Cognitive-Aware Network Virtualization Hypervisor for Efficient Resource Provisioning in Software Defined Cloud Networks

Abstract

Integration of different technologies forms an integral part of modern network engineering and 5G technology deployment. Although Software Defined Networking (SDN) and Network Functions Virtualization (NFV) function well independently, integrating these two technologies present the cooperate advantages to service providers and service users. Operations of cloud computing technologies have been enhanced with the advent of SDN and NFV for efficient solutions deployment and infrastructure management in Software Defined Cloud Datacentre Networks (SDCDCN) where dynamic controllability is indispensable for elastic service provision. The provisioning of joint compute and network resources enabled by SDCN is essential to enforce reasonable Service Level Agreements (SLAs) stating the Quality of Service (QoS) while saving energy consumption and resource wastage. This thesis presents a Cognitive- Aware Network virtualization Hypervisor which was developed from merging the programmable dynamic network control attributes of SDN and the network slicing attributes of NFV to provision joint compute and network resources in SDCDCN for QoS fulfilment and energy efficiency. It focuses on the techniques for allocating Virtual Network Requests on physical hosts and switches considering SLA, QoS, and energy efficiency aspects. The thesis advances the state-of the-art with the following key contributions: A modelling and simulation environment for Software Defined Cloud Datacentre Networks abstracting functionalities and behaviours of virtual and physical network resources. The second is a novel dynamic overbooking algorithm for energy efficiency and SLA enforcement with the migration of virtual machines and network flows. Finally, a performance-aware intelligent overbooking for predicting network resource usage and performance for the next defined time interval considering multiple performance indexes.