Network Working GroupE. Allman
Request for Comments: 4871Sendmail, Inc.
Obsoletes: 4870J. Callas
Category: Standards TrackPGP Corporation
M. Delany
M. Libbey
Yahoo! Inc
J. Fenton
M. Thomas
Cisco Systems, Inc.
May 2007

DomainKeys Identified Mail (DKIM) Signatures

Status of this Memo

This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the “Internet Official Protocol Standards” (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.

Copyright Notice

Copyright © The IETF Trust (2007). All Rights Reserved.

Abstract

DomainKeys Identified Mail (DKIM) defines a domain-level authentication framework for email using public-key cryptography and key server technology to permit verification of the source and contents of messages by either Mail Transfer Agents (MTAs) or Mail User Agents (MUAs). The ultimate goal of this framework is to permit a signing domain to assert responsibility for a message, thus protecting message signer identity and the integrity of the messages they convey while retaining the functionality of Internet email as it is known today. Protection of email identity may assist in the global control of "spam" and "phishing".

Table of Contents

1. Introduction

DomainKeys Identified Mail (DKIM) defines a mechanism by which email messages can be cryptographically signed, permitting a signing domain to claim responsibility for the introduction of a message into the mail stream. Message recipients can verify the signature by querying the signer's domain directly to retrieve the appropriate public key, and thereby confirm that the message was attested to by a party in possession of the private key for the signing domain.

The approach taken by DKIM differs from previous approaches to message signing (e.g., Secure/Multipurpose Internet Mail Extensions (S/MIME) [RFC1847], OpenPGP [RFC2440]) in that:

DKIM:

1.1 Signing Identity

DKIM separates the question of the identity of the signer of the message from the purported author of the message. In particular, a signature includes the identity of the signer. Verifiers can use the signing information to decide how they want to process the message. The signing identity is included as part of the signature header field.

INFORMATIVE RATIONALE: The signing identity specified by a DKIM signature is not required to match an address in any particular header field because of the broad methods of interpretation by recipient mail systems, including MUAs.

1.2 Scalability

DKIM is designed to support the extreme scalability requirements that characterize the email identification problem. There are currently over 70 million domains and a much larger number of individual addresses. DKIM seeks to preserve the positive aspects of the current email infrastructure, such as the ability for anyone to communicate with anyone else without introduction.

1.3 Simple Key Management

DKIM differs from traditional hierarchical public-key systems in that no Certificate Authority infrastructure is required; the verifier requests the public key from a repository in the domain of the claimed signer directly rather than from a third party.

The DNS is proposed as the initial mechanism for the public keys. Thus, DKIM currently depends on DNS administration and the security of the DNS system. DKIM is designed to be extensible to other key fetching services as they become available.

2. Terminology and Definitions

This section defines terms used in the rest of the document. Syntax descriptions use the form described in Augmented BNF for Syntax Specifications [RFC4234].

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].

2.1 Signers

Elements in the mail system that sign messages on behalf of a domain are referred to as signers. These may be MUAs (Mail User Agents), MSAs (Mail Submission Agents), MTAs (Mail Transfer Agents), or other agents such as mailing list exploders. In general, any signer will be involved in the injection of a message into the message system in some way. The key issue is that a message must be signed before it leaves the administrative domain of the signer.

2.2 Verifiers

Elements in the mail system that verify signatures are referred to as verifiers. These may be MTAs, Mail Delivery Agents (MDAs), or MUAs. In most cases it is expected that verifiers will be close to an end user (reader) of the message or some consuming agent such as a mailing list exploder.

2.3 Whitespace

There are three forms of whitespace:

The formal ABNF for these are (WSP and LWSP are given for information only):

    
    WSP =   SP / HTAB
    LWSP =  *(WSP / CRLF WSP)
    FWS =   [*WSP CRLF] 1*WSP

The definition of FWS is identical to that in [RFC2822] except for the exclusion of obs-FWS.

2.4 Common ABNF Tokens

The following ABNF tokens are used elsewhere in this document:

    
    hyphenated-word =  ALPHA [ *(ALPHA / DIGIT / "-") (ALPHA / DIGIT) ]
    base64string =     1*(ALPHA / DIGIT / "+" / "/" / [FWS])
                       [ "=" [FWS] [ "=" [FWS] ] ]

2.5 Imported ABNF Tokens

The following tokens are imported from other RFCs as noted. Those RFCs should be considered definitive.

The following tokens are imported from [RFC2821]:

The following tokens are imported from [RFC2822]:

The following tokens are imported from [RFC2045]:

INFORMATIVE NOTE: Be aware that the ABNF in RFC 2045 does not obey the rules of RFC 4234 and must be interpreted accordingly, particularly as regards case folding.

Other tokens not defined herein are imported from [RFC4234]. These are intuitive primitives such as SP, HTAB, WSP, ALPHA, DIGIT, CRLF, etc.

2.6 DKIM-Quoted-Printable

The DKIM-Quoted-Printable encoding syntax resembles that described in Quoted-Printable [RFC2045], Section 6.7: any character MAY be encoded as an "=" followed by two hexadecimal digits from the alphabet "0123456789ABCDEF" (no lowercase characters permitted) representing the hexadecimal-encoded integer value of that character. All control characters (those with values < %x20), 8-bit characters (values > %x7F), and the characters DEL (%x7F), SPACE (%x20), and semicolon (";", %x3B) MUST be encoded. Note that all whitespace, including SPACE, CR, and LF characters, MUST be encoded. After encoding, FWS MAY be added at arbitrary locations in order to avoid excessively long lines; such whitespace is NOT part of the value, and MUST be removed before decoding.

ABNF:

    
    dkim-quoted-printable =
                       *(FWS / hex-octet / dkim-safe-char)
                  ; hex-octet is from RFC 2045
    dkim-safe-char =   %x21-3A / %x3C / %x3E-7E
                  ; '!' - ':', '<', '>' - '~'
                  ; Characters not listed as "mail-safe" in
                  ; RFC 2049 are also not recommended.
INFORMATIVE NOTE: DKIM-Quoted-Printable differs from Quoted-Printable as defined in RFC 2045 in several important ways:
  1. Whitespace in the input text, including CR and LF, must be encoded. RFC 2045 does not require such encoding, and does not permit encoding of CR or LF characters that are part of a CRLF line break.
  2. Whitespace in the encoded text is ignored. This is to allow tags encoded using DKIM-Quoted-Printable to be wrapped as needed. In particular, RFC 2045 requires that line breaks in the input be represented as physical line breaks; that is not the case here.
  3. The "soft line break" syntax ("=" as the last non-whitespace character on the line) does not apply.
  4. DKIM-Quoted-Printable does not require that encoded lines be no more than 76 characters long (although there may be other requirements depending on the context in which the encoded text is being used).

3. Protocol Elements

Protocol Elements are conceptual parts of the protocol that are not specific to either signers or verifiers. The protocol descriptions for signers and verifiers are described in later sections (Signer Actions (Section 5) and Verifier Actions (Section 6)). NOTE: This section must be read in the context of those sections.

3.1 Selectors

To support multiple concurrent public keys per signing domain, the key namespace is subdivided using "selectors". For example, selectors might indicate the names of office locations (e.g., "sanfrancisco", "coolumbeach", and "reykjavik"), the signing date (e.g., "january2005", "february2005", etc.), or even the individual user.

Selectors are needed to support some important use cases. For example:

Periods are allowed in selectors and are component separators. When keys are retrieved from the DNS, periods in selectors define DNS label boundaries in a manner similar to the conventional use in domain names. Selector components might be used to combine dates with locations, for example, "march2005.reykjavik". In a DNS implementation, this can be used to allow delegation of a portion of the selector namespace.

ABNF:

     
   selector =   sub-domain *( "." sub-domain )

The number of public keys and corresponding selectors for each domain is determined by the domain owner. Many domain owners will be satisfied with just one selector, whereas administratively distributed organizations may choose to manage disparate selectors and key pairs in different regions or on different email servers.

Beyond administrative convenience, selectors make it possible to seamlessly replace public keys on a routine basis. If a domain wishes to change from using a public key associated with selector "january2005" to a public key associated with selector "february2005", it merely makes sure that both public keys are advertised in the public-key repository concurrently for the transition period during which email may be in transit prior to verification. At the start of the transition period, the outbound email servers are configured to sign with the "february2005" private key. At the end of the transition period, the "january2005" public key is removed from the public-key repository.

INFORMATIVE NOTE: A key may also be revoked as described below. The distinction between revoking and removing a key selector record is subtle. When phasing out keys as described above, a signing domain would probably simply remove the key record after the transition period. However, a signing domain could elect to revoke the key (but maintain the key record) for a further period. There is no defined semantic difference between a revoked key and a removed key.

While some domains may wish to make selector values well known, others will want to take care not to allocate selector names in a way that allows harvesting of data by outside parties. For example, if per-user keys are issued, the domain owner will need to make the decision as to whether to associate this selector directly with the user name, or make it some unassociated random value, such as a fingerprint of the public key.

