randomization   Traffic randomization / shaping

AnyTLS's default padding scheme operates across 8 levels (stop=8), with initial padding fixed at 30 bytes, small-data padding 100–400 bytes, and medium-to-large data padding chains of 400–500 bytes continuing through multiple 500–1000 byte segments. The 'c' (continue) marker allows multi-stage padding sequences within a single connection burst.

2026-anon-anytls-anytls-sing-box-2026 §2.3 defense

AnyTLS is a TLS-based proxy protocol maintained by the sing-box team, designed in 2024 and first released in the sing-box dev-next branch. Its core mechanism wraps arbitrary proxy traffic in standard TLS and applies a configurable padding scheme (Padding Scheme) to enhance traffic concealment while maintaining compatibility with standard TLS infrastructure.

2026-anon-anytls-anytls-sing-box-2026 §1.1–1.3 defense

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 evaluation

Padding-based client-side defenses including WTF-PAD and Walkie-Talkie are insufficient against active bandwidth perturbation: they reshape packet timing and burst structure but cannot remove the upstream rate limit imposed by the gateway shaper. BM-Net trained on a defense-aware dataset containing both undefended and WTF-PAD/Walkie-Talkie traces still achieves 99.65% F1, and the paper explicitly notes that 'client-side padding and burst reshaping may alter the logical traffic pattern, but they do not directly remove the rate limit imposed by the upstream bottleneck.'

2026-fan-activeflowmark-assessing-tor §VI.E.2, §VII detection

Adversarial pre-padding — prepending stochastic byte noise to packets — degrades ET-BERT encrypted traffic classification accuracy from >99% to 25.68%, exposing a structural vulnerability in all payload-byte-dependent detection systems. White-box adversarial attacks (Ayaka AH-MSI) additionally achieve evasion rates exceeding 99.5% against standard continuous-time sequence models via Manifold Shattering, where adversaries align malicious temporal distributions with benign baselines.

2026-ferrel-aegis-adversarial-entropy-guided §II, §VI-A detection

AEGIS, a flow-physics-only ML classifier using a Hyperbolic Liquid State Space Model evaluated on a 400GB adversarial corpus including VLESS Reality, GhostBear, and AMOI-morphed traffic, achieves F1-score 0.9952, 99.50% TPR, and 0.2141% FPR at 262.27 µs inference latency on an RTX 4090. The system discards all payload bytes and classifies traffic exclusively on 6-dimensional flow physics: packet size, inter-arrival time, directionality, TCP window size, TCP flags, and payload ratio.

2026-ferrel-aegis-adversarial-entropy-guided §V, Table III evaluation

Gaussian noise injection stress testing shows AEGIS maintains F1-scores of 0.9913 at 5% IAT noise and 0.9753 at 10% IAT noise, but degrades to 0.5939 at 15% Gaussian noise — establishing the 'Manifold Shattering Threshold.' The paper asserts that sustaining 15% IAT noise in practice corrupts the adversary's own C2 channel integrity, making this threshold operationally unachievable for high-throughput tunnels.

2026-ferrel-aegis-adversarial-entropy-guided §V-F detection

Routing-guided conditional aggregation (CA) that dynamically weights header versus payload contributions using per-sample MoE routing probabilities outperforms static fusion on all six datasets, demonstrating that the relative discriminative utility of headers versus payloads varies by application type — and that classifiers can adaptively shift reliance to whichever modality is less obfuscated.

2026-he-trafficmoe-heterogeneity-aware-mixture §III-E, §IV-B detection

Explicitly disentangling packet headers (structured, low-entropy) from encrypted payloads (high-entropy, stochastic) into separate MoE branches yields consistent gains across six datasets: 86.85% F1 on 120-class TLS 1.3 traffic (CSTNET-TLS), 97.88% F1 on USTC-TFC2016 malware/benign flows, and 92.65% F1 on imbalanced IoT traffic (CIC-IoT2022), demonstrating that headers and payloads carry fundamentally different and independently exploitable discriminative signals.

2026-he-trafficmoe-heterogeneity-aware-mixture §III-A, §IV-B, Tables II–III detection

An uncertainty-aware filtering (UF) mechanism quantifies per-token reliability via Shannon entropy of the cross-modal header–payload attention matrix, finding that encrypted payloads still contain low-entropy tokens with stable cross-modal alignment that serve as reliable classification anchors — demonstrating that nominally randomized byte streams retain exploitable low-entropy structure.

2026-he-trafficmoe-heterogeneity-aware-mixture §III-D detection

Incoming traffic is the dominant website fingerprinting leakage vector in Nym because the network requester does not apply Poisson-rate cover traffic by default. The Features In feature set achieves 72.1% random-forest accuracy on Nym Testnet — exceeding Tor's 70.2% — while outgoing Features Out leaks only 43.9% (vs. 65.0% for Tor). The combined Features In/Out (70.0%) provides no additional information beyond Features In alone, confirming the inbound channel is the bottleneck. Enabling Poisson-rate cover traffic on the network requester side (lambda_QR = 10-20 ms) lowers the Tik-Tok F1 score from 0.87 to 0.75-0.64 but raises latency overhead to 7.32-12.66x (Table 4, configs 10-11).

2026-joll-s-website-fingerprinting-nym §5.1, Table 3 detection

Increasing Nym's per-hop mixing delay (λM) counterintuitively worsens WF resistance: without cover traffic, raising λM to 100 ms raises Tik-Tok F1 from 0.91 (no mixing) to 0.98. Larger delays create a visible temporal gap between outgoing request bursts and incoming response bursts, making the Transposition feature set leak 95.5% classifier accuracy versus 88.5% with no delay at all.

2026-joll-s-website-fingerprinting-nym §5.2, Table 4, Table 5 evaluation

Routing web traffic through Nym's constant-rate traffic stream (POISS-OFF) reduces Tik-Tok F1 to 0.06, with all feature-set leakage at ≤9.5% accuracy (Table 8), at the cost of latency overhead 7.32 and bandwidth overhead 8.06 (Table 7). Nym's 2,413-byte packet size means each cover packet incurs 4.7× the overhead of a Tor cover cell, making constant-rate defenses significantly more expensive in this architecture. This matches near-ideal Tamaraw protection on Tor, against WTF4Nym's F1 = 0.39 at 3.78x latency and 4.22x bandwidth.

2026-joll-s-website-fingerprinting-nym §6, Tables 7–8 defense

WTF4Nym, a burst-aware cover-traffic defense adapted to Nym's Poisson-rate architecture, reduces the Tik-Tok attack F1 score from 0.87 to 0.39 at moderate overhead (latency overhead 3.78×, bandwidth overhead 4.22×). This substantially outperforms directly applying WTF-PAD to Nym (F1=0.87), applying FRONT to Nym (F1=0.63), and the F1=0.65 achieved by FRONT on Tor at comparable overhead, by learning per-network burst distributions and injecting fake bursts that match those statistics. FRONT on Tor reaches F1 = 0.65 at 2.79x latency and 1.84x bandwidth, and under WTF4Nym the Time and N-gram feature leakage fall to 26.6% and 29.0% (Table 8).

2026-joll-s-website-fingerprinting-nym §6.2, Table 7 defense

Without chunk-based padding, an XGBoost classifier identifies the target website from covert data-chunk sizes with 91% accuracy (Tranco top-100). Chunking at 2 MB reduces accuracy to 12% at a 21.3% bandwidth overhead, while 16 MB chunks reduce accuracy to near random guessing at a 480.3% overhead. Chunks as small as 64 KB already reduce accuracy to 64%, demonstrating a monotonic fingerprinting–overhead tradeoff.

2026-kamali-huma §V-C, Figure 4 evaluation

Huma's deferred-reply / double-request receive (DRR) protocol reduces a traffic-fingerprinting XGBoost classifier's accuracy to at most 54% (near random guessing) across geographically distributed clients (San Francisco, Frankfurt, Bangalore). A Kolmogorov-Smirnov test on absolute page-load timing distributions yields D=0.03, p=0.98 for U.S. clients — substantially tighter than Waterfall of Liberty's D=0.11 at p=0.5 — confirming that Huma flows are statistically indistinguishable from benign HTTPS fetches.

2026-kamali-huma §V-D, Table II evaluation

Striding with factor 4 (early downsampling) produces the largest single-factor degradation in the ablation study: average macro-F1 drops from 0.9909 to 0.9772 and cross-dataset variance increases from 4.77×10⁻⁵ to 4.51×10⁻⁴, with worst-case dataset performance falling to MIN 0.9524. Fine-grained byte order and short-range structure — protocol headers, payload signatures, repeated byte motifs — carry essential discriminative signal that stride-based aggregation destroys. This is a far larger degradation than any other design choice tested, including Mamba-1 versus Mamba-2.

2026-kulatilleke-mambanetburst-direct-byte-level §V-B, Table IV detection

Using the CBOR-based DNS format (application/dns+cbor) reduces DNS message sizes — CBOR DNS queries are at most 136 bytes versus classic DNS queries at 83 bytes for the 99th percentile — but smaller, more uniform packet sizes inadvertently improve classifier accuracy for DNS frame identification, counteracting obfuscation. Reducing packet size and reducing detectability are conflicting goals in packet-format obfuscation.

