Efficient Network Transmission Based On Network Coding

Network coding, first proposed by Ahlswede et al. in 2000, has attracted a great deal of research interests. Owing to the encoding ability of intermediate nodes, the bandwidth can be shared among sinks, and the performance of information dissemination over the network is improved by the network coding. It is shown that the upper bound of the multicast capacity is equal to the size of the minimum cut separating the source and sinks, and that this capacity can be achieved by linear network coding. However, without network coding, the upper bound of the multicast capacity may not be achieved without network coding.In this dissertation, we focus on how to improve the transmission efficiency based on network coding in the networks, such as the retransmission strategies in the network-coding-based network transmission, the cross-layer optimization and the bandwidth allocation scheme in the wireless real-time video stream transmission based on network error correction code, and the network selection method with network coding in heterogeneous wireless networks. The contributions of this dissertation are summarized as follows:1. The retransmission strategies of the generation-based network coding in packet networks are investigated. We propose two retransmission strategies, the Packet-Loss-Edge-based ReTransmission strategy (PLERT) and the Minimum ReTransmission strategy (MRT), which optimize the retransmission efficiency without the constraint on the size of the encoding field. In the PLERT strategy, since the retransmission requests are sent by the intermediate nodes, the overhead of the feedback is reduced. In the MRT strategy, a deterministic algorithm is designed to construct the retransmitted packets, and the best performance is achieved. Simulation results show that the PLERT strategy works well when the packet loss rate is small. The performance of the MRT strategy is the best at the cost of the high complexity which is still polynomial.2. The inter-generation encoding retransmission is proposed to improve the generation-based retransmission. The retransmitted packets of different generations are allowed to be encoded. Under this way, the gain of network coding is improved. However, the decoding delay is increased.3. Due to the unreliable conditions of wireless channels, the limited capacity and the strict qualityof service (QoS) requirements of application, transmitting real-time video stream over wireless networks is a challenging issue. By adding unequal protection for different frames using network error correction codes, the transmission quality could be enhanced. To impove the network coding efficiency, we propose a cross-layer model to optimize the end-to-end video quality based on network coding according to the channel conditions and the GOP structure. In the proposed model, the adaptive modulation and coding parameters at the physical layer, the ARQ parameters at the data link layer and the application-layer forward error correction (A-FEC) parameters at the application layer are jointly designed. Moreover, a low-complexity bandwidth allocation algorithm for the A-FEC is proposed.4. For the video stream transmission with network error correction code, a bandwidth allocation strategy is required to make the cross-layer optimization efficient. We proposed a fair bandwidth allocation scheme to allocate the system bandwidth for multiple users based on game theory. In our proposed shceme, the channel condition and video stream feature are considered. The users with different channel conditions achieve the same QoS if they have the same priority level.5. The network selection based on networkcoding in heterogeneous wireless networks is investigated. For the network-coding-based information exchange, an adaptive network selection strategy is proposed. In the proposed strategy, both of the access rate and the network coding gain are considered to select network, and the transmission efficiency is improved. For the network-coding-based data broadcasting, to evaluate the available network coding opportunity when a new user joins, the network coding opportunity factor is introduced. Using network coding opportunity factor, the encoding opportunity is estimated. And according to the encoding opportunity, the cost of the new access is obtained. The network with the minimum access cost is selected. Our proposed strategy is simple and efficient, and is adapt to various application conditions.