An adaptive scheme of Systematic Random Network Coding on interfered satellite link

Abstract

Systematic Network Coding (SNC) has been studied recent years and it got perfect effect on stable satellite link. However, its performance on interfered satellite link is unsatisfactory according to our research. This paper firstly proposes a novel architecture solution named K-SNC on Link Layer (LL) to improve the goodput of the communication on interfered link. In K-SNC, a Variant of Kalman Filter (VKF) is designed to estimate Packet Erase Rate (PER) in real time and a method is introduced to adjust Encoding Window (ECW) adaptively. A mathematical model is developed as the foundation of VKF about PER changing trend with time lapsing. The relationship is analyzed between the appropriate size of ECW in SNC and the PER on satellite link. Simulation works are implemented with NS-2. In comparison with SNC, K-SNC gets impressive improvement on stable satellite link and provides up to more than 30% higher goodput on the satellite link suffering unpredictable interference.