idnits 2.17.1 draft-ietf-dkim-rfc4871bis-06.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** There are 2 instances of too long lines in the document, the longest one being 9 characters in excess of 72. -- The draft header indicates that this document obsoletes RFC4871, but the abstract doesn't seem to mention this, which it should. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 1320 has weird spacing: '... email elec...' -- The document date (April 12, 2011) is 4763 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'FWS' is mentioned on line 963, but not defined == Unused Reference: 'ITU-X660-1997' is defined on line 2877, but no explicit reference was found in the text -- Possible downref: Non-RFC (?) normative reference: ref. 'FIPS-180-2-2002' -- Possible downref: Non-RFC (?) normative reference: ref. 'ITU-X660-1997' ** Obsolete normative reference: RFC 3447 (Obsoleted by RFC 8017) ** Obsolete normative reference: RFC 3490 (Obsoleted by RFC 5890, RFC 5891) ** Obsolete normative reference: RFC 4234 (ref. 'RFC5234') (Obsoleted by RFC 5234) ** Downref: Normative reference to an Informational RFC: RFC 5598 -- Obsolete informational reference (is this intentional?): RFC 4409 (Obsoleted by RFC 6409) -- Obsolete informational reference (is this intentional?): RFC 4870 (Obsoleted by RFC 4871) -- Obsolete informational reference (is this intentional?): RFC 4871 (Obsoleted by RFC 6376) -- Obsolete informational reference (is this intentional?): RFC 5226 (Obsoleted by RFC 8126) -- Obsolete informational reference (is this intentional?): RFC 5451 (Obsoleted by RFC 7001) -- Obsolete informational reference (is this intentional?): RFC 5751 (Obsoleted by RFC 8551) Summary: 5 errors (**), 0 flaws (~~), 4 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group D. Crocker, Ed. 3 Internet-Draft Brandenburg InternetWorking 4 Obsoletes: 4871 (if approved) T. Hansen, Ed. 5 Intended status: Standards Track AT&T Laboratories 6 Expires: October 14, 2011 M. Kucherawy, Ed. 7 Cloudmark 8 April 12, 2011 10 DomainKeys Identified Mail (DKIM) Signatures 11 draft-ietf-dkim-rfc4871bis-06 13 Abstract 15 DomainKeys Identified Mail (DKIM) permits a person, role, or 16 organization that owns the signing domain to claim some 17 responsibility for a message by associating the domain with the 18 message. This can be an author's organization, an operational relay 19 or one of their agents. DKIM separates the question of the identity 20 of the signer of the message from the purported author of the 21 message. Assertion of responsibility is validated through a 22 cryptographic signature and querying the signer's domain directly to 23 retrieve the appropriate public key. Message transit from author to 24 recipient is through relays that typically make no substantive change 25 to the message content and thus preserve the DKIM signature. 27 Status of this Memo 29 This Internet-Draft is submitted in full conformance with the 30 provisions of BCP 78 and BCP 79. 32 Internet-Drafts are working documents of the Internet Engineering 33 Task Force (IETF). Note that other groups may also distribute 34 working documents as Internet-Drafts. The list of current Internet- 35 Drafts is at http://datatracker.ietf.org/drafts/current/. 37 Internet-Drafts are draft documents valid for a maximum of six months 38 and may be updated, replaced, or obsoleted by other documents at any 39 time. It is inappropriate to use Internet-Drafts as reference 40 material or to cite them other than as "work in progress." 42 This Internet-Draft will expire on October 14, 2011. 44 Copyright Notice 46 Copyright (c) 2011 IETF Trust and the persons identified as the 47 document authors. All rights reserved. 49 This document is subject to BCP 78 and the IETF Trust's Legal 50 Provisions Relating to IETF Documents 51 (http://trustee.ietf.org/license-info) in effect on the date of 52 publication of this document. Please review these documents 53 carefully, as they describe your rights and restrictions with respect 54 to this document. Code Components extracted from this document must 55 include Simplified BSD License text as described in Section 4.e of 56 the Trust Legal Provisions and are provided without warranty as 57 described in the Simplified BSD License. 59 Table of Contents 61 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 62 1.1. Signing Identity . . . . . . . . . . . . . . . . . . . . . 6 63 1.2. Scalability . . . . . . . . . . . . . . . . . . . . . . . 6 64 1.3. Simple Key Management . . . . . . . . . . . . . . . . . . 6 65 1.4. Data Integrity . . . . . . . . . . . . . . . . . . . . . . 6 66 2. Terminology and Definitions . . . . . . . . . . . . . . . . . 7 67 2.1. Signers . . . . . . . . . . . . . . . . . . . . . . . . . 7 68 2.2. Verifiers . . . . . . . . . . . . . . . . . . . . . . . . 7 69 2.3. Identity . . . . . . . . . . . . . . . . . . . . . . . . . 7 70 2.4. Identifier . . . . . . . . . . . . . . . . . . . . . . . . 7 71 2.5. Signing Domain Identifier (SDID) . . . . . . . . . . . . . 7 72 2.6. Agent or User Identifier (AUID) . . . . . . . . . . . . . 8 73 2.7. Identity Assessor . . . . . . . . . . . . . . . . . . . . 8 74 2.8. Whitespace . . . . . . . . . . . . . . . . . . . . . . . . 8 75 2.9. Imported ABNF Tokens . . . . . . . . . . . . . . . . . . . 9 76 2.10. Common ABNF Tokens . . . . . . . . . . . . . . . . . . . . 9 77 2.11. DKIM-Quoted-Printable . . . . . . . . . . . . . . . . . . 9 78 3. Protocol Elements . . . . . . . . . . . . . . . . . . . . . . 10 79 3.1. Selectors . . . . . . . . . . . . . . . . . . . . . . . . 11 80 3.2. Tag=Value Lists . . . . . . . . . . . . . . . . . . . . . 12 81 3.3. Signing and Verification Algorithms . . . . . . . . . . . 13 82 3.4. Canonicalization . . . . . . . . . . . . . . . . . . . . . 15 83 3.5. The DKIM-Signature Header Field . . . . . . . . . . . . . 19 84 3.6. Key Management and Representation . . . . . . . . . . . . 28 85 3.7. Computing the Message Hashes . . . . . . . . . . . . . . . 33 86 3.8. Input Requirements . . . . . . . . . . . . . . . . . . . . 35 87 3.9. Signing by Parent Domains . . . . . . . . . . . . . . . . 35 88 3.10. Relationship between SDID and AUID . . . . . . . . . . . . 36 89 4. Semantics of Multiple Signatures . . . . . . . . . . . . . . . 36 90 4.1. Example Scenarios . . . . . . . . . . . . . . . . . . . . 37 91 4.2. Interpretation . . . . . . . . . . . . . . . . . . . . . . 38 92 5. Signer Actions . . . . . . . . . . . . . . . . . . . . . . . . 39 93 5.1. Determine Whether the Email Should Be Signed and by 94 Whom . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 95 5.2. Select a Private Key and Corresponding Selector 96 Information . . . . . . . . . . . . . . . . . . . . . . . 40 97 5.3. Normalize the Message to Prevent Transport Conversions . . 40 98 5.4. Determine the Header Fields to Sign . . . . . . . . . . . 41 99 5.5. Recommended Signature Content . . . . . . . . . . . . . . 43 100 5.6. Compute the Message Hash and Signature . . . . . . . . . . 44 101 5.7. Insert the DKIM-Signature Header Field . . . . . . . . . . 45 102 6. Verifier Actions . . . . . . . . . . . . . . . . . . . . . . . 45 103 6.1. Extract Signatures from the Message . . . . . . . . . . . 46 104 6.2. Communicate Verification Results . . . . . . . . . . . . . 51 105 6.3. Interpret Results/Apply Local Policy . . . . . . . . . . . 52 106 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 53 107 7.1. DKIM-Signature Tag Specifications . . . . . . . . . . . . 53 108 7.2. DKIM-Signature Query Method Registry . . . . . . . . . . . 54 109 7.3. DKIM-Signature Canonicalization Registry . . . . . . . . . 54 110 7.4. _domainkey DNS TXT Record Tag Specifications . . . . . . . 55 111 7.5. DKIM Key Type Registry . . . . . . . . . . . . . . . . . . 56 112 7.6. DKIM Hash Algorithms Registry . . . . . . . . . . . . . . 56 113 7.7. DKIM Service Types Registry . . . . . . . . . . . . . . . 56 114 7.8. DKIM Selector Flags Registry . . . . . . . . . . . . . . . 57 115 7.9. DKIM-Signature Header Field . . . . . . . . . . . . . . . 57 116 8. Security Considerations . . . . . . . . . . . . . . . . . . . 57 117 8.1. Misuse of Body Length Limits ("l=" Tag) . . . . . . . . . 57 118 8.2. Misappropriated Private Key . . . . . . . . . . . . . . . 58 119 8.3. Key Server Denial-of-Service Attacks . . . . . . . . . . . 59 120 8.4. Attacks Against the DNS . . . . . . . . . . . . . . . . . 59 121 8.5. Replay Attacks . . . . . . . . . . . . . . . . . . . . . . 60 122 8.6. Limits on Revoking Keys . . . . . . . . . . . . . . . . . 61 123 8.7. Intentionally Malformed Key Records . . . . . . . . . . . 61 124 8.8. Intentionally Malformed DKIM-Signature Header Fields . . . 61 125 8.9. Information Leakage . . . . . . . . . . . . . . . . . . . 61 126 8.10. Remote Timing Attacks . . . . . . . . . . . . . . . . . . 61 127 8.11. Reordered Header Fields . . . . . . . . . . . . . . . . . 61 128 8.12. RSA Attacks . . . . . . . . . . . . . . . . . . . . . . . 62 129 8.13. Inappropriate Signing by Parent Domains . . . . . . . . . 62 130 8.14. Attacks Involving Addition of Header Fields . . . . . . . 62 131 8.15. Malformed Inputs . . . . . . . . . . . . . . . . . . . . . 63 132 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 64 133 9.1. Normative References . . . . . . . . . . . . . . . . . . . 64 134 9.2. Informative References . . . . . . . . . . . . . . . . . . 65 135 Appendix A. Example of Use (INFORMATIVE) . . . . . . . . . . . . 66 136 A.1. The User Composes an Email . . . . . . . . . . . . . . . . 66 137 A.2. The Email is Signed . . . . . . . . . . . . . . . . . . . 67 138 A.3. The Email Signature is Verified . . . . . . . . . . . . . 68 139 Appendix B. Usage Examples (INFORMATIVE) . . . . . . . . . . . . 69 140 B.1. Alternate Submission Scenarios . . . . . . . . . . . . . . 69 141 B.2. Alternate Delivery Scenarios . . . . . . . . . . . . . . . 71 142 Appendix C. Creating a Public Key (INFORMATIVE) . . . . . . . . . 73 143 C.1. Compatibility with DomainKeys Key Records . . . . . . . . 74 145 Appendix D. MUA Considerations . . . . . . . . . . . . . . . . . 74 146 Appendix E. Acknowledgements . . . . . . . . . . . . . . . . . . 75 147 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 76 149 1. Introduction 151 DomainKeys Identified Mail (DKIM) permits a person, role, or 152 organization to claim some responsibility for a message by 153 associating a domain name [RFC1034] with the message [RFC5322], which 154 they are authorized to use. This can be an author's organization, an 155 operational relay or one of their agents. Assertion of 156 responsibility is validated through a cryptographic signature and 157 querying the signer's domain directly to retrieve the appropriate 158 public key. Message transit from author to recipient is through 159 relays that typically make no substantive change to the message 160 content and thus preserve the DKIM signature. A message can contain 161 multiple signatures, from the same or different organizations 162 involved with the message. 164 The approach taken by DKIM differs from previous approaches to 165 message signing (e.g., Secure/Multipurpose Internet Mail Extensions 166 (S/MIME) [RFC1847], OpenPGP [RFC4880]) in that: 168 o the message signature is written as a message header field so that 169 neither human recipients nor existing MUA (Mail User Agent) 170 software is confused by signature-related content appearing in the 171 message body; 173 o there is no dependency on public and private key pairs being 174 issued by well-known, trusted certificate authorities; 176 o there is no dependency on the deployment of any new Internet 177 protocols or services for public key distribution or revocation; 179 o signature verification failure does not force rejection of the 180 message; 182 o no attempt is made to include encryption as part of the mechanism; 184 o message archiving is not a design goal. 186 DKIM: 188 o is compatible with the existing email infrastructure and 189 transparent to the fullest extent possible; 191 o requires minimal new infrastructure; 193 o can be implemented independently of clients in order to reduce 194 deployment time; 196 o can be deployed incrementally; 198 o allows delegation of signing to third parties. 200 1.1. Signing Identity 202 DKIM separates the question of the identity of the signer of the 203 message from the purported author of the message. In particular, a 204 signature includes the identity of the signer. Verifiers can use the 205 signing information to decide how they want to process the message. 206 The signing identity is included as part of the signature header 207 field. 209 INFORMATIVE RATIONALE: The signing identity specified by a DKIM 210 signature is not required to match an address in any particular 211 header field because of the broad methods of interpretation by 212 recipient mail systems, including MUAs. 214 1.2. Scalability 216 DKIM is designed to support the extreme scalability requirements that 217 characterize the email identification problem. There are currently 218 over 70 million domains and a much larger number of individual 219 addresses. DKIM seeks to preserve the positive aspects of the 220 current email infrastructure, such as the ability for anyone to 221 communicate with anyone else without introduction. 223 1.3. Simple Key Management 225 DKIM differs from traditional hierarchical public-key systems in that 226 no Certificate Authority infrastructure is required; the verifier 227 requests the public key from a repository in the domain of the 228 claimed signer directly rather than from a third party. 