INFORMATIVE OPERATIONS NOTE: Reusing a selector with a new key (for example, changing the key associated with a user's name) makes it impossible to tell the difference between a message that didn't verify because the key is no longer valid versus a message that is actually forged. For this reason, signers are ill-advised to reuse selectors for new keys. A better strategy is to assign new keys to new selectors.

3.2 Tag=Value Lists

DKIM uses a simple "tag=value" syntax in several contexts, including in messages and domain signature records.

Values are a series of strings containing either plain text, base64 text (as defined in [RFC2045], Section 6.8), qp-section (ibid, Section 6.7), or dkim-quoted-printable (as defined in Section 2.6). The name of the tag will determine the encoding of each value. Unencoded semicolon (";") characters MUST NOT occur in the tag value, since that separates tag-specs.

INFORMATIVE IMPLEMENTATION NOTE: Although the "plain text" defined below (as tag-value) only includes 7-bit characters, an implementation that wished to anticipate future standards would be advised not to preclude the use of UTF8-encoded text in tag=value lists.

Formally, the syntax rules are as follows:

     
     tag-list  =  tag-spec 0*( ";" tag-spec ) [ ";" ]
     tag-spec  =  [FWS] tag-name [FWS] "=" [FWS] tag-value [FWS]
     tag-name  =  ALPHA 0*ALNUMPUNC
     tag-value =  [ tval 0*( 1*(WSP / FWS) tval ) ]
                       ; WSP and FWS prohibited at beginning and end
     tval      =  1*VALCHAR
     VALCHAR   =  %x21-3A / %x3C-7E
                       ; EXCLAMATION to TILDE except SEMICOLON
     ALNUMPUNC =  ALPHA / DIGIT / "_"

Note that WSP is allowed anywhere around tags. In particular, any WSP after the "=" and any WSP before the terminating ";" is not part of the value; however, WSP inside the value is significant.

Tags MUST be interpreted in a case-sensitive manner. Values MUST be processed as case sensitive unless the specific tag description of semantics specifies case insensitivity.

Tags with duplicate names MUST NOT occur within a single tag-list; if a tag name does occur more than once, the entire tag-list is invalid.

Whitespace within a value MUST be retained unless explicitly excluded by the specific tag description.

Tag=value pairs that represent the default value MAY be included to aid legibility.

Unrecognized tags MUST be ignored.

Tags that have an empty value are not the same as omitted tags. An omitted tag is treated as having the default value; a tag with an empty value explicitly designates the empty string as the value. For example, "g=" does not mean "g=*", even though "g=*" is the default for that tag.

3.3 Signing and Verification Algorithms

DKIM supports multiple digital signature algorithms. Two algorithms are defined by this specification at this time: rsa-sha1 and rsa-sha256. The rsa-sha256 algorithm is the default if no algorithm is specified. Verifiers MUST implement both rsa-sha1 and rsa-sha256. Signers MUST implement and SHOULD sign using rsa-sha256.

INFORMATIVE NOTE: Although sha256 is strongly encouraged, some senders of low-security messages (such as routine newsletters) may prefer to use sha1 because of reduced CPU requirements to compute a sha1 hash. In general, sha256 should always be used whenever possible.

3.3.1 The rsa-sha1 Signing Algorithm

The rsa-sha1 Signing Algorithm computes a message hash as described in Section 3.7 below using SHA-1 [FIPS.180-2.2002] as the hash-alg. That hash is then signed by the signer using the RSA algorithm (defined in PKCS#1 version 1.5 [RFC3447]) as the crypt-alg and the signer's private key. The hash MUST NOT be truncated or converted into any form other than the native binary form before being signed. The signing algorithm SHOULD use a public exponent of 65537.

3.3.2 The rsa-sha256 Signing Algorithm

The rsa-sha256 Signing Algorithm computes a message hash as described in Section 3.7 below using SHA-256 [FIPS.180-2.2002] as the hash-alg. That hash is then signed by the signer using the RSA algorithm (defined in PKCS#1 version 1.5 [RFC3447]) as the crypt-alg and the signer's private key. The hash MUST NOT be truncated or converted into any form other than the native binary form before being signed.

3.3.3 Key Sizes

Selecting appropriate key sizes is a trade-off between cost, performance, and risk. Since short RSA keys more easily succumb to off-line attacks, signers MUST use RSA keys of at least 1024 bits for long-lived keys. Verifiers MUST be able to validate signatures with keys ranging from 512 bits to 2048 bits, and they MAY be able to validate signatures with larger keys. Verifier policies may use the length of the signing key as one metric for determining whether a signature is acceptable.

Factors that should influence the key size choice include the following:

See [RFC3766] for further discussion on selecting key sizes.

3.3.4 Other Algorithms

Other algorithms MAY be defined in the future. Verifiers MUST ignore any signatures using algorithms that they do not implement.

3.4 Canonicalization

Empirical evidence demonstrates that some mail servers and relay systems modify email in transit, potentially invalidating a signature. There are two competing perspectives on such modifications. For most signers, mild modification of email is immaterial to the authentication status of the email. For such signers, a canonicalization algorithm that survives modest in-transit modification is preferred.

Other signers demand that any modification of the email, however minor, result in a signature verification failure. These signers prefer a canonicalization algorithm that does not tolerate in-transit modification of the signed email.

Some signers may be willing to accept modifications to header fields that are within the bounds of email standards such as [RFC2822], but are unwilling to accept any modification to the body of messages.

To satisfy all requirements, two canonicalization algorithms are defined for each of the header and the body: a "simple" algorithm that tolerates almost no modification and a "relaxed" algorithm that tolerates common modifications such as whitespace replacement and header field line rewrapping. A signer MAY specify either algorithm for header or body when signing an email. If no canonicalization algorithm is specified by the signer, the "simple" algorithm defaults for both header and body. Verifiers MUST implement both canonicalization algorithms. Note that the header and body may use different canonicalization algorithms. Further canonicalization algorithms MAY be defined in the future; verifiers MUST ignore any signatures that use unrecognized canonicalization algorithms.

Canonicalization simply prepares the email for presentation to the signing or verification algorithm. It MUST NOT change the transmitted data in any way. Canonicalization of header fields and body are described below.

NOTE: This section assumes that the message is already in "network normal" format (text is ASCII encoded, lines are separated with CRLF characters, etc.). See also Section 5.3 for information about normalizing the message.

3.4.1 The "simple" Header Canonicalization Algorithm

The "simple" header canonicalization algorithm does not change header fields in any way. Header fields MUST be presented to the signing or verification algorithm exactly as they are in the message being signed or verified. In particular, header field names MUST NOT be case folded and whitespace MUST NOT be changed.

3.4.2 The "relaxed" Header Canonicalization Algorithm

The "relaxed" header canonicalization algorithm MUST apply the following steps in order:

3.4.3 The "simple" Body Canonicalization Algorithm

The "simple" body canonicalization algorithm ignores all empty lines at the end of the message body. An empty line is a line of zero length after removal of the line terminator. If there is no body or no trailing CRLF on the message body, a CRLF is added. It makes no

Note that a completely empty or missing body is canonicalized as a single "CRLF"; that is, the canonicalized length will be 2 octets.

3.4.4 The "relaxed" Body Canonicalization Algorithm

The "relaxed" body canonicalization algorithm:

INFORMATIVE NOTE: It should be noted that the relaxed body canonicalization algorithm may enable certain types of extremely crude "ASCII Art" attacks where a message may be conveyed by adjusting the spacing between words. If this is a concern, the "simple" body canonicalization algorithm should be used instead.

3.4.5 Body Length Limits

A body length count MAY be specified to limit the signature calculation to an initial prefix of the body text, measured in octets. If the body length count is not specified, the entire message body is signed.

INFORMATIVE RATIONALE: This capability is provided because it is very common for mailing lists to add trailers to messages (e.g., instructions how to get off the list). Until those messages are also signed, the body length count is a useful tool for the verifier since it may as a matter of policy accept messages having valid signatures with extraneous data.
INFORMATIVE IMPLEMENTATION NOTE: Using body length limits enables an attack in which an attacker modifies a message to include content that solely benefits the attacker. It is possible for the appended content to completely replace the original content in the end recipient's eyes and to defeat duplicate message detection algorithms. To avoid this attack, signers should be wary of using this tag, and verifiers might wish to ignore the tag or remove text that appears after the specified content length, perhaps based on other criteria.

The body length count allows the signer of a message to permit data to be appended to the end of the body of a signed message. The body length count MUST be calculated following the canonicalization algorithm; for example, any whitespace ignored by a canonicalization algorithm is not included as part of the body length count. Signers of MIME messages that include a body length count SHOULD be sure that the length extends to the closing MIME boundary string.

INFORMATIVE IMPLEMENTATION NOTE: A signer wishing to ensure that the only acceptable modifications are to add to the MIME postlude would use a body length count encompassing the entire final MIME boundary string, including the final "--CRLF". A signer wishing to allow additional MIME parts but not modification of existing parts would use a body length count extending through the final MIME boundary string, omitting the final "--CRLF". Note that this only works for some MIME types, e.g., multipart/mixed but not multipart/signed.

A body length count of zero means that the body is completely unsigned.

