A Practical Man-In-The-Middle Attack on Signal-Based Key Generation Protocols

Abstract

Generating secret keys using physical properties of the wireless channel has recently become a popular research area. The main security assumption of these protocols is that a sufficiently distant adversary is unable to guess a generated secret due to the unpredictable behavior of multipath signal propagation. In this paper, we introduce a practical and efficient man-in-the-middle attack against such protocols. Using this attack, we demonstrate: (i) intentional sabotaging of key generation schemes, which leads to a high key disagreement rate, and (ii) a key recovery that reveals up to 47 % of the generated secret bits. We analyze statistical countermeasures (often proposed in related work) and show that attempting to detect such attacks results in a high false positive rate, questioning the overall benefit of such schemes. We implement and experimentally validate the attacks using off-the-shelf hardware, without assuming any technological advantage for the adversary.

On the Security of Key Extraction From Measuring Physical Quantities

“…We note that our work focuses on demonstrating feasibility of analyzing security for physical layer key extraction from experimentally falsifiable assumptions; while our instantiation is practical, we leave for future work the further optimization of efficiency within our framework. Finally we stress that a more ambitious objective would be to provide a framework for security in the active adversarial model; indeed, in key extraction protocols certain types of active attacks have been demonstrated, e.g., [12], [36], [37]. While this is beyond the scope of the present work, a framework such as ours that provides a way to lower bound the conditional entropy available to the two transceivers can be a fundamental intermediate step towards a formal treatment of security in the active model.…”

Proximity-Based Security Techniques for Mobile Users in Wireless Networks

“…As shown in [17], existing key generation strategies that are based on the RSSI and channel impulse response (CIR) [23], [29]- [32] or the phase [33] are vulnerable to the manin-the-middle attacks. For instance, eavesdroppers can reveal 40% to 50% of the keys, and attackers can sabotage the key agreements with 95% confidence by injecting spoofing signals during less than 4% of the overall communication duration [17].…”

“…on the received signal strength (RSS) of a single radio source, many of the proximity tests suffer from the limited proximity range and the authentication accuracy is not high in both stationary and fast changing radio environments [14], [15]. Moreover, a recent study has shown that the RSSbased strategies are vulnerable to man-in-the-middle attacks [17]. To address this problem, Zheng et al have proposed a location tag-based proximity test, which exploits the contents of ambient radio signals to improve the authentication accuracy and provides flexible range control [16].…”

A Study of Injection and Jamming Attacks in Wireless Secret Sharing Systems

“…MiM in the form of injection type of attacks constitute one of the most serious limitations in SKG systems extracting secret keys from RSS measurements [5,11,12] (it is yet unknown whether this attack can be launched to systems using CSI or the phase of the received signal [13]). Various possible approaches have so far surfaced on how to launch injection attacks; in [5] the attack consisted in controlling the movement of intermediate objects in the wireless medium, thus generating predictable changes in the received RSS (e.g., by obstructing or not a LOS), while in [11] whenever similar channel envelope measurements were received from Alice and Bob, Mallory spoofed the SKG process by injecting a MiM signal W .…”

“…However, neither the optimality of employing constant jamming signals nor the scenario of an adversary with imperfect estimate of the main channel CSI were addressed. Furthermore, in [11] and it was shown that injection type of attacks allow an active adversary to act as a man in the middle (MiM) and potentially control (a large) part of the generated key. A simple heuristic approach to defend against injection type of attacks was presented in [12] by multiplying the received signals with independent zero-mean random signals, locally generated at the legitimate nodes.…”