230 The DNS is proposed as the initial mechanism for the public keys. 231 Thus, DKIM currently depends on DNS administration and the security 232 of the DNS system. DKIM is designed to be extensible to other key 233 fetching services as they become available. 235 1.4. Data Integrity 237 A DKIM signature associates the d= name with the computed hash of 238 some or all of the message (see Section 3.7) in order to prevent the 239 re-use of the signature with different messages. Verifying the 240 signature asserts that the hashed content has not changed since it 241 was signed, and asserts nothing else about "protecting" the end-to- 242 end integrity of the message. 244 2. Terminology and Definitions 246 This section defines terms used in the rest of the document. 248 DKIM is designed to operate within the Internet Mail service, as 249 defined in [RFC5598]. Basic email terminology is taken from that 250 specification. 252 Syntax descriptions use Augmented BNF (ABNF) [RFC5234]. 254 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 255 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 256 document are to be interpreted as described in [RFC2119]. 258 2.1. Signers 260 Elements in the mail system that sign messages on behalf of a domain 261 are referred to as signers. These may be MUAs (Mail User Agents), 262 MSAs (Mail Submission Agents), MTAs (Mail Transfer Agents), or other 263 agents such as mailing list exploders. In general, any signer will 264 be involved in the injection of a message into the message system in 265 some way. The key issue is that a message must be signed before it 266 leaves the administrative domain of the signer. 268 2.2. Verifiers 270 Elements in the mail system that verify signatures are referred to as 271 verifiers. These may be MTAs, Mail Delivery Agents (MDAs), or MUAs. 272 In most cases it is expected that verifiers will be close to an end 273 user (reader) of the message or some consuming agent such as a 274 mailing list exploder. 276 2.3. Identity 278 A person, role, or organization. In the context of DKIM, examples 279 include the author, the author's organization, an ISP along the 280 handling path, an independent trust assessment service, and a mailing 281 list operator. 283 2.4. Identifier 285 A label that refers to an identity. 287 2.5. Signing Domain Identifier (SDID) 289 A single domain name that is the mandatory payload output of DKIM and 290 that refers to the identity claiming responsibility for introduction 291 of a message into the mail stream. For DKIM processing, the name has 292 only basic domain name semantics; any possible owner-specific 293 semantics are outside the scope of DKIM. It is specified in 294 Section 3.5. 296 2.6. Agent or User Identifier (AUID) 298 A single identifier that refers to the agent or user on behalf of 299 whom the Signing Domain Identifier (SDID) has taken responsibility. 300 The AUID comprises a domain name and an optional . The 301 domain name is the same as that used for the SDID or is a sub-domain 302 of it. For DKIM processing, the domain name portion of the AUID has 303 only basic domain name semantics; any possible owner-specific 304 semantics are outside the scope of DKIM. It is specified in 305 Section 3.5 . 307 2.7. Identity Assessor 309 A module that consumes DKIM's mandatory payload, which is the 310 responsible Signing Domain Identifier (SDID). The module is 311 dedicated to the assessment of the delivered identifier. Other DKIM 312 (and non-DKIM) values can also be delivered to this module as well as 313 to a more general message evaluation filtering engine. However, this 314 additional activity is outside the scope of the DKIM signature 315 specification. 317 2.8. Whitespace 319 There are three forms of whitespace: 321 o WSP represents simple whitespace, i.e., a space or a tab character 322 (formal definition in [RFC5234]). 324 o LWSP is linear whitespace, defined as WSP plus CRLF (formal 325 definition in [RFC5234]). 327 o FWS is folding whitespace. It allows multiple lines separated by 328 CRLF followed by at least one whitespace, to be joined. 330 The formal ABNF for these are (WSP and LWSP are given for information 331 only): 332 WSP = SP / HTAB 333 LWSP = *(WSP / CRLF WSP) 334 FWS = [*WSP CRLF] 1*WSP 336 The definition of FWS is identical to that in [RFC5322] except for 337 the exclusion of obs-FWS. 339 2.9. Imported ABNF Tokens 341 The following tokens are imported from other RFCs as noted. Those 342 RFCs should be considered definitive. 344 The following tokens are imported from [RFC5321]: 346 o "Local-part" (implementation warning: this permits quoted strings) 348 o "sub-domain" 350 The following tokens are imported from [RFC5322]: 352 o "field-name" (name of a header field) 354 o "dot-atom-text" (in the Local-part of an email address) 356 The following tokens are imported from [RFC2045]: 358 o "qp-section" (a single line of quoted-printable-encoded text) 360 o "hex-octet" (a quoted-printable encoded octet) 362 INFORMATIVE NOTE: Be aware that the ABNF in [RFC2045] does not 363 obey the rules of [RFC5234] and must be interpreted accordingly, 364 particularly as regards case folding. 366 Other tokens not defined herein are imported from [RFC5234]. These 367 are intuitive primitives such as SP, HTAB, WSP, ALPHA, DIGIT, CRLF, 368 etc. 370 2.10. Common ABNF Tokens 372 The following ABNF tokens are used elsewhere in this document: 373 hyphenated-word = ALPHA [ *(ALPHA / DIGIT / "-") (ALPHA / DIGIT) ] 374 ALPHADIGITPS = (ALPHA / DIGIT / "+" / "/") 375 base64string = ALPHADIGITPS *([FWS] ALPHADIGITPS) 376 [ [FWS] "=" [ [FWS] "=" ] ] 377 hdr-name = field-name 378 qp-hdr-value = dkim-quoted-printable ; with "|" encoded 380 2.11. DKIM-Quoted-Printable 382 The DKIM-Quoted-Printable encoding syntax resembles that described in 383 Quoted-Printable [RFC2045], Section 6.7: any character MAY be encoded 384 as an "=" followed by two hexadecimal digits from the alphabet 385 "0123456789ABCDEF" (no lowercase characters permitted) representing 386 the hexadecimal-encoded integer value of that character. All control 387 characters (those with values < %x20), 8-bit characters (values > 388 %x7F), and the characters DEL (%x7F), SPACE (%x20), and semicolon 389 (";", %x3B) MUST be encoded. Note that all whitespace, including 390 SPACE, CR, and LF characters, MUST be encoded. After encoding, FWS 391 MAY be added at arbitrary locations in order to avoid excessively 392 long lines; such whitespace is NOT part of the value, and MUST be 393 removed before decoding. 395 ABNF: 397 dkim-quoted-printable = *(FWS / hex-octet / dkim-safe-char) 398 ; hex-octet is from RFC2045 399 dkim-safe-char = %x21-3A / %x3C / %x3E-7E 400 ; '!' - ':', '<', '>' - '~' 401 ; Characters not listed as "mail-safe" in 402 ; [RFC2049] are also not recommended. 404 INFORMATIVE NOTE: DKIM-Quoted-Printable differs from Quoted- 405 Printable as defined in [RFC2045] in several important ways: 407 1. Whitespace in the input text, including CR and LF, must be 408 encoded. [RFC2045] does not require such encoding, and does 409 not permit encoding of CR or LF characters that are part of a 410 CRLF line break. 412 2. Whitespace in the encoded text is ignored. This is to allow 413 tags encoded using DKIM-Quoted-Printable to be wrapped as 414 needed. In particular, [RFC2045] requires that line breaks in 415 the input be represented as physical line breaks; that is not 416 the case here. 418 3. The "soft line break" syntax ("=" as the last non-whitespace 419 character on the line) does not apply. 421 4. DKIM-Quoted-Printable does not require that encoded lines be 422 no more than 76 characters long (although there may be other 423 requirements depending on the context in which the encoded 424 text is being used). 426 3. Protocol Elements 428 Protocol Elements are conceptual parts of the protocol that are not 429 specific to either signers or verifiers. The protocol descriptions 430 for signers and verifiers are described in later sections (Signer 431 Actions (Section 5) and Verifier Actions (Section 6)). NOTE: This 432 section must be read in the context of those sections. 434 3.1. Selectors 436 To support multiple concurrent public keys per signing domain, the 437 key namespace is subdivided using "selectors". For example, 438 selectors might indicate the names of office locations (e.g., 439 "sanfrancisco", "coolumbeach", and "reykjavik"), the signing date 440 (e.g., "january2005", "february2005", etc.), or even an individual 441 user. 443 Selectors are needed to support some important use cases. For 444 example: 446 o Domains that want to delegate signing capability for a specific 447 address for a given duration to a partner, such as an advertising 448 provider or other outsourced function. 450 o Domains that want to allow frequent travelers to send messages 451 locally without the need to connect with a particular MSA. 453 o "Affinity" domains (e.g., college alumni associations) that 454 provide forwarding of incoming mail, but that do not operate a 455 mail submission agent for outgoing mail. 457 Periods are allowed in selectors and are component separators. When 458 keys are retrieved from the DNS, periods in selectors define DNS 459 label boundaries in a manner similar to the conventional use in 460 domain names. Selector components might be used to combine dates 461 with locations, for example, "march2005.reykjavik". In a DNS 462 implementation, this can be used to allow delegation of a portion of 463 the selector namespace. 465 ABNF: 466 selector = sub-domain *( "." sub-domain ) 468 The number of public keys and corresponding selectors for each domain 469 is determined by the domain owner. Many domain owners will be 470 satisfied with just one selector, whereas administratively 471 distributed organizations may choose to manage disparate selectors 472 and key pairs in different regions or on different email servers. 474 Beyond administrative convenience, selectors make it possible to 475 seamlessly replace public keys on a routine basis. If a domain 476 wishes to change from using a public key associated with selector 477 "january2005" to a public key associated with selector 478 "february2005", it merely makes sure that both public keys are 479 advertised in the public-key repository concurrently for the 480 transition period during which email may be in transit prior to 481 verification. At the start of the transition period, the outbound 482 email servers are configured to sign with the "february2005" private 483 key. At the end of the transition period, the "january2005" public 484 key is removed from the public-key repository. 486 INFORMATIVE NOTE: A key may also be revoked as described below. 487 The distinction between revoking and removing a key selector 488 record is subtle. When phasing out keys as described above, a 489 signing domain would probably simply remove the key record after 490 the transition period. However, a signing domain could elect to 491 revoke the key (but maintain the key record) for a further period. 492 There is no defined semantic difference between a revoked key and 493 a removed key. 495 While some domains may wish to make selector values well known, 496 others will want to take care not to allocate selector names in a way 497 that allows harvesting of data by outside parties. For example, if 498 per-user keys are issued, the domain owner will need to make the 499 decision as to whether to associate this selector directly with the 500 name of a registered end-user, or make it some unassociated random 501 value, such as a fingerprint of the public key. 503 INFORMATIVE OPERATIONS NOTE: Reusing a selector with a new key 504 (for example, changing the key associated with a user's name) 505 makes it impossible to tell the difference between a message that 506 didn't verify because the key is no longer valid versus a message 507 that is actually forged. For this reason, signers are ill-advised 508 to reuse selectors for new keys. A better strategy is to assign 509 new keys to new selectors. 511 3.2. Tag=Value Lists 513 DKIM uses a simple "tag=value" syntax in several contexts, including 514 in messages and domain signature records. 516 Values are a series of strings containing either plain text, "base64" 517 text (as defined in [RFC2045], Section 6.8), "qp-section" (ibid, 518 Section 6.7), or "dkim-quoted-printable" (as defined in 519 Section 2.11). The name of the tag will determine the encoding of 520 each value. Unencoded semicolon (";") characters MUST NOT occur in 521 the tag value, since that separates tag-specs. 523 INFORMATIVE IMPLEMENTATION NOTE: Although the "plain text" defined 524 below (as "tag-value") only includes 7-bit characters, an 525 implementation that wished to anticipate future standards would be 526 advised not to preclude the use of UTF8-encoded text in tag=value 527 lists. 529 Formally, the ABNF syntax rules are as follows: 530 tag-list = tag-spec 0*( ";" tag-spec ) [ ";" ] 531 tag-spec = [FWS] tag-name [FWS] "=" [FWS] tag-value [FWS] 532 tag-name = ALPHA 0*ALNUMPUNC 533 tag-value = [ tval 0*( 1*(WSP / FWS) tval ) ] 534 ; WSP and FWS prohibited at beginning and end 535 tval = 1*VALCHAR 536 VALCHAR = %x21-3A / %x3C-7E 537 ; EXCLAMATION to TILDE except SEMICOLON 538 ALNUMPUNC = ALPHA / DIGIT / "_" 540 Note that WSP is allowed anywhere around tags. In particular, any 541 WSP after the "=" and any WSP before the terminating ";" is not part 542 of the value; however, WSP inside the value is significant. 544 Tags MUST be interpreted in a case-sensitive manner. Values MUST be 545 processed as case sensitive unless the specific tag description of 546 semantics specifies case insensitivity. 