Signers wishing to ensure that no modification of any sort can occur should specify the "simple" canonicalization algorithm for both header and body and omit the body length count.

3.4.6 Canonicalization Examples (INFORMATIVE)

In the following examples, actual whitespace is used only for clarity. The actual input and output text is designated using bracketed descriptors: "<SP>" for a space character, "<HTAB>" for a tab character, and "<CRLF>" for a carriage-return/line-feed sequence. For example, "X <SP> Y" and "X<SP>Y" represent the same three characters.

Example 1: A message reading:

    
    A: <SP> X <CRLF>
    B <SP> : <SP> Y <HTAB><CRLF>
    <HTAB> Z <SP><SP><CRLF>
    <CRLF>
    <SP> C <SP><CRLF>
    D <SP><HTAB><SP> E <CRLF>
    <CRLF>
    <CRLF>

when canonicalized using relaxed canonicalization for both header and body results in a header reading:

    
    a:X <CRLF>
    b:Y <SP> Z <CRLF>

and a body reading:

    
    <SP> C <CRLF>
    D <SP> E <CRLF>

Example 2: The same message canonicalized using simple canonicalization for both header and body results in a header reading:

    
    A: <SP> X <CRLF>
    B <SP> : <SP> Y <HTAB><CRLF>
    <HTAB> Z <SP><SP><CRLF>

and a body reading:

    
    <SP> C <SP><CRLF>
    D <SP><HTAB><SP> E <CRLF>

Example 3: When processed using relaxed header canonicalization and simple body canonicalization, the canonicalized version has a header of:

    
    a:X <CRLF>
    b:Y <SP> Z <CRLF>

and a body reading:

    
    <SP> C <SP><CRLF>
    D <SP><HTAB><SP> E <CRLF>

3.5 The DKIM-Signature Header Field

The signature of the email is stored in the DKIM-Signature header field. This header field contains all of the signature and key-fetching data. The DKIM-Signature value is a tag-list as described in Section 3.2.

The DKIM-Signature header field SHOULD be treated as though it were a trace header field as defined in Section 3.6 of [RFC2822], and hence SHOULD NOT be reordered and SHOULD be prepended to the message.

The DKIM-Signature header field being created or verified is always included in the signature calculation, after the rest of the header fields being signed; however, when calculating or verifying the signature, the value of the "b=" tag (signature value) of that DKIM-Signature header field MUST be treated as though it were an empty string. Unknown tags in the DKIM-Signature header field MUST be included in the signature calculation but MUST be otherwise ignored by verifiers. Other DKIM-Signature header fields that are included in the signature should be treated as normal header fields; in particular, the "b=" tag is not treated specially.

The encodings for each field type are listed below. Tags described as qp-section are encoded as described in Section 6.7 of MIME Part One [RFC2045], with the additional conversion of semicolon characters to "=3B"; intuitively, this is one line of quoted-printable encoded text. The dkim-quoted-printable syntax is defined in Section 2.6.

Tags on the DKIM-Signature header field along with their type and requirement status are shown below. Unrecognized tags MUST be ignored.

