Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE Transactions on Information Theory

On Secrecy Capacity Scaling in Wireless Networks

O. Ozan Koyluoglu, C. Emre Koksal, Hesham El Gamal

Published Jan 31, 2012• 188 citations
Publisher

Abstract

This paper studies the achievable secure rate per source-destination pair in wireless networks. First, a path loss model is considered, where the legitimate and eavesdropper nodes are assumed to be placed according to Poisson point processes with intensities λ and λe, respectively. It is shown that, as long as λe/λ =o((logn)-2), almost all of the nodes achieve a perfectly secure rate of Ω(1/√n) for the extended and dense network models. Therefore, under these assumptions, securing the network does not entail a loss in the per-node throughput. The achievability argument is based on a novel multihop forwarding scheme where randomization is added in every hop to ensure maximal ambiguity at the eavesdropper(s). Second, an ergodic fading model withnsource-destination pairs and neeavesdroppers is considered. Employing the ergodic interference alignment scheme with an appropriate secrecy precoding, each user is shown to achieve a constant positive secret rate for sufficiently largen. Remarkably, the scheme does not require eavesdropper CSI (only the statistical knowledge is assumed) and the secure throughput per node increases as we add more legitimate users to the network in this setting. Finally, the effect of eavesdropper collusion on the performance of the proposed schemes is characterized.

DOI
10.1109/tit.2012.2184692

Related papers

On Secrecy Capacity Scaling in Wireless Networks

O. Ozan Koyluoglu, C. Emre Koksal, Hesham El Gamal

This work studies the achievable secure rate per source-destination pair in wireless networks. First, a path loss model is considered, where the legitimate and eavesdropper nodes are assumed to be placed according to Poisson point processes with intensities $λ$ and $λ_e$, respectively. It is shown that, as long as $λ_e/λ=o((\log n)^{-2})$, almost all of the nodes achieve a perfectly secure rate of $Ω(\frac{1}{\sqrt{n}})$ for the extended and dense network models. Therefore, under these assumptions, securing the network does not entail a loss in the per-node throughput. The achievability argument is based on a novel multi-hop forwarding scheme where randomization is added in every hop to ensure maximal ambiguity at the eavesdropper(s). Secondly, an ergodic fading model with $n$ source-destination pairs and $n_e$ eavesdroppers is considered. Employing the ergodic interference alignment scheme with an appropriate secrecy pre-coding, each user is shown to achieve a constant positive secret rate for sufficiently large $n$. Remarkably, the scheme does not require eavesdropper CSI (only the statistical knowledge is assumed) and the secure throughput per node increases as we add more legitimate users to the network in this setting. Finally, the effect of eavesdropper collusion on the performance of the proposed schemes is characterized.

arXiv:0908.0898•Aug 6, 2009•8 citations

On secrecy capacity scaling in wireless networks

O. Ozan Koyluoglu, C. Emre Koksal, Hesham El Gamal

We study a random extended network, where the legitimate and eavesdropper nodes are assumed to be placed according to Poisson point processes in a square region of area n. It is shown that, when the legitimate nodes have unit intensity, ¿ = 1, and the eavesdroppers have an intensity of ¿e= O((log n)-2), almost all of the nodes achieve a perfectly secure rate of ¿ (1/¿n). The achievability argument is based on a novel multi-hop forwarding scheme where randomization is added in every hop to ensure maximal ambiguity at the eavesdropper(s). Remarkable, under these assumptions, securing the transmissions of nodes does not entail a loss in the per-node throughput in terms of scaling.

Information Theory and Applications Workshop (ITA)•Jan 1, 2010•46 citations

On the effect of colluding eavesdroppers on secrecy capacity scaling

O. Ozan Koyluoglu, C. Emre Koksal, Hesham El Gamal

In a powerful secrecy attack, eavesdroppers can collude, i.e., they can share their observations. Securing information in such a scenario will be an even more challenging task compared to non-colluding case. We here analyze the effect of eavesdropper collusion on the achievable performance in both the path loss and ergodic multi-path fading models. We provide two results: 1) For the Poisson point process model in a random extended network, if the legitimate nodes have unit intensity (λ = 1) and the colluding eavesdroppers have an intensity of λe= O((log n)-(2+p)) for any p > 0, almost all of the nodes can achieve a secure rate of Ω(1/(√n)); I and 2) In the K-user Gaussian interference channel with E external colluding eavesdroppers, a secure degrees of freedom of η =[1/2-E/K]+per frequency-time slot is achievable for each user in the ergodic setting (in the absence of the eavesdropper channel state information).

European Wireless Conference•Jan 1, 2010•12 citations

Security-capacity trade-off in large wireless networks using keyless secrecy

Sudarshan Vasudevan, Dennis Goeckel, Don Towsley

We investigate the scalability of a class of algorithms that exploit the dynamics of wireless fading channels to achieve secret communication in a large wireless network of n randomly located nodes. We describe a construction in which nodes transmit artificial noise to suppress eavesdroppers whose locations are unknown and ensure secrecy of messages transported across the network. Under a model in which eavesdroppers operate independently and under appropriate conditions on the achievable per-node throughput Ψ(n), we show that the network can tolerate Ω((1⁄√n(n))2c) eavesdroppers while ensuring that the aggregate rate at which eavesdroppers intercept packets goes to 0, where c is a constant such that 0 0. We also establish sufficient conditions on the number of eavesdroppers to achieve a non-zero throughput in our construction.

