Open Container Initiative Image Format Specification

Open Container Initiative

Image Format Specification

This specification defines an OCI Image, consisting of an image manifest, an image index (optional), a set of filesystem layers, and a configuration.

The goal of this specification is to enable the creation of interoperable tools for building, transporting, and preparing a container image to run.

Table of Contents

Notational Conventions

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" are to be interpreted as described in RFC 2119 (Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997).

The key words "unspecified", "undefined", and "implementation-defined" are to be interpreted as described in the rationale for the C99 standard.

An implementation is not compliant if it fails to satisfy one or more of the MUST, MUST NOT, REQUIRED, SHALL, or SHALL NOT requirements for the protocols it implements. An implementation is compliant if it satisfies all the MUST, MUST NOT, REQUIRED, SHALL, and SHALL NOT requirements for the protocols it implements.

Overview

At a high level the image manifest contains metadata about the contents and dependencies of the image including the content-addressable identity of one or more filesystem layer changeset archives that will be unpacked to make up the final runnable filesystem. The image configuration includes information such as application arguments, environments, etc. The image index is a higher-level manifest which points to a list of manifests and descriptors. Typically, these manifests may provide different implementations of the image, possibly varying by platform or other attributes.

build diagram

Once built the OCI Image can then be discovered by name, downloaded, verified by hash, trusted through a signature, and unpacked into an OCI Runtime Bundle.

runtime diagram

Understanding the Specification

The OCI Image Media Types document is a starting point to understanding the overall structure of the specification.

The high-level components of the spec include:

Future versions of this specification may include the following OPTIONAL features:

OCI Image Media Types

The following media types identify the formats described here and their referenced resources:

The following media types identify a "Layer" with distribution restrictions, but are deprecated and not recommended for future use:

Media Type Conflicts

Blob retrieval methods MAY return media type metadata. For example, a HTTP response might return a manifest with the Content-Type header set to application/vnd.oci.image.manifest.v1+json. Implementations MAY also have expectations for the blob's media type and digest (e.g. from a descriptor referencing the blob).

Compatibility Matrix

The OCI Image Specification strives to be backwards and forwards compatible when possible. Breaking compatibility with existing systems creates a burden on users whether they are build systems, distribution systems, container engines, etc. This section shows where the OCI Image Specification is compatible with formats external to the OCI Image and different versions of this specification.

application/vnd.oci.image.index.v1+json

Similar/related schema:

application/vnd.oci.image.manifest.v1+json

Similar/related schema:

application/vnd.oci.image.layer.v1.tar+gzip

Interchangeable and fully compatible mime-types:

application/vnd.oci.image.config.v1+json

Similar/related schema:

.config.StopSignal and .config.Labels are accidentally undocumented in Docker Image Spec v1.2, but these fields are not OCI-specific concepts.

Relations

The following figure shows how the above media types reference each other:

media types

Descriptors are used for all references. The image-index being a "fat manifest" references a list of image manifests per target platform. An image manifest references exactly one target configuration and possibly many layers.

OCI Content Descriptors

This section defines the application/vnd.oci.descriptor.v1+json media type.

Properties

A descriptor consists of a set of properties encapsulated in key-value fields.

The following fields contain the primary properties that constitute a Descriptor:

Descriptors pointing to application/vnd.oci.image.manifest.v1+json SHOULD include the extended field platform, see Image Index Property Descriptions for details.

Reserved

Extended Descriptor field additions proposed in other OCI specifications SHOULD first be considered for addition into this specification.

Digests

The digest property of a Descriptor acts as a content identifier, enabling content addressability. It uniquely identifies content by taking a collision-resistant hash of the bytes. If the digest can be communicated in a secure manner, one can verify content from an insecure source by recalculating the digest independently, ensuring the content has not been modified.

The value of the digest property is a string consisting of an algorithm portion and an encoded portion. The algorithm specifies the cryptographic hash function and encoding used for the digest; the encoded portion contains the encoded result of the hash function.

A digest string MUST match the following grammar:

digest                ::= algorithm ":" encoded
algorithm             ::= algorithm-component (algorithm-separator algorithm-component)*
algorithm-component   ::= [a-z0-9]+
algorithm-separator   ::= [+._-]
encoded               ::= [a-zA-Z0-9=_-]+

Note that algorithm MAY impose algorithm-specific restriction on the grammar of the encoded portion. See also Registered Algorithms.

Some example digest strings include the following:

digest algorithm Registered
sha256:6c3c624b58dbbcd3c0dd82b4c53f04194d1247c6eebdaab7c610cf7d66709b3b SHA-256 Yes
sha512:401b09eab3c013d4ca54922bb802bec8fd5318192b0a75f201d8b372742... SHA-512 Yes
multihash+base58:QmRZxt2b1FVZPNqd8hsiykDL3TdBDeTSPX9Kv46HmX4Gx8 Multihash No
sha256+b64u:LCa0a2j_xo_5m0U8HTBBNBNCLXBkg7-g-YpeiGJm564 SHA-256 with urlsafe base64 No

Please see Registered Algorithms for a list of registered algorithms.

Implementations SHOULD allow digests with unrecognized algorithms to pass validation if they comply with the above grammar. While sha256 will only use hex encoded digests, separators in algorithm and alphanumerics in encoded are included to allow for extensions. As an example, we can parameterize the encoding and algorithm as multihash+base58:QmRZxt2b1FVZPNqd8hsiykDL3TdBDeTSPX9Kv46HmX4Gx8, which would be considered valid but unregistered by this specification.

Verification

Before consuming content targeted by a descriptor from untrusted sources, the byte content SHOULD be verified against the digest string. Before calculating the digest, the size of the content SHOULD be verified to reduce hash collision space. Heavy processing before calculating a hash SHOULD be avoided. Implementations MAY employ canonicalization of the underlying content to ensure stable content identifiers.

