Fast and scalable secret key generation exploiting channel phase randomness in wireless networks

Abstract

Recently, there has been great interest in physical layer security techniques that exploit the randomness of wireless channels for securely extracting cryptographic keys. Several interesting approaches have been developed and demonstrated for their feasibility. The state-of-the-art, however, still has much room for improving their practicality. This is because i) the key bit generation rate supported by most existing approaches is very low which significantly limits their practical usage given the intermittent connectivity in mobile environments; ii) existing approaches suffer from the scalability and flexibility issues, i.e., they cannot be directly extended to support efficient group key generation and do not suit for static environments. With these observations in mind, we present a new secret key generation approach that utilizes the uniformly distributed phase information of channel responses to extract shared cryptographic keys under narrowband multipath fading models. The proposed approach enjoys a high key bit generation rate due to its efficient introduction of multiple randomized phase information within a single coherence time interval as the keying sources. The proposed approach also provides scalability and flexibility because it relies only on the transmission of periodical extensions of unmodulated sinusoidal beacons, which allows effective accumulation of channel phases across multiple nodes. The proposed scheme is thoroughly evaluated through both analytical and simulation studies. Compared to existing work that focus on pairwise key generation, our approach is highly scalable and can improve the analytical key bit generation rate by a couple of orders of magnitude.

On the Security of Key Extraction From Measuring Physical Quantities

“…Some other works made some further steps to analyze security. For instance, [23], [24] consider performing randomness tests but these are insufficient as security guarantees (such "random" sequences are not necessarily unpredictable). In [20] a type of attacks based on blind deconvolution is considered and it is argued experimentally that such attacks are unlikely to apply; however, this does not preclude other types of attacks.…”

Efficient Physical-Layer Secret Key Generation and Authentication Schemes Based on Wireless Channel-Phase

“…In the simulations, the signalto-noise ratio (SNR) is defined as 2 ℎ / 2 . The secret key capacity of the proposed secret key generation scheme is compared with the existing channelphase based secret key generation scheme [15] and channelgain based secret key generation scheme [14] in Figure 4. During channel probing, the channel condition of these three schemes is identical.…”

“…In addition, the broadcasting nature of the wireless channels causes the wireless communication channel easily to be eavesdropped on or intercepted by an adversary [1,2]. Therefore, the interest in exploiting the characteristics of the wireless channels at the PHYlayer to enhance and complement the conventional security mechanisms is growing, such as the secret key extraction from the characteristics of wireless channels [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] and PHYlayer authentication [20][21][22][23][24][25][26][27].…”

“…However, these methods suffer from scalability and low secret key generation rate (KGR) which is defined as the average amount of secret key bits produced in one measurement/second [10][11][12]. To resolve these issues, researchers exploit the channel state information (CSI) to extract secret keys [13][14][15]. Besides, to increase the KGR, the orthogonal frequency-division multiplexing (OFDM) [18,19] and multi-input-multi-output (MIMO) techniques in the PHY-layer are adopted.…”

Group Secret Key Generation in Wireless Networks: Algorithms and Rate Optimization

“…Since then, several tree-based algorithms have been developed to achieve the group secret key capacity for the multi-terminal pairwise independent network (PIN) [11]- [15]. On the other hand, effective group key generation algorithms have been proposed for wireless networks by exploiting channel characteristics in [28]- [30]. These algorithms are more practical for real systems at the expense of some scarification in the group key rate.…”