Performance Analysis of Sliding Window Based Erasure Correction for Real-Time Traffic

We propose a fountain-coding-like erasure correction method for streaming traffic with real-time requirements. A sliding window defines the range of non-expired data. Each new block entering the window is once sent as such, followed by probabilistically sending a repair packet, formed as a random combination of the blocks in the current window using a degree distribution as in LT-coding. The decoding of repair packets is based on an iterative algorithm. The performance of the method with a given channel loss probability is analyzed using an exact Markov chain model. The state space, however, has to be truncated for computational tractability. The truncation error is verified to be small enough by simulations. By using the analytical model the optimal degree distribution is found to be of single-degree type. Both analytical and simulation results on the performance with the optimal repair packet degree are presented and the effectiveness of the method demonstrated.