What should we learn from... Secure Communications via Physical-Layer and Information-Theoretic Techniques

The wireless revolution is rightfully hailed for facilitating massive data exchanges, ranging from conversations, text, and e-mail at the personal level, to financial information, utility resource allocation, support for emergency services, and medical diagnostics at the institutional level. Such connectivity also facilitates security breaches, ranging from passive eavesdropping to active Byzantine attacks, due to the open nature of wireless channels. A future characterized by an explosion of network connected devices, comprising smart grids, smart transportation systems, smart health, and more generally the Internet of Things, opens new doors to potentially disastrous scenarios if security is not built in from the ground up: The allure of instant gratification from anywhere anytime connectivity must be tempered by the risks of unprotected exchanges with infected or harmful agents.

An intriguing result by Wyner from the 1970s showed how secure communication may be achieved at the physical layer using coding, by exploiting channel impairments that constitute a physical reality of communication systems. The standard setting is the so-called wiretap channel: This scheme has a sender (‘‘Alice’’) encode information in a manner that allows the legitimate receiver (‘‘Bob’’) to reliably decode the message, yet hides information from an eavesdropper (‘‘Eve’’). The beauty of the scheme is twofold: 1) it is keyless, in that Alice and Bob may dispense with the need to share a secret key prior to message transmission yet still achieve secure message transmission, even when Eve knows all details about the code employed; and 2) it offers stronger secrecy than standard cryptography, as it appeals to the perfect secrecy condition established by Shannon in the 1940s, sometimes called information-theoretic secrecy. Perfect secrecy, in its simplest terms, is achieved when the intercepted communication conveys no more information on the transmitted message than a random guess. In such a scenario, any computational advantage of an adversary proves irrelevant. This basic result has fueled much recent work reexamining the combination of coding and cryptographic primitives all operating at the physical layer, and appropriately dubbed ‘‘physical-layer security.’’ The intent of the special issue of the Proceedings of the IEEE, October 2015, is to highlight recent advances in the field along with remaining challenges.