Information-theoretic security is the science of safeguarding information across a communication network based on the use of concepts and techniques in information theory. It encompasses protection against eavesdropping, impersonation, and substitution attacks made by potential adversaries present in the network. Whereas eavesdropping and impersonation attacks (and solutions) are well studied, substitution attacks, in which the adversary replaces messages from a source by different valid messages from the same source, present new risks in an increasingly networked world. This work explores how it is possible to make use of side information that is often available in a typical communication network to check against substitution attacks made directly on raw messages by potential adversaries in the network. This frees the process of guaranteeing information integrity of the messages from the requirements of secrecy and authentication, and better isolates the mathematical structure that leads to information integrity in relation to the network topology.
The separation principle developed in this work then allows one to more easily incorporate secrecy and authentication back into the overall information-theoretic security design. The broader implications of information-theoretic network security are increasingly manifest in electronic commerce, reliable communications, and confidentiality of legal, proprietary, diplomatic, and medical information.