Network coding-based protection

This paper serves as a tutorial lecture on the use of network coding to provide resource efficient and agile proactive protection. Network coding, which was introduced in Ahlswede et al. (2000) [1], allows intermediate network nodes to form linear combinations of packets received on different input links. The use of network coding results in capacity enhancement. This capacity enhancement is used to provide protection channels which are used to carry combinations of redundant data, and are solved by the receivers in order to recover data lost due to network failures.The paper starts by addressing network coding-based protection of bidirectional unicast connections, and explains the use of p-Cycles to carry linear combinations of the redundant data units. The paper also discusses an earlier protection strategy which is based on diversity coding, in which the linear combinations are formed at special nodes, including sources, and is used to protect unidirectional connections. A generalized network coding-based protection which uses a tree to carry the linear combinations will be presented. Protection of multicast connections using network coding is also explained.