2026-lenders-secrets-best-not §2 / §4 / §6 detection

CBOR-based DNS encoding (application/dns+cbor) reduces DNS message sizes—CBOR DNS queries max out at 136 bytes vs. 150 bytes for classic DNS, and responses at 388 bytes vs. 498 bytes—but paradoxically makes DoC traffic easier to identify within CoAP data, because smaller, more distinctive message sizes create a more separable feature distribution for classifiers.

2026-lenders-secrets-best-not §2 / §6 defense

DNS over CoAP with equalized packet lengths (64-byte block-wise transfer) and SCHC header compression reduces Random Forest classification accuracy for identifying DNS frames from a ~89% baseline to 86%; adding payload compression (application/dns+cbor) further reduces accuracy to 77%. This outperforms DNS over HTTPS in constrained IoT settings across 296 evaluated deployment scenarios.

2026-lenders-secrets-best-not Abstract / §5–6 evaluation

Using a peer-based SCHC rule set — where the device ID is the XOR of source and destination link-layer addresses rather than just the device address — allows a single rule ID to serve multiple destinations and enables eliding (rather than merely compressing) flow-identifying header fields including IPv6 prefix, device IID, and CoAP Uri-Host. This prevents the rule ID itself from becoming a per-destination flow identifier observable to an eavesdropper.

2026-lenders-secrets-best-not §2 / §6.3 defense

In SCHC header compression, even a compact rule ID (8 bits in the paper's configuration) can function as a persistent flow identifier, leaking the per-device communication pattern to an eavesdropper. Standard per-device SCHC rule assignments allow an adversary to associate rule IDs with individual constrained IoT devices across sessions.

2026-lenders-secrets-best-not §2 / §6.3 detection

Adding artificial inter-arrival time delays — a standard DNS obfuscation technique in unconstrained networks — is infeasible for constrained IoT devices due to long inherent delays and high packet loss rates on constrained wireless links; the paper explicitly excludes timing-based obfuscation for this reason. EDNS(0) padding at recommended multiples of 128 bytes similarly inflates packet sizes to the point of forcing fragmentation, multiplying loss in constrained networks.

2026-lenders-secrets-best-not §1 / §3.1 evaluation

Encrypted traffic exhibits a 'full-frequency' spectral property where both low- and high-frequency components are highly active with comparable intensity, unlike natural images which are dominated by low-frequency components. Fourier Transform analysis across CIC-IoT2023, DoHBrw2020, and ISCX-Tor2016 confirms this distinction is pervasive. This signature is an inherent consequence of encryption disrupting byte-level semantics into a visually disordered, noise-like spatial pattern.

2026-lian-decompose-understand-fuse §II detection

Obfs4's payload randomization creates a high-entropy statistical profile that a Decision Tree detects with TPR ≈ 1.00 and FPR ≈ 0.00 across all four out-of-distribution protocols tested. UPGen's procedural protocol assembly similarly yields TPR ≈ 1.00 for most classifiers, confirming that both entropy randomization and rigid structural assembly leave artifacts detectable by even simple statistical classifiers operating under strict false-positive constraints.

2026-ling-one-prompt-censorship-evasion §5.1.3, Table 2 detection

Injecting uniformly sampled random delays between agent actions substantially degrades an unadapted XGBoost classifier, but a classifier retrained on delayed traces largely recovers performance across all four datasets. Under 5-second delay injection, the classifier shifts weight onto structural features (click-coordinate dispersion, structural key ratio, link-click ratio) that survive timing perturbation.

2026-lugoloobi-known-their-actions §6.1, Figure 5 evaluation

Address blocking is a losing strategy for the censor: with n=8 endpoints and µ/λa = 1 (discovery as fast as rotation), address-denial probability falls below 10^-2, and even when the censor discovers twice as fast as the defender rotates (µ/λa = 0.5), 7 endpoints keep address denial below 10^-2 (Lemma 1, Figure 1). The collateral budget γ means the censor cannot even block all addresses it discovers, and Corollary 1 gives the explicit threshold: n ≥ ln(1/ε) / ln((λa+µ)/λa) endpoints suffice for any target ε.

2026-maiti-block-a-mole-sustainability-frontier §5, §8.1, §8.10, Lemma 1, Corollary 1, Figure 1 evaluation

When the domain burn rate β = λdisc/λintro exceeds 1, time-average availability is bounded by A ≤ 1/β regardless of rotation speed µ, endpoint count n, or domain buffer kmax (Theorem 2). Rotation speed µ appears only in the address factor, which redundancy already saturates, and is entirely absent from the binding domain layer — proving that 'rotate faster' is categorically the wrong design lever under a domain-filtering censor.

2026-maiti-block-a-mole-sustainability-frontier §6, Theorem 2 evaluation

In April 2024 the GFW began decrypting QUIC Initial packets at scale and blocking connections by domain rather than by IP address, targeting the exact transport layer that modern moving-target circumvention systems rely on for session continuity. Because blocking a registrable domain costs the censor essentially zero collateral (unlike blocking a hyperscaler IP range), the censor's optimal Stackelberg budget split allocates f* = 0.97 of its discovery budget to the domain channel once the defender runs n ≥ 2 endpoints (§8.10, Figure 7). GFWatch measurements show hundreds of thousands of domains persistently blocked.

2026-maiti-block-a-mole-sustainability-frontier §1, §3, §6, §8.10, Figure 7 detection

Simulator validation shows a sharp phase transition in (α,T)-availability at the sustainability frontier β* ≤ 1. For a canonical configuration of n=8 endpoints and buffer kmax=8, the interval frontier is β* ≈ 0.65, while the time-average crossing is near β ≈ 0.8. Sweeping rotation speed µ/λa from 1 to >100 does not move the frontier — the availability contour over the (β, µ/λa) plane is entirely vertical (Figure 3). The closed-form availability law (Theorem 1) matches simulation to three significant figures across all tested β (Table 2). A rotation-only defender at beta = 1.5 stays below 0.25 availability even at the largest buffer and rotation speed tested, while a domain-aware defender at beta = 0.6 with the same n, mu and pool exceeds 0.95.

2026-maiti-block-a-mole-sustainability-frontier §8.2, §8.5, Figure 2, Figure 3, Table 2 evaluation

Ablation experiments show that replacing ESPRESSO's transformer backbone with a CNN ('Modified DCF') while retaining time-aligned interval features achieves performance competitive with the full ESPRESSO model across most protocols (e.g., SOCAT network-mode pAUC 0.997 vs. 0.989 at FPR ≤ 10⁻³), demonstrating that the time-interval feature representation—not the transformer architecture—is the primary driver of correlation accuracy.

2026-mathews-tracing-chain-deep §V-B, Table III evaluation

A systematic robustness evaluation found that ESPRESSO is highly robust to packet padding alone but that even modest artificial timing jitter causes significant performance degradation, identifying timing-based perturbations as the primary vulnerability of correlation-based stepping-stone (and by extension, anonymity-network) detectors.

2026-mathews-tracing-chain-deep §I (contributions), Abstract detection

Ephemeral defenses were integrated with a WireGuard fork and deployed as Mullvad VPN's 'DAITA' (Defense Against AI-guided Traffic Analysis) opt-in feature across Android, iOS, macOS, Linux, and Windows for over one year, serving a growing number of thousands of daily users. Individual defenses are derived deterministically from seeds in 43.6 ± 4.7 ms on a commodity laptop, making per-connection unique defenses practical at VPN scale.

2026-pulls-ephemeral-network-layer-fingerprinting §1, §7 deployment

Ephemeral blocking defenses reduce DF accuracy from 89.0% (undefended) to 10.2% and RF from 90.1% to 14.7% with standard 30-epoch training, at 97.5% bandwidth and 68.4% delay overhead; under infinite training, DF rises to only 29.2% and RF to 24.3%, still far below undefended baselines of 92.7% and 94.7%. Defenses are tunable at deployment time by adjusting Maybenot framework-wide limits, enabling overhead-vs-protection trade-offs without redeployment.

2026-pulls-ephemeral-network-layer-fingerprinting §5.2, §5.3, Table 1, Table 2 defense

The ephemeral property — using a unique seed-derived defense per connection — prevents attackers from training classifiers on the exact deployed defense variant. Stacked combinations with height H=5 from N=1,000 base defenses yield 6.88×10^25 unique defenses (polynomial growth O(N^{2H})). Attacks trained on ephemeral defenses also generalize significantly better across other randomized defense families than attacks trained on static defenses.

2026-pulls-ephemeral-network-layer-fingerprinting §3.3, §5.4 defense

Padding-only defenses that inject bursty traffic cause severe additional delay under realistic network bottlenecks: Break-Pad's delay overhead increases from 0% to 332.6% and FRONT's from 0% to 111.2% under a per-trace simulated PPS bottleneck. Even ephemeral padding defenses induce 43.9% delay overhead under bottleneck conditions, compared to 0% without a bottleneck, due to congestion from dummy packets.

2026-pulls-ephemeral-network-layer-fingerprinting §5.2, Table 1 evaluation

With infinite training time, Laserbeak achieves 93.5%, 95.9%, and 95.9% accuracy against ephemeral padding, FRONT, and Interspace respectively, compared to 96.5% undefended — confirming that padding-only defenses provide no meaningful protection against a sufficiently trained deep-learning WF adversary. Only ephemeral blocking defenses retain measurable protection, reducing Laserbeak to 71.8% accuracy under infinite training versus 96.5% undefended.

