[Paper Review] Information-Theoretic Security or Covert Communication

Moslem Forouzesh, Paeiz Azmi+3 more|arXiv (Cornell University)|Mar 18, 2018

Wireless Communication Security Techniques24 references8 citations

TL;DR

This paper investigates information-theoretic security and covert communication under a unified framework using channel distribution information (CDI) only. It formulates power allocation problems for secrecy and covert rates via Successive Convex Approximation (SCA) and convex relaxation, showing that covert communication achieves higher rates than secrecy when Bob is far from Alice, despite stricter detection constraints.

ABSTRACT

Information-theoretic secrecy, in particular the wiretap channel formulation, provides protection against interception of a message by adversary Eve and has been widely studied in the last two decades. In contrast, covert communications under an analogous formulation provides protection against even the detection of the presence of the message by an adversary, and it has drawn significant interest recently. These two security topics are generally applicable in different scenarios; however, here we explore what can be learned by studying them under a common framework. Under a similar but not identical mathematical formulation, we introduce power optimization problems for each of the secrecy and the covert communications scenario, and we exploit common aspects of the problems to employ similar tools in their respective optimizations. Moreover, due to the practical limitations, we assume only channel

Motivation & Objective

• To analyze and compare information-theoretic security and covert communication under the same 4-node wiretap channel model with a jammer.
• To address the practical challenge of optimizing secrecy and covert rates when only channel distribution information (CDI) is available, not full CSI.
• To quantify the performance trade-off between achieving secrecy (hiding message content) and covertness (hiding message existence) under realistic channel uncertainty.
• To evaluate the sensitivity of both schemes to the location of the eavesdropper (Eve) and the jammer, particularly in terms of achievable rates.

Proposed method

• Formulates an ergodic secrecy rate maximization problem under a transmit power constraint, using a tight lower bound on the ergodic secrecy rate to enable optimization.
• Applies the Successive Convex Approximation (SCA) method to iteratively solve the non-convex secrecy rate maximization problem under CDI.
• For covert communication, introduces an auxiliary variable to transform the non-convex optimization into a convex problem, enabling efficient solution via convex relaxation.
• Models the detection error probability at Eve (the warden) as a function of transmit power, distance, and path loss, with a constraint on the minimum detection error probability (e.g., 90%, 99%, 99.9%).
• Uses a Rayleigh fading channel model with path loss and shadowing, and assumes only CDI is available at the transmitter (Alice), not instantaneous CSI.
• Employs numerical simulations to evaluate the ergodic secrecy and covert rates across varying distances between Alice-Bob, Alice-Eve, and jammer-Eve.

Experimental results

Research questions

• RQ1 How does the achievable rate of covert communication compare to that of information-theoretic secrecy when only CDI is available?
• RQ2 What is the impact of increasing the required detection error probability at Eve (i.e., making detection harder) on the achievable covert rate?
• RQ3 How sensitive are the secrecy and covert rates to the location of Eve relative to Alice and Bob?
• RQ4 How does the performance of the proposed SCA and convex relaxation methods compare to exhaustive search in terms of optimality gap?
• RQ5 Under what conditions does covert communication outperform secrecy in terms of achievable rate?

Key findings

• When Bob is near Alice (e.g., 5 m), information-theoretic secrecy achieves a higher rate than covert communication due to lower interference and power constraints.
• When Bob is far from Alice (e.g., 7–10 m), the covert communication scheme outperforms secrecy, indicating a reversal of performance depending on distance.
• Increasing the required detection error probability at Eve from 90% to 99.9% reduces the covert rate by a factor of 1.76, showing a significant performance penalty for higher covertness requirements.
• The secrecy rate is more sensitive to the location of Eve than the covert rate; small changes in Eve’s distance to Alice significantly affect secrecy rate, while covert rate is more robust.
• The proposed SCA and convex relaxation methods achieve a small optimality gap compared to exhaustive search, validating their effectiveness in practical power allocation.
• The distance between the jammer and Eve has less impact on covert rate than the distance between Alice and Eve, indicating that Alice-Eve distance is the dominant factor in covertness.