Distributed gabidulin codes for multiple-source network error correction

We consider the problem of noncoherent error correction in multiple-source multicast networks. We show that a distributed network error-correcting code can be designed as a subcode of a Gabidulin code. This allows the destination nodes to utilize any decoder designed using rank-metric. We show that this achieves the full capacity region of any network with up to three messages, with optimal scaling in field size.