2026-pulls-ephemeral-network-layer-fingerprinting §5.3, Table 2 evaluation

MIRAGE's differentially private routing function provably bounds adversary inference: for a routing protocol satisfying ε-DP with ε = ln(4), any hypothesis test achieving a true positive rate of 80% necessarily incurs a false positive rate of at least 20%. The TPR-to-FPR ratio is bounded by e^ε for any ε-DP routing function, providing a formal privacy guarantee against routing-level statistical disclosure attacks.

2026-ratliff-mirage §V–VI defense

Embedding explicit TTL values in mesh-routed messages leaks proximity information — a recipient can infer that a high-TTL message originator was recently nearby. MIRAGE mitigates this with memoryless TTLs: carriers independently discard messages with probability q per epoch, implementing a branching process with replication factor R ≤ nmax·(1−q). Setting q > 1 − 1/nmax ensures sub-critical message extinction with expected lifetime ≈ −ln(nmax)/ln(R) epochs.

2026-ratliff-mirage §VI-C defense

Applying Fano's inequality, the paper proves Pe ≥ (H(X)−1)/log|Θ|, showing that deanonymization error rate approaches 1 (perfect anonymity) when the anonymity set |Θ| is large and mutual information leakage I(X;Y) between observed traffic Y and target identity X is minimized. A uniform default tunnel length of 3 hops across all nodes, for example, contributes no differential leakage because p(y=3)=1, illustrating that standardized network parameters reduce identifiability.

2026-rohrer-convolutional-neural-networks-deanonymisation-i2p §III-A, Equation (2) detection

Scylla, a modified Sphinx mix format, applies OAEP (an all-or-nothing transform) to the concatenated recipient address and message before splitting into fragments, so the final hop cannot recover the recipient or message without all fragments. Scylla adds only 39 bytes of header overhead over Sphinx and ~247 µs per fragment of computation time—negligible compared to Nym's default 15 ms per-hop mixing delay.

2026-schadt-breaking-partially-fixing §5, §7.1, §7.4, Figure 11 defense

Countermeasures against fingerprinting can paradoxically increase a user's identifiability when deployed by a small population: spoofing-introduced attribute inconsistencies and unusual configurations (e.g., Tor Browser's letterboxing applied to an uncommon monitor size) produce a rarer, higher-entropy fingerprint than baseline. This 'information paradox' means that countermeasure effectiveness is a property of the anonymity set size, not of individual tool configuration.

2026-schramm-analyzing-societal-awareness §2.2 detection

Brave's built-in strict fingerprinting protection was discontinued in January 2024 because of low adoption and frequent web breakages caused by restricting JavaScript APIs; browser vendors are broadly reluctant to implement robust built-in countermeasures for the same compatibility reason. Spoofing-based extensions (HTTP-header spoofers, Canvas Defender) offer only partial protection and can degrade site functionality through overlapping layouts, unreadable fonts, or bot misclassification.

2026-schramm-analyzing-societal-awareness §2.2, §3.1.4 evaluation

Using only 1,000-packet windows of signed packet lengths and inter-arrival times (no payload, no URLs, no cookies), a passive adversary achieves approximately 84% accuracy at inferring behavioral persona in a mixed-site open-world setting spanning 10 modern websites and 15 canonical personas plus an open-world class. Per-site persona macro-F1 typically ranges from about 0.78 to 0.91 across representative platforms including Bilibili, eBay, Yahoo, Zhihu, and LinkedIn.

2026-song-personafingerprint-measuring-persona §5.3, §5.5, Abstract detection

Attack accuracy scales steeply with persona-labeled training data: mixed-site open-world persona accuracy rises from 55.0% at 500 windows/persona to 65.0% at 1,000, 76.0% at 2,000, and 84.0% at 5,000 windows/persona across 10 sites (results consistent across 3 random seeds with std ≤1.0%). LLM-driven browsing agents make large-scale persona-labeled traffic generation practical for adversaries.

2026-song-personafingerprint-measuring-persona §5.6, Table 3 evaluation

In the boundless (boundary-free, session-unsegmented) setting with Human+Claude training, TrafficSliver reduces WFNet accuracy from 0.760 to 0.592, Cluster Anonymization (Palette) to 0.647, and Zero-delay (FRONT) to 0.706 at 2000 training samples. All three defenses assume clean session or page-load boundaries and are less effective against an attacker classifying arbitrary 5,000-packet continuous segments.

2026-song-redefining-website-fingerprinting §5.7 / Table 2 evaluation

Dodge-mimic reduces vRF deep-learning classifier accuracy to 10.6–15.6% for k=10 anonymity sets, matching the theoretical k-anonymity bound of (100/k)%=10%. The heuristic Walls Have Ears attack also fails: WHE-seg peaks at ≤8.5% and WHE-cyk at 9.7%, confirming that equalized cycle sizes leak no per-video information.

2026-witwer-dodge-client-side-framework §5.2, Table 2 defense

In live deployment with k=10 anonymity sets, video+audio+ABR, Dodge-mimic causes zero playback stalls in approximately 68% of sessions; for the remaining 32%, the mean initial stall time is 1.6 seconds — an extension of normal startup delay rather than mid-playback interruption — and all cycles complete roughly 20 seconds before playback ends even with large trailing cycles.

2026-witwer-dodge-client-side-framework §5.3, §5.3.3 evaluation

Small random per-session perturbations (±bytes per range) to Dodge-mimic extended manifests fully compensate for HPACK header compression in HTTP/2, yielding vRF accuracy of 10.5–14.2% for k=10 — indistinguishable from HTTP/1.1 results and near the theoretical bound. Because HTTP/3-QUIC uses analogous QPACK compression, the approach generalizes to the dominant future transport.

2026-witwer-dodge-client-side-framework §5.4, Table 3 defense

Even with perfectly equalized cycle sizes, a persistent intra-response timing side channel survives in Dodge: timing characteristics of individual packets within server responses cannot be controlled client-side, causing vRF accuracy to remain slightly above random guessing in all tested scenarios. This residual leak is consistent and observable regardless of anonymity set size or media configuration.

2026-witwer-dodge-client-side-framework §5.2.1, §6 detection

With 32 worker threads, Dodge-mimic anonymity-set bandwidth overhead (as percentage of undefended video content) ranges from 21.0%±0.8 at k=2 to 73.1%±2.8 at k=10 and 107.4%±1.5 at k=20; beyond k=10–16 threads, additional parallelism yields diminishing returns and can increase overhead due to thread-boundary fragmentation of the optimal search space.

2026-witwer-dodge-client-side-framework §5.1, Table 1 evaluation

An experimental 'random-and-mimic' option in snowflake-proxy produced a DTLS ClientHello fingerprint distinct from any observed standard fingerprint and was not blocked by the Russian filter. The covert-dtls library under development by the Tor Anti-Censorship team systematically randomizes the DTLS ClientHello handshake to defeat JA3/JA4-based classification.

2026-wkrp-snowflake-targeted-dtls-filtering kad09 report + abstract team notes defense

DoHFuse's dual-branch architecture outperforms all five baselines (KNN, Random Forest, FCNN+GRU, LSTM, Transformer) across 100-, 200-, and 449-class closed-world settings with standard deviations under 0.5 percentage points, while the vanilla Transformer's standard deviation exceeds 5 percentage points at 449 classes (mean 61.50%). This demonstrates that attention-based models without recurrence-scale adaptations are unsuitable for the burst–gap temporal dynamics of DoH/3, and that multi-scale gating (DMAG-LSTM) is the key architectural driver of robustness.

2026-zhang-dohfuse-dual-branch-architecture §V.C; Table II; Figure 6 evaluation

Under DoH/3, EDNS(0) padding and HTTP/3 framing collapse packet-length diversity: more than 99% of request packets fall within 233–350 bytes and more than 99% of response packets fall within 555–777 bytes. Direction-signed length-sequence classifiers (RF, LSTM, CNN) achieve only 67.39%, 8.95%, and 7.43% accuracy respectively on 449 classes — near or below random — confirming that size-based traffic analysis features lose nearly all discriminative power against DoH/3.

2026-zhang-dohfuse-dual-branch-architecture §IV.D; §V.B; Table I detection

DoHFuse achieves 88.05% accuracy (precision 88.56, recall 87.96, F1 87.83) fingerprinting 449 websites from DoH/3 traffic alone, demonstrating that EDNS(0) padding and QUIC/HTTP3 multiplexing are insufficient to prevent website fingerprinting. In the open-world setting, the model achieves AUPRC 0.975 with F1 0.951 (precision 0.906, recall 1.0), meaning every monitored site is detected while maintaining a low false-positive rate.

2026-zhang-dohfuse-dual-branch-architecture Abstract; §V.C; §V.E evaluation

Inter-arrival time sequences between client DoH/3 request packets retain site-specific temporal structure despite QUIC padding and stream multiplexing, exhibiting characteristic micro-burst and inter-burst gap patterns. Using timing features alone, LSTM achieves 83.75% ± 0.92% accuracy at 449 classes, and DoHFuse with its dual-branch DMAG-LSTM reaches 89.08% ± 0.49%, while length-only LSTM degrades to 8.95% under the same conditions.

