Zombie POODLE, GOLDENDOODLE, and How TLSv1.3 Can Save Us All
summary
Describes a scanner that finds CBC padding oracles across Alexa sites, reports ~3,689 observable vulnerable domains (~1.6% of top 100k) and ~5,800 likely exploitable hosts, and demonstrates two exploits: GOLDENDOODLE (deterministic brute force that recovers any plaintext byte in ≤255 requests, often ≈95 printable chars) and zombie POODLE (inverted oracle behavior). Built a Python proof‑of‑concept using iptables and curl, validated against Cisco/SSL VPNs and vendor devices, and provides a padcheck tool to reproduce scans. TLS 1.3 migration is recommended.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:07TLS basics
- 2:13Cipher block chaining
- 3:33TLS encryption process
- 6:58Malleability and padding
- 8:19Understanding padding oracles
- 10:12The POODLE attack explained
- 15:41Testing for POODLE
- 17:36Developing a new scanner
- 20:15Designing test cases
- 22:33Finding GoldenDoodle
- 26:08Exploiting GoldenDoodle
- 29:47The Zombie POODLE finding
- 32:21Scan results and analysis
- 36:08Vendor disclosures
- 37:44Limitations and future work
- 38:49TLS 1.3 as the solution
- 40:38Q&A and conclusion
related talks
24mReturn Of Bleichenbacher’s Oracle Threat (ROBOT)
44mThe DROWN Attack
25mDROWN: Breaking TLS Using SSLv2
01h 16mClassic Vulnerabilities
01h 01mHTTPS & TLS in 2016: Security practices from the front lines
53mThe BEAST Wins Again: Why TLS Keeps Failing to Protect HTTP
13mX.509DoS: Exploiting and Detecting Denial-of-Service Vulnerabilities in
29mLike Lightning From the Cloud: Finding RCEs in an Embedded TLS Library