Skip to content

Feat/gdbreach#1

Open
jongcoding wants to merge 10 commits intomainfrom
feat/gdbreach
Open

Feat/gdbreach#1
jongcoding wants to merge 10 commits intomainfrom
feat/gdbreach

Conversation

@jongcoding
Copy link
Owner

No description provided.

jongcoding and others added 10 commits September 21, 2025 23:28
- Add binary search for O(log N) shrinkage detection
- Implement paper-style flush with FLUSH TABLES WITH READ LOCK + mtime polling
- Use ls -s --block-size=1 for accurate disk block measurement
- Add K-of-N attack implementation with multiple scoring methods
- Configure MariaDB 10.3 with compression+encryption support
- Successfully detect secrets at 92-97 bytes vs random at 100-150 bytes

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Add eval_kofn.py: configurable trials with accuracy/query metrics
- Add analyze_kofn.py: generate summary tables and comparisons
- Add run_eval_experiments.sh: automated experiment runner
- Support multiple datasets (english, random, email)
- Track precision/recall/F1 and DB query counts

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Random data: 90% accuracy (n=20, k=3), 76% (n=40, k=5)
- English words: 36.7% accuracy (pattern similarity)
- Binary search: O(log N) efficiency confirmed
- Results match G-DBREACH paper findings

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Added full parameter grid: n={20,40,100}, k={1,5,10}
- Random dataset: 50-88% accuracy across configurations
- English dataset: 40-88% accuracy
- Query efficiency: ~10-12 queries per candidate (O(log N))

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Chart.js로 4개의 그래프 추가
- 정확도 비교 (Random vs English)
- n, k 파라미터별 결과 시각화
- 논문 스타일 한국어 보고서

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Added complete English results: n={20,40,100}, k={1,5,10}
- English n=20 k=10: 81.1%, n=40 k=10: 78.8%
- Updated charts with full data
- Enhanced paper comparison table
- Added key observations matching paper findings

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant

Comments