2026-zhang-dohfuse-dual-branch-architecture §IV.D; §V.B; §V.C; Table II detection

Websites generating only one or two observable non-Google DNS queries per visit have classification accuracy below 30% in DoHFuse's closed-world evaluation, compared to the model's overall 88.05% on 449 classes. Sites such as ftc.gov and ailawandorder.com fell near zero accuracy, revealing that extremely sparse DNS activity is a natural but uncontrolled defense against timing-based fingerprinting.

2026-zhang-dohfuse-dual-branch-architecture §V.D; Figure 7 evaluation

The hash-based filtering strategy shifts covert channel security from algorithm secrecy to key secrecy per Kerckhoffs's principle: even when the covert algorithm is fully exposed, an adversary without the pre-shared key cannot identify which sparse subset (~1/2^L) of packets carries covert data, transforming the detection problem from statistical pattern recognition to exhaustive key-space search. The authors recommend a minimum Input Key length of 128 bits.

2026-zou-hiding-trees-forest §3.2, §4.1, §4.2.4 defense

At filter key length L=6, approximately 1.56% of packets are selected as covert carriers, yielding a unit covert capacity (packets required per covert bit) of 3.1–8 depending on bits embedded per packet. Each additional filter key bit halves the carrier selection rate (r = 1/2^L), so L=8 selects 0.39% of packets and pushes unit covert capacity above 32.

2026-zou-hiding-trees-forest §4.2.2, §5.3.2, Table 6 evaluation

The filtering strategy introduces no measurable increase in bit error rate (BER) or reduction in effective transmission rate (ETR) under packet loss rates up to 15%, reordering rates up to 20%, and delay jitter up to 100±50 ms, because the filter only selects carriers — data embedding and extraction logic remain identical to unfiltered channels.

2026-zou-hiding-trees-forest §5.3.1, Figure 5 evaluation

Per-packet processing overhead for SHA-256-based covert carrier filtering remains below 8 µs across all tested filter key sizes (L=1–8): storage channels (709-byte average payload) take 5.7–6.48 µs; timing channels (1146-byte average payload) take 7.44–7.87 µs, measured on a 12th-gen Intel Core i5-12400 at 2.50 GHz. This overhead does not cause packet loss or measurable end-to-end latency increase.

2026-zou-hiding-trees-forest §5.3.3, Table 5 evaluation

Introducing hash-based covert carrier filtering does not affect bit error rate (BER) or effective transmission rate (ETR) of the underlying covert channel under packet loss up to 15%, reordering up to 20%, or delays up to 100±50 ms, because the filtering layer only governs carrier selection and leaves embedding and extraction logic unchanged. Results for L=0 and L=6 were statistically indistinguishable across all tested network impairment conditions.

2026-zou-hiding-trees-forest §5.3.1, Fig.5 evaluation

Hash-based covert carrier filtering reduces ML classifier AUC for network storage covert channels from 1.0 (no filtering, L=0) to well below 0.7 when filter key size L≥6; SVM detection approaches random-guessing levels. The strategy selects a sparse, pseudo-random 1/2^L subset of packets as covert carriers using SHA-256 keyed by a pre-shared secret, diluting statistical anomalies in header field modifications with a large volume of unmodified legitimate packets.

2026-zou-hiding-trees-forest §5.3.2, Fig.6 defense

For network timing covert channels, hash-based covert carrier filtering reduces ML classifier AUC from 0.99 (L=0) to below 0.6 for all three classifiers (SVM, XGBoost, DT) at filter key size L=5, indistinguishable from the baseline normal-traffic AUC of 0.55. Sparse, aperiodic timing modulation events at L≥5 are masked by the network's inherent jitter, eliminating the autocorrelation and regularity signatures that standard timing-channel detectors exploit.

2026-zou-hiding-trees-forest §5.3.2, Fig.7 defense

The covert carrier filtering strategy shifts channel covertness from algorithm secrecy to key strength: once the embedding algorithm is exposed, an adversary must exhaustively search a 2^L key space to enumerate the covert carrier set rather than applying pattern recognition. At L=6, only 1/64 (~1.56%) of candidate packets are selected as covert carriers, yielding unit covert capacity Ucc=8 per packet for 8-bit embeddings and Ucc=3.1 for 21-bit embeddings.

2026-zou-hiding-trees-forest §3.2, §4.2.2, Table 6 defense

Drivel evaluates its design against the GFW's fully-encrypted-traffic detector (documented in Wu et al. 2023). The thesis demonstrates that switching to post-quantum primitives does not by itself change the traffic's appearance to a statistical censor classifier — the fully-encrypted detection problem is independent of the underlying cryptographic algorithm and must be addressed at the traffic-shaping layer regardless of key-exchange choice.

2025-himmelberger-drivel §3, §4 detection

Drivel is an obfs4-style fully-encrypted proxy protocol that replaces obfs4's pre-quantum cryptographic primitives with post-quantum alternatives. It is one of the first circumvention protocols explicitly designed to remain secure under a quantum adversary, addressing the forward-secrecy threat to deployed circumvention traffic recorded today for future decryption.

2025-himmelberger-drivel §1, §2 defense

Most deployed circumvention protocols (obfs4, Shadowsocks, Trojan, VMess, etc.) still rely on pre-quantum primitives (X25519, AES-GCM, ChaCha20). Drivel is the first published treatment of how to perform this migration in the specific context of a fully-encrypted pluggable transport, providing a design template and security analysis that does not exist elsewhere in the circumvention literature.

2025-himmelberger-drivel §2, §5 defense

In 24-hour live proxy deployments, covertDTLS mimicry had a 18.2% DTLS handshake failure rate (vs 12.5% baseline, 27.0% randomization, 25.8% Chrome webextension). Randomization generates ≈994 billion unique fingerprint permutations (cipher shuffling: 109,600; extension shuffling: 994,218,624,000), making blocklist-based fingerprinting infeasible, but at the cost of higher connection failures due to cipher mismatches. Mimicry of DTLS 1.2 was stable and effective; DTLS 1.3 mimicry is not yet achievable with the current Pion library.

2025-midtlien-fingerprint-resistant §4.2, Table 1, §5 evaluation

The framework confines active traffic shaping to the first N seconds of a connection (N is a user-defined parameter, e.g., N=10), after which normal unmodified traffic resumes. The authors hypothesize that this design keeps per-session throughput and latency overhead negligible, since the shaping window is a small fraction of total connection time; N can be extended to the full session if the censor is believed capable of classifying beyond early traffic.

2025-pereira-extended §2.3, §3 defense

The framework's GAN-based schedule generator trains on short session windows (e.g., the first 10 seconds) of real browsing traffic from the Tranco Top 1000 sites, learning joint distributions of packet sizes, inter-arrival times, and burst patterns to produce realistic synthetic schedules. This repurposes GAN architectures previously used for traffic analysis (e.g., GANDaLF) as a defense-side cover-traffic generator.

2025-pereira-extended §2.2 defense

The proposed framework operates as a transparent shim between application and network layers, enforcing a configurable schedule over packet size, timing, and burst patterns. The shaping logic is transport-agnostic — applicable across TCP, UDP, QUIC, and TLS — and activates only after the underlying protocol handshake completes, making it reusable across heterogeneous circumvention stacks.

2025-pereira-extended §2.1 defense

Combinations of Bayesian methods, data augmentation with mixup, and NOTA defensive padding cut the open-world false positive rate by up to 92% at 0.5 recall on HTTPS-only traffic and 75% on Tor traffic relative to the deterministic MSP baseline. Even with these improvements, sustaining a world size in the hundreds of millions (approaching YouTube-scale) requires accepting recall of 0.5–0.6 and precision of only 0.1–0.2; at precision 0.5 and recall 0.5, the maximum workable world size is only 37.5M for HTTPS-only (Table 3), far below YouTube's ~10 billion video catalog.

2025-walsh-improved-open-world-fingerprinting §4.5, Tables 1–3 evaluation

When a fingerprinting model is trained on traffic collected from one geographic vantage point and tested on traffic from a different continent, the HTTPS-only open-world FPR at 0.5 recall increased by factors ranging from 2.8x (EU-West-2) to 50.3x (Africa) relative to the same-vantage baseline — despite 60-way closed-world accuracy remaining above 0.99 across all vantage-point pairs (Table 5). For Tor traffic the effect was weaker but still reached 25.2x (Asia-Pacific Southeast-1), showing path diversity also disrupts Tor-based fingerprinting.

2025-walsh-improved-open-world-fingerprinting §5.1, Table 5 evaluation

The paper establishes, for the first time in a large open-world scenario (64,000 unmonitored test videos), that HTTPS-only video stream fingerprinting is significantly easier than Tor-based fingerprinting because DASH adaptive bitrate selection introduces a second-order network-condition effect: clients request entirely different video segments at different quality levels depending on path conditions, causing traffic traces from different geographic vantage points to diverge at the application layer even when network conditions are nominally similar. This makes NOTA and synthetic training sample techniques less effective on Tor data due to inherent trace noisiness.

2025-walsh-improved-open-world-fingerprinting §4.5, §5.1 evaluation

Shaperd's adaptive blocking-detection mode can integrate with external blockage-detection tools (e.g., Troll Patrol) to detect when a constraint set is no longer effective and automatically switch to an alternate constraint set, changing packet patterns to restore connectivity without user intervention.

2025-wilson-extended Appendix A defense

Shaperd's proof-of-concept prototype (~1000 lines of Go) introduces a minimal 4.1% throughput overhead for a single entropy constraint; the first additional constraint added 5.1% overhead and the second added 5.5%, with total overhead scaling with constraint count and rigor.

