Storage Carving Tools is a project developed as part of Smart India Hackathon 2024, focused on efficient data recovery from filesystems. Written in C++ with CUDA support, it currently supports XFS and will soon add BTRFS support. This project is designed to handle data carving with enhanced performance and accuracy.
- Filesystem-Aware Carving: Parses filesystem structures (currently XFS) to focus on unallocated locations.
- Fragmentation Handling: Appropriately manages fragmented data to improve recovery precision.
- CUDA-Accelerated Processing: Uses CUDA to speed up the carving process for large datasets.
- XFS Structure Parsing: Extracts metadata and filesystem layout information to identify unallocated blocks.
- Unallocated Space Scanning: Targets only unused storage areas, reducing time and unnecessary reads.
- Parallel Data Processing: Leverages GPU parallelism to enhance performance and efficiency.
This project aims to:
- Deepen understanding of filesystem internals (XFS and BTRFS).
- Apply CUDA programming to real-world data processing challenges.
- Improve data carving techniques with a focus on performance and correctness.
- Currently limited to XFS; BTRFS support is under development.
- Limited recovery capabilities compared to full-featured file recovery suites.
- Add support for BTRFS filesystem parsing and carving.
- Implement advanced file-type-specific carving strategies.
- Optimize CUDA kernel designs for further performance gains.