548 Tags with duplicate names MUST NOT occur within a single tag-list; if 549 a tag name does occur more than once, the entire tag-list is invalid. 551 Whitespace within a value MUST be retained unless explicitly excluded 552 by the specific tag description. 554 Tag=value pairs that represent the default value MAY be included to 555 aid legibility. 557 Unrecognized tags MUST be ignored. 559 Tags that have an empty value are not the same as omitted tags. An 560 omitted tag is treated as having the default value; a tag with an 561 empty value explicitly designates the empty string as the value. 563 3.3. Signing and Verification Algorithms 565 DKIM supports multiple digital signature algorithms. Two algorithms 566 are defined by this specification at this time: rsa-sha1 and rsa- 567 sha256. Signers MUST implement and SHOULD sign using rsa-sha256. 568 Verifiers MUST implement rsa-sha256. 570 INFORMATIVE NOTE: Although sha256 is strongly encouraged, some 571 senders of low-security messages (such as routine newsletters) may 572 prefer to use sha1 because of reduced CPU requirements to compute 573 a sha1 hash. In general, sha256 should always be used whenever 574 possible. 576 3.3.1. The rsa-sha1 Signing Algorithm 578 The rsa-sha1 Signing Algorithm computes a message hash as described 579 in Section 3.7 below using SHA-1 [FIPS-180-2-2002] as the hash-alg. 580 That hash is then signed by the signer using the RSA algorithm 581 (defined in PKCS#1 version 1.5 [RFC3447]) as the crypt-alg and the 582 signer's private key. The hash MUST NOT be truncated or converted 583 into any form other than the native binary form before being signed. 584 The signing algorithm SHOULD use a public exponent of 65537. 586 3.3.2. The rsa-sha256 Signing Algorithm 588 The rsa-sha256 Signing Algorithm computes a message hash as described 589 in Section 3.7 below using SHA-256 [FIPS-180-2-2002] as the hash-alg. 590 That hash is then signed by the signer using the RSA algorithm 591 (defined in PKCS#1 version 1.5 [RFC3447]) as the crypt-alg and the 592 signer's private key. The hash MUST NOT be truncated or converted 593 into any form other than the native binary form before being signed. 595 3.3.3. Key Sizes 597 Selecting appropriate key sizes is a trade-off between cost, 598 performance, and risk. Since short RSA keys more easily succumb to 599 off-line attacks, signers MUST use RSA keys of at least 1024 bits for 600 long-lived keys. Verifiers MUST be able to validate signatures with 601 keys ranging from 512 bits to 2048 bits, and they MAY be able to 602 validate signatures with larger keys. Verifier policies may use the 603 length of the signing key as one metric for determining whether a 604 signature is acceptable. 606 Factors that should influence the key size choice include the 607 following: 609 o The practical constraint that large (e.g., 4096 bit) keys may not 610 fit within a 512-byte DNS UDP response packet 612 o The security constraint that keys smaller than 1024 bits are 613 subject to off-line attacks 615 o Larger keys impose higher CPU costs to verify and sign email 617 o Keys can be replaced on a regular basis, thus their lifetime can 618 be relatively short 620 o The security goals of this specification are modest compared to 621 typical goals of other systems that employ digital signatures 623 See [RFC3766] for further discussion on selecting key sizes. 625 3.3.4. Other Algorithms 627 Other algorithms MAY be defined in the future. Verifiers MUST ignore 628 any signatures using algorithms that they do not implement. 630 3.4. Canonicalization 632 Some mail systems modify email in transit, potentially invalidating a 633 signature. For most signers, mild modification of email is 634 immaterial to validation of the DKIM domain name's use. For such 635 signers, a canonicalization algorithm that survives modest in-transit 636 modification is preferred. 638 Other signers demand that any modification of the email, however 639 minor, result in a signature verification failure. These signers 640 prefer a canonicalization algorithm that does not tolerate in-transit 641 modification of the signed email. 643 Some signers may be willing to accept modifications to header fields 644 that are within the bounds of email standards such as [RFC5322], but 645 are unwilling to accept any modification to the body of messages. 647 To satisfy all requirements, two canonicalization algorithms are 648 defined for each of the header and the body: a "simple" algorithm 649 that tolerates almost no modification and a "relaxed" algorithm that 650 tolerates common modifications such as whitespace replacement and 651 header field line rewrapping. A signer MAY specify either algorithm 652 for header or body when signing an email. If no canonicalization 653 algorithm is specified by the signer, the "simple" algorithm defaults 654 for both header and body. Verifiers MUST implement both 655 canonicalization algorithms. Note that the header and body may use 656 different canonicalization algorithms. Further canonicalization 657 algorithms MAY be defined in the future; verifiers MUST ignore any 658 signatures that use unrecognized canonicalization algorithms. 660 Canonicalization simply prepares the email for presentation to the 661 signing or verification algorithm. It MUST NOT change the 662 transmitted data in any way. Canonicalization of header fields and 663 body are described below. 665 NOTE: This section assumes that the message is already in "network 666 normal" format (text is ASCII encoded, lines are separated with CRLF 667 characters, etc.). See also Section 5.3 for information about 668 normalizing the message. 670 3.4.1. The "simple" Header Canonicalization Algorithm 672 The "simple" header canonicalization algorithm does not change header 673 fields in any way. Header fields MUST be presented to the signing or 674 verification algorithm exactly as they are in the message being 675 signed or verified. In particular, header field names MUST NOT be 676 case folded and whitespace MUST NOT be changed. 678 3.4.2. The "relaxed" Header Canonicalization Algorithm 680 The "relaxed" header canonicalization algorithm MUST apply the 681 following steps in order: 683 o Convert all header field names (not the header field values) to 684 lowercase. For example, convert "SUBJect: AbC" to "subject: AbC". 686 o Unfold all header field continuation lines as described in 687 [RFC5322]; in particular, lines with terminators embedded in 688 continued header field values (that is, CRLF sequences followed by 689 WSP) MUST be interpreted without the CRLF. Implementations MUST 690 NOT remove the CRLF at the end of the header field value. 692 o Convert all sequences of one or more WSP characters to a single SP 693 character. WSP characters here include those before and after a 694 line folding boundary. 696 o Delete all WSP characters at the end of each unfolded header field 697 value. 699 o Delete any WSP characters remaining before and after the colon 700 separating the header field name from the header field value. The 701 colon separator MUST be retained. 703 3.4.3. The "simple" Body Canonicalization Algorithm 705 The "simple" body canonicalization algorithm ignores all empty lines 706 at the end of the message body. An empty line is a line of zero 707 length after removal of the line terminator. If there is no body or 708 no trailing CRLF on the message body, a CRLF is added. It makes no 709 other changes to the message body. In more formal terms, the 710 "simple" body canonicalization algorithm converts "0*CRLF" at the end 711 of the body to a single "CRLF". 713 Note that a completely empty or missing body is canonicalized as a 714 single "CRLF"; that is, the canonicalized length will be 2 octets. 716 The sha1 value (in base64) for an empty body (canonicalized to a 717 "CRLF") is: 718 uoq1oCgLlTqpdDX/iUbLy7J1Wic= 719 The sha256 value is: 720 frcCV1k9oG9oKj3dpUqdJg1PxRT2RSN/XKdLCPjaYaY= 722 3.4.4. The "relaxed" Body Canonicalization Algorithm 724 The "relaxed" body canonicalization algorithm MUST apply the 725 following steps (a) and (b) in order: 727 a. Reduce whitespace: 729 * Ignore all whitespace at the end of lines. Implementations 730 MUST NOT remove the CRLF at the end of the line. 732 * Reduce all sequences of WSP within a line to a single SP 733 character. 735 b. Ignore all empty lines at the end of the message body. "Empty 736 line" is defined in Section 3.4.3. If the body is non-empty, but 737 does not end with a CRLF, a CRLF is added. (For email, this is 738 only possible when using extensions to SMTP or non-SMTP transport 739 mechanisms.) 741 The sha1 value (in base64) for an empty body (canonicalized to a null 742 input) is: 743 2jmj7l5rSw0yVb/vlWAYkK/YBwk= 744 The sha256 value is: 745 47DEQpj8HBSa+/TImW+5JCeuQeRkm5NMpJWZG3hSuFU= 747 INFORMATIVE NOTE: It should be noted that the relaxed body 748 canonicalization algorithm may enable certain types of extremely 749 crude "ASCII Art" attacks where a message may be conveyed by 750 adjusting the spacing between words. If this is a concern, the 751 "simple" body canonicalization algorithm should be used instead. 753 3.4.5. Body Length Limits 755 A body length count MAY be specified to limit the signature 756 calculation to an initial prefix of the body text, measured in 757 octets. If the body length count is not specified, the entire 758 message body is signed. 760 INFORMATIVE RATIONALE: This capability is provided because it is 761 very common for mailing lists to add trailers to messages (e.g., 762 instructions how to get off the list). Until those messages are 763 also signed, the body length count is a useful tool for the 764 verifier since it may as a matter of policy accept messages having 765 valid signatures with extraneous data. 767 INFORMATIVE IMPLEMENTATION NOTE: Using body length limits enables 768 an attack in which an attacker modifies a message to include 769 content that solely benefits the attacker. It is possible for the 770 appended content to completely replace the original content in the 771 end recipient's eyes and to defeat duplicate message detection 772 algorithms. To avoid this attack, signers should be wary of using 773 this tag, and verifiers might wish to ignore the tag or remove 774 text that appears after the specified content length, perhaps 775 based on other criteria. 777 The body length count allows the signer of a message to permit data 778 to be appended to the end of the body of a signed message. The body 779 length count MUST be calculated following the canonicalization 780 algorithm; for example, any whitespace ignored by a canonicalization 781 algorithm is not included as part of the body length count. Signers 782 of MIME messages that include a body length count SHOULD be sure that 783 the length extends to the closing MIME boundary string. 785 INFORMATIVE IMPLEMENTATION NOTE: A signer wishing to ensure that 786 the only acceptable modifications are to add to the MIME postlude 787 would use a body length count encompassing the entire final MIME 788 boundary string, including the final "--CRLF". A signer wishing 789 to allow additional MIME parts but not modification of existing 790 parts would use a body length count extending through the final 791 MIME boundary string, omitting the final "--CRLF". Note that this 792 only works for some MIME types, e.g., multipart/mixed but not 793 multipart/signed. 795 A body length count of zero means that the body is completely 796 unsigned. 798 Signers wishing to ensure that no modification of any sort can occur 799 should specify the "simple" canonicalization algorithm for both 800 header and body and omit the body length count. 802 3.4.6. Canonicalization Examples (INFORMATIVE) 804 In the following examples, actual whitespace is used only for 805 clarity. The actual input and output text is designated using 806 bracketed descriptors: "" for a space character, "" for a 807 tab character, and "" for a carriage-return/line-feed sequence. 808 For example, "X Y" and "XY" represent the same three 809 characters. 811 Example 1: A message reading: 812 A: X 813 B : Y 814 Z 815 816 C 817 D E 818 819 821 when canonicalized using relaxed canonicalization for both header and 822 body results in a header reading: 823 a:X 824 b:Y Z 826 and a body reading: 827 C 828 D E 830 Example 2: The same message canonicalized using simple 831 canonicalization for both header and body results in a header 832 reading: 833 A: X 834 B : Y 835 Z 837 and a body reading: 838 C 839 D E 841 Example 3: When processed using relaxed header canonicalization and 842 simple body canonicalization, the canonicalized version has a header 843 of: 844 a:X 845 b:Y Z 847 and a body reading: 848 C 849 D E 851 3.5. The DKIM-Signature Header Field 853 The signature of the email is stored in the DKIM-Signature header 854 field. This header field contains all of the signature and key- 855 fetching data. The DKIM-Signature value is a tag-list as described 856 in Section 3.2. 858 The DKIM-Signature header field SHOULD be treated as though it were a 859 trace header field as defined in Section 3.6 of [RFC5322], and hence 860 SHOULD NOT be reordered and SHOULD be prepended to the message. 