7 releases (stable)
|1.0.5||Jul 23, 2020|
|1.0.4||Jul 9, 2020|
|1.0.3||Mar 19, 2020|
|1.0.2||Aug 5, 2019|
|0.1.0||Feb 6, 2019|
#114 in Algorithms
215 downloads per month
Used in 7 crates (4 directly)
This crate implements the "FastCDC" content defined chunking algorithm in pure Rust. A critical aspect of its behavior is that it returns exactly the same results for the same input. To learn more about content defined chunking and its applications, see the reference material linked below.
- Rust stable (2018 edition)
$ cargo clean $ cargo build $ cargo test
An example can be found in the
examples directory of the source repository,
which demonstrates reading files of arbitrary size into a memory-mapped buffer
and passing them through the chunker (and computing the SHA256 hash digest of
$ cargo run --example dedupe -- --size 32768 test/fixtures/SekienAkashita.jpg Finished dev [unoptimized + debuginfo] target(s) in 0.03s Running `target/debug/examples/dedupe --size 32768 test/fixtures/SekienAkashita.jpg` hash=5a80871bad4588c7278d39707fe68b8b174b1aa54c59169d3c2c72f1e16ef46d offset=0 size=32857 hash=13f6a4c6d42df2b76c138c13e86e1379c203445055c2b5f043a5f6c291fa520d offset=32857 size=16408 hash=0fe7305ba21a5a5ca9f89962c5a6f3e29cd3e2b36f00e565858e0012e5f8df36 offset=49265 size=60201
The unit tests also have some short examples of using the chunker, of which this code snippet is an example:
let read_result = fs::read("test/fixtures/SekienAkashita.jpg"); assert!(read_result.is_ok()); let contents = read_result.unwrap(); let chunker = FastCDC::new(&contents, 16384, 32768, 65536); let results: Vec<Chunk> = chunker.collect(); assert_eq!(results.len(), 3); assert_eq!(results.offset, 0); assert_eq!(results.length, 32857); assert_eq!(results.offset, 32857); assert_eq!(results.length, 16408); assert_eq!(results.offset, 49265); assert_eq!(results.length, 60201);
The algorithm is as described in "FastCDC: a Fast and Efficient Content-Defined Chunking Approach for Data Deduplication"; see the paper, and presentation for details. There are some minor differences, as described below.
The explanation below is copied from ronomon/deduplication since this codebase is little more than a translation of that implementation:
The following optimizations and variations on FastCDC are involved in the chunking algorithm:
- A right shift instead of a left shift to remove the need for an additional modulus operator, which would otherwise have been necessary to prevent overflow.
- Masks are no longer zero-padded since a right shift is used instead of a left shift.
- A more adaptive threshold based on a combination of average and minimum chunk size (rather than just average chunk size) to decide the pivot point at which to switch masks. A larger minimum chunk size now switches from the strict mask to the eager mask earlier.
- Masks use 1 bit of chunk size normalization instead of 2 bits of chunk size normalization.
The primary objective of this codebase was to have a Rust implementation with a permissive license, which could be used for new projects, without concern for data parity with existing implementations.
This crate is little more than a rewrite of the implementation by Joran Dirk Greef (see the ronomon link below), in Rust, and greatly simplified in usage. One significant difference is that the chunker in this crate does not calculate a hash digest of the chunks.
- rdedup_cdc at docs.rs
- An alternative implementation of FastCDC to the one in this crate.
- Similar but slightly earlier algorithm by some of the same researchers.
- Pure Python port of FastCDC. Compatible with this implementation.