v=
Version (MUST be included). This tag defines the version of this specification that applies to the signature record. It MUST have the value "1". Note that verifiers must do a string comparison on this value; for example, "1" is not the same as "1.0".
ABNF:
    sig-v-tag   = %x76 [FWS] "=" [FWS] "1"
INFORMATIVE NOTE: DKIM-Signature version numbers are expected to increase arithmetically as new versions of this specification are released.
a=
The algorithm used to generate the signature (plain-text; REQUIRED). Verifiers MUST support "rsa-sha1" and "rsa-sha256"; signers SHOULD sign using "rsa-sha256". See Section 3.3 for a description of algorithms.
ABNF:
    sig-a-tag       = %x61 [FWS] "=" [FWS] sig-a-tag-alg
    sig-a-tag-alg   = sig-a-tag-k "-" sig-a-tag-h
    sig-a-tag-k     = "rsa" / x-sig-a-tag-k
    sig-a-tag-h     = "sha1" / "sha256" / x-sig-a-tag-h
    x-sig-a-tag-k   = ALPHA *(ALPHA / DIGIT)   ; for later extension
    x-sig-a-tag-h   = ALPHA *(ALPHA / DIGIT)   ; for later extension
b=
The signature data (base64; REQUIRED). Whitespace is ignored in this value and MUST be ignored when reassembling the original signature. In particular, the signing process can safely insert FWS in this value in arbitrary places to conform to line-length limits. See Signer Actions (Section 5) for how the signature is computed.
ABNF:
    sig-b-tag       = %x62 [FWS] "=" [FWS] sig-b-tag-data
    sig-b-tag-data  = base64string
bh=
The hash of the canonicalized body part of the message as limited by the "l=" tag (base64; REQUIRED). Whitespace is ignored in this value and MUST be ignored when reassembling the original signature. In particular, the signing process can safely insert FWS in this value in arbitrary places to conform to line-length limits. See Section 3.7 for how the body hash is computed.
ABNF:
    sig-bh-tag      = %x62 %x68 [FWS] "=" [FWS] sig-bh-tag-data
    sig-bh-tag-data = base64string
c=
Message canonicalization (plain-text; OPTIONAL, default is "simple/simple"). This tag informs the verifier of the type of canonicalization used to prepare the message for signing. It consists of two names separated by a "slash" (%d47) character, corresponding to the header and body canonicalization algorithms respectively. These algorithms are described in Section 3.4. If only one algorithm is named, that algorithm is used for the header and "simple" is used for the body. For example, "c=relaxed" is treated the same as "c=relaxed/simple".
ABNF:
    sig-c-tag       = %x63 [FWS] "=" [FWS] sig-c-tag-alg
                  ["/" sig-c-tag-alg]
    sig-c-tag-alg   = "simple" / "relaxed" / x-sig-c-tag-alg
    x-sig-c-tag-alg = hyphenated-word    ; for later extension
d=
The domain of the signing entity (plain-text; REQUIRED). This is the domain that will be queried for the public key. This domain MUST be the same as or a parent domain of the "i=" tag (the signing identity, as described below), or it MUST meet the requirements for parent domain signing described in Section 3.8. When presented with a signature that does not meet these requirement, verifiers MUST consider the signature invalid.
Internationalized domain names MUST be encoded as described in [RFC3490].
ABNF:
    sig-d-tag       = %x64 [FWS] "=" [FWS] domain-name
    domain-name     = sub-domain 1*("." sub-domain)
             ; from RFC 2821 Domain, but excluding address-literal
h=
Signed header fields (plain-text, but see description; REQUIRED). A colon-separated list of header field names that identify the header fields presented to the signing algorithm. The field MUST contain the complete list of header fields in the order presented to the signing algorithm. The field MAY contain names of header fields that do not exist when signed; nonexistent header fields do not contribute to the signature computation (that is, they are treated as the null input, including the header field name, the separating colon, the header field value, and any CRLF terminator). The field MUST NOT include the DKIM-Signature header field that is being created or verified, but may include others. Folding whitespace (FWS) MAY be included on either side of the colon separator. Header field names MUST be compared against actual header field names in a case-insensitive manner. This list MUST NOT be empty. See Section 5.4 for a discussion of choosing header fields to sign.
ABNF:
    sig-h-tag       = %x68 [FWS] "=" [FWS] hdr-name
                  0*( *FWS ":" *FWS hdr-name )
    hdr-name        = field-name
INFORMATIVE EXPLANATION: By "signing" header fields that do not actually exist, a signer can prevent insertion of those header fields before verification. However, since a signer cannot possibly know what header fields might be created in the future, and that some MUAs might present header fields that are embedded inside a message (e.g., as a message/rfc822 content type), the security of this solution is not total.
INFORMATIVE EXPLANATION: The exclusion of the header field name and colon as well as the header field value for non-existent header fields prevents an attacker from inserting an actual header field with a null value.
i=
Identity of the user or agent (e.g., a mailing list manager) on behalf of which this message is signed (dkim-quoted-printable; OPTIONAL, default is an empty Local-part followed by an "@" followed by the domain from the "d=" tag). The syntax is a standard email address where the Local-part MAY be omitted. The domain part of the address MUST be the same as or a subdomain of the value of the "d=" tag.
Internationalized domain names MUST be converted using the steps listed in Section 4 of [RFC3490] using the ToASCII function.
ABNF:
    sig-i-tag =   %x69 [FWS] "=" [FWS] [ Local-part ] "@" domain-name
INFORMATIVE NOTE: The local-part of the "i=" tag is optional because in some cases a signer may not be able to establish a verified individual identity. In such cases, the signer may wish to assert that although it is willing to go as far as signing for the domain, it is unable or unwilling to commit to an individual user name within their domain. It can do so by including the domain part but not the local-part of the identity.
INFORMATIVE DISCUSSION: This document does not require the value of the "i=" tag to match the identity in any message header fields. This is considered to be a verifier policy issue. Constraints between the value of the "i=" tag and other identities in other header fields seek to apply basic authentication into the semantics of trust associated with a role such as content author. Trust is a broad and complex topic and trust mechanisms are subject to highly creative attacks. The real-world efficacy of any but the most basic bindings between the "i=" value and other identities is not well established, nor is its vulnerability to subversion by an attacker. Hence reliance on the use of these options should be strictly limited. In particular, it is not at all clear to what extent a typical end-user recipient can rely on any assurances that might be made by successful use of the "i=" options.
l=
Body length count (plain-text unsigned decimal integer; OPTIONAL, default is entire body). This tag informs the verifier of the number of octets in the body of the email after canonicalization included in the cryptographic hash, starting from 0 immediately following the CRLF preceding the body. This value MUST NOT be larger than the actual number of octets in the canonicalized message body.
INFORMATIVE IMPLEMENTATION WARNING: Use of the "l=" tag might allow display of fraudulent content without appropriate warning to end users. The "l=" tag is intended for increasing signature robustness when sending to mailing lists that both modify their content and do not sign their messages. However, using the "l=" tag enables attacks in which an intermediary with malicious intent modifies a message to include content that solely benefits the attacker. It is possible for the appended content to completely replace the original content in the end recipient's eyes and to defeat duplicate message detection algorithms. Examples are described in Security Considerations (Section 8). To avoid this attack, signers should be extremely wary of using this tag, and verifiers might wish to ignore the tag or remove text that appears after the specified content length.
INFORMATIVE NOTE: The value of the "l=" tag is constrained to 76 decimal digits. This constraint is not intended to predict the size of future messages or to require implementations to use an integer representation large enough to represent the maximum possible value, but is intended to remind the implementer to check the length of this and all other tags during verification and to test for integer overflow when decoding the value. Implementers may need to limit the actual value expressed to a value smaller than 10^76, e.g., to allow a message to fit within the available storage space.
ABNF:
sig-l-tag    = %x6c [FWS] "=" [FWS] 1*76DIGIT
q=
A colon-separated list of query methods used to retrieve the public key (plain-text; OPTIONAL, default is "dns/txt"). Each query method is of the form "type[/options]", where the syntax and semantics of the options depend on the type and specified options. If there are multiple query mechanisms listed, the choice of query mechanism MUST NOT change the interpretation of the signature. Implementations MUST use the recognized query mechanisms in the order presented.
Currently, the only valid value is "dns/txt", which defines the DNS TXT record lookup algorithm described elsewhere in this document. The only option defined for the "dns" query type is "txt", which MUST be included. Verifiers and signers MUST support "dns/txt".
ABNF:
    sig-q-tag        = %x71 [FWS] "=" [FWS] sig-q-tag-method
                   *([FWS] ":" [FWS] sig-q-tag-method)
    sig-q-tag-method = "dns/txt" / x-sig-q-tag-type
                   ["/" x-sig-q-tag-args]
    x-sig-q-tag-type = hyphenated-word  ; for future extension
    x-sig-q-tag-args = qp-hdr-value
s=
The selector subdividing the namespace for the "d=" (domain) tag (plain-text; REQUIRED).
ABNF:
    sig-s-tag    = %x73 [FWS] "=" [FWS] selector
t=
Signature Timestamp (plain-text unsigned decimal integer; RECOMMENDED, default is an unknown creation time). The time that this signature was created. The format is the number of seconds since 00:00:00 on January 1, 1970 in the UTC time zone. The value is expressed as an unsigned integer in decimal ASCII. This value is not constrained to fit into a 31- or 32-bit integer. Implementations SHOULD be prepared to handle values up to at least 10^12 (until approximately AD 200,000; this fits into 40 bits). To avoid denial-of-service attacks, implementations MAY consider any value longer than 12 digits to be infinite. Leap seconds are not counted. Implementations MAY ignore signatures that have a timestamp in the future.
ABNF:
    sig-t-tag    = %x74 [FWS] "=" [FWS] 1*12DIGIT
x=
Signature Expiration (plain-text unsigned decimal integer; RECOMMENDED, default is no expiration). The format is the same as in the "t=" tag, represented as an absolute date, not as a time delta from the signing timestamp. The value is expressed as an unsigned integer in decimal ASCII, with the same constraints on the value in the "t=" tag. Signatures MAY be considered invalid if the verification time at the verifier is past the expiration date. The verification time should be the time that the message was first received at the administrative domain of the verifier if that time is reliably available; otherwise the current time should be used. The value of the "x=" tag MUST be greater than the value of the "t=" tag if both are present.
INFORMATIVE NOTE: The "x=" tag is not intended as an anti-replay defense.
ABNF:
    sig-x-tag    = %x78 [FWS] "=" [FWS] 1*12DIGIT
z=
Copied header fields (dkim-quoted-printable, but see description; OPTIONAL, default is null). A vertical-bar-separated list of selected header fields present when the message was signed, including both the field name and value. It is not required to include all header fields present at the time of signing. This field need not contain the same header fields listed in the "h=" tag. The header field text itself must encode the vertical bar ("|", %x7C) character (i.e., vertical bars in the "z=" text are metacharacters, and any actual vertical bar characters in a copied header field must be encoded). Note that all whitespace must be encoded, including whitespace between the colon and the header field value. After encoding, FWS MAY be added at arbitrary locations in order to avoid excessively long lines; such whitespace is NOT part of the value of the header field, and MUST be removed before decoding.
The header fields referenced by the "h=" tag refer to the fields in the RFC 2822 header of the message, not to any copied fields in the "z=" tag. Copied header field values are for diagnostic use.
Header fields with characters requiring conversion (perhaps from legacy MTAs that are not [RFC2822] compliant) SHOULD be converted as described in MIME Part Three [RFC2047].
ABNF:
    sig-z-tag      = %x7A [FWS] "=" [FWS] sig-z-tag-copy
                 *( [FWS] "|" sig-z-tag-copy )
sig-z-tag-copy = hdr-name ":" qp-hdr-value
qp-hdr-value   = dkim-quoted-printable    ; with "|" encoded
INFORMATIVE EXAMPLE of a signature header field spread across multiple continuation lines: 
DKIM-Signature: a=rsa-sha256; d=example.net; s=brisbane;
   c=simple; q=dns/txt; i=@eng.example.net;
   t=1117574938; x=1118006938;
   h=from:to:subject:date;
   z=From:foo@eng.example.net|To:joe@example.com|
     Subject:demo=20run|Date:July=205,=202005=203:44:08=20PM=20-0700;
   bh=MTIzNDU2Nzg5MDEyMzQ1Njc4OTAxMjM0NTY3ODkwMTI=;
   b=dzdVyOfAKCdLXdJOc9G2q8LoXSlEniSbav+yuU4zGeeruD00lszZ
            VoG4ZHRNiYzR

3.6 Key Management and Representation

Signature applications require some level of assurance that the verification public key is associated with the claimed signer. Many applications achieve this by using public key certificates issued by a trusted third party. However, DKIM can achieve a sufficient level of security, with significantly enhanced scalability, by simply having the verifier query the purported signer's DNS entry (or some security-equivalent) in order to retrieve the public key.

DKIM keys can potentially be stored in multiple types of key servers and in multiple formats. The storage and format of keys are irrelevant to the remainder of the DKIM algorithm.

Parameters to the key lookup algorithm are the type of the lookup (the "q=" tag), the domain of the signer (the "d=" tag of the DKIM-Signature header field), and the selector (the "s=" tag).

    public_key = dkim_find_key(q_val, d_val, s_val)

This document defines a single binding, using DNS TXT records to distribute the keys. Other bindings may be defined in the future.

3.6.1 Textual Representation

It is expected that many key servers will choose to present the keys in an otherwise unstructured text format (for example, an XML form would not be considered to be unstructured text for this purpose). The following definition MUST be used for any DKIM key represented in an otherwise unstructured textual form.

The overall syntax is a tag-list as described in Section 3.2. The current valid tags are described below. Other tags MAY be present and MUST be ignored by any implementation that does not understand them.

v=
Version of the DKIM key record (plain-text; RECOMMENDED, default is "DKIM1"). If specified, this tag MUST be set to "DKIM1" (without the quotes). This tag MUST be the first tag in the record. Records beginning with a "v=" tag with any other value MUST be discarded. Note that verifiers must do a string comparison on this value; for example, "DKIM1" is not the same as "DKIM1.0".

ABNF:
    key-v-tag    = %x76 [FWS] "=" [FWS] "DKIM1"
g=
Granularity of the key (plain-text; OPTIONAL, default is "*"). This value MUST match the Local-part of the "i=" tag of the DKIM-Signature header field (or its default value of the empty string if "i=" is not specified), with a single, optional "*" character matching a sequence of zero or more arbitrary characters ("wildcarding"). An email with a signing address that does not match the value of this tag constitutes a failed verification. The intent of this tag is to constrain which signing address can legitimately use this selector, for example, when delegating a key to a third party that should only be used for special purposes. Wildcarding allows matching for addresses such as "user+*" or "*-offer". An empty "g=" value never matches any addresses.
ABNF:
    key-g-tag       = %x67 [FWS] "=" [FWS] key-g-tag-lpart
    key-g-tag-lpart = [dot-atom-text] ["*" [dot-atom-text] ]
h=
Acceptable hash algorithms (plain-text; OPTIONAL, defaults to allowing all algorithms). A colon-separated list of hash algorithms that might be used. Signers and Verifiers MUST support the "sha256" hash algorithm. Verifiers MUST also support the "sha1" hash algorithm.

