Performance evaluation of constant dimension subspace code for error and erasures correcting in linear network coding

In random linear network coding (R.L.N.C.), intermediate node has the opportunity to sent a packet that is a random F-linear combination of packet it received. Thus, a corrupt packet may contaminate other packets when coded with them and a phenomenon of error propagation was introduced. Coding theories for error correcting started with N.Cai and R.Yeung and a code construction was proposed by Kotter and Kschischang (said KK code) in the context of a noncoherent transmission model for random linear network coding that captures the effects both of errors and of erasures. In this paper, we evaluate the performance of such constant dimension subspace code for random and deterministic network coding in terms of packet error rate (P.E.R.).