Conditional Random Network Coding in the Case of Transmission over q-ary Gilbert Elliot Channel

Abstract: The benefit of random linear network coding (RLNC) has been studied in wired and wireless networks. Furthermore, it has been shown that RLNC improves throughput, bandwidth and robustness to network changes or link failures, however it is highly susceptible to error propagation. As existing error correcting codes are not adapted to such errors, recently Kötter and Kschischang introduced a new subspace code in the context of a noncoherent transmission model. In this paper, we evaluate the performance of such code in the case of transmission over a network where each link is a q-ary Gilbert Elliot channel. We consider the conditional RLNC over Fq where, if all random coefficients are nulls, they have to be regenerated until at least one of them is not null. Simulation results verify theoretical results of [8] for different channel and coder parameters