2025-wilson-extended §2 evaluation

Shaperd introduces a constraint-agnostic traffic shaping system that operates on both packet content and timing in real time, designed for drop-in integration with any existing FEP. The system uses a four-component constraint definition (function, value, comparison operator, target packets) capable of expressing any rule based on a computable deterministic function over packet contents.

2025-wilson-extended §2 defense

The GFW's QUIC censor does not reassemble QUIC client Initial packets that are split across multiple UDP datagrams, nor does it reassemble QUIC CRYPTO frames split within a single datagram. Three practical bypasses follow: (1) send any UDP datagram with a random payload before the QUIC Initial—the GFW uses 60-second UDP flow state and won't inspect a mid-flow packet; (2) fragment the TLS ClientHello SNI across multiple QUIC CRYPTO frames; (3) use an unknown QUIC version number in the first packet (Version Negotiation bypass, payload undecryptable). Chrome independently exploits (2) through its Chaos Protection feature (since 2021) and post-quantum Kyber key-agreement (since v124, Sep 2024), whose larger key sizes force fragmentation across UDP datagrams. As of January 2025, the GFW also does not block ECH-containing QUIC payloads unless the outer (cleartext) SNI is on the blocklist.

2025-zohaib-quic-sni §3.2 / §7 defense

The threat model requires no DPI and was fully implemented as a Linux kernel module on a NETGEAR R6120 with only a 580 MHz processor, 16 MB ROM, and 64 MB RAM, adding negligible overhead. Unlike ML-based or DPI-based VPN classifiers, the statistical model operates pre-NAT on per-device private IP flows, making it immune to obfuscation techniques that alter packet payloads or disguise protocol handshakes.

2024-almutairi-fingerprinting §I, §IV detection

The authors propose two countermeasures: (1) widespread adoption of traffic splitting so not all user traffic is routed through a single VPN tunnel, neutralizing the single-destination session signature; and (2) VPN servers should rotate at random intervals so that no prolonged session to one IP accumulates enough packets to trigger the threshold T.

2024-almutairi-fingerprinting §VI defense

Oscur0 eliminates Conjure's separate registration phase by steganographically encoding ECDH public key, phantom IP, and transport parameters into the encrypted application data of the first UDP (DTLS 1.2 with Connection ID) packet sent to the phantom IP, using Elligator encoding to make the public key indistinguishable from random bytes. This removes several round trips — registration, TCP handshake, and application handshake — compared to standard Conjure, and means censors cannot block the scheme by blocking registration alone.

2024-chen-extended §3 Design defense

The encapsulated TCP three-way handshake (3WHS) is detected in 80.59% of VPN flows but only 0.33% of plain UDP flows, making it—on its own—a near-practical VPN detector with 0.33% FPR; its presence is required by the classifier regardless of the compliance-rate threshold t.

2024-hanlon-detecting §4.2 Feature Importance detection

Random padding alone raises the classifier FPR only slightly (0.11% to 0.15%), and connection multiplexing alone raises it to 0.53%; however, combining both defenses raises FPR to 2.57%, making the detector impractical for a real-world censor and yielding TPR of 93.40%.

2024-hanlon-detecting §4.2 Robustness of 3WHS / Table 3 defense

A protocol-agnostic classifier that identifies RFC-mandated TCP behaviors (three-way handshake, 500ms ACK, 2×RMSS acknowledgement) leaking through UDP-based VPN tunnels achieves a false positive rate of 0.11–0.29% on real campus traffic, an order of magnitude lower than ML-based VPN detection techniques (FPR 1.4–5.5%) and on par with the GFW's estimated heuristic FPR of 0.6%.

2024-hanlon-detecting §3–§4.2 detection

An attacker who generates 10 defended copies of each training trace (re-sampling noise each time) improves Tik-Tok accuracy against DeTorrent from 31.9% to 48.2%, demonstrating that dataset augmentation with multiple defended samples is a practical countermeasure against randomized padding defenses including DeTorrent and FRONT.

2024-holland-detorrent §6.4 detection

Against the state-of-the-art DeepCoFFEA flow-correlation attacker, FC-DeTorrent reduces the true positive rate at a 10^-5 false positive rate to approximately 0.12 — less than half that of the next-best defense Decaf (TPR ≈ 0.29) — while using 97.3% bandwidth overhead, without delaying any real traffic packets.

2024-holland-detorrent §4.2, §6, Table 5 evaluation

DeTorrent exhibits strong diminishing returns in the bandwidth-performance tradeoff: increasing the dummy-download budget from N=1,000 to N=3,000 reduces Tik-Tok accuracy by ~19.1 percentage points, while a further increase from N=5,000 to N=7,000 yields only an additional 4.9-point reduction (accuracy floor near 20.8% at ~210% overhead). At the lowest tested budget (~40% overhead) Tik-Tok accuracy is still only 52.8%.

2024-holland-detorrent §6.2, Figure 6 evaluation

DeTorrent is implemented as a Tor pluggable transport on top of the WFPadTools/Obfsproxy framework and deployed against live Tor traffic; a modest VPS with 4 GB RAM and 2 vCPUs running at under 50% CPU utilization can defend five simultaneous connections in real time with no GPU required. Performance drops only 0.7% when the generator is trained on one dataset partition and tested on another.

2024-holland-detorrent §5.4 deployment

DeTorrent reduces closed-world Tik-Tok attack accuracy from 93.4% to 31.9% on the BE dataset — 10.5 percentage points better than the next-best padding-only defense (FRONT at 42.4%) — and reduces Deep Fingerprinting accuracy from 94.3% to 30.0%, at a bandwidth overhead of 98.9%. On the larger DF dataset, Tik-Tok accuracy falls from 97.7% to 79.5%.

2024-holland-detorrent §6.1, Table 4 evaluation

When a user splits traffic across N paths, a censor observing a single path sees only a partial trace, substantially reducing the accuracy of classifiers trained on complete network traces. Prior Tor traffic-splitting work (TrafficSliver, CoMPS, multipath Tor studies) has validated this defense against website fingerprinting outside the PT context.

2024-lorimer-extended §1, §4.2 defense

obfs4 and obfs⋆ produce characteristic wire patterns—bursts of roughly MTU-sized payloads followed by a randomly-sized chaff packet—that CNN classifiers detect purely from packet-size sequences without payload inspection. A trivial per-bridge entropy-biasing re-encoding (obfs⋆) completely defeats the hand-tuned decision tree (0% precision, 0% recall) but does not reduce CNN detectability, because the CNN generalizes across size-distribution variants.

2024-wails-precisely §V-E, §IV-D-3, Figure 4 detection

Achieving active security (FEP-CCFA) requires that on any AEAD decryption failure a fully encrypted protocol silently return the empty string and keep the channel open indefinitely, never emitting a channel-closure signal. Any observable behavioral difference — including connection termination timing — leaks information about ciphertext-boundary locations to an active adversary.

2023-fenske-security §2.3 defense

No existing fully encrypted protocol — including Obfs4, Shadowsocks, VMess, and Obfuscated OpenSSH — simultaneously satisfies passive indistinguishability (FEP-CPFA), active-manipulation resistance (FEP-CCFA), and output-length shaping. The paper presents a novel stream-based construction that provably satisfies all three using AEAD-authenticated length blocks, an output buffer supporting arbitrary fragmentation, and a padding mechanism allowing the sender to emit exactly p output bytes on demand.

2023-fenske-security §2–3 defense

Censors optimize for utility under asymmetric misclassification costs rather than raw accuracy: false positives (blocking legitimate traffic) carry economic and political costs that make censors conservative about deploying classifiers with high false-positive rates. Multi-flow stateful classifiers — such as the obfs4 Elligator probabilistic distinguisher, which requires correlating observations across multiple connections — are operationally more expensive than single-packet or connection-initiation classifiers, which the author suggests explains why probabilistic multi-flow distinguishers have not been exploited in practice even when theoretically available.

2023-fifield-comments §5 evaluation

Voiceover's DCGAN, trained on ~400 hours of two-person telephone conversations, generates conversation timing templates that constrain when the tunnel transmits audio. This reduces ML classifier performance from auROC 0.981/aucPR 0.959 (unshaped baseline) to auROC 0.682/aucPR 0.482, and the improvement holds at 500-packet windows (auROC 0.68/aucPR 0.50), suggesting robustness to memory-limited adversaries.

2023-jia-voiceover §3, §4.2, Table 2 defense

Voiceover achieves 31.16 bytes/s goodput with default parameters—roughly half the 62.32 bytes/s of the unshaped baseline—because GAN-imposed silence periods reduce transmission time. Skype's OPUS codec bounds the theoretical ceiling at 750–63,750 bytes/s, so all multimedia tunnels over this path are constrained to low-bandwidth use cases; the authors explicitly position Voiceover as an out-of-band channel for sharing secret keys rather than a general-purpose data path.

2023-jia-voiceover §4.5, Table 3 evaluation

Protocol fingerprinting — including DPI-based identification of VPNs, circumvention tools, and E2EE messengers — was active in only 6% of countries during the measurement period (13% all-time), but all confirmed instances came from focused individual studies, not from mass measurement platforms like OONI or Censored Planet. The authors flag encrypted traffic analysis (ETA) tools and next-generation firewalls (NGFWs) capable of blocking Signal or Tor Browser as an emerging threat to freedom of expression.