Digest calculations

A digest is calculated by the following pseudo-code, where H is the selected hash algorithm, identified by string <alg>:

let ID(C) = Descriptor.digest
let C = <bytes>
let D = '<alg>:' + Encode(H(C))
let verified = ID(C) == D

Above, we define the content identifier as ID(C), extracted from the Descriptor.digest field. Content C is a string of bytes. Function H returns the hash of C in bytes and is passed to function Encode and prefixed with the algorithm to obtain the digest. The result verified is true if ID(C) is equal to D, confirming that C is the content identified by D. After verification, the following is true:

D == ID(C) == '<alg>:' + Encode(H(C))

The digest is confirmed as the content identifier by independently calculating the digest.

Registered algorithms

While the algorithm component of the digest string allows the use of a variety of cryptographic algorithms, compliant implementations SHOULD use SHA-256.

The following algorithm identifiers are currently defined by this specification:

algorithm identifier algorithm
sha256 SHA-256
sha512 SHA-512

If a useful algorithm is not included in the above table, it SHOULD be submitted to this specification for registration.

SHA-256

SHA-256 is a collision-resistant hash function, chosen for ubiquity, reasonable size and secure characteristics. Implementations MUST implement SHA-256 digest verification for use in descriptors.

When the algorithm identifier is sha256, the encoded portion MUST match /[a-f0-9]{64}/. Note that [A-F] MUST NOT be used here.

SHA-512

SHA-512 is a collision-resistant hash function which may be more performant than SHA-256 on some CPUs. Implementations MAY implement SHA-512 digest verification for use in descriptors.

When the algorithm identifier is sha512, the encoded portion MUST match /[a-f0-9]{128}/. Note that [A-F] MUST NOT be used here.

Embedded Content

In many contexts, such as when downloading content over a network, resolving a descriptor to its content has a measurable fixed "roundtrip" latency cost. For large blobs, the fixed cost is usually inconsequential, as the majority of time will be spent actually fetching the content. For very small blobs, the fixed cost can be quite significant.

Implementations MAY choose to embed small pieces of content directly within a descriptor to avoid roundtrips.

Implementations MUST NOT populate the data field in situations where doing so would modify existing content identifiers. For example, a registry MUST NOT arbitrarily populate data fields within uploaded manifests, as that would modify the content identifier of those manifests. In contrast, a client MAY populate the data field before uploading a manifest, because the manifest would not yet have a content identifier in the registry.

Implementations SHOULD consider portability when deciding whether to embed data, as some providers are known to refuse to accept or parse manifests that exceed a certain size.

Examples

The following example describes a Manifest with a content identifier of "sha256:5b0bcabd1ed22e9fb1310cf6c2dec7cdef19f0ad69efa1f392e94a4333501270" and a size of 7682 bytes:

{
  "mediaType": "application/vnd.oci.image.manifest.v1+json",
  "size": 7682,
  "digest": "sha256:5b0bcabd1ed22e9fb1310cf6c2dec7cdef19f0ad69efa1f392e94a4333501270"
}

In the following example, the descriptor indicates that the referenced manifest is retrievable from a particular URL:

{
  "mediaType": "application/vnd.oci.image.manifest.v1+json",
  "size": 7682,
  "digest": "sha256:5b0bcabd1ed22e9fb1310cf6c2dec7cdef19f0ad69efa1f392e94a4333501270",
  "urls": [
    "https://example.com/example-manifest"
  ]
}

In the following example, the descriptor indicates the type of artifact it is referencing:

{
  "mediaType": "application/vnd.oci.image.manifest.v1+json",
  "size": 123,
  "digest": "sha256:87923725d74f4bfb94c9e86d64170f7521aad8221a5de834851470ca142da630",
  "artifactType": "application/vnd.example.sbom.v1"
}

OCI Image Layout Specification

Given an image layout and a ref, a tool can create an OCI Runtime Specification bundle by:

Content

The image layout is as follows:

Example Layout

This is an example image layout:

$ cd example.com/app/
$ find . -type f
./index.json
./oci-layout
./blobs/sha256/3588d02542238316759cbf24502f4344ffcc8a60c803870022f335d1390c13b4
./blobs/sha256/4b0bc1c4050b03c95ef2a8e36e25feac42fd31283e8c30b3ee5df6b043155d3c
./blobs/sha256/7968321274dc6b6171697c33df7815310468e694ac5be0ec03ff053bb135e768

Blobs are named by their contents:

$ shasum -a 256 ./blobs/sha256/afff3924849e458c5ef237db5f89539274d5e609db5db935ed3959c90f1f2d51
afff3924849e458c5ef237db5f89539274d5e609db5db935ed3959c90f1f2d51 ./blobs/sha256/afff3924849e458c5ef237db5f89539274d5e609db5db935ed3959c90f1f2d51

Blobs

Example Blobs

$ cat ./blobs/sha256/9b97579de92b1c195b85bb42a11011378ee549b02d7fe9c17bf2a6b35d5cb079 | jq
{
  "schemaVersion": 2,
  "manifests": [
    {
      "mediaType": "application/vnd.oci.image.manifest.v1+json",
      "size": 7143,
      "digest": "sha256:afff3924849e458c5ef237db5f89539274d5e609db5db935ed3959c90f1f2d51",
      "platform": {
        "architecture": "ppc64le",
        "os": "linux"
      }
    },
...

$ cat ./blobs/sha256/afff3924849e458c5ef237db5f89539274d5e609db5db935ed3959c90f1f2d51 | jq
{
  "schemaVersion": 2,
  "config": {
    "mediaType": "application/vnd.oci.image.config.v1+json",
    "size": 7023,
    "digest": "sha256:5b0bcabd1ed22e9fb1310cf6c2dec7cdef19f0ad69efa1f392e94a4333501270"
  },
  "layers": [
    {
      "mediaType": "application/vnd.oci.image.layer.v1.tar+gzip",
      "size": 32654,
      "digest": "sha256:9834876dcfb05cb167a5c24953eba58c4ac89b1adf57f28f2f9d09af107ee8f0"
    },
...

