FEC in optical communications - A tutorial overview on the evolution of architectures and the future prospects of outband and inband FEC for optical communications

There are three representative architectures of 1/2/3-g outband and inband forward error correction (FEC) schemes. The first generation outband FEC is Reed-Solomon (RS) codes which are Maximum Distance Separable (MDS) and suitable for burst-form errors, because of their nonbinary structure. An extensive experience on the corrective properties of RS codes (compact disks, wireless/satellite links, deep space missions) made them a natural first choice for the enhancement of optical links. A widely used approach is RS coding is the code-symbol interleaving of a number of individual RS-codes. The reason for interleaving is the distribution of channel errors to all partial codes in the FEC scheme so that no single one of them is overwhelmed by these errors. The second generation came from the maturity of DWDM systems, diverse FEC-concentration schemes made their appearance in the market. The clear performance-advantage over classic FEC and the common element of serial FEC concatenation put these schemes under the umbrella of second generation FEC systems for optical communications. It was soon realized that second generation FEC schemes don't achieve the required system margin for some important ULH links. Third-generation FEC mostly rely on Turbo and low-density parity-check (LDPC) coding concepts, they also leverage iterative decoding to obtain realizable receivers and achieve very high coding-gains.