I'm Unique, Just Like You: Human Side-Channels and Their Implications for Security and Privacy
summary
Defines human side-channels as unintended behavioral identifiers (keystroke timing, mouse cadence, interaction rhythms) and shows concrete attacks that turn those signals into persistent IDs using in-browser JavaScript collectors and server-side correlation. Demonstrations include proof-of-concept scripts and a proof-of-concept classifier trained on interaction traces that successfully re-identifies users across sessions, plus defensive experiments that add noise, throttle events, and require permission prompts which materially reduce tracking. Practical mitigations and API controls are shown to work.
AI-written summary · about
chapters (12)
- 0:00Introduction and welcome
- 1:06Kubernetes architecture
- 4:48Attack vectors in components
- 9:03API server interactions
- 14:06Direct scheduling demonstration
- 19:14The dangers of host paths
- 21:30Docker and container breakouts
- 25:59Linux primitives and containers
- 30:56Advanced namespace manipulation
- 35:02Mitigation with admission control
- 38:46Policy enforcement demonstration
- 42:13Final takeaways and resources
related talks
49mWeb Side Story: A Privacy Odyssey through Tracking and Fingerprinting
27mJavaScript, Cyber-abuse and the Illusion of Privacy
14mBrowser Fingerprinting: Stalking With a Personal Touch
36mIt will slip by, or about Side Channels
40mModern Single Sign-On: On the Security of Single Sign-On ...
44m📗 Threat Chords: Tuning into Persistent Patterns in Adversary Behavior 🔥
26mLoophole: Timing Attacks on Shared Event Loops in Chrome
21mTuesday Lightning Talks 2