$ cat ./blobs/sha256/5b0bcabd1ed22e9fb1310cf6c2dec7cdef19f0ad69efa1f392e94a4333501270 | jq
{
  "architecture": "amd64",
  "author": "Alyssa P. Hacker <alyspdev@example.com>",
  "config": {
    "Hostname": "8dfe43d80430",
    "Domainname": "",
    "User": "",
    "AttachStdin": false,
    "AttachStdout": false,
    "AttachStderr": false,
    "Tty": false,
    "OpenStdin": false,
    "StdinOnce": false,
    "Env": [
      "PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin"
    ],
    "Cmd": null,
    "Image": "sha256:6986ae504bbf843512d680cc959484452034965db15f75ee8bdd1b107f61500b",
...

$ cat ./blobs/sha256/9834876dcfb05cb167a5c24953eba58c4ac89b1adf57f28f2f9d09af107ee8f0
[gzipped tar stream]

oci-layout file

This JSON object serves as a marker for the base of an Open Container Image Layout and to provide the version of the image-layout in use. The imageLayoutVersion value will align with the OCI Image Specification version at the time changes to the layout are made, and will pin a given version until changes to the image layout are required. This section defines the application/vnd.oci.layout.header.v1+json media type.

oci-layout Example

{
    "imageLayoutVersion": "1.0.0"
}

index.json file

This REQUIRED file is the entry point for references and descriptors of the image-layout. The image index is a multi-descriptor entry point.

This index provides an established path (/index.json) to have an entry point for an image-layout and to discover auxiliary descriptors.

Implementor's Note: A common use case of descriptors with a "org.opencontainers.image.ref.name" annotation is representing a "tag" for a container image. For example, an image may have a tag for different versions or builds of the software. In the wild you often see "tags" like "v1.0.0-vendor.0", "2.0.0-debug", etc. Those tags will often be represented in an image-layout repository with matching "org.opencontainers.image.ref.name" annotations like "v1.0.0-vendor.0", "2.0.0-debug", etc.

Index Example

{
  "schemaVersion": 2,
  "mediaType": "application/vnd.oci.image.index.v1+json",
  "manifests": [
    {
      "mediaType": "application/vnd.oci.image.index.v1+json",
      "size": 7143,
      "digest": "sha256:0228f90e926ba6b96e4f39cf294b2586d38fbb5a1e385c05cd1ee40ea54fe7fd",
      "annotations": {
        "org.opencontainers.image.ref.name": "stable-release"
      }
    },
    {
      "mediaType": "application/vnd.oci.image.manifest.v1+json",
      "size": 7143,
      "digest": "sha256:e692418e4cbaf90ca69d05a66403747baa33ee08806650b51fab815ad7fc331f",
      "platform": {
        "architecture": "ppc64le",
        "os": "linux"
      },
      "annotations": {
        "org.opencontainers.image.ref.name": "v1.0"
      }
    },
    {
      "mediaType": "application/xml",
      "size": 7143,
      "digest": "sha256:b3d63d132d21c3ff4c35a061adf23cf43da8ae054247e32faa95494d904a007e",
      "annotations": {
        "org.freedesktop.specifications.metainfo.version": "1.0",
        "org.freedesktop.specifications.metainfo.type": "AppStream"
      }
    }
  ],
  "annotations": {
    "com.example.index.revision": "r124356"
  }
}

This illustrates an index that provides two named references and an auxiliary mediatype for this image layout.

The first named reference (stable-release) points to another index that might contain multiple references with distinct platforms and annotations. Note that the org.opencontainers.image.ref.name annotation SHOULD only be considered valid when on descriptors on index.json.

The second named reference (v1.0) points to a manifest that is specific to the linux/ppc64le platform.

OCI Image Manifest Specification

There are three main goals of the Image Manifest Specification. The first goal is content-addressable images, by supporting an image model where the image's configuration can be hashed to generate a unique ID for the image and its components. The second goal is to allow multi-architecture images, through a "fat manifest" which references image manifests for platform-specific versions of an image. In OCI, this is codified in an image index. The third goal is to be translatable to the OCI Runtime Specification.

This section defines the application/vnd.oci.image.manifest.v1+json media type. For the media type(s) that this is compatible with see the matrix.

Image Manifest

Unlike the image index, which contains information about a set of images that can span a variety of architectures and operating systems, an image manifest provides a configuration and set of layers for a single container image for a specific architecture and operating system.

Image Manifest Property Descriptions

Example Image Manifest

Example showing an image manifest:

{
  "schemaVersion": 2,
  "mediaType": "application/vnd.oci.image.manifest.v1+json",
  "config": {
    "mediaType": "application/vnd.oci.image.config.v1+json",
    "digest": "sha256:b5b2b2c507a0944348e0303114d8d93aaaa081732b86451d9bce1f432a537bc7",
    "size": 7023
  },
  "layers": [
    {
      "mediaType": "application/vnd.oci.image.layer.v1.tar+gzip",
      "digest": "sha256:9834876dcfb05cb167a5c24953eba58c4ac89b1adf57f28f2f9d09af107ee8f0",
      "size": 32654
    },
    {
      "mediaType": "application/vnd.oci.image.layer.v1.tar+gzip",
      "digest": "sha256:3c3a4604a545cdc127456d94e421cd355bca5b528f4a9c1905b15da2eb4a4c6b",
      "size": 16724
    },
    {
      "mediaType": "application/vnd.oci.image.layer.v1.tar+gzip",
      "digest": "sha256:ec4b8955958665577945c89419d1af06b5f7636b4ac3da7f12184802ad867736",
      "size": 73109
    }
  ],
  "subject": {
    "mediaType": "application/vnd.oci.image.manifest.v1+json",
    "digest": "sha256:5b0bcabd1ed22e9fb1310cf6c2dec7cdef19f0ad69efa1f392e94a4333501270",
    "size": 7682
  },
  "annotations": {
    "com.example.key1": "value1",
    "com.example.key2": "value2"
  }
}

Guidance for an Empty Descriptor

Implementers note: The following is considered GUIDANCE for portability.

