MurmurHash

MurmurHash is a family of non-cryptographic hash functions designed for fast, non-secure hashing of data. Developed by Austin Appleby and released in 2008, the algorithms are intended for hash-based data structures such as hash tables and for quick data indexing. They emphasize speed, good distribution, and simplicity across platforms.

The lineage includes MurmurHash1 and MurmurHash2, followed by MurmurHash3, which introduced several variants optimized for different architectures. MurmurHash3 provides 32-bit and 128-bit output forms, with separate implementations for x86 and x64 platforms. The 32-bit variant is known for speed in small-key scenarios, while the 128-bit versions offer better collision resistance for large sets.

In operation, MurmurHash processes input in fixed-size blocks, applying multiplications, rotations, and XOR operations to mix bits thoroughly. The 32-bit version uses a sequence of mixing steps on 4-byte blocks, while the 128-bit variants operate on 16-byte blocks. A finalization step ensures diffusion of all input bits.

Applications include non-secure hashing for hash tables, caches, and distributed systems, as well as general-purpose data indexing in programming languages. It is widely implemented and ported across languages (C, C++, Java, Go, Python, and more).

Although fast and well-distributed, MurmurHash is not suitable for cryptographic purposes. It is vulnerable to collision and preimage attacks, and should not be used for password hashing or security-sensitive tasks. Users should select a variant and output size appropriate to their data size and performance needs.