BLAKE3
Generated by GPT-5-miniExpansion Funnel Raw 113 → Dedup 0 → NER 0 → Enqueued 0
| BLAKE3 | |
|---|---|
| Name | BLAKE3 |
| Designers | Hugo Krawczyk; Jean-Philippe Aumasson; Sam Lang; Jack O'Connor |
| Published | 2020 |
| Type | Cryptographic hash function |
| Digest size | 32 bytes |
| Block size | 64 bytes |
BLAKE3 is a cryptographic hash function designed for high performance, parallelism, and modern platform support. It was introduced by a team including Hugo Krawczyk, Jean-Philippe Aumasson, Sam Lang, and Jack O'Connor, and aims to combine the design lineage of earlier hash functions with advances in concurrency and extendable-output functionality. The algorithm emphasizes speed on desktop, server, mobile, and embedded environments while maintaining a conservative security posture influenced by prior competitions and analysis.
History
BLAKE3 emerged from a lineage of designs and evaluations rooted in projects and events such as the Advanced Encryption Standard, the SHA-3 competition, the NIST hash standardization efforts, and earlier algorithms like MD5, SHA-1, SHA-2, and SHA-3. Its authors drew on work from the BLAKE family, which itself was a finalist in the SHA-3 competition alongside submissions like Keccak, Grøstl, and Skein. Influences and cross-references include academic groups at institutions like ETH Zurich, MIT, Stanford University, University of California, Berkeley, and research labs at Google, Microsoft, Apple Inc., and ARM Ltd.. Public discussion occurred on platforms used by cryptographers such as the IACR conferences, CRYPTO, Eurocrypt, and workshops linked to USENIX and RSA Conference. The release and announcement were disseminated through repositories and communities like GitHub, Reddit, and mailing lists used by contributors to projects like OpenSSL, LibreSSL, BoringSSL, and GnuPG. Adoption considerations referenced standards bodies such as IETF and organizations like NIST and ISO for later evaluation.
Design and algorithm
The design of the function builds on primitives and ideas used by predecessors including the ChaCha stream cipher, permutation concepts discussed in papers from Ronald Rivest, Adi Shamir, and Leonard Adleman, and the sponge construction popularized by Keccak. Its internal mixing function leverages a keyed compression function related to constructs evaluated in submissions from teams such as Jean-Philippe Aumasson's prior work and algorithms used in Poly1305 designs. The algorithm introduces a tree hashing mode inspired by parallel constructions in systems developed at Google for large-scale data processing and by distributed storage systems like Hadoop, Ceph, and Amazon S3. It supports extendable-output features akin to designs in standards produced by NIST and research from groups at Carnegie Mellon University and ETH Zurich. The core uses a 64-byte block schedule with rotation and addition operations similar to constructions analyzed by cryptographers including Daniel J. Bernstein, Tanja Lange, and Peter Schwabe.
Performance and benchmarks
Benchmarks for the function have been published by implementers and researchers at organizations such as Google, Microsoft Research, Amazon Web Services, Mozilla, and academic groups from University of Illinois Urbana–Champaign and ETH Zurich. Comparative reports often include references to performance of SHA-256, SHA-3 (Keccak), BLAKE2, and implementations in standard libraries like those maintained by OpenSSL, LibreSSL, BoringSSL, and language ecosystems from Python, Rust, Go, Java, and Node.js. Test vectors and throughput studies have been conducted on hardware from Intel Corporation, AMD, ARM Ltd., and accelerators from NVIDIA, using toolchains like GCC, Clang, and MSVC. Results cited by cloud providers including Google Cloud Platform, Microsoft Azure, and Amazon Web Services show strong single-thread and multi-thread scalability, with parallelism leveraged in systems like Kubernetes, Docker, and OpenStack for fast integrity checks.
Security and cryptanalysis
Security evaluation references include analysis by researchers published in venues such as IACR ePrint Archive, CRYPTO, Eurocrypt, and journals associated with ACM and IEEE. Comparisons often cite the cryptanalytic history of algorithms like MD5, SHA-1, SHA-2, and earlier BLAKE designs, and draw on techniques developed by teams including those of Wang Xiaoyun, Xiaoyun Wang, Alex Biryukov, Luca Henzen, and Florian Mendel. Formal proofs and reductions reference methods used in works from Hugo Krawczyk and standards discussions at IETF and NIST, while security margins are explored in collaborations including researchers from ETH Zurich, INRIA, and Technische Universität Darmstadt. No catastrophic weaknesses comparable to collision attacks on MD5 or SHA-1 have been published; continued scrutiny by the cryptographic community and institutions such as IACR remains active.
Implementations and libraries
Production and reference implementations exist across ecosystems maintained by projects and organizations like GitHub, OpenSSL, Mozilla Foundation, Rust Foundation, The Linux Foundation, and contributors from Google, Microsoft, Amazon Web Services, and Cloudflare. Implementations are available in languages and runtimes including C, C++, Rust, Go, Python, Java, JavaScript, and WebAssembly for use in environments such as Node.js, Electron, and Firefox. Integrations have been made in tools and systems like git, rsync, OpenSSH, WireGuard, and backup solutions maintained by communities around BorgBackup, Restic, and Duplicacy. Commercial vendors in cloud, CDN, and security sectors including Cloudflare, Akamai, Fastly, Amazon Web Services, and Microsoft Azure have explored or adopted implementations for integrity checks and content-addressable storage.
Applications and use cases
Typical use cases span file integrity verification in distributed systems such as Ceph, Hadoop, GlusterFS, and IPFS, cryptographic libraries in projects like OpenSSL and GnuPG, and package management ecosystems including Debian, Fedora Project, Homebrew, and npm. It is used for deduplication and chunking in backup systems developed by teams at Restic, BorgBackup, and enterprise vendors including Veeam. Developers in cloud platforms such as Google Cloud Platform, Microsoft Azure, and Amazon Web Services use it for fast checksumming in container registries like Docker Hub and orchestration platforms like Kubernetes. Additional applications include content-addressable storage in projects like Git, databases and blockchains explored by research groups at Ethereum Foundation, Hyperledger, and startups funded by Y Combinator, where fast hashing accelerates transaction processing and state synchronization.