862 The DKIM-Signature header field being created or verified is always 863 included in the signature calculation, after the rest of the header 864 fields being signed; however, when calculating or verifying the 865 signature, the value of the "b=" tag (signature value) of that DKIM- 866 Signature header field MUST be treated as though it were an empty 867 string. Unknown tags in the DKIM-Signature header field MUST be 868 included in the signature calculation but MUST be otherwise ignored 869 by verifiers. Other DKIM-Signature header fields that are included 870 in the signature should be treated as normal header fields; in 871 particular, the "b=" tag is not treated specially. 873 The encodings for each field type are listed below. Tags described 874 as qp-section are encoded as described in Section 6.7 of MIME Part 875 One [RFC2045], with the additional conversion of semicolon characters 876 to "=3B"; intuitively, this is one line of quoted-printable encoded 877 text. The dkim-quoted-printable syntax is defined in Section 2.11. 879 Tags on the DKIM-Signature header field along with their type and 880 requirement status are shown below. Unrecognized tags MUST be 881 ignored. 883 v= Version (MUST be included). This tag defines the version of this 884 specification that applies to the signature record. It MUST have 885 the value "1". Note that verifiers must do a string comparison on 886 this value; for example, "1" is not the same as "1.0". 888 ABNF: 889 sig-v-tag = %x76 [FWS] "=" [FWS] "1" 891 INFORMATIVE NOTE: DKIM-Signature version numbers are expected 892 to increase arithmetically as new versions of this 893 specification are released. 895 a= The algorithm used to generate the signature (plain-text; 896 REQUIRED). Verifiers MUST support "rsa-sha1" and "rsa-sha256"; 897 signers SHOULD sign using "rsa-sha256". 899 ABNF: 901 sig-a-tag = %x61 [FWS] "=" [FWS] sig-a-tag-alg 902 sig-a-tag-alg = sig-a-tag-k "-" sig-a-tag-h 903 sig-a-tag-k = "rsa" / x-sig-a-tag-k 904 sig-a-tag-h = "sha1" / "sha256" / x-sig-a-tag-h 905 x-sig-a-tag-k = ALPHA *(ALPHA / DIGIT) 906 ; for later extension 907 x-sig-a-tag-h = ALPHA *(ALPHA / DIGIT) 908 ; for later extension 910 b= The signature data (base64; REQUIRED). Whitespace is ignored in 911 this value and MUST be ignored when reassembling the original 912 signature. In particular, the signing process can safely insert 913 FWS in this value in arbitrary places to conform to line-length 914 limits. See Signer Actions (Section 5) for how the signature is 915 computed. 917 ABNF: 918 sig-b-tag = %x62 [FWS] "=" [FWS] sig-b-tag-data 919 sig-b-tag-data = base64string 921 bh= The hash of the canonicalized body part of the message as 922 limited by the "l=" tag (base64; REQUIRED). Whitespace is ignored 923 in this value and MUST be ignored when reassembling the original 924 signature. In particular, the signing process can safely insert 925 FWS in this value in arbitrary places to conform to line-length 926 limits. See Section 3.7 for how the body hash is computed. 928 ABNF: 929 sig-bh-tag = %x62 %x68 [FWS] "=" [FWS] sig-bh-tag-data 930 sig-bh-tag-data = base64string 932 c= Message canonicalization (plain-text; OPTIONAL, default is 933 "simple/simple"). This tag informs the verifier of the type of 934 canonicalization used to prepare the message for signing. It 935 consists of two names separated by a "slash" (%d47) character, 936 corresponding to the header and body canonicalization algorithms 937 respectively. These algorithms are described in Section 3.4. If 938 only one algorithm is named, that algorithm is used for the header 939 and "simple" is used for the body. For example, "c=relaxed" is 940 treated the same as "c=relaxed/simple". 942 ABNF: 943 sig-c-tag = %x63 [FWS] "=" [FWS] sig-c-tag-alg 944 ["/" sig-c-tag-alg] 945 sig-c-tag-alg = "simple" / "relaxed" / x-sig-c-tag-alg 946 x-sig-c-tag-alg = hyphenated-word ; for later extension 948 d= The SDID claiming responsibility for an introduction of a message 949 into the mail stream (plain-text; REQUIRED). Hence, the SDID 950 value is used to form the query for the public key. The SDID MUST 951 correspond to a valid DNS name under which the DKIM key record is 952 published. The conventions and semantics used by a signer to 953 create and use a specific SDID are outside the scope of the DKIM 954 Signing specification, as is any use of those conventions and 955 semantics. When presented with a signature that does not meet 956 these requirements, verifiers MUST consider the signature invalid. 958 Internationalized domain names MUST be encoded as described in 959 [RFC3490]. 961 ABNF: 963 sig-d-tag = %x64 [FWS] "=" [FWS] domain-name 964 domain-name = sub-domain 1*("." sub-domain) 965 ; from RFC 5321 Domain, but excluding address-literal 967 h= Signed header fields (plain-text, but see description; REQUIRED). 968 A colon-separated list of header field names that identify the 969 header fields presented to the signing algorithm. The field MUST 970 contain the complete list of header fields in the order presented 971 to the signing algorithm. The field MAY contain names of header 972 fields that do not exist when signed; nonexistent header fields do 973 not contribute to the signature computation (that is, they are 974 treated as the null input, including the header field name, the 975 separating colon, the header field value, and any CRLF 976 terminator). The field MUST NOT include the DKIM-Signature header 977 field that is being created or verified, but may include others. 978 Folding whitespace (FWS) MAY be included on either side of the 979 colon separator. Header field names MUST be compared against 980 actual header field names in a case-insensitive manner. This list 981 MUST NOT be empty. See Section 5.4 for a discussion of choosing 982 header fields to sign. 984 ABNF: 985 sig-h-tag = %x68 [FWS] "=" [FWS] hdr-name 986 0*( [FWS] ":" [FWS] hdr-name ) 988 INFORMATIVE EXPLANATION: By "signing" header fields that do not 989 actually exist, a signer can prevent insertion of those header 990 fields before verification. However, since a signer cannot 991 possibly know what header fields might be created in the 992 future, and that some MUAs might present header fields that are 993 embedded inside a message (e.g., as a message/rfc822 content 994 type), the security of this solution is not total. 996 INFORMATIVE EXPLANATION: The exclusion of the header field name 997 and colon as well as the header field value for non-existent 998 header fields prevents an attacker from inserting an actual 999 header field with a null value. 1001 i= The Agent or User Identifier (AUID) on behalf of which the SDID is 1002 taking responsibility (dkim-quoted-printable; OPTIONAL, default is 1003 an empty Local-part followed by an "@" followed by the domain from 1004 the "d=" tag). 1006 The syntax is a standard email address where the Local-part MAY be 1007 omitted. The domain part of the address MUST be the same as, or a 1008 subdomain of, the value of the "d=" tag. 1010 Internationalized domain names MUST be converted using the steps 1011 listed in Section 4 of [RFC3490] using the "ToASCII" function. 1013 ABNF: 1015 sig-i-tag = %x69 [FWS] "=" [FWS] [ Local-part ] 1016 "@" domain-name 1018 The AUID is specified as having the same syntax as an email 1019 address, but is not required to have the same semantics. Notably, 1020 the domain name is not required to be registered in the DNS -- so 1021 it might not resolve in a query -- and the Local-part MAY be drawn 1022 from a namespace unrelated to any mailbox. The details of the 1023 structure and semantics for the namespace are determined by the 1024 Signer. Any knowledge or use of those details by verifiers or 1025 assessors is outside the scope of the DKIM Signing specification. 1026 The Signer MAY choose to use the same namespace for its AUIDs as 1027 its users' email addresses or MAY choose other means of 1028 representing its users. However, the signer SHOULD use the same 1029 AUID for each message intended to be evaluated as being within the 1030 same sphere of responsibility, if it wishes to offer receivers the 1031 option of using the AUID as a stable identifier that is finer 1032 grained than the SDID. 1034 INFORMATIVE NOTE: The Local-part of the "i=" tag is optional 1035 because in some cases a signer may not be able to establish a 1036 verified individual identity. In such cases, the signer might 1037 wish to assert that although it is willing to go as far as 1038 signing for the domain, it is unable or unwilling to commit to 1039 an individual user name within their domain. It can do so by 1040 including the domain part but not the Local-part of the 1041 identity. 1043 INFORMATIVE DISCUSSION: This specification does not require the 1044 value of the "i=" tag to match the identity in any message 1045 header fields. This is considered to be a verifier policy 1046 issue. Constraints between the value of the "i=" tag and other 1047 identities in other header fields seek to apply basic 1048 authentication into the semantics of trust associated with a 1049 role such as content author. Trust is a broad and complex 1050 topic and trust mechanisms are subject to highly creative 1051 attacks. The real-world efficacy of any but the most basic 1052 bindings between the "i=" value and other identities is not 1053 well established, nor is its vulnerability to subversion by an 1054 attacker. Hence reliance on the use of these options should be 1055 strictly limited. In particular, it is not at all clear to 1056 what extent a typical end-user recipient can rely on any 1057 assurances that might be made by successful use of the "i=" 1058 options. 1060 l= Body length count (plain-text unsigned decimal integer; OPTIONAL, 1061 default is entire body). This tag informs the verifier of the 1062 number of octets in the body of the email after canonicalization 1063 included in the cryptographic hash, starting from 0 immediately 1064 following the CRLF preceding the body. This value MUST NOT be 1065 larger than the actual number of octets in the canonicalized 1066 message body. 1068 INFORMATIVE IMPLEMENTATION WARNING: Use of the "l=" tag might 1069 allow display of fraudulent content without appropriate warning 1070 to end users. The "l=" tag is intended for increasing 1071 signature robustness when sending to mailing lists that both 1072 modify their content and do not sign their messages. However, 1073 using the "l=" tag enables attacks in which an intermediary 1074 with malicious intent modifies a message to include content 1075 that solely benefits the attacker. It is possible for the 1076 appended content to completely replace the original content in 1077 the end recipient's eyes and to defeat duplicate message 1078 detection algorithms. Examples are described in Security 1079 Considerations Section 8. To avoid this attack, signers should 1080 be extremely wary of using this tag, and verifiers might wish 1081 to ignore the tag or remove text that appears after the 1082 specified content length. 1084 INFORMATIVE NOTE: The value of the "l=" tag is constrained to 1085 76 decimal digits. This constraint is not intended to predict 1086 the size of future messages or to require implementations to 1087 use an integer representation large enough to represent the 1088 maximum possible value, but is intended to remind the 1089 implementer to check the length of this and all other tags 1090 during verification and to test for integer overflow when 1091 decoding the value. Implementers may need to limit the actual 1092 value expressed to a value smaller than 10^76, e.g., to allow a 1093 message to fit within the available storage space. 1095 ABNF: 1096 sig-l-tag = %x6c [FWS] "=" [FWS] 1097 1*76DIGIT 1099 q= A colon-separated list of query methods used to retrieve the 1100 public key (plain-text; OPTIONAL, default is "dns/txt"). Each 1101 query method is of the form "type[/options]", where the syntax and 1102 semantics of the options depend on the type and specified options. 1103 If there are multiple query mechanisms listed, the choice of query 1104 mechanism MUST NOT change the interpretation of the signature. 1105 Implementations MUST use the recognized query mechanisms in the 1106 order presented. Unrecognized query mechanisms MUST be ignored. 1108 Currently, the only valid value is "dns/txt", which defines the 1109 DNS TXT record lookup algorithm described elsewhere in this 1110 document. The only option defined for the "dns" query type is 1111 "txt", which MUST be included. Verifiers and signers MUST support 1112 "dns/txt". 1114 ABNF: 1115 sig-q-tag = %x71 [FWS] "=" [FWS] sig-q-tag-method 1116 *([FWS] ":" [FWS] sig-q-tag-method) 1117 sig-q-tag-method = "dns/txt" / x-sig-q-tag-type 1118 ["/" x-sig-q-tag-args] 1119 x-sig-q-tag-type = hyphenated-word ; for future extension 1120 x-sig-q-tag-args = qp-hdr-value 1122 s= The selector subdividing the namespace for the "d=" (domain) tag 1123 (plain-text; REQUIRED). 1125 ABNF: 1126 sig-s-tag = %x73 [FWS] "=" [FWS] selector 1128 t= Signature Timestamp (plain-text unsigned decimal integer; 1129 RECOMMENDED, default is an unknown creation time). The time that 1130 this signature was created. The format is the number of seconds 1131 since 00:00:00 on January 1, 1970 in the UTC time zone. The value 1132 is expressed as an unsigned integer in decimal ASCII. This value 1133 is not constrained to fit into a 31- or 32-bit integer. 1134 Implementations SHOULD be prepared to handle values up to at least 1135 10^12 (until approximately AD 200,000; this fits into 40 bits). 1136 To avoid denial-of-service attacks, implementations MAY consider 1137 any value longer than 12 digits to be infinite. Leap seconds are 1138 not counted. Implementations MAY ignore signatures that have a 1139 timestamp in the future. 