The networking community has recognized the benefit of the new form of telecommunication networks triggered by Software Defined Networking (SDN) and Network Function Virtualization (NFV). According to these heavily developed approaches, a future network shall be constructed by inexpensive, general purpose servers, which provide a standardized and universal platform for various software-based network functions e.g., switching, routing, firewall or network address translation. These network functions shall be run in virtual environments, since the fast, dinamic and adaptive operation of network can be guaranteed nothing but that way. Consequently, network operations will be virtual network functions (VNF), thus modules can be run in any site and arbitrary service chains can be configured easily.
Hence, the performance analysis of such virtual network functions became important for the early adopter enterprises along with the cost and power effectiveness. Since these developed approaches are relative new and bringing a multi layer architecture, the effects of the different virtualization techniques and these layers impact to each other are not known yet. Analyzing the effects of virtualized environments on the performance of software-based switches is an essential problem.
The goal of this master thesis is to present, improve and optimize these new diversified aspects of telecommunication networks. For this purpose, it is necessary to introduce and analyze the wildly used, open source virtualization techniques along with the production ready software-based switches. I map and use the techniques of performance impovement and optimization for cooperation between virtualization and software-based switches. It is necessary to classify the technologies by parameters based on transmission capacity, packet delay and power consumption. For the precise measurements, I process a measurement methodology, which is capable of comparing different technologies mentioned above.