Machine learning techniques for optimizing the provision of storage resources in cloud computing infrastructure as a service (iaas): a comparative study
Addero, Edgar Otieno
MetadataShow full item record
Cloud computing is a very popular field at present which is growing very fast and the future of the field seems really wide. With progressive spotlight on cloud computing as a possible solution for a flexible, on-demand computing infrastructure for lots of applications, many companies and unions have started using it. Obviously, cloud computing has been recognized as a model for supporting infrastructure, platform and software services. Within cloud systems, massive distributed data center infrastructure, virtualized physical resources, virtualized middleware platform such as VMware as well as applications are all being provided and consumed as services. The cloud clients should get good and reliable services from a provider and the provider should allocate the resources in a proper way so as to render good services to a client. This brings about the problem of optimization where clients request for more services than they actually require leading to wastage of the cloud storage resource .This demands for optimization both on the part of the client and the cloud service provider. This has lead to increased research in the various techniques that can be used for resource allocation within cloud services. This research focuses on the analysis of machine learning as a technique that can be used to predict the cloud storage service request patterns from the clients. The research focuses on a review of machine learning as a technique that can be used to predict and therefore optimize the user storage resource demand and usage for the case of cloud computing storage IaaS. Data on cloud storage resource usage was subjected to experiments using machine learning techniques so as to determine which give the most accurate prediction. Some of these machine learning techniques to be reviewed in this research include linear regression, artificial neural networks (ANN), support vector machines (SVM).From the experiments done in this research, it can be concluded that the use of support vector machine algorithm (SVM) proves to be the best algorithm for learning the storage resource usage patterns and predicting their future usage so as to enable better resource budgeting.
University of Nairobi
Thesis Master of Science in Computer Science