2023-master-worldwide §4.3 detection

The GFW detects Shadowsocks by flagging apparently high-entropy connections that are not TLS or HTTP, but this detection is brittle: connections are explicitly allowed if the first 6 bytes of the first packet of a flow are all printable ASCII characters (range 0x20–0x7E). Adding a 6-byte alphanumeric preamble to the Shadowsocks message definition is sufficient to bypass this heuristic and requires only a short patch to the protocol specification file.

2023-wails-proteus §3.2 detection

OpenVPN's application-layer P_ACK packets — uniform in size and concentrated only in the handshake phase — provide a timing and count fingerprint detectable via threshold comparison over 10-packet bins. Tunnel-based obfuscation wrappers (Stunnel, SSH, obfs2/3, Shadowsocks) that do not add random padding preserve the 1:1 packet correspondence with the underlying OpenVPN stream, leaving 16 of 20 tested tunnel-based obfuscated configurations vulnerable to ACK fingerprinting.

2022-xue-openvpn §6.2, §9.1 detection

34 of 41 obfuscated OpenVPN configurations and 18 of 20 UDP configurations were co-located with vanilla TCP OpenVPN servers within the same /29 subnet; probing the /29 subnet of a suspected obfuscated or UDP endpoint revealed nearby vanilla TCP servers, enabling confirmation by 'guilt by association' even when the obfuscated endpoint itself resisted direct fingerprinting. Some providers additionally share infrastructure across different VPN brands, further compounding exposure.

2022-xue-openvpn §7.5, §9.1 detection

A two-phase passive-filter-plus-active-probing framework deployed at a 1-million-user ISP identified 85.90% of vanilla OpenVPN flows (1,718/2,000) and 72.67% of obfuscated flows (1,468/2,020), with an upper-bound false positive rate of 0.0039% across over 10 million flows — three orders of magnitude lower than prior ML-based approaches (1.4–5.5%). The system processed 15 TB and 2 billion flows per day on a single commodity server.

2022-xue-openvpn §9, Table 3 evaluation

OpenVPN's unencrypted opcode header byte is exploited to fingerprint vanilla and XOR-obfuscated flows: the XOR patch specification excludes the first buffer byte (the opcode) from reversal, so opcodes are always XOR-ed with the same key byte and map deterministically to fixed ciphertext values. All 4 of the top-5 VPN providers that offer obfuscated services use XOR-based obfuscation, and all were flagged by opcode fingerprinting over 90% of the time.

2022-xue-openvpn §6.1, §9.1 detection

MIMIQ requires modifications only to a single trusted network (an ISP or enterprise AS): an address allocation server and several critical edge switches. Remote QUIC servers and the wider Internet require no changes. The authors argue ISPs have financial incentives to deploy MIMIQ as a privacy-enhancing service, and that QUIC's rapid adoption (600K+ QUIC-enabled domains, 1.6M QUIC-capable IPs as of 2020) means server-side support is increasingly given.

2020-govil-mimiq §5 deployment

At a round-trip time of 50 ms, MIMIQ incurs only a 10% throughput drop when migrating every 25 packets (frequency = 4 per 100 packets). Per-migration latency ranges from 7–64 ms at 10 ms RTT and 43–99 ms at 50 ms RTT as migration frequency scales from 0 to 50 migrations per 100 packets.

2020-govil-mimiq §4.2 evaluation

MIMIQ leverages QUIC's connection migration to change a client's IP address mid-connection without disrupting ongoing transfers. QUIC's specification requires that endpoints never reuse connection IDs during migration and that migration tokens are encrypted; this makes migration events indistinguishable from a second distinct client initiating a new connection from a fresh IP address.

2020-govil-mimiq §3.1 defense

Against censors that detect blacklisted application protocols by examining only the first 30 packets of a flow (e.g., the technique in Wang et al. 2015), a single IP migration after 30 packets have been exchanged is sufficient to defeat detection while incurring minimal performance overhead—the client continues the connection normally on the new address.

2020-govil-mimiq §5 defense

Migrating the client IP address every 25–100 packets reduces state-of-the-art website fingerprinting attack accuracy to below 10% in the closed-world setting, outperforming advanced dedicated defenses such as HyWF multipathing. The mechanism works because most fingerprinting classifiers rely on as many packets per flow as possible, and flow splitting degrades feature quality.

2020-govil-mimiq §5 defense

An adaptive censor that retrains classifiers on both unmodified and GAN-transformed Meek traffic ('informed NN') partially recovers detection capability: informed NN achieves a PR-AUC of 0.440 against modified traffic versus 0.309 for the naive NN, and achieves FPR of 0.667 versus 1.000 for the naive NN. However, the informed NN suffers from catastrophic interference and performs worse on FPR than the naive classifier on unmodified data (0.545 vs. 0.002).

2019-sheffey-improving §5 Results, §6 Discussion detection

A GAN-based adversarial transformer applied to Meek traffic signatures increases mean classifier FPR from 0.183 to 0.834 and decreases mean area under the precision-recall curve (PR-AUC) from 0.990 to 0.414 across naive neural network, informed neural network, and CART decision tree classifiers evaluated on three geographically distinct datasets (residential, university, AWS).

2019-sheffey-improving §5 Results, Tables 1–2 evaluation

The paper identifies that Meek traffic is compared against average HTTPS traffic across all domains rather than against traffic specific to the CDN fronting host (e.g., ajax.aspnetcdn.com for meek-azure), meaning a transformed signature that mimics generic HTTPS may still appear anomalous relative to expected traffic to that specific CDN host. This dataset construction limitation means real-world GAN-guided shaping must target host-specific traffic baselines, not population-wide HTTPS baselines.

2019-sheffey-improving §6 Discussion defense

Incorporating perturbation loss — the mean absolute difference between original and transformed traffic signatures — into the GAN's training objective constrains the transformer to make minimal modifications, reducing the implementation overhead a real-time traffic shaper would require. The perturbation loss is weighted at 10× relative to classification losses, enforcing sparse modifications while still fooling the discriminator.

2019-sheffey-improving §4.2 Training defense

Feature importance analysis of XGBoost models reveals that Facet covert channels are identifiable primarily through packets in the 115–195 byte range (dominated by Skype audio packets), while DeltaShaper is identifiable through two distinct packet-length clusters: 85–100 bytes and 1105–1205 bytes. XGBoost assigns non-zero importance to only ~58% of the 300 quantized packet-length bins for Facet and ~42% of 600 bins for DeltaShaper, indicating that leakage is concentrated in a narrow portion of the packet-size distribution.

2018-barradas-effective §4.4, Figure 4 detection

I2P obfuscates payload content to prevent protocol identification, but flow analysis can still fingerprint I2P traffic because the first four handshake messages between I2P routers have fixed lengths of exactly 288, 304, 448, and 48 bytes. The I2P team acknowledged this and was developing an authenticated key agreement protocol to resist automated identification.

2018-hoang-empirical §2 detection

Shadowsocks traffic appears as ordinary TCP with no payload keywords or obvious protocol markers because the entire payload is encrypted; firewalls cannot distinguish it from generic TLS without behavioral flow analysis. This makes signature- and keyword-based detection ineffective against it.

2017-deng-random §III.A detection

For complete blockage (>99%) over 10 hours, the adversary requires a swarming ratio of 12.8, translating to 128,000 censors against a single server with 10,000 CoAs. Scaling to a 10-server, 10-interface deployment forces the adversary to operate 106,700 humans in parallel; with a 5-minute CAPTCHA registration and a 12-hour reset cycle, achieving complete blockage within 10 hours requires 1,067 non-stop human operators in the first two hours.

2017-heydari-scalable §IV-A, §IV-B evaluation

A credit-based accounting method dynamically assigns users to larger groups as their trust score accumulates (credit increases by G−1 per unblocked interval), requiring a user's credit to be twice the group's risk before joining. This reduces the total number of CoAs needed while making it costly for censor agents to infiltrate large groups, since they must wait through many clean intervals before the group reaches exploitable size.

2017-heydari-scalable §IV-D defense

A proof-of-concept Linux prototype using UMIP (open-source MIPv6) with three routers and five commodity machines (2.4GHz Intel Core 2 Duo, 4GB RAM) demonstrated correct CoA rotation every 10 seconds. Signaling overhead was reduced to one-third of standard MIPv6 by eliminating return routability messages; per-packet transmission overhead was 24 bytes (IPsec ESP), identical to the baseline secure-channel cost, yielding zero net overhead attributable to the MTD mechanism.

2017-heydari-scalable §IV-E evaluation

The MI-MTD framework uses Mobile IPv6 Care-of Addresses (CoAs) rotated among randomized user groups every shuffling interval. With 1,000,000 users, 5,000 censors, and 10,000 CoAs (swarming ratio φ=0.5), per-interval access probability is 60.88%; over one minute with 10-second shuffling intervals, blocking probability drops to approximately 0.358%, meaning users retain ~99.6% chance of access.

2017-heydari-scalable §IV-A defense

HTTP GET fuzzing via subtle token modifications bypassed large fractions of filters: removing the `\r\n` before the Host header bypassed 36–38 of 44 Host-header filters; embedding the censored URL in the middle of a long hostname string bypassed 33–35 filters; placing the URL in an after-Host field with a non-empty Host bypassed 29–36 filters. Blacklist coverage was also weak: no filter blocked all 100 of the Alexa top adult sites, and some blocked as few as 31.