Parts of the spec necessitate including a descriptor to a blob where some implementations of artifacts do not have associated content. While an empty blob (size of 0) may be preferable, practice has shown that not to be ubiquitously supported. The media type application/vnd.oci.empty.v1+json (MediaTypeEmptyJSON) has been specified for a descriptor that has no content for the implementation. The blob payload is the most minimal content that is still a valid JSON object: {} (size of 2). The blob digest of {} is sha256:44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a. The data field is optional, and if included is the base64 encoding of {}: e30=.

The resulting descriptor shown here is also defined in reference code as DescriptorEmptyJSON:

{
  "mediaType": "application/vnd.oci.empty.v1+json",
  "digest": "sha256:44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a",
  "size": 2,
  "data": "e30="
}

Guidelines for Artifact Usage

Content other than OCI container images MAY be packaged using the image manifest. When this is done, the config.mediaType value MUST be set to a value specific to the artifact type or the empty value. If the config.mediaType is set to the empty value, the artifactType MUST be defined. If the artifact does not need layers, a single layer SHOULD be included with a non-zero size. The suggested content for an unused layers array is the empty descriptor.

The design of the artifact depends on what content is being packaged with the artifact. The decision tree below and the associated examples MAY be used to design new artifacts:

  1. Does the artifact consist of at least one file or blob? If yes, continue to 2. If no, specify the artifactType, and set the config and a single layers element to the empty descriptor value. Here is an example of this with annotations included:

    {
  "schemaVersion": 2,
  "mediaType": "application/vnd.oci.image.manifest.v1+json",
  "artifactType": "application/vnd.example+type",
  "config": {
    "mediaType": "application/vnd.oci.empty.v1+json",
    "digest": "sha256:44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a",
    "size": 2
  },
  "layers": [
    {
      "mediaType": "application/vnd.oci.empty.v1+json",
      "digest": "sha256:44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a",
      "size": 2
    }
  ],
  "annotations": {
    "oci.opencontainers.image.created": "2023-01-02T03:04:05Z",
    "com.example.data": "payload"
  }
}

  2. Does the artifact have additional JSON formatted metadata as configuration? If yes, continue to 3. If no, specify the artifactType, include the artifact in the layers, and set config to the empty descriptor value. Here is an example of this with a single layer:

    {
  "schemaVersion": 2,
  "mediaType": "application/vnd.oci.image.manifest.v1+json",
  "artifactType": "application/vnd.example+type",
  "config": {
    "mediaType": "application/vnd.oci.empty.v1+json",
    "digest": "sha256:44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a",
    "size": 2
  },
  "layers": [
    {
      "mediaType": "application/vnd.example+type",
      "digest": "sha256:e258d248fda94c63753607f7c4494ee0fcbe92f1a76bfdac795c9d84101eb317",
      "size": 1234
    }
  ]
}

  3. For artifacts with a config blob, specify the artifactType to a common value for your artifact tooling, specify the config with the metadata for this artifact, and include the artifact in the layers. Here is an example of this:

    {
  "schemaVersion": 2,
  "mediaType": "application/vnd.oci.image.manifest.v1+json",
  "artifactType": "application/vnd.example+type",
  "config": {
    "mediaType": "application/vnd.example.config.v1+json",
    "digest": "sha256:5891b5b522d5df086d0ff0b110fbd9d21bb4fc7163af34d08286a2e846f6be03",
    "size": 123
  },
  "layers": [
    {
      "mediaType": "application/vnd.example.data.v1.tar+gzip",
      "digest": "sha256:e258d248fda94c63753607f7c4494ee0fcbe92f1a76bfdac795c9d84101eb317",
      "size": 1234
    }
  ]
}

Implementers note: artifacts have historically been created without an artifactType field, and tooling to work with artifacts should fallback to the config.mediaType value.

OCI Image Index Specification

The image index is a higher-level manifest which points to specific image manifests, ideal for one or more platforms. While the use of an image index is OPTIONAL for image providers, image consumers SHOULD be prepared to process them.

This section defines the application/vnd.oci.image.index.v1+json media type.

For the media type(s) that this document is compatible with, see the matrix.

Image Index Property Descriptions

Platform Variants

When the variant of the CPU is not listed in the table, values are implementation-defined and SHOULD be submitted to this specification for standardization.

ISA/ABI architecture variant
ARM 32-bit, v6 arm v6
ARM 32-bit, v7 arm v7
ARM 32-bit, v8 arm v8
ARM 64-bit, v8 arm64 v8

Example Image Index

Example showing a simple image index pointing to image manifests for two platforms:

{
  "schemaVersion": 2,
  "mediaType": "application/vnd.oci.image.index.v1+json",
  "manifests": [
    {
      "mediaType": "application/vnd.oci.image.manifest.v1+json",
      "size": 7143,
      "digest": "sha256:e692418e4cbaf90ca69d05a66403747baa33ee08806650b51fab815ad7fc331f",
      "platform": {
        "architecture": "ppc64le",
        "os": "linux"
      }
    },
    {
      "mediaType": "application/vnd.oci.image.manifest.v1+json",
      "size": 7682,
      "digest": "sha256:5b0bcabd1ed22e9fb1310cf6c2dec7cdef19f0ad69efa1f392e94a4333501270",
      "platform": {
        "architecture": "amd64",
        "os": "linux"
      }
    }
  ],
  "annotations": {
    "com.example.key1": "value1",
    "com.example.key2": "value2"
  }
}

Example Image Index with multiple media types

Example showing an image index pointing to manifests with multiple media types:

{
  "schemaVersion": 2,
  "mediaType": "application/vnd.oci.image.index.v1+json",
  "manifests": [
    {
      "mediaType": "application/vnd.oci.image.manifest.v1+json",
      "size": 7143,
      "digest": "sha256:e692418e4cbaf90ca69d05a66403747baa33ee08806650b51fab815ad7fc331f",
      "platform": {
        "architecture": "ppc64le",
        "os": "linux"
      }
    },
    {
      "mediaType": "application/vnd.oci.image.index.v1+json",
      "size": 7682,
      "digest": "sha256:601570aaff1b68a61eb9c85b8beca1644e698003e0cdb5bce960f193d265a8b7"
    }
  ],
  "annotations": {
    "com.example.key1": "value1",
    "com.example.key2": "value2"
  }
}

Image Layer Filesystem Changeset

This document describes how to serialize a filesystem and filesystem changes like removed files into a blob called a layer. One or more layers are applied on top of each other to create a complete filesystem. This document will use a concrete example to illustrate how to create and consume these filesystem layers.