ACM International Symposium on Mobile Ad Hoc Networking and Computing (MobiHoc)•Sep 20, 2010•85 citations

Cooperation for Secrecy in Wireless Networks

Frédéric Gabry

The growth of wireless networks has been considerable over the last decades. Due to the broadcast nature of these networks, security issues have taken a critical role in today’s communications. A promising direction towards achieving secure communications is information theoretic secrecy, which is an approach exploiting the randomness of the channels to ensure secrecy. Based on this approach, there has been a recent surge of interest in a potential cooperation between users to enhance the secrecy of communications. In this thesis we investigate the interaction between cooperation and secrecy. In particular the contributions of the thesis can be divided into two parts. In the first part, we study cooperative strategies for secrecy for wireless channels. Our goal is to evaluate the effect of fading and limited CSI on the eavesdropper’s channels. In that purpose we consider a scenario where a helper aims at increasing the secrecy of the communication between a source and destination in the presence of an eavesdropper. Several strategies are discussed for the helper, namely decode-and-forward, amplify-and-forward, and cooperative jamming. We introduce the secrecy outage probability, the conditional secrecy outage probability and the secure throughput as secrecy measures. For each measure, we investigate and compare the secrecy performance of cooperation. We furthermore elaborate a system optimization in terms of strategy selection, node positioning, power allocation and rate design. In the second part, we consider cooperation in the 4-node scenario against a more sophisticated adversary: an active eavesdropper, which can either passively eavesdrop, or jam the transmission. A game-theoretic perspective is a natural way to analyze the competitive interaction between the helper and the eavesdropper. Therefore we define several secrecy games, for which we find the Nash and Stackelberg equilibria as well as the corresponding secrecy rate outcomes. Another important consideration in this scenario is the interaction between the source and the helper, which we model and solve as a Stackelberg game, and we illustrate its impact on the achievable secrecy rates.

•Jan 1, 2012•5 citations

Secrecy transmission capacity of decentralized wireless networks

Xiangyun Zhou, Radha Krishna Ganti, Jeffrey G. Andrews +1 more

Secure communication over large-scale decentralized wireless networks is an extremely challenging task due to the cost and difficulty in establishing secret keys among all the nodes in a distributed manner. For this reason, the notion of physical layer security has recently drawn significant attention, which may assist with key exchange and provide an additional layer of protection in such networks. In this paper, we investigate how the physical layer security constraints affect the network throughput. We consider a random network in which the legitimate and eavesdropper nodes are located according to independent Poisson point processes. We introduce a new metric "secrecy transmission capacity" to characterize the network throughput in terms of the area spectral efficiency of secure transmissions, subject to constraints on both the quality of service and the level of security. This capacity framework allows us to quantitatively study the throughput cost of physical layer security constraints. We observe that the throughput cost of achieving a moderate level of security is quite low, while throughput must be significantly sacrificed to realize a highly secure network.