2017-jermyn-autosonda §4.1 / Appendix A evaluation

Among the 44 non-DNS filters, 11 did not reassemble TCP segments and 7 did not reassemble IP fragments before inspection, meaning a censored URL split across segment or fragment boundaries evaded detection. Five filters applied fragment/segment reassembly timeouts of under 2 seconds despite maintaining HTTP request state for more than 8.5 seconds, creating a window where a deliberately fragmented flow with artificial delay avoids inspection entirely.

2017-jermyn-autosonda §4.1 Mechanism detection

ScholarCloud's 'message blinding' — a non-public byte mapping (f: [0, 2^8) → [0, 2^8)) applied between domestic and remote proxy — successfully evades GFW deep packet inspection with 0.22% average packet loss rate, statistically indistinguishable from native VPN (0.21%). The paper reports that even this simple encoding suffices because the GFW cannot classify the traffic; confidentiality of the algorithm is the operative property, not cryptographic strength. Because the operator controls both proxy endpoints, the blinding scheme can be rotated at any time without requiring client-side updates.

2017-lu-accessing §3 defense

A naive active-probing resistance scheme that embeds a fixed-length token in the initial request is vulnerable to flow fingerprinting because the censor can detect connections that always begin with a fixed byte count; pseudo-random padding removes this length-based signature. Separately, obfuscating-service schemes that reveal server aliveness by completing TCP expose the server IP to enumeration even before the application-layer challenge fires.

2016-khattak-sok §5.5 defense

Wiley's Bayesian classifier against obfuscated protocols (Dust, SSL, obfs-openssh) found that entropy detection achieved 94% accuracy using only the first packet, timing-based detection achieved 89% accuracy over entire packet streams, and length-based detection achieved only 16% accuracy.

2016-khattak-sok §2.4.1 detection

CDNReaper's Scrambler defeats domain-based and Wang et al. k-NN fingerprinting by injecting decoy requests uniformly distributed across ndom popular domains and dropping ~24% of advertisement/analytics requests (which constitute on average 24% of top-1000 Alexa page requests); even at low traffic overheads, fingerprinting accuracy drops significantly from the 0.991/0.94 baseline, with dropping traffic providing more benefit at lower overhead budgets.

2016-zolfaghari-practical §4.4 defense

The survey identifies 'soft censorship' — including throttling, packet-loss injection, and quality-of-experience degradation — as detected by only 2 of 13 surveyed platforms (rTurtle and UBICA) as of 2015. The paper explicitly flags this as a measurement gap, noting that soft censorship symptoms are indistinguishable from ordinary network congestion without ground-truth probes placed outside the censor's network.

2015-aceto-internet §3 / Table 2 detection

The paper formally defines circumvention as either preventing the trigger from being seen by the surveillance device, or countering the effects of the censoring action. This two-path decomposition — hide the trigger vs. nullify the enforcement — provides a clean design framework: a circumvention tool can succeed by making traffic unrecognizable (no trigger fires) or by routing around the blocking device (action nullified).

2015-aceto-internet §2.1 defense

Marionette is the first programmable obfuscation system to simultaneously satisfy all five threat-model dimensions evaluated in Figure 2: resistance to blacklist DPI, whitelist DPI, statistical-test DPI, protocol-enforcing proxy traversal, and multi-layer traffic control, while sustaining throughput above 1 Mbps (up to 6.7 Mbps). Every prior system (obfs4, ScrambleSuit, SkypeMorph, StegoTorus, FTE, JumpBox, etc.) fails at least one dimension, most commonly stateful proxy traversal or statistical-feature control.

2015-dyer-marionette §2, Figure 2 defense

Randomization-based obfuscation systems (obfs2/3, obfs4, ScrambleSuit, Dust) resist blacklist DPI but fail entirely under protocol-whitelist filtering, as explicitly demonstrated during the Iranian elections where censors permitted only known-good protocols. Pure randomization provides no signal of being a permitted protocol, making it trivially blockable under any whitelist regime.

2015-dyer-marionette §1, §2 detection

Beverly et al. found that 77% of Internet clients can spoof source addresses within their own /24 and 11% can spoof within their own /16, with these characteristics holding across a wide range of countries and regions. The authors use this result to argue that IP-spoofed cover traffic — where measurement probes appear to originate from many hosts in the same AS — is broadly feasible in practice.

2015-jones-can §4.2 defense

The paper demonstrates that 'having no fingerprint is itself a fingerprint': randomizing obfuscators that emit uniformly random bytes from the first packet are detectable precisely because conventional protocols (TLS, SSH, HTTP) always begin with fixed plaintext headers. This structural distinction requires no deep payload parsing — the attack operates on only the first TCP packet — and achieves TPR=1.0 / FPR=0.002 against obfsproxy3/4 using commodity-implementable statistics.

2015-wang-seeing §1, §5.1 defense

Obfsproxy3 and obfsproxy4 are reliably detected by an entropy-distribution test (KS test, block size k=8) applied to the first 2,048 bytes of the first client-to-server packet, combined with a minimum payload-length check of 149 bytes. On three university campus datasets totaling over 14 million TCP flows, the test achieves TPR=1.0 with FPR ranging from 0.24% to 0.33%. Omitting the length check raises the SSL/TLS false-positive rate to approximately 23%.

2015-wang-seeing §5.1, Table 5 detection

Facet's video morphing — embedding the requested video in a fraction s of H.264 macroblocks within a randomly chosen chat video — raises the censor's required false positive rate dramatically. At steganography level s=0.125, blocking 90% of Facet connections requires disrupting over 40% of genuine videoconferencing traffic; blocking 80% requires disrupting at least 20% of legitimate calls.

2014-li-facet §7.2, §9.3–9.4, Figure 9 defense

GNS encrypts all DHT queries and responses using a zone-private-key-derived symmetric key (h = x·l mod n; query = H(hG)) such that a passive DHT observer can only mount a confirmation attack — requiring simultaneous knowledge of both the zone's public key and the specific label. Without both values, an adversary observing DHT traffic cannot determine the label, zone, or record data; even fully participating malicious DHT nodes see only opaque signed blobs unlinkable to their originating query.

2014-wachs-censorship-resistant §4 defense

GoHop without traffic shaping achieved 76.8–78.5 Mbps (virtual NIC) on a 1 Gbps LAN; traffic shaping reduced this to 58.1 Mbps (~26% overhead from fragmentation). In a Beijing-to-Seattle real-world download test, GoHop delivered 960–999 KB/s against a 1,544 KB/s direct baseline, with the 96.7 Mbps WAN link—not GoHop—as the bottleneck. This compares to Tor's 40–300 KB/s (30–80 KB/s with obfuscation plugins such as SkypeMorph).

2014-wang-gohop §V.B, Table II, Table III evaluation

Packet padding alone is insufficient to defeat statistical traffic analysis unless every packet is padded to MTU; small-size padding has minimal effect on classifier accuracy (citing Hjelmvik & John 2010). Traffic shaping that also fragments large packets—transforming the full packet-size CDF to match a target distribution rather than merely inflating small packets—is required to statistically impersonate a target traffic class.

2014-wang-gohop §III.B defense

A pre-shared key enables encrypting the entire GoHop packet—header, payload, and padding bytes—achieving true randomness in the full byte stream. Standard VPN protocols such as OpenVPN encrypt only the payload while leaving headers in plaintext, exposing protocol-identifying fields to DPI without payload inspection. This design choice is a prerequisite for defeating header-based fingerprinting.

2014-wang-gohop §III.A defense

Spreading UDP datagrams across a randomized port range breaks traditional 5-tuple-based session tracking, randomizes per-port inter-arrival times, and reduces per-port throughput to a small fraction of the aggregate—making per-flow statistical analysis significantly harder. Critically, the number of random ports does not reduce aggregate throughput: GoHop measured 76.8 Mbps (1 port) versus 78.5 Mbps (100 ports) at the virtual NIC.

2014-wang-gohop §III.C, §V.A, Table II defense

GoHop's naïve traffic shaping targeting a uniform packet-size distribution (0–MTU) successfully morphed both HTTP and SSH flows: K-S test D values were 0.019 (HTTP) and 0.022 (SSH), both below the 0.025 rejection threshold, with p-values of 0.20 and 0.11 respectively. After shaping, packet-size CDFs and statistical metrics (mean ~782–783 bytes, variance ~163,600) for both protocols became nearly identical, eliminating the size signals that distinguish them.

2014-wang-gohop §V.A, Table I evaluation

The IBST construction is provably secure under the bilinear decisional Diffie-Hellman (BDDH) assumption in the random oracle model. Any adversary with advantage ε(λ) against IBST indistinguishability implies an adversary against BDDH with advantage at least ε(λ)/e(1+qE), where qE is the number of private-key extraction queries. Tags produced by the scheme are computationally indistinguishable from uniform random bitstrings for any party lacking the recipient's private key.

2013-ruffing-identity-based §3.2 Construction / Theorem 1 defense

ScrambleSuit achieves polymorphism by seeding each server's PRNG with a randomly generated 256-bit value, which generates server-specific probability distributions over packet lengths (up to 100 bins) and inter-arrival times (bins in [0, 10) ms). The seed is shared with clients after authentication, so both sides shape traffic identically; a censor monitoring two distinct ScrambleSuit servers observes different distributions and cannot build a single universal classifier.