1141 ABNF: 1142 sig-t-tag = %x74 [FWS] "=" [FWS] 1*12DIGIT 1143 x= Signature Expiration (plain-text unsigned decimal integer; 1144 RECOMMENDED, default is no expiration). The format is the same as 1145 in the "t=" tag, represented as an absolute date, not as a time 1146 delta from the signing timestamp. The value is expressed as an 1147 unsigned integer in decimal ASCII, with the same constraints on 1148 the value in the "t=" tag. Signatures MAY be considered invalid 1149 if the verification time at the verifier is past the expiration 1150 date. The verification time should be the time that the message 1151 was first received at the administrative domain of the verifier if 1152 that time is reliably available; otherwise the current time should 1153 be used. The value of the "x=" tag MUST be greater than the value 1154 of the "t=" tag if both are present. 1156 INFORMATIVE NOTE: The "x=" tag is not intended as an anti- 1157 replay defense. 1159 INFORMATIVE NOTE: Due to clock drift, the receiver's notion of 1160 when to consider the signature expired may not match exactly 1161 when the sender is expecting. Receivers MAY add a 'fudge 1162 factor' to allow for such possible drift. 1164 ABNF: 1165 sig-x-tag = %x78 [FWS] "=" [FWS] 1166 1*12DIGIT 1168 z= Copied header fields (dkim-quoted-printable, but see description; 1169 OPTIONAL, default is null). A vertical-bar-separated list of 1170 selected header fields present when the message was signed, 1171 including both the field name and value. It is not required to 1172 include all header fields present at the time of signing. This 1173 field need not contain the same header fields listed in the "h=" 1174 tag. The header field text itself must encode the vertical bar 1175 ("|", %x7C) character (i.e., vertical bars in the "z=" text are 1176 meta-characters, and any actual vertical bar characters in a 1177 copied header field must be encoded). Note that all whitespace 1178 must be encoded, including whitespace between the colon and the 1179 header field value. After encoding, FWS MAY be added at arbitrary 1180 locations in order to avoid excessively long lines; such 1181 whitespace is NOT part of the value of the header field, and MUST 1182 be removed before decoding. 1184 The header fields referenced by the "h=" tag refer to the fields 1185 in the [RFC5322] header of the message, not to any copied fields 1186 in the "z=" tag. Copied header field values are for diagnostic 1187 use. 1189 Header fields with characters requiring conversion (perhaps from 1190 legacy MTAs that are not [RFC5322] compliant) SHOULD be converted 1191 as described in MIME Part Three [RFC2047]. 1193 ABNF: 1194 sig-z-tag = %x7A [FWS] "=" [FWS] sig-z-tag-copy 1195 *( "|" [FWS] sig-z-tag-copy ) 1196 sig-z-tag-copy = hdr-name [FWS] ":" qp-hdr-value 1198 INFORMATIVE EXAMPLE of a signature header field spread across 1199 multiple continuation lines: 1200 DKIM-Signature: v=1; a=rsa-sha256; d=example.net; s=brisbane; 1201 c=simple; q=dns/txt; i=@eng.example.net; 1202 t=1117574938; x=1118006938; 1203 h=from:to:subject:date; 1204 z=From:foo@eng.example.net|To:joe@example.com| 1205 Subject:demo=20run|Date:July=205,=202005=203:44:08=20PM=20-0700; 1206 bh=MTIzNDU2Nzg5MDEyMzQ1Njc4OTAxMjM0NTY3ODkwMTI=; 1207 b=dzdVyOfAKCdLXdJOc9G2q8LoXSlEniSbav+yuU4zGeeruD00lszZVoG4ZHRNiYzR 1209 3.6. Key Management and Representation 1211 Signature applications require some level of assurance that the 1212 verification public key is associated with the claimed signer. Many 1213 applications achieve this by using public key certificates issued by 1214 a trusted third party. However, DKIM can achieve a sufficient level 1215 of security, with significantly enhanced scalability, by simply 1216 having the verifier query the purported signer's DNS entry (or some 1217 security-equivalent) in order to retrieve the public key. 1219 DKIM keys can potentially be stored in multiple types of key servers 1220 and in multiple formats. The storage and format of keys are 1221 irrelevant to the remainder of the DKIM algorithm. 1223 Parameters to the key lookup algorithm are the type of the lookup 1224 (the "q=" tag), the domain of the signer (the "d=" tag of the DKIM- 1225 Signature header field), and the selector (the "s=" tag). 1227 public_key = dkim_find_key(q_val, d_val, s_val) 1229 This document defines a single binding, using DNS TXT records to 1230 distribute the keys. Other bindings may be defined in the future. 1232 3.6.1. Textual Representation 1234 It is expected that many key servers will choose to present the keys 1235 in an otherwise unstructured text format (for example, an XML form 1236 would not be considered to be unstructured text for this purpose). 1237 The following definition MUST be used for any DKIM key represented in 1238 an otherwise unstructured textual form. 1240 The overall syntax is a tag-list as described in Section 3.2. The 1241 current valid tags are described below. Other tags MAY be present 1242 and MUST be ignored by any implementation that does not understand 1243 them. 1245 v= Version of the DKIM key record (plain-text; RECOMMENDED, default 1246 is "DKIM1"). If specified, this tag MUST be set to "DKIM1" 1247 (without the quotes). This tag MUST be the first tag in the 1248 record. Records beginning with a "v=" tag with any other value 1249 MUST be discarded. Note that verifiers must do a string 1250 comparison on this value; for example, "DKIM1" is not the same as 1251 "DKIM1.0". 1253 ABNF: 1254 key-v-tag = %x76 [FWS] "=" [FWS] %x44 %x4B %x49 %x4D %x31 1256 h= Acceptable hash algorithms (plain-text; OPTIONAL, defaults to 1257 allowing all algorithms). A colon-separated list of hash 1258 algorithms that might be used. Unrecognized algorithms MUST be 1259 ignored. Refer to Section 3.3 for a discussion of the hash 1260 algorithms implemented by Signers and Verifiers. . The set of 1261 algorithms listed in this tag in each record is an operational 1262 choice made by the Signer. 1264 ABNF: 1265 key-h-tag = %x68 [FWS] "=" [FWS] key-h-tag-alg 1266 0*( [FWS] ":" [FWS] key-h-tag-alg ) 1267 key-h-tag-alg = "sha1" / "sha256" / x-key-h-tag-alg 1268 x-key-h-tag-alg = hyphenated-word ; for future extension 1269 k= Key type (plain-text; OPTIONAL, default is "rsa"). Signers and 1270 verifiers MUST support the "rsa" key type. The "rsa" key type 1271 indicates that an ASN.1 DER-encoded [ITU-X660-1997] RSAPublicKey 1272 [RFC3447] (see Sections Section 3.1 and A.1.1) is being used in 1273 the "p=" tag. (Note: the "p=" tag further encodes the value using 1274 the base64 algorithm.) Unrecognized key types MUST be ignored. 1276 ABNF: 1277 key-k-tag = %x76 [FWS] "=" [FWS] key-k-tag-type 1278 key-k-tag-type = "rsa" / x-key-k-tag-type 1279 x-key-k-tag-type = hyphenated-word ; for future extension 1281 n= Notes that might be of interest to a human (qp-section; OPTIONAL, 1282 default is empty). No interpretation is made by any program. 1283 This tag should be used sparingly in any key server mechanism that 1284 has space limitations (notably DNS). This is intended for use by 1285 administrators, not end users. 1287 ABNF: 1288 key-n-tag = %x6e [FWS] "=" [FWS] qp-section 1290 p= Public-key data (base64; REQUIRED). An empty value means that 1291 this public key has been revoked. The syntax and semantics of 1292 this tag value before being encoded in base64 are defined by the 1293 "k=" tag. 1295 INFORMATIVE RATIONALE: If a private key has been compromised or 1296 otherwise disabled (e.g., an outsourcing contract has been 1297 terminated), a signer might want to explicitly state that it 1298 knows about the selector, but all messages using that selector 1299 should fail verification. Verifiers should ignore any DKIM- 1300 Signature header fields with a selector referencing a revoked 1301 key. 1303 ABNF: 1304 key-p-tag = %x70 [FWS] "=" [ [FWS] base64string] 1306 INFORMATIVE NOTE: A base64string is permitted to include white 1307 space (FWS) at arbitrary places; however, any CRLFs must be 1308 followed by at least one WSP character. Implementors and 1309 administrators are cautioned to ensure that selector TXT 1310 records conform to this specification. 1312 s= Service Type (plain-text; OPTIONAL; default is "*"). A colon- 1313 separated list of service types to which this record applies. 1314 Verifiers for a given service type MUST ignore this record if the 1315 appropriate type is not listed. Unrecognized service types MUST 1316 be ignored. Currently defined service types are as follows: 1318 * matches all service types 1320 email electronic mail (not necessarily limited to SMTP) 1322 This tag is intended to constrain the use of keys for other 1323 purposes, should use of DKIM be defined by other services in the 1324 future. 1326 ABNF: 1327 key-s-tag = %x73 [FWS] "=" [FWS] key-s-tag-type 1328 0*( [FWS] ":" [FWS] key-s-tag-type ) 1329 key-s-tag-type = "email" / "*" / x-key-s-tag-type 1330 x-key-s-tag-type = hyphenated-word ; for future extension 1332 t= Flags, represented as a colon-separated list of names (plain- 1333 text; OPTIONAL, default is no flags set). Unrecognized flags MUST 1334 be ignored. The defined flags are as follows: 1336 y This domain is testing DKIM. Verifiers MUST NOT treat messages 1337 from signers in testing mode differently from unsigned email, even 1338 should the signature fail to verify. Verifiers MAY wish to track 1339 testing mode results to assist the signer. 1341 s Any DKIM-Signature header fields using the "i=" tag MUST have the 1342 same domain value on the right-hand side of the "@" in the "i=" 1343 tag and the value of the "d=" tag. That is, the "i=" domain MUST 1344 NOT be a subdomain of "d=". Use of this flag is RECOMMENDED 1345 unless subdomaining is required. 1347 ABNF: 1348 key-t-tag = %x74 [FWS] "=" [FWS] key-t-tag-flag 1349 0*( [FWS] ":" [FWS] key-t-tag-flag ) 1350 key-t-tag-flag = "y" / "s" / x-key-t-tag-flag 1351 x-key-t-tag-flag = hyphenated-word ; for future extension 1353 Unrecognized flags MUST be ignored. 1355 3.6.2. DNS Binding 1357 A binding using DNS TXT records as a key service is hereby defined. 1358 All implementations MUST support this binding. 1360 3.6.2.1. Namespace 1362 All DKIM keys are stored in a subdomain named "_domainkey". Given a 1363 DKIM-Signature field with a "d=" tag of "example.com" and an "s=" tag 1364 of "foo.bar", the DNS query will be for 1365 "foo.bar._domainkey.example.com". 1367 INFORMATIVE OPERATIONAL NOTE: Wildcard DNS records (e.g., 1368 *.bar._domainkey.example.com) do not make sense in this context 1369 and should not be used. Note also that wildcards within domains 1370 (e.g., s._domainkey.*.example.com) are not supported by the DNS. 1372 3.6.2.2. Resource Record Types for Key Storage 1374 The DNS Resource Record type used is specified by an option to the 1375 query-type ("q=") tag. The only option defined in this base 1376 specification is "txt", indicating the use of a TXT Resource Record 1377 (RR). A later extension of this standard may define another RR type. 1379 Strings in a TXT RR MUST be concatenated together before use with no 1380 intervening whitespace. TXT RRs MUST be unique for a particular 1381 selector name; that is, if there are multiple records in an RRset, 1382 the results are undefined. 1384 TXT RRs are encoded as described in Section 3.6.1 1386 3.7. Computing the Message Hashes 1388 Both signing and verifying message signatures start with a step of 1389 computing two cryptographic hashes over the message. Signers will 1390 choose the parameters of the signature as described in Signer Actions 1391 Section 5; verifiers will use the parameters specified in the DKIM- 1392 Signature header field being verified. In the following discussion, 1393 the names of the tags in the DKIM-Signature header field that either 1394 exists (when verifying) or will be created (when signing) are used. 1395 Note that canonicalization (Section 3.4) is only used to prepare the 1396 email for signing or verifying; it does not affect the transmitted 1397 email in any way. 1399 The signer/verifier MUST compute two hashes, one over the body of the 1400 message and one over the selected header fields of the message. 1402 Signers MUST compute them in the order shown. Verifiers MAY compute 1403 them in any order convenient to the verifier, provided that the 1404 result is semantically identical to the semantics that would be the 1405 case had they been computed in this order. 1407 In hash step 1, the signer/verifier MUST hash the message body, 1408 canonicalized using the body canonicalization algorithm specified in 1409 the "c=" tag and then truncated to the length specified in the "l=" 1410 tag. That hash value is then converted to base64 form and inserted 1411 into (signers) or compared to (verifiers) the "bh=" tag of the DKIM- 1412 Signature header field. 1414 In hash step 2, the signer/verifier MUST pass the following to the 1415 hash algorithm in the indicated order. 1417 1. The header fields specified by the "h=" tag, in the order 1418 specified in that tag, and canonicalized using the header 1419 canonicalization algorithm specified in the "c=" tag. Each 1420 header field MUST be terminated with a single CRLF. 1422 2. The DKIM-Signature header field that exists (verifying) or will 1423 be inserted (signing) in the message, with the value of the "b=" 1424 tag (including all surrounding whitespace) deleted (i.e., treated 1425 as the empty string), canonicalized using the header 1426 canonicalization algorithm specified in the "c=" tag, and without 1427 a trailing CRLF. 1429 All tags and their values in the DKIM-Signature header field are 1430 included in the cryptographic hash with the sole exception of the 1431 value portion of the "b=" (signature) tag, which MUST be treated as 1432 the null string. All tags MUST be included even if they might not be 1433 understood by the verifier. The header field MUST be presented to 1434 the hash algorithm after the body of the message rather than with the 1435 rest of the header fields and MUST be canonicalized as specified in 1436 the "c=" (canonicalization) tag. The DKIM-Signature header field 1437 MUST NOT be included in its own h= tag, although other DKIM-Signature 1438 header fields MAY be signed (see Section 4). 