Abstract
As wireless networks evolve toward 5G and beyond, the dual requirements of Physical Layer Security (PLS) and low-latency Quality of Service (QoS) often come into conflict. This paper introduces the concept of Effective Secrecy Rate (ESR) to characterize secure communication in Non-Orthogonal Multiple Access (NOMA) networks under strict delay constraints. By merging secrecy rate analysis with Effective Capacity (EC) theory, the authors provide a mathematical framework to determine the maximum secure data rate achievable when packets must meet specific statistical delay targets.
Introduction
Non-Orthogonal Multiple Access (NOMA) is a cornerstone technology for enhancing spectral efficiency in modern networks. However, its reliance on Successive Interference Cancellation (SIC) creates inherent security vulnerabilities; users with superior channel conditions must decode the signals of "weaker" users, making them potential internal eavesdroppers. While traditional Shannon secrecy capacity provides a theoretical upper bound for security, it fails to account for the queuing delays inherent in fading channels. This research addresses this gap by evaluating how delay sensitivity impacts the secrecy performance of NOMA systems.
Methodology
The authors define the Effective Secrecy Rate (ESR) as the maximum constant arrival rate that can be securely served while satisfying a predefined delay outage probability. The study employs a downlink NOMA model consisting of a Base Station (BS), multiple legitimate users, and potential eavesdroppers.
Two primary scenarios are analyzed:
- Internal Eavesdropper: A scenario where every NOMA user is treated as a potential passive eavesdropper.
- External Eavesdropper: A scenario where NOMA users are trusted, but a malicious outsider attempts to intercept the transmission.
The mathematical framework utilizes Order Statistics to rank users by channel gain, allowing for the derivation of closed-form ESR expressions, particularly in the high Signal-to-Noise Ratio (SNR) regime. Delay stringency is modeled using a delay exponent ($\theta$), where a higher value indicates a lower tolerance for latency.
Key Insights and Results
The research reveals a fundamental trade-off between delay and secrecy: the ESR is a monotonically decreasing function of the delay exponent. As the requirement for "fresh" data increases, the amount of data that can be transmitted securely decreases.
Impact of SNR and User Density
A critical finding is the behavior of the ESR at high SNR. Unlike traditional capacity, the ESR for the strongest user approaches a constant ceiling rather than growing indefinitely with power. This saturation occurs because increasing transmit power simultaneously improves the signal quality for both the legitimate user and the eavesdropper.
Furthermore, the size of the NOMA user cluster ($M$) significantly impacts performance. As shown in Figure 1, increasing the number of users in the cluster reduces the ESR. This is attributed to the increased probability that an eavesdropper will possess a channel gain comparable to the legitimate user, thereby narrowing the secrecy gap.
The Role of Channel State Information (CSI)
The study also highlights that for external eavesdroppers, the absence of Channel State Information (CSI) at the attacker allows legitimate users to maintain positive secure rates even under strict delay constraints. In internal scenarios, however, security is more precarious, with only the user with the absolute best channel gain typically achieving a positive ESR.
Conclusion
This work demonstrates that while NOMA networks face unique security challenges due to their "leaky" nature, they can still provide robust physical layer security at high SNR levels. The introduction of the ESR metric provides a vital tool for network designers to balance the competing demands of security and low-latency communication. Future work may explore optimal power allocation strategies and the impact of imperfect CSI or multi-antenna configurations on the ESR.
Summary updated