ABNF:
    key-h-tag       = %x68 [FWS] "=" [FWS] key-h-tag-alg
                  0*( [FWS] ":" [FWS] key-h-tag-alg )
    key-h-tag-alg   = "sha1" / "sha256" / x-key-h-tag-alg
    x-key-h-tag-alg = hyphenated-word   ; for future extension
k=
Key type (plain-text; OPTIONAL, default is "rsa"). Signers and verifiers MUST support the "rsa" key type. The "rsa" key type indicates that an ASN.1 DER-encoded [ITU.X660.1997] RSAPublicKey [RFC3447] (see Sections 3.1 and A.1.1) is being used in the "p=" tag. (Note: the "p=" tag further encodes the value using the base64 algorithm.)

ABNF:
    key-k-tag        = %x76 [FWS] "=" [FWS] key-k-tag-type
    key-k-tag-type   = "rsa" / x-key-k-tag-type
    x-key-k-tag-type = hyphenated-word   ; for future extension
n=
Notes that might be of interest to a human (qp-section; OPTIONAL, default is empty). No interpretation is made by any program. This tag should be used sparingly in any key server mechanism that has space limitations (notably DNS). This is intended for use by administrators, not end users.
ABNF:
    key-n-tag    = %x6e [FWS] "=" [FWS] qp-section
p=
Public-key data (base64; REQUIRED). An empty value means that this public key has been revoked. The syntax and semantics of this tag value before being encoded in base64 are defined by the "k=" tag.
INFORMATIVE RATIONALE: If a private key has been compromised or otherwise disabled (e.g., an outsourcing contract has been terminated), a signer might want to explicitly state that it knows about the selector, but all messages using that selector should fail verification. Verifiers should ignore any DKIM-Signature header fields with a selector referencing a revoked key.
ABNF:
    key-p-tag    = %x70 [FWS] "=" [ [FWS] base64string ]
INFORMATIVE NOTE: A base64string is permitted to include white space (FWS) at arbitrary places; however, any CRLFs must be followed by at least one WSP character. Implementors and administrators are cautioned to ensure that selector TXT records conform to this specification.
s=
Service Type (plain-text; OPTIONAL; default is "*"). A colon-separated list of service types to which this record applies. Verifiers for a given service type MUST ignore this record if the appropriate type is not listed. Currently defined service types are as follows:
*
matches all service types
email
electronic mail (not necessarily limited to SMTP)
This tag is intended to constrain the use of keys for other purposes, should use of DKIM be defined by other services in the future.
ABNF: 
    key-s-tag        = %x73 [FWS] "=" [FWS] key-s-tag-type
                    0*( [FWS] ":" [FWS] key-s-tag-type )
    key-s-tag-type   = "email" / "*" / x-key-s-tag-type
    x-key-s-tag-type = hyphenated-word   ; for future extension
t=
Flags, represented as a colon-separated list of names (plain-text; OPTIONAL, default is no flags set). The defined flags are as follows:
y
This domain is testing DKIM. Verifiers MUST NOT treat messages from signers in testing mode differently from unsigned email, even should the signature fail to verify. Verifiers MAY wish to track testing mode results to assist the signer.
s
Any DKIM-Signature header fields using the "i=" tag MUST have the same domain value on the right-hand side of the "@" in the "i=" tag and the value of the "d=" tag. That is, the "i=" domain MUST NOT be a subdomain of "d=". Use of this flag is RECOMMENDED unless subdomaining is required.
ABNF: 
    key-t-tag        = %x74 [FWS] "=" [FWS] key-t-tag-flag
                   0*( [FWS] ":" [FWS] key-t-tag-flag )
    key-t-tag-flag   = "y" / "s" / x-key-t-tag-flag
    x-key-t-tag-flag = hyphenated-word   ; for future extension
Unrecognized flags MUST be ignored.

3.6.2 DNS Binding

A binding using DNS TXT records as a key service is hereby defined. All implementations MUST support this binding.

3.6.2.1 Namespace

All DKIM keys are stored in a subdomain named _domainkey. Given a DKIM-Signature field with a "d=" tag of example.com and an "s=" tag of foo.bar, the DNS query will be for foo.bar._domainkey.example.com.

INFORMATIVE OPERATIONAL NOTE: Wildcard DNS records (e.g., *.bar._domainkey.example.com) do not make sense in this context and should not be used. Note also that wildcards within domains (e.g., s._domainkey.*.example.com) are not supported by the DNS.

3.6.2.2 Resource Record Types for Key Storage

The DNS Resource Record type used is specified by an option to the query-type ("q=") tag. The only option defined in this base specification is "txt", indicating the use of a TXT Resource Record (RR). A later extension of this standard may define another RR type.

Strings in a TXT RR MUST be concatenated together before use with no intervening whitespace. TXT RRs MUST be unique for a particular selector name; that is, if there are multiple records in an RRset, the results are undefined.

TXT RRs are encoded as described in Section 3.6.1.

3.7 Computing the Message Hashes

Both signing and verifying message signatures start with a step of computing two cryptographic hashes over the message. Signers will choose the parameters of the signature as described in Signer Actions (Section 5); verifiers will use the parameters specified in the DKIM-Signature header field being verified. In the following discussion, the names of the tags in the DKIM-Signature header field that either exists (when verifying) or will be created (when signing) are used. Note that canonicalization (Section 3.4) is only used to prepare the email for signing or verifying; it does not affect the transmitted email in any way.

The signer/verifier MUST compute two hashes, one over the body of the message and one over the selected header fields of the message. Signers MUST compute them in the order shown. Verifiers MAY compute them in any order convenient to the verifier, provided that the result is semantically identical to the semantics that would be the case had they been computed in this order.

In hash step 1, the signer/verifier MUST hash the message body, canonicalized using the body canonicalization algorithm specified in the "c=" tag and then truncated to the length specified in the "l=" tag. That hash value is then converted to base64 form and inserted into (signers) or compared to (verifiers) the "bh=" tag of the DKIM-Signature header field.

In hash step 2, the signer/verifier MUST pass the following to the hash algorithm in the indicated order.

  1. The header fields specified by the "h=" tag, in the order specified in that tag, and canonicalized using the header canonicalization algorithm specified in the "c=" tag. Each header field MUST be terminated with a single CRLF.
  2. The DKIM-Signature header field that exists (verifying) or will be inserted (signing) in the message, with the value of the "b=" tag deleted (i.e., treated as the empty string), canonicalized using the header canonicalization algorithm specified in the "c=" tag, and without a trailing CRLF.

All tags and their values in the DKIM-Signature header field are included in the cryptographic hash with the sole exception of the value portion of the "b=" (signature) tag, which MUST be treated as the null string. All tags MUST be included even if they might not be understood by the verifier. The header field MUST be presented to the hash algorithm after the body of the message rather than with the rest of the header fields and MUST be canonicalized as specified in the "c=" (canonicalization) tag. The DKIM-Signature header field MUST NOT be included in its own h= tag, although other DKIM-Signature header fields MAY be signed (see Section 4).

When calculating the hash on messages that will be transmitted using base64 or quoted-printable encoding, signers MUST compute the hash after the encoding. Likewise, the verifier MUST incorporate the values into the hash before decoding the base64 or quoted-printable text. However, the hash MUST be computed before transport level encodings such as SMTP "dot-stuffing" (the modification of lines beginning with a "." to avoid confusion with the SMTP end-of-message marker, as specified in [RFC2821]).

With the exception of the canonicalization procedure described in Section 3.4, the DKIM signing process treats the body of messages as simply a string of octets. DKIM messages MAY be either in plain-text or in MIME format; no special treatment is afforded to MIME content. Message attachments in MIME format MUST be included in the content that is signed.

More formally, the algorithm for the signature is as follows:

    body-hash = hash-alg(canon_body)
    header-hash = hash-alg(canon_header || DKIM-SIG)
    signature = sig-alg(header-hash, key)

where sig-alg is the signature algorithm specified by the "a=" tag, hash-alg is the hash algorithm specified by the "a=" tag, canon_header and canon_body are the canonicalized message header and body (respectively) as defined in Section 3.4 (excluding the DKIM-Signature header field), and DKIM-SIG is the canonicalized DKIM-Signature header field sans the signature value itself, but with body-hash included as the "bh=" tag.

INFORMATIVE IMPLEMENTERS' NOTE: Many digital signature APIs provide both hashing and application of the RSA private key using a single sign() primitive. When using such an API, the last two steps in the algorithm would probably be combined into a single call that would perform both the hash-alg and the sig-alg.

