the provision of computational and data storage services. Without new and efficient energy conservation approaches, energy consumption is likely to become the main bottleneck for the growth of the Internet. The core principle of the "Coal Free = Green Energy" or "Green Cloud" is to perform all cloud computing services in as energetically/economically efficient way as possible. A zero carbon approach is the target from using renewable (green) energy sources e.g. sun, wind, tide power. However, while we believe that renewable energy can help mitigate the impact of IT and cloud computing, we are far off being able to use renewable sources as the primary energy supply for the entire Internet and its associated data centres. Currently provision of public clouds is focused on mega-data centres that are maintained by a small number of large, predominantly US-based companies. Yet this model is not without flaws. These include vendor lock in, and dependency on services provided by an organisation which may be operating in a different economic and political context to that of consumers of these services. In this thesis the state-of-the-art in cloud computing was reviewed. It is hypothesised that when the energy consumed in the Internet infrastructure is taken into account, alternative cloud scenarios to the use of mega-scale data-centres may be more appropriate. This thesis focuses on modelling energy consumption at a total systems level, and how to achieve a good balance of the energy consumption between computation usage and communication usage in the cloud computing model. Consumption both within data centres and the network traffic between consumers and providers are studied using mathematical modelling techniques.