This section defines the application/vnd.oci.image.layer.v1.tar, application/vnd.oci.image.layer.v1.tar+gzip, application/vnd.oci.image.layer.v1.tar+zstd, application/vnd.oci.image.layer.nondistributable.v1.tar, application/vnd.oci.image.layer.nondistributable.v1.tar+gzip, and application/vnd.oci.image.layer.nondistributable.v1.tar+zstd media types.

+gzip Media Types

+zstd Media Types

Distributable Format

Change Types

Types of changes that can occur in a changeset are:

Additions and Modifications are represented the same in the changeset tar archive.

Removals are represented using "whiteout" file entries (See Representing Changes).

File Types

Throughout this document section, the use of word "files" or "entries" includes the following, where supported:

File Attributes

Where supported, MUST include file attributes for Additions and Modifications include:

Sparse files SHOULD NOT be used because they lack consistent support across tar implementations.

SET LinkMap to map[< Major:Minor String >]map[< inode integer >]< path string >
SET LinkNames to map[< src path string >]< dest path string >
FOR each path in root path
  IF path type is directory
    CONTINUE
  ENDIF
  SET filestat to stat(path)
  IF filestat num of links == 1
    CONTINUE
  ENDIF
  IF LinkMap[filestat device][filestat inode] is not empty
    SET LinkNames[path] to LinkMap[filestat device][filestat inode]
  ELSE
    SET LinkMap[filestat device][filestat inode] to path
  ENDIF
END FOR

With this approach, the link map and links names of a directory could be compared against that of another directory to derive additions and changes to hardlinks.

Platform-specific attributes

Implementations on Windows MUST support these additional attributes, encoded in PAX vendor extensions as follows:

Creating

Initial Root Filesystem

The initial root filesystem is the base or parent layer.

For this example, an image root filesystem has an initial state as an empty directory. The name of the directory is not relevant to the layer itself, only for the purpose of producing comparisons.

Here is an initial empty directory structure for a changeset, with a unique directory name rootfs-c9d-v1.

rootfs-c9d-v1/

Populate Initial Filesystem

Files and directories are then created:

rootfs-c9d-v1/
  etc/
    my-app-config
  bin/
    my-app-binary
    my-app-tools

The rootfs-c9d-v1 directory is then created as a plain tar archive with relative path to rootfs-c9d-v1. Entries for the following files:

./
./etc/
./etc/my-app-config
./bin/
./bin/my-app-binary
./bin/my-app-tools

Populate a Comparison Filesystem

Create a new directory and initialize it with a copy or snapshot of the prior root filesystem. Example commands that can preserve file attributes to make this copy are:

Any changes to the snapshot MUST NOT change or affect the directory it was copied from.

For example rootfs-c9d-v1.s1 is an identical snapshot of rootfs-c9d-v1. In this way rootfs-c9d-v1.s1 is prepared for updates and alterations.

Implementor's Note: a copy-on-write or union filesystem can efficiently make directory snapshots

Initial layout of the snapshot:

rootfs-c9d-v1.s1/
  etc/
    my-app-config
  bin/
    my-app-binary
    my-app-tools

See Change Types for more details on changes.

For example, add a directory at /etc/my-app.d containing a default config file, removing the existing config file. Also a change (in attribute or file content) to ./bin/my-app-tools binary to handle the config layout change.

Following these changes, the representation of the rootfs-c9d-v1.s1 directory:

rootfs-c9d-v1.s1/
  etc/
    my-app.d/
      default.cfg
  bin/
    my-app-binary
    my-app-tools

Determining Changes

When two directories are compared, the relative root is the top-level directory. The directories are compared, looking for files that have been added, modified, or removed.

For this example, rootfs-c9d-v1/ and rootfs-c9d-v1.s1/ are recursively compared, each as relative root path.

The following changeset is found:

Added:      /etc/my-app.d/
Added:      /etc/my-app.d/default.cfg
Modified:   /bin/my-app-tools
Deleted:    /etc/my-app-config

This reflects the removal of /etc/my-app-config and creation of a file and directory at /etc/my-app.d/default.cfg. /bin/my-app-tools has also been replaced with an updated version.

Representing Changes

A tar archive is then created which contains only this changeset:

The resulting tar archive for rootfs-c9d-v1.s1 has the following entries:

./etc/my-app.d/
./etc/my-app.d/default.cfg
./bin/my-app-tools
./etc/.wh.my-app-config

To signify that the resource ./etc/my-app-config MUST be removed when the changeset is applied, the basename of the entry is prefixed with .wh..

Applying Changesets

Changeset over existing files

This section specifies applying an entry from a layer changeset if the target path already exists.

If the entry and the existing path are both directories, then the existing path's attributes MUST be replaced by those of the entry in the changeset. In all other cases, the implementation MUST do the semantic equivalent of the following:

Whiteouts

The following is a base layer with several resources:

a/
a/b/
a/b/c/
a/b/c/bar

When the next layer is created, the original a/b directory is deleted and recreated with a/b/c/foo:

a/
a/.wh..wh..opq
a/b/
a/b/c/
a/b/c/foo

When processing the second layer, a/.wh..wh..opq is applied first, before creating the new version of a/b, regardless of the ordering in which the whiteout file was encountered. For example, the following layer is equivalent to the layer above:

a/
a/b/
a/b/c/
a/b/c/foo
a/.wh..wh..opq

Implementations SHOULD generate layers such that the whiteout files appear before sibling directory entries.

