Network coding mechanisms for Ka-band satellite time varying channel

Network coding over time varying channels has been investigated and a new scheme is proposed. We propose a novel model for packet transmission over time variant channels that exploits the channel delay profile and the dependency between channel states via first order auto-regression for Ka-band satellite communications. We provide an approximation of the delay induced assuming finite number of time slots to transmit a given number of packets. We also propose a novel adaptive transmission scheme that compensates for the lost degrees of freedom by tracking the packet erasures over time. Our results show that network coding non-adaptive mechanism for time variant channels has around 2 times throughput and delay performance gains for small size packets over network coding mechanisms with fixed channel erasures and similar performance gains for large size packets. In addition, it is shown that network coding non-adaptive mechanism for time variant channels has similar performance to the Selective Repeat (SR) with ARQ, and better performance when packet error probability is high, while due to better utilization of channel resources SR performance is similar or moderately better at very low erasures, i.e., at high SNR. However, our adaptive transmission scheme outperforms the network coding non-adaptive mechanism and SR with more than 7 times in throughput and delay performance gains.