2013-winter-scramblesuit §4.3 defense

Tor's traffic contains a characteristic prevalence of 586-byte packets (Tor's 512-byte cells plus TLS header overhead) that form a strong flow-level fingerprint detectable from a few dozen captured packets. ScrambleSuit's packet length morphing eliminates this signature and shifts the distribution toward MTU-sized packets, but the authors note that a censor using the VNG++ classifier — which relies on coarse features like connection duration, total bytes, and burstiness — would still require only a marginal increase in ScrambleSuit's overhead to defeat.

2013-winter-scramblesuit §4.3.1, §5.1, Figure 10 detection

SkypeMorph's packet size and inter-packet delay distributions are statistically indistinguishable from real Skype video calls: Kolmogorov-Smirnov tests on both the naïve traffic-shaping and enhanced Traffic Morphing outputs report p > 0.5, indicating no significant difference from the Skype target distribution. The original Tor traffic distribution, by contrast, is considerably different from Skype, validating the need for the morphing layer.

2012-moghaddam-skypemorph §7, Figure 6 evaluation

Encrypted channels expose only two statistical features to an external observer: packet sizes and inter-packet arrival times. Original Traffic Morphing (Wright et al. 2009) shaped only packet-size distributions, leaving inter-packet timing as an unobfuscated fingerprint identical to the source (Tor) distribution. SkypeMorph extends Traffic Morphing to jointly sample from nth-order conditional distributions of both packet sizes and inter-packet delays (tested up to n = 3), closing the timing gap.

2012-moghaddam-skypemorph §4.3, §4.4 defense

BTP's wire protocol contains no handshakes, timeouts, or plaintext headers. Connections open with a pseudo-random b-byte tag that the recipient can compute in advance from its key state, making BTP frames indistinguishable from random data to a passive observer who does not know the shared secret.

2012-rogers-secure §2, §3.2 defense

StegoTorus distributes a fixed set of packet traces and HTTP covertext databases with the software, but allows users to record their own; classifiers trained on the distributed covertext will not generalize to user-generated databases. The paper further notes that reusing a small number of traces repeatedly creates a statistical fingerprint because censors can learn conversation patterns from packet sizes and timings alone, implying that trace diversity must be maintained over time.

2012-weinberg-stegotorus §4.1.1, §5.2 defense

Even with end-to-end encrypted messages, a censor observing subscription queries can detect anomalous interest in a short tag (e.g., a sudden domestic surge in followers of a foreign pop star's hashtag) and use timing/size traffic analysis to distinguish #h00t subscriptions from ordinary hashtag follows. The paper flags this as an open threat and proposes two mitigations: (1) push cover traffic for randomly selected short tags to all clients regardless of their actual subscriptions, or (2) silently redirect normal clients' hashtag follows to the corresponding #h00t short tags.

2011-bachrach-h00t §6.1 detection

If a large site such as Google or Wikipedia scrambled all served content using a publicly known de-scrambling algorithm, the censor faces a strict all-or-nothing blocking decision: it cannot selectively filter banned scrambled content without blocking the entire site, since scrambled legitimate and banned content are computationally indistinguishable prior to running S⁻¹. This property scales the political cost of blocking proportionally to the size of the co-scrambling platform.

2011-bonneau-scrambling §1 defense

Scrambling without secret key management can frustrate DPI-based censors if the de-scrambling function satisfies 'high-inertia' — meaning an adversary computing S⁻¹ on n inputs cannot use less than Θ(n) times the resources of a single commodity-PC user, including electricity, memory, and computation time. This forces bulk censorship to become computationally infeasible without over-censoring all scrambled content.

2011-bonneau-scrambling §1–2 defense

Transmitting the de-scrambling algorithm S⁻¹ as in-page JavaScript alongside AJAX-fetched scrambled content eliminates the need for special client software installation or trusted public-key distribution, removing the primary bootstrapping vulnerability that cryptographic censorship-resistance schemes (including Tor) share — a vulnerability exploited when Iran blocked Tor by filtering its Diffie-Hellman parameter bit sequence.

2011-bonneau-scrambling §1, §3 defense

The proposed multi-stage scrambling composes four orthogonal layers: (a) 128-bit AES with 20 bits stripped, requiring brute-force search; (b) an AES key derived from a CAPTCHA solution; (c) a memory-bound function key; and (d) blocks whose de-scrambling exploits JavaScript floating-point and string-processing quirks. Each layer independently forces a censor to build or emulate a distinct acceleration environment, multiplying total reverse-engineering cost.

2011-bonneau-scrambling §3 defense

Applying a BEAR all-or-nothing package transform (using a zero key) to message blocks forces any censor attempting to scan content to cache all blocks from all active concurrent transfers simultaneously, since no individual block reveals any information about the original message until all blocks are received. Artificially delaying block transmission amplifies censor state requirements proportionally.

2011-bonneau-scrambling §3 defense

Dust defeats DPI fingerprinting by constructing all packets from entirely encrypted or single-use random bytes (defeating static string matching), appending a random number of random padding bytes to every packet (defeating length matching), and permitting a complete client–server conversation to be encoded in a single UDP or TCP packet (defeating timing analysis for sufficiently small payloads).

2011-wiley-dust §3, §3 Discussion defense

Dust eliminates the in-band key-exchange fingerprint surface via an out-of-band half-handshake: the server's public key, IP, port, and a single-use secret are bundled into a PBKDF-encrypted invite packet transmitted out-of-band; only the decryption password (not the server IP) appears in plaintext, defeating the email/IM IP-address blocking attacks documented against prior systems.

2011-wiley-dust §3, §3.1 defense

Censors responding to encryption-based circumvention have two escalation options: block all encrypted connections outright, or identify the underlying protocol via traffic signatures that persist even inside encrypted tunnels. The paper frames these as the two dominant censor responses to DPI being defeated by encryption.

2011-wright-fine-grained §3 detection

Global anonymity is maximized when the anonymity set is large and behavior is uniformly distributed: 'global anonymity is maximal iff all subjects within the anonymity set are equally likely.' Strong global anonymity does not protect individual 'likely suspects' — even in a strong-anonymity system, one user with distinctive behavior may have weak individual anonymity. Strong or even maximal global anonymity does not imply strong anonymity of each particular subject.

2010-pfitzmann-terminology §3 defense

Adding dummy traffic to any anonymity mechanism yields the corresponding kind of unobservability: 'A mechanism to achieve some kind of anonymity appropriately combined with dummy traffic yields the corresponding kind of unobservability.' DC-nets achieve sender anonymity and MIX-nets achieve relationship anonymity; with dummy traffic both achieve the corresponding sender and relationship unobservability respectively.

2010-pfitzmann-terminology §8 defense

Undetectability of a message requires that it be indistinguishable from 'random noise' — an attacker cannot sufficiently distinguish whether the message exists or not. This is distinct from anonymity, which protects only the relationship between an IOI and a subject, not the IOI's existence itself. Undetectability is possible only for subjects not involved in the IOI; senders and recipients cannot achieve it against each other.

2010-pfitzmann-terminology §6 defense

The paper establishes a strict property hierarchy: unobservability ⇒ anonymity, and sender/recipient anonymity ⇒ relationship anonymity. Unobservability is strictly stronger than anonymity because it additionally requires undetectability against all uninvolved subjects — the IOI's very existence must be hidden — while anonymity only hides the subject's relationship to the IOI.

2010-pfitzmann-terminology §7 defense

When the GFC keyword blacklist is known, multiple server-side-only evasion techniques become viable requiring no client modification: IP packet fragmentation to split keywords across MTU boundaries, HTML comment injection mid-keyword (e.g., 'Fa<!- Comment ->lun Gong'), alternative URL percent-encodings (e.g., 'F%61lun Gong'), and spam-style character substitution ('F@1un G0-ng'); the GFC implementation was observed not to check control characters in URL requests.

2007-crandall-conceptdoppler §5 defense

Tor encrypts payload but does not obscure traffic volume, leaving a residual publisher-vs-reader asymmetry: a user publishing a home video generates a markedly different upload/download ratio than one reading news. The paper also notes that website fingerprinting attacks — where the adversary pre-downloads hundreds of popular sites and matches traffic patterns to a Tor client's stream — remain possible even through bridge circuits, and are exacerbated by Tor's varying supported protocols (web vs. IM produce different timing signatures).

2006-dingledine-design §8.2 detection

Dagster requires both clients and servers to enforce a randomness predicate rand?(x) on every block before storage or forwarding, ensuring all server-stored data is statistically indistinguishable from uniform random noise. This provides server deniability — the operator can credibly deny knowledge of content — and also closes the attack present in Publius and Freenet where a malicious client could post plaintext, potentially exposing the operator for 'knowingly' hosting illegal content.

2001-stubblefield-dagster §4.2, §5.3 defense

Traffic analysis is identified as the primary threat to location secrecy in a distributed anonymous storage system: if an adversary can correlate inter-server communications or link requests to stored file locations, it can target physical seizure. The paper proposes mix-nets (Chaum 1981) for user-facing file delivery and dining-cryptographers ring protocols for inter-server communications, supplemented by traffic padding, so that even traffic analysis yields no actionable location information.

1996-anderson-eternity §4.5 detection