Throughput Scaling of Covert Communication over Wireless Adhoc Networks

Kang-Hee Cho, Si-Hyeon Lee, Vincent Y. F. Tan

Abstract

This paper investigates the fundamental limits of throughput scaling in wireless adhoc networks under strict covertness constraints. We consider a network of $n$ legitimate nodes (LNs) and $n^{\kappa}$ ($0 < \kappa < 1$) non-communicating warden nodes (WNs) randomly distributed in a unit area. The objective is to maximize throughput while ensuring that WNs cannot detect the presence of communication. By introducing "preservation regions" around wardens and employing modified multi-hop (MH) and hierarchical cooperation (HC) schemes, we establish matching upper and lower bounds for throughput scaling across different physical environments.

Introduction

In traditional wireless networking, the primary goal is to maximize aggregate throughput as the number of nodes $n$ grows. However, in security-sensitive applications, communication must often be covert, meaning the very existence of a signal must be hidden from potential adversaries (wardens). While the "square-root law" for covert communication is well-established for point-to-point links—limiting transmission to $O(\sqrt{l})$ bits over $l$ channel uses—its application to large-scale adhoc networks introduces significant complexity. The challenge lies in balancing the signal strength required for multi-hop relaying with the need for near-silence to avoid detection by wardens distributed throughout the network.

Proposed Communication Architectures

To achieve covertness, the authors propose a framework that limits transmit power based on the proximity of wardens. The core of this strategy involves Preservation Regions: square "no-transmit" zones defined around every warden node. These regions prevent any LN from being caught by a nearby warden and allow LNs outside these regions to utilize higher transmit power than would otherwise be possible.

Illustration of the preservation region around a warden node (WN) and the resulting detour path for legitimate nodes (LN).

Because these preservation regions act as obstacles, the authors introduce two primary modified schemes:

  1. Detoured Multi-hop (MH): In standard MH, data travels along direct paths. In this covert model, data paths are modified to take a "detour" around preservation regions. To avoid creating bottlenecks, these detours are distributed evenly over a wider area, preventing any single node from bearing an excessive relaying burden.
  2. Modified Hierarchical Cooperation (HC): This scheme leverages nodes to form "virtual MIMO" arrays. To maintain covertness, the authors strictly control symbol power and schedule transmissions to ensure the aggregate signal at any warden remains statistically indistinguishable from background noise (measured via Kullback-Leibler divergence).

Key Results and Operating Regimes

The research identifies four distinct operating regimes for the network, primarily determined by the path loss exponent ($\alpha$) and the warden's observation capabilities.

Summary of throughput scaling regimes based on the path loss exponent $\alpha$ and the vanishing/non-vanishing nature of the short-range SNR.

The fundamental constraint on transmit power scales with the term $n^{(1/2 - \kappa/2)(\alpha-2)} / \sqrt{l}$, where $l$ is the number of channel uses available to the warden. Key findings include:

  • Weak Path Loss ($2 < \alpha \le 3$): The modified Hierarchical Cooperation scheme is optimal, as the benefits of virtual MIMO outweigh the detection risks when signals do not attenuate rapidly.
  • Strong Path Loss ($\alpha > 3$): The optimal scheme depends on the warden's observation window. If the warden has many channel uses to observe, the Detoured Multi-hop scheme becomes more efficient.
  • Power Boosting: Counter-intuitively, the exclusion of LNs from preservation regions is what enables the rest of the network to function. By ensuring no LN is "too close" to a warden, the entire network can lift its average transmit power without triggering an alarm.

Conclusion and Future Work

This work establishes that covert communication in adhoc networks follows a modified scaling law that accounts for warden density and path loss. While the results provide a rigorous theoretical foundation, they assume wardens do not cooperate and that their locations are known to the legitimate network. Future research may explore scenarios where warden locations are uncertain or where wardens can share information to improve their detection capabilities, which would likely impose even stricter limits on network throughput.

Summary updated

Abstract

We consider the problem of covert communication over wireless adhoc networks in which (roughly) $n$ legitimate nodes (LNs) and $n^{\kappa}$ for $0<\kappa<1$ non-communicating warden nodes (WNs) are randomly distributed in a square of unit area. Each legitimate source wants to communicate with its intended destination node while ensuring that every WN is unable to detect the presence of the communication. In this scenario, we study the throughput scaling law. Due the covert communication constraint, the transmit powers are necessarily limited. Under this condition, we introduce a preservation region around each WN. This region serves to prevent transmission from the LNs and to increase the transmit power of the LNs outside the preservation regions. For the achievability results, multi-hop (MH), hierarchical cooperation (HC), and hybrid HC-MH schemes are utilized with some appropriate modifications. In the proposed MH and hybrid schemes, because the preservation regions may impede communication along direct data paths, the data paths are suitably modified by taking a detour around each preservation region. To avoid the concentration of detours resulting extra relaying burdens, we distribute the detours evenly over a wide region. In the proposed HC scheme, we control the symbol power and the scheduling of distributed multiple-input multiple-output transmission. We also present matching upper bounds on the throughput scaling under the assumption that every active LN consumes the same average transmit power over the time period in which the WNs observe the channel outputs.
BibTeX
@inproceedings{cho2019throughput,
  author = {Kang-Hee Cho and Si-Hyeon Lee and Vincent Y. F. Tan},
  title = {{Throughput Scaling of Covert Communication over Wireless Adhoc Networks}},
  booktitle = {ISIT},
  year = {2019},
  doi = {10.1109/isit.2019.8849852},
}
Figure