FINDING · EVALUATION
Protozoa's encoded media tunneling achieves an AUC of 0.59 against a state-of-the-art ML traffic classifier using packet-size and inter-arrival-time features—near the 0.5 random-guessing baseline—compared to >99% detection rates for prior tools such as Facet and DeltaShaper. To block 80% of Protozoa flows (TPR=0.8), a censor would erroneously flag approximately 60% of legitimate WebRTC flows (FPR=0.6). This resistance holds across trace durations from 10–60 seconds (AUC range 0.56–0.61) and across RTT, bandwidth, and packet-loss variations.
From 2020-barradas-poking — Poking a Hole in the Wall: Efficient Censorship-Resistant Internet Communications by Parasitizing on WebRTC · §6.2, Table 1, Figure 8 · 2020 · Computer and Communications Security
Implications
- Design covert channels that embed data after codec compression (not into raw pixels) to avoid distorting packet-size distributions that ML classifiers exploit.
- Maintain continuous dummy traffic ('chaff') during idle periods so the classifier cannot distinguish active from idle covert sessions.
Tags
Extracted by claude-sonnet-4-6 — review before relying.