Multipath Adaptive Video Streaming over Multipath TCP

Multimedia streaming over the Internet has become popular over the years and expected to constitute 90% of Internet traffic in the near future. However, state-of-the-art solutions for multimedia streaming are inherently limited by their use of a single network path and a lack of efficient adaptation algorithms. Dynamic Adaptive Streaming over HTTP (DASH) is a new streaming technology that uses streaming-rate adaptation to accommodate bandwidth variations and utilize the available bandwidth more efficiently. Multipath TCP (MPTCP) is a new transport protocol that enables the use of multiple paths for data transfer. In this thesis, we design and implement a framework for adaptive multimedia streaming using MPTCP, which in many scenarios provides a better streaming performance and robustness to path failures. At the core of our solution are efficient rate-adaptation algorithms that are applicable to both MPTCP and TCP streaming. With these algorithms, the performance of adaptive MPTCP streaming is evaluated under different network conditions to demonstrate the suitability of the proposed framework.