Analysis of Short Blocklength Codes for Secrecy

In this paper we provide secrecy metrics applicable to physical-layer coding techniques with finite blocklengths over Gaussian and fading wiretap channel models. Our metrics go beyond some of the known practical secrecy measures, such as bit error rate and security gap, so as to make lower bound probabilistic guarantees on error rates over short blocklengths both preceding and following a secrecy decoder. Our techniques are especially useful in cases where application of traditional information-theoretic security measures is either impractical or simply not yet understood. The metrics can aid both practical system analysis, and practical system design for physical-layer security codes. Furthermore, these new measures fill a void in the current landscape of practical security measures for physical-layer security coding, and may assist in the wide-scale adoption of physical-layer techniques for security in real-world systems. We also show how the new metrics provide techniques for reducing realistic channel models to simpler discrete memoryless wiretap channel equivalents over which existing secrecy code designs may achieve information-theoretic security.

💡 Research Summary

The paper tackles the practical problem of evaluating physical‑layer security for coding schemes that operate with short blocklengths over realistic Gaussian and fading wiretap channels. Traditional information‑theoretic secrecy metrics—perfect secrecy, weak secrecy, strong secrecy, and semantic secrecy—are asymptotic in nature and assume infinitely long codes, making them unsuitable for many modern low‑latency or IoT applications. Likewise, the widely used security‑gap metric, which is based on average bit‑error‑rate (BER) differences between the legitimate receiver and the eavesdropper, fails to capture the variability of error performance on a per‑packet basis.

To address these gaps, the authors introduce two novel secrecy measures that are built on the cumulative distribution function (CDF) of the block‑error rate rather than on its mean. The first metric, BE‑CDF₍bc₎ (Before‑code CDF), quantifies the probability that the raw error proportion (bits in error divided by block length) before the outer secrecy encoder exceeds a threshold of 0.5 − δ. The second metric, BER‑CDF₍ac₎ (After‑code CDF), does the same for the error proportion after the outer decoder. By focusing on specific percentiles (e.g., the 1 % or 10 % worst‑case blocks), these metrics provide probabilistic guarantees that even short blocks will appear essentially random to an eavesdropper, i.e., their error rate will be close to 0.5.

The paper illustrates the concepts with a simple concatenated scheme: a binary scrambler (outer code) followed by a BCH(127, 64) inner error‑correcting code. Simulations show how varying δ (0.05, 0.10, 0.15) changes the curves of Pr(ĤP_b > 0.5 − δ) as a function of Eve’s E_b/N_0. For δ = 0.15, the probability approaches one when Eve’s SNR is no better than 3 dB, indicating that the scrambler successfully propagates errors so that Eve’s decoded bits are essentially random. The authors also discuss the non‑independence of errors after decoding and propose corrections using Markov‑chain approximations and bootstrap resampling.

Beyond the toy example, the authors apply the CDF‑based metrics to a more sophisticated concatenated architecture that may include interleaving, puncturing, or multiple secrecy layers, with an inner LDPC or turbo code for reliability. By evaluating BE‑CDF₍bc₎ and BER‑CDF₍ac₎ at each stage, designers can allocate SNR margins and power budgets to meet a desired security percentile. Crucially, the paper shows how to map a continuous‑valued Gaussian or fading wiretap channel onto an equivalent discrete‑memoryless wiretap channel (DMC) by partitioning the SNR space and estimating transition probabilities. Once the DMC model is obtained, any existing strong‑secrecy code (e.g., polar or lattice‑based constructions) can be applied, thereby achieving information‑theoretic secrecy while still benefiting from the short‑block, high‑percentile guarantees provided by the CDF metrics.

In summary, the work contributes a practical, statistically rigorous framework for assessing and designing physical‑layer security in the short‑block regime. The CDF‑based metrics fill a void left by asymptotic information‑theoretic measures and average‑BER security gaps, offering designers concrete, probabilistic guarantees on a per‑packet basis. The paper also outlines future directions, including extensions to multi‑antenna, multi‑user scenarios, and hardware‑level validation, positioning the proposed methodology as a bridge between theoretical secrecy concepts and real‑world secure communication systems.