3.8 Signing by Parent Domains

In some circumstances, it is desirable for a domain to apply a signature on behalf of any of its subdomains without the need to maintain separate selectors (key records) in each subdomain. By default, private keys corresponding to key records can be used to sign messages for any subdomain of the domain in which they reside; e.g., a key record for the domain example.com can be used to verify messages where the signing identity ("i=" tag of the signature) is sub.example.com, or even sub1.sub2.example.com. In order to limit the capability of such keys when this is not intended, the s flag may be set in the "t=" tag of the key record to constrain the validity of the record to exactly the domain of the signing identity. If the referenced key record contains the s flag as part of the "t=" tag, the domain of the signing identity ("i=" flag) MUST be the same as that of the d= domain. If this flag is absent, the domain of the signing identity MUST be the same as, or a subdomain of, the d= domain. Key records that are not intended for use with subdomains SHOULD specify the s flag in the "t=" tag.

4. Semantics of Multiple Signatures

4.1 Example Scenarios

There are many reasons why a message might have multiple signatures. For example, a given signer might sign multiple times, perhaps with different hashing or signing algorithms during a transition phase.

INFORMATIVE EXAMPLE: Suppose SHA-256 is in the future found to be insufficiently strong, and DKIM usage transitions to SHA-1024. A signer might immediately sign using the newer algorithm, but continue to sign using the older algorithm for interoperability with verifiers that had not yet upgraded. The signer would do this by adding two DKIM-Signature header fields, one using each algorithm. Older verifiers that did not recognize SHA-1024 as an acceptable algorithm would skip that signature and use the older algorithm; newer verifiers could use either signature at their option, and all other things being equal might not even attempt to verify the other signature.

Similarly, a signer might sign a message including all headers and no "l=" tag (to satisfy strict verifiers) and a second time with a limited set of headers and an "l=" tag (in anticipation of possible message modifications in route to other verifiers). Verifiers could then choose which signature they preferred.

INFORMATIVE EXAMPLE: A verifier might receive a message with two signatures, one covering more of the message than the other. If the signature covering more of the message verified, then the verifier could make one set of policy decisions; if that signature failed but the signature covering less of the message verified, the verifier might make a different set of policy decisions.

Of course, a message might also have multiple signatures because it passed through multiple signers. A common case is expected to be that of a signed message that passes through a mailing list that also signs all messages. Assuming both of those signatures verify, a recipient might choose to accept the message if either of those signatures were known to come from trusted sources.

INFORMATIVE EXAMPLE: Recipients might choose to whitelist mailing lists to which they have subscribed and that have acceptable anti-abuse policies so as to accept messages sent to that list even from unknown authors. They might also subscribe to less trusted mailing lists (e.g., those without anti-abuse protection) and be willing to accept all messages from specific authors, but insist on doing additional abuse scanning for other messages.

Another related example of multiple signers might be forwarding services, such as those commonly associated with academic alumni sites.

INFORMATIVE EXAMPLE: A recipient might have an address at members.example.org, a site that has anti-abuse protection that is somewhat less effective than the recipient would prefer. Such a recipient might have specific authors whose messages would be trusted absolutely, but messages from unknown authors that had passed the forwarder's scrutiny would have only medium trust.

4.2 Interpretation

A signer that is adding a signature to a message merely creates a new DKIM-Signature header, using the usual semantics of the h= option. A signer MAY sign previously existing DKIM-Signature header fields using the method described in Section 5.4 to sign trace header fields.

INFORMATIVE NOTE: Signers should be cognizant that signing DKIM-Signature header fields may result in signature failures with intermediaries that do not recognize that DKIM-Signature header fields are trace header fields and unwittingly reorder them, thus breaking such signatures. For this reason, signing existing DKIM-Signature header fields is unadvised, albeit legal.
INFORMATIVE NOTE: If a header field with multiple instances is signed, those header fields are always signed from the bottom up. Thus, it is not possible to sign only specific DKIM-Signature header fields. For example, if the message being signed already contains three DKIM-Signature header fields A, B, and C, it is possible to sign all of them, B and C only, or C only, but not A only, B only, A and B only, or A and C only.

A signer MAY add more than one DKIM-Signature header field using different parameters. For example, during a transition period a signer might want to produce signatures using two different hash algorithms.

Signers SHOULD NOT remove any DKIM-Signature header fields from messages they are signing, even if they know that the signatures cannot be verified.

When evaluating a message with multiple signatures, a verifier SHOULD evaluate signatures independently and on their own merits. For example, a verifier that by policy chooses not to accept signatures with deprecated cryptographic algorithms would consider such signatures invalid. Verifiers MAY process signatures in any order of their choice; for example, some verifiers might choose to process signatures corresponding to the From field in the message header before other signatures. See Section 6.1 for more information about signature choices.

INFORMATIVE IMPLEMENTATION NOTE: Verifier attempts to correlate valid signatures with invalid signatures in an attempt to guess why a signature failed are ill-advised. In particular, there is no general way that a verifier can determine that an invalid signature was ever valid.

Verifiers SHOULD ignore failed signatures as though they were not present in the message. Verifiers SHOULD continue to check signatures until a signature successfully verifies to the satisfaction of the verifier. To limit potential denial-of-service attacks, verifiers MAY limit the total number of signatures they will attempt to verify.

5. Signer Actions

The following steps are performed in order by signers.

5.1 Determine Whether the Email Should Be Signed and by Whom

A signer can obviously only sign email for domains for which it has a private key and the necessary knowledge of the corresponding public key and selector information. However, there are a number of other reasons beyond the lack of a private key why a signer could choose not to sign an email.

INFORMATIVE NOTE: Signing modules may be incorporated into any portion of the mail system as deemed appropriate, including an MUA, a SUBMISSION server, or an MTA. Wherever implemented, signers should beware of signing (and thereby asserting responsibility for) messages that may be problematic. In particular, within a trusted enclave the signing address might be derived from the header according to local policy; SUBMISSION servers might only sign messages from users that are properly authenticated and authorized.
INFORMATIVE IMPLEMENTER ADVICE: SUBMISSION servers should not sign Received header fields if the outgoing gateway MTA obfuscates Received header fields, for example, to hide the details of internal topology.

If an email cannot be signed for some reason, it is a local policy decision as to what to do with that email.

5.2 Select a Private Key and Corresponding Selector Information

This specification does not define the basis by which a signer should choose which private key and selector information to use. Currently, all selectors are equal as far as this specification is concerned, so the decision should largely be a matter of administrative convenience. Distribution and management of private keys is also outside the scope of this document.

INFORMATIVE OPERATIONS ADVICE: A signer should not sign with a private key when the selector containing the corresponding public key is expected to be revoked or removed before the verifier has an opportunity to validate the signature. The signer should anticipate that verifiers may choose to defer validation, perhaps until the message is actually read by the final recipient. In particular, when rotating to a new key pair, signing should immediately commence with the new private key and the old public key should be retained for a reasonable validation interval before being removed from the key server.

5.3 Normalize the Message to Prevent Transport Conversions

Some messages, particularly those using 8-bit characters, are subject to modification during transit, notably conversion to 7-bit form. Such conversions will break DKIM signatures. In order to minimize the chances of such breakage, signers SHOULD convert the message to a suitable MIME content transfer encoding such as quoted-printable or base64 as described in MIME Part One [RFC2045] before signing. Such conversion is outside the scope of DKIM; the actual message SHOULD be converted to 7-bit MIME by an MUA or MSA prior to presentation to the DKIM algorithm.

If the message is submitted to the signer with any local encoding that will be modified before transmission, that modification to canonical [RFC2822] form MUST be done before signing. In particular, bare CR or LF characters (used by some systems as a local line separator convention) MUST be converted to the SMTP-standard CRLF sequence before the message is signed. Any conversion of this sort SHOULD be applied to the message actually sent to the recipient(s), not just to the version presented to the signing algorithm.

More generally, the signer MUST sign the message as it is expected to be received by the verifier rather than in some local or internal form.

5.4 Determine the Header Fields to Sign

The From header field MUST be signed (that is, included in the "h=" tag of the resulting DKIM-Signature header field). Signers SHOULD NOT sign an existing header field likely to be legitimately modified or removed in transit. In particular, [RFC2821] explicitly permits modification or removal of the Return-Path header field in transit. Signers MAY include any other header fields present at the time of signing at the discretion of the signer.

INFORMATIVE OPERATIONS NOTE: The choice of which header fields to sign is non-obvious. One strategy is to sign all existing, non-repeatable header fields. An alternative strategy is to sign only header fields that are likely to be displayed to or otherwise be likely to affect the processing of the message at the receiver. A third strategy is to sign only "well known" headers. Note that verifiers may treat unsigned header fields with extreme skepticism, including refusing to display them to the end user or even ignoring the signature if it does not cover certain header fields. For this reason, signing fields present in the message such as Date, Subject, Reply-To, Sender, and all MIME header fields are highly advised.

The DKIM-Signature header field is always implicitly signed and MUST NOT be included in the "h=" tag except to indicate that other preexisting signatures are also signed.

Signers MAY claim to have signed header fields that do not exist (that is, signers MAY include the header field name in the "h=" tag even if that header field does not exist in the message). When computing the signature, the non-existing header field MUST be treated as the null string (including the header field name, header field value, all punctuation, and the trailing CRLF).

