Improving Bandwidth Allocation in Cloud Computing Environments via "Bandwidth as a Service" Partitioning Scheme
MetadataShow full metadata
Cloud computing is a relatively new idea where virtualized computers commu- nicating with each other perform a multitude of services such as general purpose computing, cloud services to local applications for storage or processing power, and hosting services for entire applications executed within the cloud. This thesis focuses on network bandwidth utilization within the cloud-computing environment and proposes the concept of providing bandwidth control as a service (Bandwidth as a Service) within the cloud. Prior methods of providing control over cloud band- width utilization are discussed, and a novel algorithm is proposed for dynamically shaping cloud bandwidth into percentage groups containing individual transport- level connections, maximizing bandwidth efficiency while maintaining limitations on network resources. An analysis of inter-virtual machine bandwidth performance is discussed along with a survey of throughput efficiency and improvements within the hypervisor.