An Energy-Efficient Multi-Cloud Service Broker for Green Cloud Computing Environment

The heavy demands on cloud computing resources have led to a substantial growth in energy consumption of the data transferred between cloud computing parties (i.e., providers, datacentres, users, and services) and in datacentre’s services due to the increasing loads on these services. From one hand, routing and transferring large amounts of data into a datacentre located far from the user’s geographical location consume more energy than just processing and storing the same data on the cloud datacentre. On the other hand, when a cloud user submits a job (in the form of a set of functional and non-functional requirements) to a cloud service provider (aka, datacentre) via a cloud services broker; the broker becomes responsible to find the best-fit service to the user request based mainly on the user’s requirements and Quality of Service (QoS) (i.e., response time, latency). Hence, it becomes a high necessity to locate the lowest energy consumption route between the user and the designated datacentre; and the minimum possible number of most energy efficient services that satisfy the user request. In fact, finding the most energy-efficient route to the datacentre, and most energy efficient service(s) to the user are the biggest challenges of multi-cloud broker’s environment. This thesis presents and evaluates a novel multi-cloud broker solution that contains three innovative models and their associated algorithms. The first one is aimed at finding the most energy efficient route, among multiple possible routes, between the user and cloud datacentre. The second model is to find and provide the lowest possible number of most energy efficient services in order to minimise data exchange based on a bin-packing approach. The third model creates an energy-aware composition plan by integrating the most energy efficient services, in order to fulfil user requirements. The results demonstrated a favourable performance of these models in terms of selecting the most energy efficient route and reaching the least possible number of services for an optimum and energy efficient composition