SDN aware DASH client architecture

Dynamic adaptive video streaming applications are one of the most popular applications of the last decade. In such applications, the basic architecture allows users to change quality as they desire. Evaluating Quality of Experience (QoE) effectively for Transmission Control Protocol (TCP) video streaming with network performance parameters is a challenging task. For this reason, the use of flexible systems such as Software Defined Network (SDN) to adapt to variable network conditions has increased, rather than traditional methods, in the development of these applications. Due to the varying conditions of the network, the use of multi-path and methods to improve downloaded segments provide great advantages for a stable and high-quality monitoring experience. Using multiple paths maintains the stability of the system where segment sizes differ or where short-term bandwidth decreases. Hence, using Scalable Video Codec (SVC) to improve the downloaded segments prevents bandwidth wastage because segment quality is provided by decoding layers. In this thesis, we propose a QoE aware adaptive streaming over Hypertext Transfer Protocol (HTTP), named "mpBack^2", that requests segments of an SVC coded video content by using multiple paths by connecting an SDN Controller through OpenFlow protocol. Simulation results show that users achieve higher QoE by using our approach when compared to conventional rate adaptation algorithms. Keywords: Video Streaming over HTTP, Scalable Video Coding, Multiple Path Rate Adaptation, Software Defined Networks, Quality of Experience