SYSTEM Cited by 1 source

Snappy compression¶

What it is¶

Snappy is a fast compression/decompression library originally developed at Google, optimised for high throughput and low CPU cost rather than maximum compression ratio. Widely used in storage engines (LevelDB, RocksDB, Cassandra) and columnar formats (Parquet) where the trade-off is "save bandwidth / disk I/O without paying expensive CPU."

Why it shows up on this wiki¶

Canonical citation is as the client-side wire compressor in PlanetScale's HTTP API (Source: sources/2026-04-21-planetscale-faster-mysql-with-http3). The experimental Go database/sql-compatible driver encodes queries with protobuf and Snappy- compresses the payload client-side before sending — enabling client-side query compression, which is structurally impossible over the MySQL binary protocol.

Architectural properties¶

Fast compression/decompression — designed for the throughput-sensitive case (many small payloads / hot path).
Moderate compression ratio — not as compact as zstd or LZ4HC; close to LZ4-fast.
Minimal metadata overhead — suitable for per-message wire-format compression without a shared dictionary.
Google origin — used internally at Google for years before open-sourcing; battle-tested.

Stub¶

Minimum viable page. Expand when a source is ingested that discloses Snappy-specific production retrospective material.

Seen in¶

sources/2026-04-21-planetscale-faster-mysql-with-http3 — client-side wire compressor in the experimental HTTP-API Go driver. "Snappy for the compression." Canonical wiki instance of client-side query compression in a database- transport context.