A new algorithm on variable-rate convolutional broadcast for network coding in cyclic networks

Variable rate is a significant property of many networks nowadays due to traffic congestion, link failures, etc. To adapt to variable source rate in networks, variable-rate broadcast network coding is of the essence. Many efforts have been put into the issue in cyclic networks which contributes to plenty of achievements with some limitation. In the paper, we propose a new construction algorithm of variable-rate convolutional broadcast to adjust heterogeneous cyclic networks to possible link failures. Different from constructing a new local encoding kernel in the non-source node with other approaches, our algorithm only needs to eliminate the last k rows of global encoding kernels when link failures happen. The construction algorithm is explicitly adopted within polynomial time.