INFORMATIVE RATIONALE: This allows signers to explicitly assert the absence of a header field; if that header field is added later the signature will fail.
INFORMATIVE NOTE: A header field name need only be listed once more than the actual number of that header field in a message at the time of signing in order to prevent any further additions. For example, if there is a single Comments header field at the time of signing, listing Comments twice in the "h=" tag is sufficient to prevent any number of Comments header fields from being appended; it is not necessary (but is legal) to list Comments three or more times in the "h=" tag.

Signers choosing to sign an existing header field that occurs more than once in the message (such as Received) MUST sign the physically last instance of that header field in the header block. Signers wishing to sign multiple instances of such a header field MUST include the header field name multiple times in the h= tag of the DKIM-Signature header field, and MUST sign such header fields in order from the bottom of the header field block to the top. The signer MAY include more instances of a header field name in h= than there are actual corresponding header fields to indicate that additional header fields of that name SHOULD NOT be added.

INFORMATIVE EXAMPLE:
If the signer wishes to sign two existing Received header fields, and the existing header contains:
    Received: <A>
    Received: <B>
    Received: <C>
then the resulting DKIM-Signature header field should read: 
    DKIM-Signature: ... h=Received : Received : ...
and Received header fields <C> and <B> will be signed in that order.

Signers should be careful of signing header fields that might have additional instances added later in the delivery process, since such header fields might be inserted after the signed instance or otherwise reordered. Trace header fields (such as Received) and Resent-* blocks are the only fields prohibited by [RFC2822] from being reordered. In particular, since DKIM-Signature header fields may be reordered by some intermediate MTAs, signing existing DKIM-Signature header fields is error-prone.

INFORMATIVE ADMONITION: Despite the fact that [RFC2822] permits header fields to be reordered (with the exception of Received header fields), reordering of signed header fields with multiple instances by intermediate MTAs will cause DKIM signatures to be broken; such anti-social behavior should be avoided.
INFORMATIVE IMPLEMENTER'S NOTE: Although not required by this specification, all end-user visible header fields should be signed to avoid possible "indirect spamming". For example, if the Subject header field is not signed, a spammer can resend a previously signed mail, replacing the legitimate subject with a one-line spam.

5.5 Recommended Signature Content

In order to maximize compatibility with a variety of verifiers, it is recommended that signers follow the practices outlined in this section when signing a message. However, these are generic recommendations applying to the general case; specific senders may wish to modify these guidelines as required by their unique situations. Verifiers MUST be capable of verifying signatures even if one or more of the recommended header fields is not signed (with the exception of From, which must always be signed) or if one or more of the disrecommended header fields is signed. Note that verifiers do have the option of ignoring signatures that do not cover a sufficient portion of the header or body, just as they may ignore signatures from an identity they do not trust.

The following header fields SHOULD be included in the signature, if they are present in the message being signed:

The following header fields SHOULD NOT be included in the signature:

Optional header fields (those not mentioned above) normally SHOULD NOT be included in the signature, because of the potential for additional header fields of the same name to be legitimately added or reordered prior to verification. There are likely to be legitimate exceptions to this rule, because of the wide variety of application-specific header fields that may be applied to a message, some of which are unlikely to be duplicated, modified, or reordered.

Signers SHOULD choose canonicalization algorithms based on the types of messages they process and their aversion to risk. For example, e-commerce sites sending primarily purchase receipts, which are not expected to be processed by mailing lists or other software likely to modify messages, will generally prefer "simple" canonicalization. Sites sending primarily person-to-person email will likely prefer to be more resilient to modification during transport by using "relaxed" canonicalization.

Signers SHOULD NOT use "l=" unless they intend to accommodate intermediate mail processors that append text to a message. For example, many mailing list processors append "unsubscribe" information to message bodies. If signers use "l=", they SHOULD include the entire message body existing at the time of signing in computing the count. In particular, signers SHOULD NOT specify a body length of 0 since this may be interpreted as a meaningless signature by some verifiers.

5.6 Compute the Message Hash and Signature

The signer MUST compute the message hash as described in Section 3.7 and then sign it using the selected public-key algorithm. This will result in a DKIM-Signature header field that will include the body hash and a signature of the header hash, where that header includes the DKIM-Signature header field itself.

Entities such as mailing list managers that implement DKIM and that modify the message or a header field (for example, inserting unsubscribe information) before retransmitting the message SHOULD check any existing signature on input and MUST make such modifications before re-signing the message.

The signer MAY elect to limit the number of bytes of the body that will be included in the hash and hence signed. The length actually hashed should be inserted in the "l=" tag of the DKIM-Signature header field.

5.7 Insert the DKIM-Signature Header Field

Finally, the signer MUST insert the DKIM-Signature header field created in the previous step prior to transmitting the email. The DKIM-Signature header field MUST be the same as used to compute the hash as described above, except that the value of the "b=" tag MUST be the appropriately signed hash computed in the previous step, signed using the algorithm specified in the "a=" tag of the DKIM-Signature header field and using the private key corresponding to the selector given in the "s=" tag of the DKIM-Signature header field, as chosen above in Section 5.2

The DKIM-Signature header field MUST be inserted before any other DKIM-Signature fields in the header block.

INFORMATIVE IMPLEMENTATION NOTE: The easiest way to achieve this is to insert the DKIM-Signature header field at the beginning of the header block. In particular, it may be placed before any existing Received header fields. This is consistent with treating DKIM-Signature as a trace header field.

6. Verifier Actions

Since a signer MAY remove or revoke a public key at any time, it is recommended that verification occur in a timely manner. In many configurations, the most timely place is during acceptance by the border MTA or shortly thereafter. In particular, deferring verification until the message is accessed by the end user is discouraged.

A border or intermediate MTA MAY verify the message signature(s). An MTA who has performed verification MAY communicate the result of that verification by adding a verification header field to incoming messages. This considerably simplifies things for the user, who can now use an existing mail user agent. Most MUAs have the ability to filter messages based on message header fields or content; these filters would be used to implement whatever policy the user wishes with respect to unsigned mail.

A verifying MTA MAY implement a policy with respect to unverifiable mail, regardless of whether or not it applies the verification header field to signed messages.

Verifiers MUST produce a result that is semantically equivalent to applying the following steps in the order listed. In practice, several of these steps can be performed in parallel in order to improve performance.

6.1 Extract Signatures from the Message

The order in which verifiers try DKIM-Signature header fields is not defined; verifiers MAY try signatures in any order they like. For example, one implementation might try the signatures in textual order, whereas another might try signatures by identities that match the contents of the From header field before before trying other signatures. Verifiers MUST NOT attribute ultimate meaning to the order of multiple DKIM-Signature header fields. In particular, there is reason to believe that some relays will reorder the header fields in potentially arbitrary ways.

INFORMATIVE IMPLEMENTATION NOTE: Verifiers might use the order as a clue to signing order in the absence of any other information. However, other clues as to the semantics of multiple signatures (such as correlating the signing host with Received header fields) may also be considered.

A verifier SHOULD NOT treat a message that has one or more bad signatures and no good signatures differently from a message with no signature at all; such treatment is a matter of local policy and is beyond the scope of this document.

When a signature successfully verifies, a verifier will either stop processing or attempt to verify any other signatures, at the discretion of the implementation. A verifier MAY limit the number of signatures it tries to avoid denial-of-service attacks.

INFORMATIVE NOTE: An attacker could send messages with large numbers of faulty signatures, each of which would require a DNS lookup and corresponding CPU time to verify the message. This could be an attack on the domain that receives the message, by slowing down the verifier by requiring it to do a large number of DNS lookups and/or signature verifications. It could also be an attack against the domains listed in the signatures, essentially by enlisting innocent verifiers in launching an attack against the DNS servers of the actual victim.

In the following description, text reading "return status (explanation)" (where "status" is one of "PERMFAIL" or "TEMPFAIL") means that the verifier MUST immediately cease processing that signature. The verifier SHOULD proceed to the next signature, if any is present, and completely ignore the bad signature. If the status is "PERMFAIL", the signature failed and should not be reconsidered. If the status is "TEMPFAIL", the signature could not be verified at this time but may be tried again later. A verifier MAY either defer the message for later processing, perhaps by queueing it locally or issuing a 451/4.7.5 SMTP reply, or try another signature; if no good signature is found and any of the signatures resulted in a TEMPFAIL status, the verifier MAY save the message for later processing. The "(explanation)" is not normative text; it is provided solely for clarification.

Verifiers SHOULD ignore any DKIM-Signature header fields where the signature does not validate. Verifiers that are prepared to validate multiple signature header fields SHOULD proceed to the next signature header field, should it exist. However, verifiers MAY make note of the fact that an invalid signature was present for consideration at a later step.

INFORMATIVE NOTE: The rationale of this requirement is to permit messages that have invalid signatures but also a valid signature to work. For example, a mailing list exploder might opt to leave the original submitter signature in place even though the exploder knows that it is modifying the message in some way that will break that signature, and the exploder inserts its own signature. In this case, the message should succeed even in the presence of the known-broken signature.

For each signature to be validated, the following steps should be performed in such a manner as to produce a result that is semantically equivalent to performing them in the indicated order.

