الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 181 |  |  |
| | | from | 181 | | |
Abstract
Network Function Virtualization (NFV) stands out quickly as a promising innovation in telecommunication networks. It leverages the concept of cloud technology and virtualization techniques. The widespread adoption of NFV leads to providing network services through a chain of Virtual Network Functions (VNFs). This architecture is called Service Function Chain (SFC) or Virtual Network Function Forwarding Graph (VNF-FG), which can be hosted on top of commodity servers and switches located at the cloud. Meanwhile, Software-Defined Networking (SDN) can be utilized to manage VNFs to handle traffic flows through SFC/VNF-FG. One of the most critical issues that needs to be addressed in NFV is VNF placement that optimizes physical link bandwidth consumption. VNF chaining and placement can be considered as the VNF-FG mapped to a cloud provider infrastructure also called NFV Infrastructure (NFVI). Moreover, deploying SFCs enables service providers to consider different goals such as minimizing overall cost, energy consumption, and service response time. Furthermore, the continuity of cloud network services has become an essential availability requirement for NFV. Failure of VNFs may cause critical quality assurance problems for such network services. However, this mapping is addressed while neglecting the massive link utilization and bandwidth consumption that can be encountered during VNF recovery phase.
In this thesis, we try to optimize VNF placement for SFC/VNF-FG in software-defined cloud computing environments. The optimization is achieved in the normal situation by introducing a load-balancing approach. Moreover, a complementary approach has been introduced to facilitate high availability that is utilized when failure occurs. The first approach is named Virtual Network Functions and their Replica Placement (VNFRP), It tries to achieve load balancing over the core links while considering multiple resource constraints. Meanwhile, the second approach is named Redundant VNF-FG Deployment (RVNF-FGD) that attempts to find a near-optimal solution to achieve a trade-off between availability and scalability with a reasonable convergence time. Thus, facilitating the practical deployment of the proposed approach.
The contribution of this work is twofold: First, optimizing link bandwidth utilization in normal situations, and second, optimizing link bandwidth utilization when node/VM failures occur. In both cases, an Integer Linear Programming (ILP) model for the optimization problem is introduced. Then, VNFRP algorithm is proposed to find a near-optimal solution for this optimization problem. VNFRP aims to minimize link bandwidth consumption, energy consumption, and SFC placement cost. On the other hand, RVNF-FGD algorithm attempts to achieve high availability against node/VM failures, reduce link utilization, and minimize bandwidth consumption across the core layer of cloud network. In addition, it reduces VNF-FG communication cost and overall energy consumption. Furthermore, it takes VNF migration into consideration.
Simulation studies are conducted to evaluate the performance of the proposed approaches. Simulation results show that VNFRP can significantly improve load balancing by 80% when the number of replicas is increased. Additionally, VNFRP provides more than 54% reduction in network energy consumption. Furthermore, it can efficiently reduce the SFC placement cost by more than 67%. Moreover, with the advantages of a fast response time and rapid convergence, VNFRP can be considered as a scalable solution for large networking environments.
Meanwhile, simulation results confirm that RVNF-FGD algorithm is capable of simultaneously reducing link utilization and bandwidth consumption across the core layer of the cloud network. Additionally, it reduces VNF-FG communication cost and overall energy consumption. The convergence time of RVNF-FGD algorithm is assessed by applying it to broader cloud network architectures. This assessment indicates the viability of the proposed approach in responding quickly to failures.
While all these findings are encouraging, this work can be seen as a significant move towards SFCs/VNF-FGs placement. Mainly because it takes into consideration the volatility of network traffic among VNFs in cloud computing environments.