SecureMatch: Scalable Authentication and Key Relegation for IoT Using Physical-Layer Techniques

Conventional authentication approaches, whether based on (preconfigured) symmetric or asymmetric keys, cannot react timely to large-scale attacks on Internet of Things (IoT) devices because they need to manage (i.e., create, certify, distribute, and revoke) long lists of keys or device identities. In particular, existing key revocation schemes are inept at handling the extreme scale of IoT networks. In this paper, we propose a new concept called relegation to facilitate developing SecureMatch, a protocol to cope with scalability while incurring minimal time and/or space complexity. Relegation is like a fuzzy key revocation by which only the trustworthiness level of a vulnerable but not compromised device is reduced. Inspired by lightweight identity-based signatures, SecureMatch uses public features of a class of devices as their public keys to characterize and relegate all similar devices that share a common vulnerability. Further, the proposed protocol encodes those public features, which are often collected from different layers of the protocol stack, into a device tag, and then transmits this tag in the physical layer using a reliable preamble-embedding technique. SecureMatch has low overhead and enables an authenticator to quickly authenticate a set of features as long as the devices exhibit the valid credentials. The performance of tag extraction was evaluated through LabVIEW and USRP experiments using commodity devices as transmitters.