Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article

Degree

Doctor of Philosophy

Program

Electrical and Computer Engineering

Supervisor

Shami, Abdallah

Abstract

Network virtualization has become a key approach for Network Service Providers (NSPs) to mitigate the challenge of the continually increasing demands for network services. Tightly coupled with their software components, legacy network devices are difficult to upgrade or modify to meet the dynamically changing end-user needs. To virtualize their infrastructure and mitigate those challenges, NSPs have started to adopt Software Defined Networking (SDN) and Network Function Virtualization (NFV). To this end, this thesis addresses the challenges faced on the road of transforming the legacy networking infrastructure to a more dynamic and agile virtualized environment to meet the rapidly increasing demand for network services and serve as an enabler for key emerging technologies such as the Internet of Things (IoT) and 5G networking. The thesis considers different approaches and platforms to serve as an NFV/SDN based cloud applications while closely considering how such an environment deploys its virtualized services to optimize the network and reducing their costs. The thesis starts first by defining the standards of adopting microservices as architecture for NFV. Then, it focuses on the latency-aware deployment approach of virtual network functions (VNFs) forming service function chains (SFC) in a cloud environment. This approach ensures that NSPs still meet their strict quality of service and service level agreements while considering both functional and non-functional constraints of the NFV-based applications such as, delay, resource allocation, and intercorrelation between VNF instances. In addition, the thesis proposes a detailed approach on recovering and handling of those instances by optimizing the decision of migrating or re-instantiating the virtualized services upon a sudden event (failure/overload…). All the proposed approaches contribute to the orchestration of NFV applications to meet the requirements of the IoT and NGNs era.

Summary for Lay Audience

Network virtualization has become a key approach for Network Service Providers (NSPs) to mitigate the challenge of the continually increasing demands for network services. Tightly coupled with their software components, legacy network devices are difficult to upgrade or modify to meet the dynamically changing end-user needs. To virtualize their infrastructure and mitigate those challenges, NSPs have started to adopt Software Defined Networking (SDN) and Network Function Virtualization (NFV). To this end, this thesis addresses the challenges faced on the road of transforming the legacy networking infrastructure to a more dynamic and agile virtualized environment to meet the rapidly increasing demand for network services and serve as an enabler for key emerging technologies such as the Internet of Things (IoT) and 5G networking. The thesis considers different approaches and platforms to serve as an NFV/SDN based cloud applications while closely considering how such an environment deploys its virtualized services to optimize the network and reducing their costs. The thesis starts first by defining the standards of adopting microservices as architecture for NFV. Then, it focuses on the latency-aware deployment approach of virtual network functions (VNFs) forming service function chains (SFC) in a cloud environment. This approach ensures that NSPs still meet their strict quality of service and service level agreements while considering both functional and non-functional constraints of the NFV-based applications such as, delay, resource allocation, and intercorrelation between VNF instances. In addition, the thesis proposes a detailed approach on recovering and handling of those instances by optimizing the decision of migrating or re-instantiating the virtualized services upon a sudden event (failure/overload…). All the proposed approaches contribute to the orchestration of NFV applications to meet the requirements of the IoT and NGNs era.

Available for download on Friday, December 31, 2021

Share

COinS