AbstractCloud computing has become a profound name and also a solid bedrock for new emerging technology. Cloud technology fully supports the dynamic provisioning of computing resources as a utility service on a pay-as-you-go approach. Its numerous benefits (such as rapid elasticity, flexibility, network resource pooling) empower small, medium, and large enterprises to use other technologies through the Internet's flexibility. Despite the benefits cloud technology offers, the challenge of the high power consumption rate it incurs as it leverages its promised attributes remains a major concern. Academic research has shown that a typical 500 square meter data centre consumes about 27,048 kilowatts per hour of power per day regardless of whether it is active or not. Over the years, the most dominant energy efficient techniques for managing data centre has been Dynamic voltage frequency scaling and virtual machine consolidation, which has had a significant setback due to its inability to manage systems on overload state. Therefore, a novel paradigm based on an intelligent mobile agent approach has been proposed. This proposed approach is highly intelligent can easily detect underutilised and overloaded components of the data centre due to its unique feature. Agent technique has successfully shown it can prevent and manage overloading issues due to change in workloads and achieve a more efficient load balancing with a low power consumption rate. The mobile agent was embedded into servers and switches to regulate their activities and then shut down underutilised components. Mobile agent (Java agent) is the first of its kind used in a cloud environment. This research proposal saves a significant amount of energy and improves the entire system performance.
In this thesis, the intelligent-based Agent approach is used to address energy efficiency and cost-aware related problems. Agent approach helps facilitate resource management, allocation of cloud data centre components with a significant reduction in energy usage rate with a more efficient system performance while maintaining a highly reliable system as promised by service providers
|Date of Award||Jan 2021|
|Supervisor||Mo Adda (Supervisor) & Alexander Gegov (Supervisor)|