1440 When calculating the hash on messages that will be transmitted using 1441 base64 or quoted-printable encoding, signers MUST compute the hash 1442 after the encoding. Likewise, the verifier MUST incorporate the 1443 values into the hash before decoding the base64 or quoted-printable 1444 text. However, the hash MUST be computed before transport level 1445 encodings such as SMTP "dot-stuffing" (the modification of lines 1446 beginning with a "." to avoid confusion with the SMTP end-of-message 1447 marker, as specified in [RFC5321]). 1449 With the exception of the canonicalization procedure described in 1450 Section 3.4, the DKIM signing process treats the body of messages as 1451 simply a string of octets. DKIM messages MAY be either in plain-text 1452 or in MIME format; no special treatment is afforded to MIME content. 1453 Message attachments in MIME format MUST be included in the content 1454 that is signed. 1456 More formally, pseudo-code for the signature algorithm is: 1457 body-hash = hash-alg (canon-body, l-param) 1458 data-hash = hash-alg (h-headers, D-SIG, content-hash) 1459 signature = sig-alg (d-domain, selector, data-hash) 1461 where: 1463 body-hash: is the output from hashing the body, using hash-alg. 1465 hash-alg: is the hashing algorithm specified in the "a" 1466 parameter. 1468 canon-body: is a canonicalized representation of the body, 1469 produced by using the body algorithm specified in the "c" 1470 parameter, as defined in Section 3.4 and excluding the 1471 DKIM-Signature field. 1473 l-param: is the length-of-body value of the "l" parameter. 1475 data-hash: is the output from using the hash-alg algorithm, to 1476 hash the header including the DKIM-Signature header, and the 1477 body hash. 1479 h-headers: is the list of headers to be signed, as specified in 1480 the "h" parameter. 1482 D-SIG: is the canonicalized DKIM-Signature field without the 1483 signature value portion of the parameter, itself; that is, an 1484 empty parameter value. 1486 signature: is the signature value produced by the signing 1487 algorithm. 1489 sig-alg: is the signature algorithm specified by the "a" 1490 parameter. 1492 d-domain: is the domain name specified in the "d" parameter. 1494 selector: is the selector value specified in the "s" parameter. 1496 NOTE: Many digital signature APIs provide both hashing and 1497 application of the RSA private key using a single "sign()" 1498 primitive. When using such an API, the last two steps in the 1499 algorithm would probably be combined into a single call that would 1500 perform both the "a-hash-alg" and the "sig-alg". 1502 3.8. Input Requirements 1504 DKIM's design is predicated on valid input. Therefore, signers and 1505 verifiers SHOULD take reasonable steps to ensure that the messages 1506 they are processing are valid according to [RFC5322], [RFC2045], and 1507 any other relevant message format standards. See Section 8.15 for 1508 additional discussion and references. 1510 3.9. Signing by Parent Domains 1512 In some circumstances, it is desirable for a domain to apply a 1513 signature on behalf of any of its subdomains without the need to 1514 maintain separate selectors (key records) in each subdomain. By 1515 default, private keys corresponding to key records can be used to 1516 sign messages for any subdomain of the domain in which they reside; 1517 for example, a key record for the domain example.com can be used to 1518 verify messages where the AUID ("i=" tag of the signature) is 1519 sub.example.com, or even sub1.sub2.example.com. In order to limit 1520 the capability of such keys when this is not intended, the "s" flag 1521 MAY be set in the "t=" tag of the key record, to constrain the 1522 validity of the domain of the AUID. If the referenced key record 1523 contains the "s" flag as part of the "t=" tag, the domain of the AUID 1524 ("i=" flag) MUST be the same as that of the SDID (d=) domain. If 1525 this flag is absent, the domain of the AUID MUST be the same as, or a 1526 subdomain of, the SDID. 1528 3.10. Relationship between SDID and AUID 1530 DKIM's primary task is to communicate from the Signer to a recipient- 1531 side Identity Assessor a single Signing Domain Identifier (SDID) that 1532 refers to a responsible identity. DKIM MAY optionally provide a 1533 single responsible Agent or User Identifier (AUID). 1535 Hence, DKIM's mandatory output to a receive-side Identity Assessor is 1536 a single domain name. Within the scope of its use as DKIM output, 1537 the name has only basic domain name semantics; any possible owner- 1538 specific semantics are outside the scope of DKIM. That is, within 1539 its role as a DKIM identifier, additional semantics cannot be assumed 1540 by an Identity Assessor. 1542 Upon successfully verifying the signature, a receive-side DKIM 1543 verifier MUST communicate the Signing Domain Identifier (d=) to a 1544 consuming Identity Assessor module and MAY communicate the Agent or 1545 User Identifier (i=) if present. 1547 To the extent that a receiver attempts to intuit any structured 1548 semantics for either of the identifiers, this is a heuristic function 1549 that is outside the scope of DKIM's specification and semantics. 1550 Hence, it is relegated to a higher-level service, such as a delivery 1551 handling filter that integrates a variety of inputs and performs 1552 heuristic analysis of them. 1554 INFORMATIVE DISCUSSION: This document does not require the value 1555 of the SDID or AUID to match an identifier in any other message 1556 header field. This requirement is, instead, an assessor policy 1557 issue. The purpose of such a linkage would be to authenticate the 1558 value in that other header field. This, in turn, is the basis for 1559 applying a trust assessment based on the identifier value. Trust 1560 is a broad and complex topic and trust mechanisms are subject to 1561 highly creative attacks. The real-world efficacy of any but the 1562 most basic bindings between the SDID or AUID and other identities 1563 is not well established, nor is its vulnerability to subversion by 1564 an attacker. Hence, reliance on the use of such bindings should 1565 be strictly limited. In particular, it is not at all clear to 1566 what extent a typical end-user recipient can rely on any 1567 assurances that might be made by successful use of the SDID or 1568 AUID. 1570 4. Semantics of Multiple Signatures 1571 4.1. Example Scenarios 1573 There are many reasons why a message might have multiple signatures. 1574 For example, a given signer might sign multiple times, perhaps with 1575 different hashing or signing algorithms during a transition phase. 1577 INFORMATIVE EXAMPLE: Suppose SHA-256 is in the future found to be 1578 insufficiently strong, and DKIM usage transitions to SHA-1024. A 1579 signer might immediately sign using the newer algorithm, but 1580 continue to sign using the older algorithm for interoperability 1581 with verifiers that had not yet upgraded. The signer would do 1582 this by adding two DKIM-Signature header fields, one using each 1583 algorithm. Older verifiers that did not recognize SHA-1024 as an 1584 acceptable algorithm would skip that signature and use the older 1585 algorithm; newer verifiers could use either signature at their 1586 option, and all other things being equal might not even attempt to 1587 verify the other signature. 1589 Similarly, a signer might sign a message including all headers and no 1590 "l=" tag (to satisfy strict verifiers) and a second time with a 1591 limited set of headers and an "l=" tag (in anticipation of possible 1592 message modifications in route to other verifiers). Verifiers could 1593 then choose which signature they preferred. 1595 INFORMATIVE EXAMPLE: A verifier might receive a message with two 1596 signatures, one covering more of the message than the other. If 1597 the signature covering more of the message verified, then the 1598 verifier could make one set of policy decisions; if that signature 1599 failed but the signature covering less of the message verified, 1600 the verifier might make a different set of policy decisions. 1602 Of course, a message might also have multiple signatures because it 1603 passed through multiple signers. A common case is expected to be 1604 that of a signed message that passes through a mailing list that also 1605 signs all messages. Assuming both of those signatures verify, a 1606 recipient might choose to accept the message if either of those 1607 signatures were known to come from trusted sources. 1609 INFORMATIVE EXAMPLE: Recipients might choose to whitelist mailing 1610 lists to which they have subscribed and that have acceptable anti- 1611 abuse policies so as to accept messages sent to that list even 1612 from unknown authors. They might also subscribe to less trusted 1613 mailing lists (e.g., those without anti-abuse protection) and be 1614 willing to accept all messages from specific authors, but insist 1615 on doing additional abuse scanning for other messages. 1617 Another related example of multiple signers might be forwarding 1618 services, such as those commonly associated with academic alumni 1619 sites. 1621 INFORMATIVE EXAMPLE: A recipient might have an address at 1622 members.example.org, a site that has anti-abuse protection that is 1623 somewhat less effective than the recipient would prefer. Such a 1624 recipient might have specific authors whose messages would be 1625 trusted absolutely, but messages from unknown authors that had 1626 passed the forwarder's scrutiny would have only medium trust. 1628 4.2. Interpretation 1630 A signer that is adding a signature to a message merely creates a new 1631 DKIM-Signature header, using the usual semantics of the h= option. A 1632 signer MAY sign previously existing DKIM-Signature header fields 1633 using the method described in Section 5.4 to sign trace header 1634 fields. 1636 INFORMATIVE NOTE: Signers should be cognizant that signing DKIM- 1637 Signature header fields may result in signature failures with 1638 intermediaries that do not recognize that DKIM-Signature header 1639 fields are trace header fields and unwittingly reorder them, thus 1640 breaking such signatures. For this reason, signing existing DKIM- 1641 Signature header fields is unadvised, albeit legal. 1643 INFORMATIVE NOTE: If a header field with multiple instances is 1644 signed, those header fields are always signed from the bottom up. 1645 Thus, it is not possible to sign only specific DKIM-Signature 1646 header fields. For example, if the message being signed already 1647 contains three DKIM-Signature header fields A, B, and C, it is 1648 possible to sign all of them, B and C only, or C only, but not A 1649 only, B only, A and B only, or A and C only. 1651 A signer MAY add more than one DKIM-Signature header field using 1652 different parameters. For example, during a transition period a 1653 signer might want to produce signatures using two different hash 1654 algorithms. 1656 Signers SHOULD NOT remove any DKIM-Signature header fields from 1657 messages they are signing, even if they know that the signatures 1658 cannot be verified. 1660 When evaluating a message with multiple signatures, a verifier SHOULD 1661 evaluate signatures independently and on their own merits. For 1662 example, a verifier that by policy chooses not to accept signatures 1663 with deprecated cryptographic algorithms would consider such 1664 signatures invalid. Verifiers MAY process signatures in any order of 1665 their choice; for example, some verifiers might choose to process 1666 signatures corresponding to the From field in the message header 1667 before other signatures. See Section 6.1 for more information about 1668 signature choices. 1670 INFORMATIVE IMPLEMENTATION NOTE: Verifier attempts to correlate 1671 valid signatures with invalid signatures in an attempt to guess 1672 why a signature failed are ill-advised. In particular, there is 1673 no general way that a verifier can determine that an invalid 1674 signature was ever valid. 1676 Verifiers SHOULD ignore failed signatures as though they were not 1677 present in the message. Verifiers SHOULD continue to check 1678 signatures until a signature successfully verifies to the 1679 satisfaction of the verifier. To limit potential denial-of-service 1680 attacks, verifiers MAY limit the total number of signatures they will 1681 attempt to verify. 1683 5. Signer Actions 1685 The following steps are performed in order by signers. 1687 5.1. Determine Whether the Email Should Be Signed and by Whom 1689 A signer can obviously only sign email for domains for which it has a 1690 private key and the necessary knowledge of the corresponding public 1691 key and selector information. However, there are a number of other 1692 reasons beyond the lack of a private key why a signer could choose 1693 not to sign an email. 