Transmitter Selection for Secrecy Against Colluding Eavesdroppers with Backhaul Uncertainty

TL;DR

This paper introduces generalized secrecy analysis methods for multi-transmitter systems with backhaul uncertainty and colluding eavesdroppers, deriving ESR expressions and highlighting the benefits of backhaul activity knowledge.

Contribution

It proposes new transmitter selection schemes and a unified secrecy analysis framework applicable to both backhaul knowledge scenarios, enabling ESR evaluation.

Findings

Secrecy performance improves with backhaul activity knowledge.

Performance degrades as the number of eavesdroppers increases.

Asymptotic expressions reveal system parameter influences.

Abstract

Due to the exponential growth of interconnected devices and reduced cell coverage, beyond fifth-generation networks will be dense. Thus, instead of wired backhaul, wireless backhaul will be cost-effective and flexible. For security in multi-transmitter systems, sub-optimal and optimal transmitter selection schemes exist. However, including backhaul activity knowledge available (BKA) and backhaul activity knowledge unavailable (BKU) cases and transmitter selection schemes, there is no generalized secrecy analysis method. Moreover, evaluation of the ergodic secrecy rate (ESR) of the optimal selection schemes is impossible using existing solution approaches. To address these, we propose two sub-optimal and optimal transmitter selection schemes for a small-cell multi-transmitter system in BKU or BKA cases in the presence of multiple colluding eavesdroppers. We derive the distribution of the ratio of the destination channel SNR and eavesdropping channel SNR, thereby providing different secrecy performance metrics uniformly irrespective of selection schemes and BKU or BKA cases. Simplified asymptotic expressions are provided to elucidate the influence of the system parameters and of the backhaul reliability. We observe that the secrecy performance improves when the knowledge of backhaul link activity is utilized, and the improvement is most noticeable when the backhaul is highly unreliable. We also observe that while the secrecy performance degrades with an increasing number of eavesdroppers, neither the asymptotic saturation value of the secrecy outage probability nor the rate of improvement of the ESR with signal-to-noise-ratio depends on the number of eavesdroppers.