User:StefanoZacchiroli/Content deduplication

Some experiments on deduplicating contents at sub-file granularity.

Datasets

Linux kernel, Git repo

• origin: git.kernel.org, on 2018-01-06
• 1.653.941 content blobs, for a total of 19 GB (compressed)

Rabin fingerprints

• Approach: use Rabin fingerprints
• Implementation: swh-dedup-blocks.py

test 1: linux git

Rabin fingerprint parameters:

• prime: 3
• window_size: 48 KB
• min_block_size: 2 KB
• avg_block_size: 8 KB
• max_block_size: 64 KB

Results:

• original size (uncompressed): 55.89 GB
• dedup chunk size (uncompressed): 19.87 GB (35.55%)

References

• LBFS
• Hailstorm
• DCT
• ScanCode tokenizer