Perfectly Deniable Steganographic Disk Encryption
summary
A working steganographic disk system is implemented as a Linux kernel module (about 30,000 lines) plus tiny assembly bootstraps that self‑bootstrap from scattered slack sectors (bootstrap‑in‑sectors). Data is spread with randomized placement and recovered with forward error correction using LDPC, allowing recovery even after roughly 50% overwrite of slack space. It runs fast on NVMe, can host a hidden OS (even VirtualBox), and survives normal cover use; practical limits include multi‑snapshot forensics and SSD/TRIM/FTL quirks that weaken guarantees.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 1:04Outline
- 2:21Steganography
- 4:54topical applications
- 7:53Forensic Analysis Techniques
- 8:36Essential Ingredients
- 9:30Background
- 10:02Vanilla Disk Encryption
- 10:55Hidden Partition Functionality
- 12:26Veracrypt
- 13:24Practice
- 15:04Problem Solution
- 15:58Birds Eye View
- 16:52Air Correction
- 18:31Randomization
- 20:25Kernel Architecture
- 24:27In Practice
- 39:39Takeaways
- 41:51System Health
- 45:34Questions
related talks
34mAuthenticated and Encrypted Storage on Embedded Linux
45mFull Disk Encryption for Confidential Computing Guests
34mRecoverable, Tamper-resistant Full-disk Encryption at the Distributed Edge
53mSecure Embedded Linux Product - A Success Story
57mFaux Disk Encryption: Realities of Secure Storage On Mobile Devices
13mlo-fi spy: video, steganography, and you
45mNavigating Security Tradeoffs in Embedded Linux Systems
30mSecrets in the Pixels! Exploring Steganography