Opaque Whiteout

Let's take the following base layer as an example:

etc/
  my-app-config
bin/
  my-app-binary
  my-app-tools
  tools/
    my-app-tool-one

If all children of bin/ are removed, the next layer would have the following:

bin/
  .wh..wh..opq

This is called opaque whiteout format. An opaque whiteout file hides all children of the bin/ including sub-directories and all descendants. Using explicit whiteout files, this would be equivalent to the following:

bin/
  .wh.my-app-binary
  .wh.my-app-tools
  .wh.tools

In this case, a unique whiteout file is generated for each entry. If there were more children of bin/ in the base layer, there would be an entry for each. Note that this opaque file will apply to all children, including sub-directories, other resources and all descendants.

Implementations SHOULD generate layers using explicit whiteout files, but MUST accept both.

Any given image is likely to be composed of several of these Image Filesystem Changeset tar archives.

Non-Distributable Layers

NOTE: Non-distributable layers are deprecated, and not recommended for future use. Implementations SHOULD NOT produce new non-distributable layers.

Due to legal requirements, certain layers may not be regularly distributable. Such "non-distributable" layers are typically downloaded directly from a distributor but never uploaded.

Non-distributable layers SHOULD be tagged with an alternative mediatype of application/vnd.oci.image.layer.nondistributable.v1.tar. Implementations SHOULD NOT upload layers tagged with this media type; however, such a media type SHOULD NOT affect whether an implementation downloads the layer.

Descriptors referencing non-distributable layers MAY include urls for downloading these layers directly; however, the presence of the urls field SHOULD NOT be used to determine whether or not a layer is non-distributable.

OCI Image Configuration

An OCI Image is an ordered collection of root filesystem changes and the corresponding execution parameters for use within a container runtime. This specification outlines the JSON format describing images for use with a container runtime and execution tool and its relationship to filesystem changesets, described in Layers.

This section defines the application/vnd.oci.image.config.v1+json media type.

Terminology

This specification uses the following terms:

Layer

Image JSON

Layer DiffID

A layer DiffID is the digest over the layer's uncompressed tar archive and serialized in the descriptor digest format, e.g., sha256:a9561eb1b190625c9adb5a9513e72c4dedafc1cb2d4c5236c9a6957ec7dfd5a9. Layers SHOULD be packed and unpacked reproducibly to avoid changing the layer DiffID, for example by using tar-split to save the tar headers.

NOTE: Do not confuse DiffIDs with layer digests, often referenced in the manifest, which are digests over compressed or uncompressed content.

Layer ChainID

For convenience, it is sometimes useful to refer to a stack of layers with a single identifier. While a layer's DiffID identifies a single changeset, the ChainID identifies the subsequent application of those changesets. This ensures that we have handles referring to both the layer itself, as well as the result of the application of a series of changesets. Use in combination with rootfs.diff_ids while applying layers to a root filesystem to uniquely and safely identify the result.

Definition

The ChainID of an applied set of layers is defined with the following recursion:

ChainID(L₀) =  DiffID(L₀)
ChainID(L₀|...|Lₙ₋₁|Lₙ) = Digest(ChainID(L₀|...|Lₙ₋₁) + " " + DiffID(Lₙ))

For this, we define the binary | operation to be the result of applying the right operand to the left operand. For example, given base layer A and a changeset B, we refer to the result of applying B to A as A|B.

Above, we define the ChainID for a single layer (L₀) as equivalent to the DiffID for that layer. Otherwise, the ChainID for a set of applied layers (L₀|...|Lₙ₋₁|Lₙ) is defined as the recursion Digest(ChainID(L₀|...|Lₙ₋₁) + " " + DiffID(Lₙ)).

Explanation

Let's say we have layers A, B, C, ordered from bottom to top, where A is the base and C is the top. Defining | as a binary application operator, the root filesystem may be A|B|C. While it is implied that C is only useful when applied to A|B, the identifier C is insufficient to identify this result, as we'd have the equality C = A|B|C, which isn't true.

The main issue is when we have two definitions of C, C = C and C = A|B|C. If this is true (with some handwaving), C = x|C where x = any application. This means that if an attacker can define x, relying on C provides no guarantee that the layers were applied in any order.

The ChainID addresses this problem by being defined as a compound hash. We differentiate the changeset C, from the order-dependent application A|B|C by saying that the resulting rootfs is identified by ChainID(A|B|C), which can be calculated by ImageConfig.rootfs.

Let's expand the definition of ChainID(A|B|C) to explore its internal structure:

ChainID(A) = DiffID(A)
ChainID(A|B) = Digest(ChainID(A) + " " + DiffID(B))
ChainID(A|B|C) = Digest(ChainID(A|B) + " " + DiffID(C))

We can replace each definition and reduce to a single equality:

ChainID(A|B|C) = Digest(Digest(DiffID(A) + " " + DiffID(B)) + " " + DiffID(C))

Hopefully, the above is illustrative of the actual contents of the ChainID. Most importantly, we can easily see that ChainID(C) != ChainID(A|B|C), otherwise, ChainID(C) = DiffID(C), which is the base case, could not be true.

ImageID

Each image's ID is given by the SHA256 hash of its configuration JSON. It is represented as a hexadecimal encoding of 256 bits, e.g., sha256:a9561eb1b190625c9adb5a9513e72c4dedafc1cb2d4c5236c9a6957ec7dfd5a9. Since the configuration JSON that gets hashed references hashes of each layer in the image, this formulation of the ImageID makes images content-addressable.

Properties

Note: Any OPTIONAL field MAY also be set to null, which is equivalent to being absent.

Any extra fields in the Image JSON struct are considered implementation specific and MUST NOT generate an error by any implementations which are unable to interpret them.

Whitespace is OPTIONAL and implementations MAY have compact JSON with no whitespace.