6.1.1 Validate the Signature Header Field

Implementers MUST meticulously validate the format and values in the DKIM-Signature header field; any inconsistency or unexpected values MUST cause the header field to be completely ignored and the verifier to return PERMFAIL (signature syntax error). Being "liberal in what you accept" is definitely a bad strategy in this security context. Note however that this does not include the existence of unknown tags in a DKIM-Signature header field, which are explicitly permitted.

Verifiers MUST ignore DKIM-Signature header fields with a "v=" tag that is inconsistent with this specification and return PERMFAIL (incompatible version).

INFORMATIVE IMPLEMENTATION NOTE: An implementation may, of course, choose to also verify signatures generated by older versions of this specification.

If any tag listed as "required" in Section 3.5 is omitted from the DKIM-Signature header field, the verifier MUST ignore the DKIM-Signature header field and return PERMFAIL (signature missing required tag).

INFORMATIONAL NOTE: The tags listed as required in Section 3.5 are v=, a=, b=, bh=, d=, h=, and s=. Should there be a conflict between this note and Section 3.5, Section 3.5 is normative.

If the DKIM-Signature header field does not contain the "i=" tag, the verifier MUST behave as though the value of that tag were "@d", where "d" is the value from the "d=" tag.

Verifiers MUST confirm that the domain specified in the "d=" tag is the same as or a parent domain of the domain part of the "i=" tag. If not, the DKIM-Signature header field MUST be ignored and the verifier should return PERMFAIL (domain mismatch).

If the "h=" tag does not include the From header field, the verifier MUST ignore the DKIM-Signature header field and return PERMFAIL (From field not signed).

Verifiers MAY ignore the DKIM-Signature header field and return PERMFAIL (signature expired) if it contains an "x=" tag and the signature has expired.

Verifiers MAY ignore the DKIM-Signature header field if the domain used by the signer in the "d=" tag is not associated with a valid signing entity. For example, signatures with "d=" values such as "com" and "co.uk" may be ignored. The list of unacceptable domains SHOULD be configurable.

Verifiers MAY ignore the DKIM-Signature header field and return PERMFAIL (unacceptable signature header) for any other reason, for example, if the signature does not sign header fields that the verifier views to be essential. As a case in point, if MIME header fields are not signed, certain attacks may be possible that the verifier would prefer to avoid.

6.1.2 Get the Public Key

The public key for a signature is needed to complete the verification process. The process of retrieving the public key depends on the query type as defined by the "q=" tag in the DKIM-Signature header field. Obviously, a public key need only be retrieved if the process of extracting the signature information is completely successful. Details of key management and representation are described in Section 3.6. The verifier MUST validate the key record and MUST ignore any public key records that are malformed.

When validating a message, a verifier MUST perform the following steps in a manner that is semantically the same as performing them in the order indicated (in some cases, the implementation may parallelize or reorder these steps, as long as the semantics remain unchanged):

  1. Retrieve the public key as described in Section 3.6 using the algorithm in the "q=" tag, the domain from the "d=" tag, and the selector from the "s=" tag.
  2. If the query for the public key fails to respond, the verifier MAY defer acceptance of this email and return TEMPFAIL (key unavailable). If verification is occurring during the incoming SMTP session, this MAY be achieved with a 451/4.7.5 SMTP reply code. Alternatively, the verifier MAY store the message in the local queue for later trial or ignore the signature. Note that storing a message in the local queue is subject to denial-of-service attacks.
  3. If the query for the public key fails because the corresponding key record does not exist, the verifier MUST immediately return PERMFAIL (no key for signature).
  4. If the query for the public key returns multiple key records, the verifier may choose one of the key records or may cycle through the key records performing the remainder of these steps on each record at the discretion of the implementer. The order of the key records is unspecified. If the verifier chooses to cycle through the key records, then the "return ..." wording in the remainder of this section means "try the next key record, if any; if none, return to try another signature in the usual way".
  5. If the result returned from the query does not adhere to the format defined in this specification, the verifier MUST ignore the key record and return PERMFAIL (key syntax error). Verifiers are urged to validate the syntax of key records carefully to avoid attempted attacks. In particular, the verifier MUST ignore keys with a version code ("v=" tag) that they do not implement.
  6. If the "g=" tag in the public key does not match the Local-part of the "i=" tag in the message signature header field, the verifier MUST ignore the key record and return PERMFAIL (inapplicable key). If the Local-part of the "i=" tag on the message signature is not present, the "g=" tag must be "*" (valid for all addresses in the domain) or the entire g= tag must be omitted (which defaults to "g=*"), otherwise the verifier MUST ignore the key record and return PERMFAIL (inapplicable key). Other than this test, verifiers SHOULD NOT treat a message signed with a key record having a "g=" tag any differently than one without; in particular, verifiers SHOULD NOT prefer messages that seem to have an individual signature by virtue of a "g=" tag versus a domain signature.
  7. If the "h=" tag exists in the public key record and the hash algorithm implied by the a= tag in the DKIM-Signature header field is not included in the contents of the "h=" tag, the verifier MUST ignore the key record and return PERMFAIL (inappropriate hash algorithm).
  8. If the public key data (the "p=" tag) is empty, then this key has been revoked and the verifier MUST treat this as a failed signature check and return PERMFAIL (key revoked). There is no defined semantic difference between a key that has been revoked and a key record that has been removed.
  9. If the public key data is not suitable for use with the algorithm and key types defined by the "a=" and "k=" tags in the DKIM-Signature header field, the verifier MUST immediately return PERMFAIL (inappropriate key algorithm).

6.1.3 Compute the Verification

Given a signer and a public key, verifying a signature consists of actions semantically equivalent to the following steps.

  1. Based on the algorithm defined in the "c=" tag, the body length specified in the "l=" tag, and the header field names in the "h=" tag, prepare a canonicalized version of the message as is described in Section 3.7 (note that this version does not actually need to be instantiated). When matching header field names in the "h=" tag against the actual message header field, comparisons MUST be case-insensitive.
  2. Based on the algorithm indicated in the "a=" tag, compute the message hashes from the canonical copy as described in Section 3.7.
  3. Verify that the hash of the canonicalized message body computed in the previous step matches the hash value conveyed in the "bh=" tag. If the hash does not match, the verifier SHOULD ignore the signature and return PERMFAIL (body hash did not verify).
  4. Using the signature conveyed in the "b=" tag, verify the signature against the header hash using the mechanism appropriate for the public key algorithm described in the "a=" tag. If the signature does not validate, the verifier SHOULD ignore the signature and return PERMFAIL (signature did not verify).
  5. Otherwise, the signature has correctly verified.

INFORMATIVE IMPLEMENTER'S NOTE: Implementations might wish to initiate the public-key query in parallel with calculating the hash as the public key is not needed until the final decryption is calculated. Implementations may also verify the signature on the message header before validating that the message hash listed in the "bh=" tag in the DKIM-Signature header field matches that of the actual message body; however, if the body hash does not match, the entire signature must be considered to have failed.

A body length specified in the "l=" tag of the signature limits the number of bytes of the body passed to the verification algorithm. All data beyond that limit is not validated by DKIM. Hence, verifiers might treat a message that contains bytes beyond the indicated body length with suspicion, such as by truncating the message at the indicated body length, declaring the signature invalid (e.g., by returning PERMFAIL (unsigned content)), or conveying the partial verification to the policy module.

INFORMATIVE IMPLEMENTATION NOTE: Verifiers that truncate the body at the indicated body length might pass on a malformed MIME message if the signer used the "N-4" trick (omitting the final "--CRLF") described in the informative note in (Section 3.4.5). Such verifiers may wish to check for this case and include a trailing "--CRLF" to avoid breaking the MIME structure. A simple way to achieve this might be to append "--CRLF" to any "multipart" message with a body length; if the MIME structure is already correctly formed, this will appear in the postlude and will not be displayed to the end user.

6.2 Communicate Verification Results

Verifiers wishing to communicate the results of verification to other parts of the mail system may do so in whatever manner they see fit. For example, implementations might choose to add an email header field to the message before passing it on. Any such header field SHOULD be inserted before any existing DKIM-Signature or preexisting authentication status header fields in the header field block.

INFORMATIVE ADVICE to MUA filter writers: Patterns intended to search for results header fields to visibly mark authenticated mail for end users should verify that such header field was added by the appropriate verifying domain and that the verified identity matches the author identity that will be displayed by the MUA. In particular, MUA filters should not be influenced by bogus results header fields added by attackers. To circumvent this attack, verifiers may wish to delete existing results header fields after verification and before adding a new header field.

6.3 Interpret Results/Apply Local Policy

It is beyond the scope of this specification to describe what actions a verifier system should make, but an authenticated email presents an opportunity to a receiving system that unauthenticated email cannot. Specifically, an authenticated email creates a predictable identifier by which other decisions can reliably be managed, such as trust and reputation. Conversely, unauthenticated email lacks a reliable identifier that can be used to assign trust and reputation. It is reasonable to treat unauthenticated email as lacking any trust and having no positive reputation.

In general, verifiers SHOULD NOT reject messages solely on the basis of a lack of signature or an unverifiable signature; such rejection would cause severe interoperability problems. However, if the verifier does opt to reje