Zstandard Dictionary Content-Encoding

Introduction Zstandard is a compression algorithm that supports dictionaries for improved compression of small payloads. This document defines a new HTTP content coding, "zstdict", to signal Zstandard compression with a pre-shared dictionary. The dictionary is identified by its SHA-256 hash. Clients MUST verify the hash before decoding.

The "zstdict" Content Coding The "zstdict" content coding is identical to "zstd" , except that a dictionary is used for both compression and decompression. A response using "zstdict" MUST include a Content-Dict header containing the SHA-256 hash of the dictionary, in lowercase hex. Example:

The Content-Dict Header The Content-Dict header identifies the dictionary used for compression. Its value is a hash algorithm identifier followed by the hex-encoded hash.

Dictionary Distribution The dictionary MUST be available to the client before decoding. Three distribution mechanisms are defined:

Well-Known URI Servers MAY publish dictionaries at: ]]> where <hash> is the SHA-256 hash from Content-Dict. Clients MAY fetch this URI on first encounter and cache it.

Application-Bundled Clients MAY ship dictionaries with the application. In this case, the Content-Dict header is informational and the client uses the bundled dictionary matching the hash.

Reference Implementation A reference dictionary and C library implementation for this specification are available at:

Security Considerations Dictionary deserialization is a known attack vector. Clients MUST validate the dictionary hash before use. Malicious dictionaries can cause denial of service. Implementations MUST limit the maximum size of accepted dictionaries.

IANA Considerations This document registers "zstdict" in the HTTP Content Coding Registry.

Name:: zstdict
Description:: Zstandard compression with pre-shared dictionary
Reference:: this document