Example

Here is an example image configuration JSON document:

{
    "created": "2015-10-31T22:22:56.015925234Z",
    "author": "Alyssa P. Hacker <alyspdev@example.com>",
    "architecture": "amd64",
    "os": "linux",
    "config": {
        "User": "alice",
        "ExposedPorts": {
            "8080/tcp": {}
        },
        "Env": [
            "PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin",
            "FOO=oci_is_a",
            "BAR=well_written_spec"
        ],
        "Entrypoint": [
            "/bin/my-app-binary"
        ],
        "Cmd": [
            "--foreground",
            "--config",
            "/etc/my-app.d/default.cfg"
        ],
        "Volumes": {
            "/var/job-result-data": {},
            "/var/log/my-app-logs": {}
        },
        "WorkingDir": "/home/alice",
        "Labels": {
            "com.example.project.git.url": "https://example.com/project.git",
            "com.example.project.git.commit": "45a939b2999782a3f005621a8d0f29aa387e1d6b"
        }
    },
    "rootfs": {
      "diff_ids": [
        "sha256:c6f988f4874bb0add23a778f753c65efe992244e148a1d2ec2a8b664fb66bbd1",
        "sha256:5f70bf18a086007016e948b04aed3b82103a36bea41755b6cddfaf10ace3c6ef"
      ],
      "type": "layers"
    },
    "history": [
      {
        "created": "2015-10-31T22:22:54.690851953Z",
        "created_by": "/bin/sh -c #(nop) ADD file:a3bc1e842b69636f9df5256c49c5374fb4eef1e281fe3f282c65fb853ee171c5 in /"
      },
      {
        "created": "2015-10-31T22:22:55.613815829Z",
        "created_by": "/bin/sh -c #(nop) CMD [\"sh\"]",
        "empty_layer": true
      },
      {
        "created": "2015-10-31T22:22:56.329850019Z",
        "created_by": "/bin/sh -c apk add curl"
      }
    ]
}

Annotations

Several components of the specification, like Image Manifests and Descriptors, feature an optional annotations property, whose format is common and defined in this section.

This property contains arbitrary metadata.

Rules

Pre-Defined Annotation Keys

This specification defines the following annotation keys, intended for but not limited to image index, image manifest, and descriptor authors.

Back-compatibility with Label Schema

Label Schema defined a number of conventional labels for container images, and these are now superceded by annotations with keys starting org.opencontainers.image.

While users are encouraged to use the org.opencontainers.image keys, tools MAY choose to support compatible annotations using the org.label-schema prefix as follows.

org.opencontainers.image prefix org.label-schema prefix Compatibility notes
created build-date Compatible
url url Compatible
source vcs-url Compatible
version version Compatible
revision vcs-ref Compatible
vendor vendor Compatible
title name Compatible
description description Compatible
documentation usage Value is compatible if the documentation is located by a URL
authors No equivalent in Label Schema
licenses No equivalent in Label Schema
ref.name No equivalent in Label Schema
schema-version No equivalent in the OCI Image Spec
docker.*, rkt.* No equivalent in the OCI Image Spec

Conversion to OCI Runtime Configuration

When extracting an OCI Image into an OCI Runtime bundle, two orthogonal components of the extraction are relevant:

  1. Extraction of the root filesystem from the set of filesystem layers.
  2. Conversion of the image configuration blob to an OCI Runtime configuration blob.

This section defines how to convert an application/vnd.oci.image.config.v1+json blob to an OCI runtime configuration blob (the latter component of extraction). The former component of extraction is defined elsewhere and is orthogonal to configuration of a runtime bundle. The values of runtime configuration properties not specified by this document are implementation-defined.

A converter MUST rely on the OCI image configuration to build the OCI runtime configuration as described by this document; this will create the "default generated runtime configuration".

The "default generated runtime configuration" MAY be overridden or combined with externally provided inputs from the caller. In addition, a converter MAY have its own implementation-defined defaults and extensions which MAY be combined with the "default generated runtime configuration". The restrictions in this document refer only to combining implementation-defined defaults with the "default generated runtime configuration". Externally provided inputs are considered to be a modification of the application/vnd.oci.image.config.v1+json used as a source, and such modifications have no restrictions.

For example, externally provided inputs MAY cause an environment variable to be added, removed or changed. However an implementation-defined default SHOULD NOT result in an environment variable being removed or changed.

Verbatim Fields

Certain image configuration fields have an identical counterpart in the runtime configuration. Some of these are purely annotation-based fields, and have been extracted into a separate subsection. A compliant configuration converter MUST extract the following fields verbatim to the corresponding field in the generated runtime configuration:

Image Field Runtime Field Notes
Config.WorkingDir process.cwd
Config.Env process.env 1
Config.Entrypoint process.args 2
Config.Cmd process.args 2
  1. The converter MAY add additional entries to process.env but it SHOULD NOT add entries that have variable names present in Config.Env.
  2. If both Config.Entrypoint and Config.Cmd are specified, the converter MUST append the value of Config.Cmd to the value of Config.Entrypoint and set process.args to that combined value.

Annotation Fields

These fields all affect the annotations of the runtime configuration, and are thus subject to precedence.

Image Field Runtime Field Notes
os annotations 1,2
architecture annotations 1,3
variant annotations 1,4
os.version annotations 1,5
os.features annotations 1,6
author annotations 1,7
created annotations 1,8
Config.Labels annotations
Config.StopSignal annotations 1,9
  1. If a user has explicitly specified this annotation with Config.Labels, then the value specified in this field takes lower precedence and the converter MUST instead use the value from Config.Labels.
  2. The value of this field MUST be set as the value of org.opencontainers.image.os in annotations.
  3. The value of this field MUST be set as the value of org.opencontainers.image.architecture in annotations.
  4. The value of this field MUST be set as the value of org.opencontainers.image.variant in annotations.
  5. The value of this field MUST be set as the value of org.opencontainers.image.os.version in annotations.
  6. The value of this field MUST be set as the value of org.opencontainers.image.os.features in annotations.
  7. The value of this field MUST be set as the value of org.opencontainers.image.author in annotations.
  8. The value of this field MUST be set as the value of org.opencontainers.image.created in annotations.
  9. The value of this field MUST be set as the value of org.opencontainers.image.stopSignal in annotations.

Parsed Fields

Certain image configuration fields have a counterpart that must first be translated. A compliant configuration converter SHOULD parse all of these fields and set the corresponding fields in the generated runtime configuration:

Image Field Runtime Field
Config.User process.user.*

The method of parsing the above image fields are described in the following sections.

