Context-Aware Interflow Network Coding and Scheduling in Wireless Networks
Tuan Tho Tran, Thinh P. Nguyen · IEEE Transactions on Vehicular Technology · 2016
Recent approaches using network coding (NC) to mix data from different flows show significant throughput improvement in wireless networks. However, in this paper, we argue that exhaustively mixing packets from different flows may decrease network quality of service (QoS), particularly in the presence of flows with different service classes. We therefore propose a context-aware interflow network coding and scheduling (CARE) framework, which adaptively encodes data across the traffic to maximize the network QoS. First, we develop a perception-oriented QoS (PQoS) to measure the user satisfaction of different types of services. Next, based on the characteristics of the traffic, we optimally combine data across the flows and schedule the encoded packets in each time frame to maximize the PQoS at the receivers. Solving CARE is NP-hard; thus, we devise a computationally efficient approximation algorithm based on the Markov chain Monte Carlo method to approximate the optimal solution. We prove that the proposed approximation algorithm is guaranteed to converge to the optimal solution. The analytical and simulation results show that, under certain channel conditions, the proposed CARE-based schemes not only improve the network QoS but achieve high throughput across all receivers as well. Additionally, the results show that the approximation algorithm is efficient and robust to the number of data flows. In some transmission conditions, our CARE-based schemes can improve the network QoS up to 50% compared with the existing randomized NC techniques.