Offloading mechanisms for mobile networks using SDN in virtualized environments

• Autor(a) principal: Silva, Rui Miguel Morais da
• Data de Publicação: 2018
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/25031
• topic: 5G; 4G; SDN; NFV; Cloud; Virtualization; Offloading; EPC

The exploding mobile data traffic increase in recent years has been putting a high load on both mobile cells and core network, with operators facing the need to upgrade their networks. Nowadays, to do this upgrade, operators need to purchase new specialized equipment for network functions, having to cope with a high upgrade CAPEX. Furthermore, networks are deployed with a one size fits all approach, which in some cases might not satisfy the requirements of specific services. 5G aims to solve these problems by virtualizing network functions in datacenters, decoupling the software from the hardware for network functions and using general purpose hardware instead. To support this, Software Defined Networking (SDN) is introduced, which allows the network to have a higher degree of programmability, enabling new features such as higher flexibility and network slicing, where multiple virtual networks can be created and tailored to specific requirements. This thesis addresses an architecture that evolves the Evolved Packet Core (EPC) into a core network closer to 5G by virtualizing EPC’s network functions, introducing SDN and supporting 4G to Wi-Fi traffic offloading, helping to reduce the load on mobile cells by leveraging on the smartphone’s support for dual connectivity and high density of Wi-Fi access points already deployed worldwide. The proposed architecture is then evaluated and compared to a vanilla EPC whenever possible showing that, although there is an increase in latency at the virtual EPC, the bottleneck of the system resides in the air interface. Also, a use case for this architecture was defined and evaluated. The use case considered traffic offloading and dynamic Wi-Fi slice creation, with results showing that it can seamlessly offload a video stream from 4G to Wi-Fi without affecting the user’s Quality of Experience.