Config.User

If the values of user or group in Config.User are numeric (uid or gid) then the values MUST be copied verbatim to process.user.uid and process.user.gid respectively. If the values of user or group in Config.User are not numeric (user or group) then a converter SHOULD resolve the user information using a method appropriate for the container's context. For Unix-like systems, this MAY involve resolution through NSS or parsing /etc/passwd from the extracted container's root filesystem to determine the values of process.user.uid and process.user.gid.

In addition, a converter SHOULD set the value of process.user.additionalGids to a value corresponding to the user in the container's context described by Config.User. For Unix-like systems, this MAY involve resolution through NSS or parsing /etc/group and determining the group memberships of the user specified in process.user.uid. The converter SHOULD NOT modify process.user.additionalGids if the value of user in Config.User is numeric or if Config.User specifies a group.

If Config.User is not defined, the converted process.user value is implementation-defined. If Config.User does not correspond to a user in the container's context, the converter MUST return an error.

Optional Fields

Certain image configuration fields are not applicable to all conversion use cases, and thus are optional for configuration converters to implement. A compliant configuration converter SHOULD provide a way for users to extract these fields into the generated runtime configuration:

Image Field Runtime Field Notes
Config.ExposedPorts annotations 1
Config.Volumes mounts 2
  1. The runtime configuration does not have a corresponding field for this image field. However, converters SHOULD set the org.opencontainers.image.exposedPorts annotation.
  2. Implementations SHOULD provide mounts for these locations such that application data is not written to the container's root filesystem. If a converter implements conversion for this field using mountpoints, it SHOULD set the destination of the mountpoint to the value specified in Config.Volumes. An implementation MAY seed the contents of the mount with data in the image at the same location. If a new image is created from a container based on the image described by this configuration, data in these paths SHOULD NOT be included in the new image. The other mounts fields are platform and context dependent, and thus are implementation-defined. Note that the implementation of Config.Volumes need not use mountpoints, as it is effectively a mask of the filesystem.

Config.ExposedPorts

The OCI runtime configuration does not provide a way of expressing the concept of "container exposed ports". However, converters SHOULD set the org.opencontainers.image.exposedPorts annotation, unless doing so will cause a conflict.

org.opencontainers.image.exposedPorts is the list of values that correspond to the keys defined for Config.ExposedPorts (string, comma-separated values).

Annotations

There are three ways of annotating an OCI image in this specification:

  1. Config.Labels in the configuration of the image.
  2. annotations in the manifest of the image.
  3. annotations in the image index of the image.

In addition, there are also implicit annotations that are defined by this section which are determined from the values of the image configuration. A converter SHOULD NOT attempt to extract annotations from manifests or image indices. If there is a conflict (same key but different value) between an implicit annotation (or annotation in manifests or image indices) and an explicitly specified annotation in Config.Labels, the value specified in Config.Labels MUST take precedence.

A converter MAY add annotations which have keys not specified in the image. A converter MUST NOT modify the values of annotations specified in the image.

Considerations

Extensibility

Implementations storing or copying content MUST NOT modify or alter the content in a way that would change the digest of the content. Examples of these implementations include:

Implementations processing content SHOULD NOT generate an error if they encounter an unknown property in a known media type. Examples of these implementations include:

Canonicalization

JSON

JSON content SHOULD be serialized as canonical JSON. Of the OCI Image Format Specification media types, all the types ending in +json contain JSON content. Implementations:

EBNF

For field formats described in this specification, we use a limited subset of Extended Backus-Naur Form, similar to that used by the XML specification. Grammars present in the OCI specification are regular and can be converted to a single regular expressions. However, regular expressions are avoided to limit ambiguity between regular expression syntax. By defining a subset of EBNF used here, the possibility of variation, misunderstanding or ambiguities from linking to a larger specification can be avoided.

Grammars are made up of rules in the following form:

symbol ::= expression

We can say we have the production identified by symbol if the input is matched by the expression. Whitespace is completely ignored in rule definitions.

Expressions

The simplest expression is the literal, surrounded by quotes:

literal ::= "matchthis"

The above expression defines a symbol, "literal", that matches the exact input of "matchthis". Character classes are delineated by brackets ([]), describing either a set, range or multiple range of characters:

set := [abc]
range := [A-Z]

The above symbol "set" would match one character of either "a", "b" or "c". The symbol "range" would match any character, "A" to "Z", inclusive. Currently, only matching for 7-bit ascii literals and character classes is defined, as that is all that is required by this specification. Multiple character ranges and explicit characters can be specified in a single character classes, as follows:

multipleranges := [a-zA-Z=-]

The above matches the characters in the range A to Z, a to z and the individual characters - and =.

Expressions can be made up of one or more expressions, such that one must be followed by the other. This is known as an implicit concatenation operator. For example, to satisfy the following rule, both A and B must be matched to satisfy the rule:

symbol ::= A B

Each expression must be matched once and only once, A followed by B. To support the description of repetition and optional match criteria, the postfix operators * and + are defined. * indicates that the preceding expression can be matched zero or more times. + indicates that the preceding expression must be matched one or more times. These appear in the following form:

zeroormore ::= expression*
oneormore ::= expression+

Parentheses are used to group expressions into a larger expression:

group ::= (A B)

Like simpler expressions above, operators can be applied to groups, as well. To allow for alternates, we also define the infix operator |.

oneof ::= A | B

The above indicates that the expression should match one of the expressions, A or B.

Precedence

The operator precedence is in the following order:

The precedence can be better described using grouping to show equivalents. Concatenation has higher precedence than alternates, such A B | C D is equivalent to (A B) | (C D). Unary operators have higher precedence than alternates and concatenation, such that A+ | B+ is equivalent to (A+) | (B+).

Examples

The following combines the previous definitions to match a simple, relative path name, describing the individual components:

path      ::= component ("/" component)*
component ::= [a-z]+

The production "component" is one or more lowercase letters. A "path" is then at least one component, possibly followed by zero or more slash-component pairs. The above can be converted into the following regular expression:

[a-z]+(?:/[a-z]+)*

OCI Image Implementations

Projects or Companies currently adopting the OCI Image Specification

(to add your project please open a pull-request)