Allerton Conference on Communication, Control, and Computing•Sep 1, 2011•4 citations
This paper studies the achievable secure rate per source-destination pair in wireless networks. First, a path loss model is considered, where the legitimate and eavesdropper nodes are assumed to be placed according to Poisson point processes with intensities λ and λe, respectively. It is shown that, as long as λe/λ =o((logn)-2), almost all of the nodes achieve a perfectly secure rate of Ω(1/√n) for the extended and dense network models. Therefore, under these assumptions, securing the network does not entail a loss in the per-node throughput. The achievability argument is based on a novel multihop forwarding scheme where randomization is added in every hop to ensure maximal ambiguity at the eavesdropper(s). Second, an ergodic fading model withnsource-destination pairs and neeavesdroppers is considered. Employing the ergodic interference alignment scheme with an appropriate secrecy precoding, each user is shown to achieve a constant positive secret rate for sufficiently largen. Remarkably, the scheme does not require eavesdropper CSI (only the statistical knowledge is assumed) and the secure throughput per node increases as we add more legitimate users to the network in this setting. Finally, the effect of eavesdropper collusion on the performance of the proposed schemes is characterized.
This work studies the achievable secure rate per source-destination pair in wireless networks. First, a path loss model is considered, where the legitimate and eavesdropper nodes are assumed to be placed according to Poisson point processes with intensities λλλ and λeλ_eλe​, respectively. It is shown that, as long as λe/λ=o((log⁡n)−2)λ_e/λ=o((\log n)^{-2})λe​/λ=o((logn)−2), almost all of the nodes achieve a perfectly secure rate of Ω(1n)Ω(\frac{1}{\sqrt{n}})Ω(n​1​) for the extended and dense network models. Therefore, under these assumptions, securing the network does not entail a loss in the per-node throughput. The achievability argument is based on a novel multi-hop forwarding scheme where randomization is added in every hop to ensure maximal ambiguity at the eavesdropper(s). Secondly, an ergodic fading model with nnn source-destination pairs and nen_ene​ eavesdroppers is considered. Employing the ergodic interference alignment scheme with an appropriate secrecy pre-coding, each user is shown to achieve a constant positive secret rate for sufficiently large nnn. Remarkably, the scheme does not require eavesdropper CSI (only the statistical knowledge is assumed) and the secure throughput per node increases as we add more legitimate users to the network in this setting. Finally, the effect of eavesdropper collusion on the performance of the proposed schemes is characterized.
We study a random extended network, where the legitimate and eavesdropper nodes are assumed to be placed according to Poisson point processes in a square region of area n. It is shown that, when the legitimate nodes have unit intensity, ¿ = 1, and the eavesdroppers have an intensity of ¿e= O((log n)-2), almost all of the nodes achieve a perfectly secure rate of ¿ (1/¿n). The achievability argument is based on a novel multi-hop forwarding scheme where randomization is added in every hop to ensure maximal ambiguity at the eavesdropper(s). Remarkable, under these assumptions, securing the transmissions of nodes does not entail a loss in the per-node throughput in terms of scaling.
In a powerful secrecy attack, eavesdroppers can collude, i.e., they can share their observations. Securing information in such a scenario will be an even more challenging task compared to non-colluding case. We here analyze the effect of eavesdropper collusion on the achievable performance in both the path loss and ergodic multi-path fading models. We provide two results: 1) For the Poisson point process model in a random extended network, if the legitimate nodes have unit intensity (λ = 1) and the colluding eavesdroppers have an intensity of λe= O((log n)-(2+p)) for any p > 0, almost all of the nodes can achieve a secure rate of Ω(1/(√n)); I and 2) In the K-user Gaussian interference channel with E external colluding eavesdroppers, a secure degrees of freedom of η =[1/2-E/K]+per frequency-time slot is achievable for each user in the ergodic setting (in the absence of the eavesdropper channel state information).
We investigate the scalability of a class of algorithms that exploit the dynamics of wireless fading channels to achieve secret communication in a large wireless network of n randomly located nodes. We describe a construction in which nodes transmit artificial noise to suppress eavesdroppers whose locations are unknown and ensure secrecy of messages transported across the network. Under a model in which eavesdroppers operate independently and under appropriate conditions on the achievable per-node throughput Ψ(n), we show that the network can tolerate Ω((1⁄√n(n))2c) eavesdroppers while ensuring that the aggregate rate at which eavesdroppers intercept packets goes to 0, where c is a constant such that 0 0. We also establish sufficient conditions on the number of eavesdroppers to achieve a non-zero throughput in our construction.
The growth of wireless networks has been considerable over the last decades. Due to the broadcast nature of these networks, security issues have taken a critical role in today’s communications. A promising direction towards achieving secure communications is information theoretic secrecy, which is an approach exploiting the randomness of the channels to ensure secrecy. Based on this approach, there has been a recent surge of interest in a potential cooperation between users to enhance the secrecy of communications. In this thesis we investigate the interaction between cooperation and secrecy. In particular the contributions of the thesis can be divided into two parts. In the first part, we study cooperative strategies for secrecy for wireless channels. Our goal is to evaluate the effect of fading and limited CSI on the eavesdropper’s channels. In that purpose we consider a scenario where a helper aims at increasing the secrecy of the communication between a source and destination in the presence of an eavesdropper. Several strategies are discussed for the helper, namely decode-and-forward, amplify-and-forward, and cooperative jamming. We introduce the secrecy outage probability, the conditional secrecy outage probability and the secure throughput as secrecy measures. For each measure, we investigate and compare the secrecy performance of cooperation. We furthermore elaborate a system optimization in terms of strategy selection, node positioning, power allocation and rate design. In the second part, we consider cooperation in the 4-node scenario against a more sophisticated adversary: an active eavesdropper, which can either passively eavesdrop, or jam the transmission. A game-theoretic perspective is a natural way to analyze the competitive interaction between the helper and the eavesdropper. Therefore we define several secrecy games, for which we find the Nash and Stackelberg equilibria as well as the corresponding secrecy rate outcomes. Another important consideration in this scenario is the interaction between the source and the helper, which we model and solve as a Stackelberg game, and we illustrate its impact on the achievable secrecy rates.
Secure communication over large-scale decentralized wireless networks is an extremely challenging task due to the cost and difficulty in establishing secret keys among all the nodes in a distributed manner. For this reason, the notion of physical layer security has recently drawn significant attention, which may assist with key exchange and provide an additional layer of protection in such networks. In this paper, we investigate how the physical layer security constraints affect the network throughput. We consider a random network in which the legitimate and eavesdropper nodes are located according to independent Poisson point processes. We introduce a new metric "secrecy transmission capacity" to characterize the network throughput in terms of the area spectral efficiency of secure transmissions, subject to constraints on both the quality of service and the level of security. This capacity framework allows us to quantitatively study the throughput cost of physical layer security constraints. We observe that the throughput cost of achieving a moderate level of security is quite low, while throughput must be significantly sacrificed to realize a highly secure network.