1695 INFORMATIVE NOTE: Signing modules may be incorporated into any 1696 portion of the mail system as deemed appropriate, including an 1697 MUA, a SUBMISSION server, or an MTA. Wherever implemented, 1698 signers should beware of signing (and thereby asserting 1699 responsibility for) messages that may be problematic. In 1700 particular, within a trusted enclave the signing address might be 1701 derived from the header according to local policy; SUBMISSION 1702 servers might only sign messages from users that are properly 1703 authenticated and authorized. 1705 INFORMATIVE IMPLEMENTER ADVICE: SUBMISSION servers should not sign 1706 Received header fields if the outgoing gateway MTA obfuscates 1707 Received header fields, for example, to hide the details of 1708 internal topology. 1710 If an email cannot be signed for some reason, it is a local policy 1711 decision as to what to do with that email. 1713 5.2. Select a Private Key and Corresponding Selector Information 1715 This specification does not define the basis by which a signer should 1716 choose which private key and selector information to use. Currently, 1717 all selectors are equal as far as this specification is concerned, so 1718 the decision should largely be a matter of administrative 1719 convenience. Distribution and management of private keys is also 1720 outside the scope of this document. 1722 INFORMATIVE OPERATIONS ADVICE: A signer should not sign with a 1723 private key when the selector containing the corresponding public 1724 key is expected to be revoked or removed before the verifier has 1725 an opportunity to validate the signature. The signer should 1726 anticipate that verifiers may choose to defer validation, perhaps 1727 until the message is actually read by the final recipient. In 1728 particular, when rotating to a new key pair, signing should 1729 immediately commence with the new private key and the old public 1730 key should be retained for a reasonable validation interval before 1731 being removed from the key server. 1733 5.3. Normalize the Message to Prevent Transport Conversions 1735 Some messages, particularly those using 8-bit characters, are subject 1736 to modification during transit, notably conversion to 7-bit form. 1737 Such conversions will break DKIM signatures. In order to minimize 1738 the chances of such breakage, signers SHOULD convert the message to a 1739 suitable MIME content transfer encoding such as quoted-printable or 1740 base64 as described in [RFC2045] before signing. Such conversion is 1741 outside the scope of DKIM; the actual message SHOULD be converted to 1742 7-bit MIME by an MUA or MSA prior to presentation to the DKIM 1743 algorithm. 1745 Similarly, a message that is not compliant with RFC5322, RFC2045 and 1746 RFC2047 can be subject to attempts by intermediaries to correct or 1747 interpret such content. See Section 8 of [RFC4409] for examples of 1748 changes that are commonly made. Such "corrections" may break DKIM 1749 signatures or have other undesirable effects. Therefore, a verifier 1750 SHOULD NOT validate a message that is not compliant with those 1751 specifications. 1753 If the message is submitted to the signer with any local encoding 1754 that will be modified before transmission, that modification to 1755 canonical [RFC5322] form MUST be done before signing. In particular, 1756 bare CR or LF characters (used by some systems as a local line 1757 separator convention) MUST be converted to the SMTP-standard CRLF 1758 sequence before the message is signed. Any conversion of this sort 1759 SHOULD be applied to the message actually sent to the recipient(s), 1760 not just to the version presented to the signing algorithm. 1762 More generally, the signer MUST sign the message as it is expected to 1763 be received by the verifier rather than in some local or internal 1764 form. 1766 5.4. Determine the Header Fields to Sign 1768 The From header field MUST be signed (that is, included in the "h=" 1769 tag of the resulting DKIM-Signature header field). Signers SHOULD 1770 NOT sign an existing header field likely to be legitimately modified 1771 or removed in transit. In particular, [RFC5321] explicitly permits 1772 modification or removal of the Return-Path header field in transit. 1773 Signers MAY include any other header fields present at the time of 1774 signing at the discretion of the signer. 1776 INFORMATIVE OPERATIONS NOTE: The choice of which header fields to 1777 sign is non-obvious. One strategy is to sign all existing, non- 1778 repeatable header fields. An alternative strategy is to sign only 1779 header fields that are likely to be displayed to or otherwise be 1780 likely to affect the processing of the message at the receiver. A 1781 third strategy is to sign only "well known" headers. Note that 1782 verifiers may treat unsigned header fields with extreme 1783 skepticism, including refusing to display them to the end user or 1784 even ignoring the signature if it does not cover certain header 1785 fields. For this reason, signing fields present in the message 1786 such as Date, Subject, Reply-To, Sender, and all MIME header 1787 fields are highly advised. 1789 The DKIM-Signature header field is always implicitly signed and MUST 1790 NOT be included in the "h=" tag except to indicate that other 1791 preexisting signatures are also signed. 1793 Signers MAY claim to have signed header fields that do not exist 1794 (that is, signers MAY include the header field name in the "h=" tag 1795 even if that header field does not exist in the message). When 1796 computing the signature, the non-existing header field MUST be 1797 treated as the null string (including the header field name, header 1798 field value, all punctuation, and the trailing CRLF). 1800 INFORMATIVE RATIONALE: This allows signers to explicitly assert 1801 the absence of a header field; if that header field is added later 1802 the signature will fail. 1804 INFORMATIVE NOTE: A header field name need only be listed once 1805 more than the actual number of that header field in a message at 1806 the time of signing in order to prevent any further additions. 1807 For example, if there is a single Comments header field at the 1808 time of signing, listing Comments twice in the "h=" tag is 1809 sufficient to prevent any number of Comments header fields from 1810 being appended; it is not necessary (but is legal) to list 1811 Comments three or more times in the "h=" tag. 1813 Signers choosing to sign an existing header field that occurs more 1814 than once in the message (such as Received) MUST sign the physically 1815 last instance of that header field in the header block. Signers 1816 wishing to sign multiple instances of such a header field MUST 1817 include the header field name multiple times in the h= tag of the 1818 DKIM-Signature header field, and MUST sign such header fields in 1819 order from the bottom of the header field block to the top. The 1820 signer MAY include more instances of a header field name in h= than 1821 there are actual corresponding header fields to indicate that 1822 additional header fields of that name SHOULD NOT be added. 1824 INFORMATIVE EXAMPLE: 1826 If the signer wishes to sign two existing Received header fields, 1827 and the existing header contains: 1828 Received: 1829 Received: 1830 Received: 1832 then the resulting DKIM-Signature header field should read: 1834 DKIM-Signature: ... h=Received : Received :... 1835 and Received header fields and will be signed in that 1836 order. 1838 Signers should be careful of signing header fields that might have 1839 additional instances added later in the delivery process, since such 1840 header fields might be inserted after the signed instance or 1841 otherwise reordered. Trace header fields (such as Received) and 1842 Resent-* blocks are the only fields prohibited by [RFC5322] from 1843 being reordered. In particular, since DKIM-Signature header fields 1844 may be reordered by some intermediate MTAs, signing existing DKIM- 1845 Signature header fields is error-prone. 1847 INFORMATIVE ADMONITION: Despite the fact that [RFC5322] permits 1848 header fields to be reordered (with the exception of Received 1849 header fields), reordering of signed header fields with multiple 1850 instances by intermediate MTAs will cause DKIM signatures to be 1851 broken; such anti-social behavior should be avoided. 1853 INFORMATIVE IMPLEMENTER'S NOTE: Although not required by this 1854 specification, all end-user visible header fields should be signed 1855 to avoid possible "indirect spamming". For example, if the 1856 Subject header field is not signed, a spammer can resend a 1857 previously signed mail, replacing the legitimate subject with a 1858 one-line spam. 1860 5.5. Recommended Signature Content 1862 In order to maximize compatibility with a variety of verifiers, it is 1863 recommended that signers follow the practices outlined in this 1864 section when signing a message. However, these are generic 1865 recommendations applying to the general case; specific senders may 1866 wish to modify these guidelines as required by their unique 1867 situations. Verifiers MUST be capable of verifying signatures even 1868 if one or more of the recommended header fields is not signed (with 1869 the exception of From, which must always be signed) or if one or more 1870 of the dis-recommended header fields is signed. Note that verifiers 1871 do have the option of ignoring signatures that do not cover a 1872 sufficient portion of the header or body, just as they may ignore 1873 signatures from an identity they do not trust. 1875 The following header fields SHOULD be included in the signature, if 1876 they are present in the message being signed: 1878 o From (REQUIRED in all signatures) 1880 o Sender, Reply-To 1882 o Subject 1884 o Date, Message-ID 1886 o To, Cc 1888 o MIME-Version 1890 o Content-Type, Content-Transfer-Encoding, Content-ID, Content- 1891 Description 1893 o Resent-Date, Resent-From, Resent-Sender, Resent-To, Resent-Cc, 1894 Resent-Message-ID 1896 o In-Reply-To, References 1898 o List-Id, List-Help, List-Unsubscribe, List-Subscribe, List-Post, 1899 List-Owner, List-Archive 1901 The following header fields SHOULD NOT be included in the signature: 1903 o Return-Path 1904 o Received 1906 o Comments, Keywords 1908 o Bcc, Resent-Bcc 1910 o DKIM-Signature 1912 Optional header fields (those not mentioned above) normally SHOULD 1913 NOT be included in the signature, because of the potential for 1914 additional header fields of the same name to be legitimately added or 1915 reordered prior to verification. There are likely to be legitimate 1916 exceptions to this rule, because of the wide variety of application- 1917 specific header fields that may be applied to a message, some of 1918 which are unlikely to be duplicated, modified, or reordered. 1920 Signers SHOULD choose canonicalization algorithms based on the types 1921 of messages they process and their aversion to risk. For example, 1922 e-commerce sites sending primarily purchase receipts, which are not 1923 expected to be processed by mailing lists or other software likely to 1924 modify messages, will generally prefer "simple" canonicalization. 1925 Sites sending primarily person-to-person email will likely prefer to 1926 be more resilient to modification during transport by using "relaxed" 1927 canonicalization. 1929 Signers SHOULD NOT use "l=" unless they intend to accommodate 1930 intermediate mail processors that append text to a message. For 1931 example, many mailing list processors append "unsubscribe" 1932 information to message bodies. If signers use "l=", they SHOULD 1933 include the entire message body existing at the time of signing in 1934 computing the count. In particular, signers SHOULD NOT specify a 1935 body length of 0 since this may be interpreted as a meaningless 1936 signature by some verifiers. 1938 5.6. Compute the Message Hash and Signature 1940 The signer MUST compute the message hash as described in Section 3.7 1941 and then sign it using the selected public-key algorithm. This will 1942 result in a DKIM-Signature header field that will include the body 1943 hash and a signature of the header hash, where that header includes 1944 the DKIM-Signature header field itself. 1946 Entities such as mailing list managers that implement DKIM and that 1947 modify the message or a header field (for example, inserting 1948 unsubscribe information) before retransmitting the message SHOULD 1949 check any existing signature on input and MUST make such 1950 modifications before re-signing the message. 1952 The signer MAY elect to limit the number of bytes of the body that 1953 will be included in the hash and hence signed. The length actually 1954 hashed should be inserted in the "l=" tag of the DKIM-Signature 1955 header field. 1957 5.7. Insert the DKIM-Signature Header Field 1959 Finally, the signer MUST insert the DKIM-Signature header field 1960 created in the previous step prior to transmitting the email. The 1961 DKIM-Signature header field MUST be the same as used to compute the 1962 hash as described above, except that the value of the "b=" tag MUST 1963 be the appropriately signed hash computed in the previous step, 1964 signed using the algorithm specified in the "a=" tag of the DKIM- 1965 Signature header field and using the private key corresponding to the 1966 selector given in the "s=" tag of the DKIM-Signature header field, as 1967 chosen above in Section 5.2 1969 The DKIM-Signature header field MUST be inserted before any other 1970 DKIM-Signature fields in the header block. 1972 INFORMATIVE IMPLEMENTATION NOTE: The easiest way to achieve this 1973 is to insert the DKIM-Signature header field at the beginning of 1974 the header block. In particular, it may be placed before any 1975 existing Received header fields. This is consistent with treating 1976 DKIM-Signature as a trace header field. 