A censor using latency analysis to classify decoy routing sessions achieves a maximum F-score that drops to nearly 0 when the base rate of decoy routing falls below 10^-4 (one in 10,000 connections). Even at higher adoption rates the F-score remains below 0.5 for most overt sites, making reliable detection infeasible without unacceptable false-positive rates on legitimate traffic.

For China (a highly connected, routing-capable adversary), the gossip protocol combined with any symmetric decoy routing design requires only 5 heavyweight downstream stations plus 880 lightweight upstream gossip stations — versus 880 heavyweight stations for purely symmetric designs. Five downstream stations alone impact 78% of routes from Chinese users, while a single downstream station already covers nearly 25% of traffic.

Table 2 / §3.2 evaluation

An asymmetric gossip protocol adds only 1.0055× bandwidth overhead for n=5 downstream stations — approximately 11 Mb/s extra on a typical 2 Gb/s OC48 link. Upstream gossip stations require no in-line blocking and impose zero additional load on overt sites, making them substantially lighter than heavyweight symmetric relay stations that must check every TLS connection for steganographic tags.

§3.3 evaluation

Between 80% and 90% of internet routes are asymmetric, with only about 10% of flows symmetric in Tier-1 (backbone) networks and roughly 70% symmetric at the network edge. This asymmetry makes decoy routing systems requiring relay stations on both upstream and downstream paths impractical for the majority of real-world deployments.

§3 evaluation

Decoy routing systems that re-encrypt TLS application data across the relay station (Slitheen, Rebound, Waterfall) are vulnerable to nonce-reuse attacks: an adversary capable of observing traffic on both sides of the relay can exploit reused GCM nonces to decrypt or modify covert traffic. Although this falls outside the standard decoy routing threat model, it poses a concrete risk to users already under heightened surveillance who face adversaries with broad network visibility.

§5.2 detection

A three-stage detection pipeline exploiting the "dual-role" behavioral fingerprint of single-IP circumvention relays achieved 23.2% recall (96/414 ground-truth relays) with a 0.18% false-positive rate against 97,651 benign TLS servers, for an overall accuracy of 99.5%. The ground-truth set covered OpenVPN, WireGuard, and SOCKS relays identified in a 17 TB single-day backbone trace (WIDE Project, April 9, 2025).

2026-almutairi-server §3.4, Table 1 traffic-shapedpiflow-correlation

NATA (Non-invasive Active Traffic-correlation Analysis) injects low-frequency bandwidth waveforms (sinusoidal, square-wave, triangular) into Tor TCP connections at an upstream gateway without endpoint compromise, payload decryption, or Tor-browser modification. BM-Net, a selective state-space classifier trained on the exit-side observations, achieves a 99.65% binary detection F1 score distinguishing watermarked from natural traffic on a 20,000-trace real-world dataset.

2026-fan-activeflowmark-assessing-tor §IV, §VI.D, Table III flow-correlationtraffic-shapeml-classifier

BM-Net achieves a 99.65% binary detection F1 score for distinguishing bandwidth-watermarked Tor flows from natural traffic, outperforming all evaluated baselines (next best: TikTok at 75.96% F1). The accuracy gap stems from active perturbation imposing a deterministic low-frequency throughput constraint rather than relying on subtle natural metadata, making the detection task fundamentally easier than passive website fingerprinting.

2026-fan-activeflowmark-assessing-tor §VI-D, Table III–IV flow-correlationtraffic-shapeml-classifier

BM-Net achieves a 99.65% binary detection F1 score distinguishing watermarked from natural Tor flows, and a 97.5% macro-F1 score for fine-grained modulation classification across sinusoidal, square-wave, and triangular patterns. The fine-grained test set contains 201 held-out samples collected from ten clients across five geographic regions (Europe, North America, Australia, Southeast Asia, East Asia), with training traces including traffic collected under WTF-PAD and Walkie-Talkie defenses.

2026-fan-activeflowmark-assessing-tor §VI-D, Table III, Table V flow-correlationml-classifiertraffic-shape

Active bandwidth perturbation has an inherent detectability–stability trade-off: overly aggressive low-rate phases cause Tor SENDME-based flow control stalls, retransmissions, timeouts, or circuit replacement before sufficient correlation evidence is collected. The paper selects a 30-second modulation period and an empirically determined minimum shaping rate; the usable shaping range varies with relay load, path length, TCP congestion control behavior, and Tor multiplexing.

2026-fan-activeflowmark-assessing-tor §VII-A flow-correlationtraffic-shape

BM-Net achieves a 97.5% macro-F1 score for fine-grained classification of three modulation geometries (sinusoidal, square-wave, triangular) from noisy exit-side Tor observations using only 201 labeled test samples collected across cross-continental Tor paths. Residual errors concentrate between natural traffic and square-wave modulation, as abrupt low-rate transitions are partially smoothed by Tor multiplexing and network jitter.

2026-fan-activeflowmark-assessing-tor §VI-D, Table V flow-correlationml-classifiertraffic-shape