Zstandard Seekable Format

May 4, 2025 ยท View on GitHub

Notices

The initial version of this document, as well as the copyright notice, are copied from version 0.1.0 of the seekable format specification of zstd.

Copyright (c) Meta Platforms, Inc. and affiliates.

Permission is granted to copy and distribute this document for any purpose and without charge, including translations into other languages and incorporation into compilations, provided that the copyright notice and this notice are preserved, and that any substantive changes or deletions from the original are clearly marked. Distribution of this document is unlimited.

Version

0.1.1 (02/05/25)

Introduction

This document defines a format for compressed data to be stored so that subranges of the data can be efficiently decompressed without requiring the entire document to be decompressed. This is done by splitting up the input data into frames, each of which are compressed independently, and so can be decompressed independently. Decompression then takes advantage of a provided 'seek table', which allows the decompressor to immediately jump to the desired data. This is done in a way that is compatible with the original Zstandard format by placing the seek table in a Zstandard skippable frame.

Overall conventions

In this document:

  • square brackets i.e. [ and ] are used to indicate optional fields or parameters.
  • the naming convention for identifiers is Mixed_Case_With_Underscores
  • All numeric fields are little-endian unless specified otherwise

Format

The format consists of a number of frames (Zstandard compressed frames and skippable frames), and a skippable frame containing the seek table. The seek table frame can either be appended to the end of all other frames, or written to a separate file.

Seek Table Format

The seek table can be structured in two different formats, i.e. the classic Foot format and the Head format, which was added in version 0.1.1 of this document. Seekable decoders start reading Foot seek tables from the end, while Head seek tables can be read from the beginning without needing to seek to the end of the file. The classic Foot format is designed to be placed at the end of seekable archives. The Head format, on the other hand, should be placed in stand-alone files.

In version 0.1.0 of this document, Seek_Table_Entries contained an optional checksum. Placing checksums in the seek table is deprecated in version 0.1.1, the Content_Checksum field of regular Zstandard frames is used instead. However, seek tables that contain checksums can still be decoded successfully, although any checksum data is ignored.

Foot

The seek table integrity field is at the end of the skippable frame, after any frame data. This format is the same as specified in version 0.1.0 of this document.

Skippable_Magic_NumberFrame_Size[Seek_Table_Entries]Seek_Table_Integrity
4 bytes4 bytes8 bytes each9 bytes

Head

The seek table integrity field is placed directly after the skippable header, before any frame data. This format was added in version 0.1.1 of this document and is incompatible with prior decoders.

Skippable_Magic_NumberFrame_SizeSeek_Table_Integrity[Seek_Table_Entries]
4 bytes4 bytes9 bytes8 bytes each

Skippable_Magic_Number

Value : 0x184D2A5E. This is for compatibility with Zstandard skippable frames. Since it is legal for other Zstandard skippable frames to use the same magic number, it is not recommended for a decoder to recognize frames solely on this.

Frame_Size

The total size of the skippable frame, not including the Skippable_Magic_Number or Frame_Size. This is for compatibility with Zstandard skippable frames.

Seek_Table_Integrity

The seek table integrity format is as follows:

Number_Of_FramesSeek_Table_DescriptorSeekable_Magic_Number
4 bytes1 byte4 bytes

Seek_Table_Integrity is called Seek_Table_Footer in version 0.1.0 of this document.

Seekable_Magic_Number

Value : 0x8F92EAB1. This value must be the last bytes present in the integrity field so that decoders can efficiently find it and determine if there is an actual seek table present.

Number_Of_Frames

The number of stored frames in the data.

Seek_Table_Descriptor

A bitfield describing the format of the seek table.

Bit numberField name
7Checksum_Flag
6-2Reserved_Bits
1-0Unused_Bits

While only Checksum_Flag currently exists, there are 7 other bits in this field that can be used for future changes to the format, for example the addition of inline dictionaries.

Checksum_Flag

The checksum flag is not actively used in version 0.1.1. Checksum data will not be placed in the seek table, instead the Content_Checksum field of regular Zstandard frames is used. Every seek table created with version 0.1.1 sets Checksum_Flag to zero.

Reserved_Bits are not currently used but may be used in the future for breaking changes, so a compliant decoder should ensure they are set to 0. Unused_Bits may be used in the future for non-breaking changes, so a compliant decoder should not interpret these bits.

Seek_Table_Entries

Seek_Table_Entries consists of Number_Of_Frames (one for each frame in the data, not including the seek table frame) entries of the following form, in sequence:

Compressed_SizeDecompressed_Size[Checksum]
4 bytes4 bytes4 bytes

Compressed_Size

The compressed size of the frame. The cumulative sum of the Compressed_Size fields of frames 0 to i gives the offset in the compressed file of frame i+1.

Decompressed_Size

The size of the decompressed data contained in the frame. For skippable or otherwise empty frames, this value is 0.

Checksum

Not used in version 0.1.1, created seek tables will never contain checksums. Legacy seek tables that have Checksum_Flag set and contain checksums can be decoded, but the checksums are ignored. The Content_Checksum field of regular Zstandard frames is used instead.

Version Changes

  • 0.1.1: add Foot and Head seek table formats, deprecate checksum data in seek table
  • 0.1.0: initial version