| It has long been believed that physical layer primitives extracted from wireless channel are not breakable given certain physical constraints. This belief is built upon the old faith in wireless channel's uncontrollability and unpredictability, and it is a cornerstone of physical layer security. Through our study, however, we find that the trust on an unauthenticated wireless channel's integrity is not valid. Here we propose various attacks and counterattacks on physical layer primitives including physical features about the environment and secret keys, and provide strong evidence that physical layer security solutions which purely rely on extracting these features from wireless channel need to be redesigned.;In the thesis, we start with our finding that an in-band full-duplex forwarder can change wireless channel in fine granularity. We propose PhyCloak, which is a fine-grained channel controller. It works against a popular set of applications that relies on physical primitive extracted from the channel called communication based sensing. This work is the first to counter the threat of unwanted or even malicious communication based sensing: it proposes a blackbox sensor obfuscation technique which distorts only the physical information in the communication signal that leaks privacy. Moreover, the design allows coupling the PhyCloak module with legitimate sensors, so that their sensing is preserved, while that of illegitimate sensors is obfuscated.;In the second part of the thesis, we focus on another popular set of applications that rely on physical primitives: channel based secret sharing. First, we introduce a channel based secret key extraction approach called Puzzle. In this part, we show how physical primitives can be extracted from the wireless channel, and be exploited to build a security solution. In addition, we will see more clearly what are the difficulties to break such channel based secret key extraction approaches. Next, in an effort to preserve channel reciprocity, we propose a channel controller called Channel Spoofer to force the two communicating ends to produce highly predictable secrets. This work is the first to demonstrate breaking of all extant channel based secret key extraction protocols, with no unrealistic assumptions on the environment. In the end, we propose a scheme to overcome the vulnerability of unauthenticated wireless channel by combining both physical layer solutions and traditional cryptographic solutions. We present a generic hop-to-hop secure communication protocol atop a physical layer that adapts various physical layer parameters.;We believe that given the vulnerability revealed by the proposed channel controllers, systems which purely rely on secret primitives extracted from physical layer are not secure anymore, and therefore, combining traditional cryptographic primitives and physical primitive together is more sound. |
