Correcting Erasures and Errors in Random Network Coding

Rank-metric codes in matrix representation are used by Koetter, Kschischang, Silva in their theory of random network coding. They showed that the decoding procedure can be reduced to decoding of rank codes. In this paper, we analyzed situations under different conditions at the decoder and establish if there are errors only or some type of mixture of errors and erasures. For correcting we used Gabidulin decoding algorithms. In many situations this analysis helps to choose a suitable algorithm and to eliminate some of computing operations. An example is given.