1978 6. Verifier Actions 1980 Since a signer MAY remove or revoke a public key at any time, it is 1981 recommended that verification occur in a timely manner. In many 1982 configurations, the most timely place is during acceptance by the 1983 border MTA or shortly thereafter. In particular, deferring 1984 verification until the message is accessed by the end user is 1985 discouraged. 1987 A border or intermediate MTA MAY verify the message signature(s). An 1988 MTA who has performed verification MAY communicate the result of that 1989 verification by adding a verification header field to incoming 1990 messages. This considerably simplifies things for the user, who can 1991 now use an existing mail user agent. Most MUAs have the ability to 1992 filter messages based on message header fields or content; these 1993 filters would be used to implement whatever policy the user wishes 1994 with respect to unsigned mail. 1996 A verifying MTA MAY implement a policy with respect to unverifiable 1997 mail, regardless of whether or not it applies the verification header 1998 field to signed messages. 2000 Verifiers MUST produce a result that is semantically equivalent to 2001 applying the following steps in the order listed. In practice, 2002 several of these steps can be performed in parallel in order to 2003 improve performance. 2005 6.1. Extract Signatures from the Message 2007 The order in which verifiers try DKIM-Signature header fields is not 2008 defined; verifiers MAY try signatures in any order they like. For 2009 example, one implementation might try the signatures in textual 2010 order, whereas another might try signatures by identities that match 2011 the contents of the From header field before trying other signatures. 2012 Verifiers MUST NOT attribute ultimate meaning to the order of 2013 multiple DKIM-Signature header fields. In particular, there is 2014 reason to believe that some relays will reorder the header fields in 2015 potentially arbitrary ways. 2017 INFORMATIVE IMPLEMENTATION NOTE: Verifiers might use the order as 2018 a clue to signing order in the absence of any other information. 2019 However, other clues as to the semantics of multiple signatures 2020 (such as correlating the signing host with Received header fields) 2021 may also be considered. 2023 A verifier SHOULD NOT treat a message that has one or more bad 2024 signatures and no good signatures differently from a message with no 2025 signature at all; such treatment is a matter of local policy and is 2026 beyond the scope of this document. 2028 When a signature successfully verifies, a verifier will either stop 2029 processing or attempt to verify any other signatures, at the 2030 discretion of the implementation. A verifier MAY limit the number of 2031 signatures it tries to avoid denial-of-service attacks. 2033 INFORMATIVE NOTE: An attacker could send messages with large 2034 numbers of faulty signatures, each of which would require a DNS 2035 lookup and corresponding CPU time to verify the message. This 2036 could be an attack on the domain that receives the message, by 2037 slowing down the verifier by requiring it to do a large number of 2038 DNS lookups and/or signature verifications. It could also be an 2039 attack against the domains listed in the signatures, essentially 2040 by enlisting innocent verifiers in launching an attack against the 2041 DNS servers of the actual victim. 2043 In the following description, text reading "return status 2044 (explanation)" (where "status" is one of "PERMFAIL" or "TEMPFAIL") 2045 means that the verifier MUST immediately cease processing that 2046 signature. The verifier SHOULD proceed to the next signature, if any 2047 is present, and completely ignore the bad signature. If the status 2048 is "PERMFAIL", the signature failed and should not be reconsidered. 2049 If the status is "TEMPFAIL", the signature could not be verified at 2050 this time but may be tried again later. A verifier MAY either defer 2051 the message for later processing, perhaps by queueing it locally or 2052 issuing a 451/4.7.5 SMTP reply, or try another signature; if no good 2053 signature is found and any of the signatures resulted in a TEMPFAIL 2054 status, the verifier MAY save the message for later processing. The 2055 "(explanation)" is not normative text; it is provided solely for 2056 clarification. 2058 Verifiers SHOULD ignore any DKIM-Signature header fields where the 2059 signature does not validate. Verifiers that are prepared to validate 2060 multiple signature header fields SHOULD proceed to the next signature 2061 header field, should it exist. However, verifiers MAY make note of 2062 the fact that an invalid signature was present for consideration at a 2063 later step. 2065 INFORMATIVE NOTE: The rationale of this requirement is to permit 2066 messages that have invalid signatures but also a valid signature 2067 to work. For example, a mailing list exploder might opt to leave 2068 the original submitter signature in place even though the exploder 2069 knows that it is modifying the message in some way that will break 2070 that signature, and the exploder inserts its own signature. In 2071 this case, the message should succeed even in the presence of the 2072 known-broken signature. 2074 For each signature to be validated, the following steps should be 2075 performed in such a manner as to produce a result that is 2076 semantically equivalent to performing them in the indicated order. 2078 6.1.1. Validate the Signature Header Field 2080 Implementers MUST meticulously validate the format and values in the 2081 DKIM-Signature header field; any inconsistency or unexpected values 2082 MUST cause the header field to be completely ignored and the verifier 2083 to return PERMFAIL (signature syntax error). Being "liberal in what 2084 you accept" is definitely a bad strategy in this security context. 2085 Note however that this does not include the existence of unknown tags 2086 in a DKIM-Signature header field, which are explicitly permitted. 2087 Verifiers MUST ignore DKIM-Signature header fields with a "v=" tag 2088 that is inconsistent with this specification and return PERMFAIL 2089 (incompatible version). 2091 INFORMATIVE IMPLEMENTATION NOTE: An implementation may, of course, 2092 choose to also verify signatures generated by older versions of 2093 this specification. 2095 If any tag listed as "required" in Section 3.5 is omitted from the 2096 DKIM-Signature header field, the verifier MUST ignore the DKIM- 2097 Signature header field and return PERMFAIL (signature missing 2098 required tag). 2100 INFORMATIONAL NOTE: The tags listed as required in Section 3.5 are 2101 "v=", "a=", "b=", "bh=", "d=", "h=", and "s=". Should there be a 2102 conflict between this note and Section 3.5, Section 3.5 is 2103 normative. 2105 If the DKIM-Signature header field does not contain the "i=" tag, the 2106 verifier MUST behave as though the value of that tag were "@d", where 2107 "d" is the value from the "d=" tag. 2109 Verifiers MUST confirm that the domain specified in the "d=" tag is 2110 the same as or a parent domain of the domain part of the "i=" tag. 2111 If not, the DKIM-Signature header field MUST be ignored and the 2112 verifier should return PERMFAIL (domain mismatch). 2114 If the "h=" tag does not include the From header field, the verifier 2115 MUST ignore the DKIM-Signature header field and return PERMFAIL (From 2116 field not signed). 2118 Verifiers MAY ignore the DKIM-Signature header field and return 2119 PERMFAIL (signature expired) if it contains an "x=" tag and the 2120 signature has expired. 2122 Verifiers MAY ignore the DKIM-Signature header field if the domain 2123 used by the signer in the "d=" tag is not associated with a valid 2124 signing entity. For example, signatures with "d=" values such as 2125 "com" and "co.uk" may be ignored. The list of unacceptable domains 2126 SHOULD be configurable. 2128 Verifiers MAY ignore the DKIM-Signature header field and return 2129 PERMFAIL (unacceptable signature header) for any other reason, for 2130 example, if the signature does not sign header fields that the 2131 verifier views to be essential. As a case in point, if MIME header 2132 fields are not signed, certain attacks may be possible that the 2133 verifier would prefer to avoid. 2135 6.1.2. Get the Public Key 2137 The public key for a signature is needed to complete the verification 2138 process. The process of retrieving the public key depends on the 2139 query type as defined by the "q=" tag in the DKIM-Signature header 2140 field. Obviously, a public key need only be retrieved if the process 2141 of extracting the signature information is completely successful. 2142 Details of key management and representation are described in 2143 Section 3.6. The verifier MUST validate the key record and MUST 2144 ignore any public key records that are malformed. 2146 NOTE: The use of a wildcard TXT record that covers a queried DKIM 2147 domain name will produce a response to a DKIM query that is 2148 unlikely to be valid DKIM key record. This problem is not 2149 specific to DKIM and applies to many other types of queries. 2150 Client software that processes DNS responses needs to take this 2151 problem into account. 2153 When validating a message, a verifier MUST perform the following 2154 steps in a manner that is semantically the same as performing them in 2155 the order indicated -- in some cases the implementation may 2156 parallelize or reorder these steps, as long as the semantics remain 2157 unchanged: 2159 1. Retrieve the public key as described in Section 3.6 using the 2160 algorithm in the "q=" tag, the domain from the "d=" tag, and the 2161 selector from the "s=" tag. 2163 2. If the query for the public key fails to respond, the verifier 2164 MAY defer acceptance of this email and return TEMPFAIL (key 2165 unavailable). If verification is occurring during the incoming 2166 SMTP session, this MAY be achieved with a 451/4.7.5 SMTP reply 2167 code. Alternatively, the verifier MAY store the message in the 2168 local queue for later trial or ignore the signature. Note that 2169 storing a message in the local queue is subject to denial-of- 2170 service attacks. 2172 3. If the query for the public key fails because the corresponding 2173 key record does not exist, the verifier MUST immediately return 2174 PERMFAIL (no key for signature). 2176 4. If the query for the public key returns multiple key records, the 2177 verifier may choose one of the key records or may cycle through 2178 the key records performing the remainder of these steps on each 2179 record at the discretion of the implementer. The order of the 2180 key records is unspecified. If the verifier chooses to cycle 2181 through the key records, then the "return ..." wording in the 2182 remainder of this section means "try the next key record, if any; 2183 if none, return to try another signature in the usual way". 2185 5. If the result returned from the query does not adhere to the 2186 format defined in this specification, the verifier MUST ignore 2187 the key record and return PERMFAIL (key syntax error). Verifiers 2188 are urged to validate the syntax of key records carefully to 2189 avoid attempted attacks. In particular, the verifier MUST ignore 2190 keys with a version code ("v=" tag) that they do not implement. 2192 6. If the "h=" tag exists in the public key record and the hash 2193 algorithm implied by the "a=" tag in the DKIM-Signature header 2194 field is not included in the contents of the "h=" tag, the 2195 verifier MUST ignore the key record and return PERMFAIL 2196 (inappropriate hash algorithm). 2198 7. If the public key data (the "p=" tag) is empty, then this key has 2199 been revoked and the verifier MUST treat this as a failed 2200 signature check and return PERMFAIL (key revoked). There is no 2201 defined semantic difference between a key that has been revoked 2202 and a key record that has been removed. 2204 8. If the public key data is not suitable for use with the algorithm 2205 and key types defined by the "a=" and "k=" tags in the DKIM- 2206 Signature header field, the verifier MUST immediately return 2207 PERMFAIL (inappropriate key algorithm). 2209 6.1.3. Compute the Verification 2211 Given a signer and a public key, verifying a signature consists of 2212 actions semantically equivalent to the following steps. 2214 1. Based on the algorithm defined in the "c=" tag, the body length 2215 specified in the "l=" tag, and the header field names in the "h=" 2216 tag, prepare a canonicalized version of the message as is 2217 described in Section 3.7 (note that this version does not 2218 actually need to be instantiated). When matching header field 2219 names in the "h=" tag against the actual message header field, 2220 comparisons MUST be case-insensitive. 2222 2. Based on the algorithm indicated in the "a=" tag, compute the 2223 message hashes from the canonical copy as described in 2224 Section 3.7. 2226 3. Verify that the hash of the canonicalized message body computed 2227 in the previous step matches the hash value conveyed in the "bh=" 2228 tag. If the hash does not match, the verifier SHOULD ignore the 2229 signature and return PERMFAIL (body hash did not verify). 2231 4. Using the signature conveyed in the "b=" tag, verify the 2232 signature against the header hash using the mechanism appropriate 2233 for the public key algorithm described in the "a=" tag. If the 2234 signature does not validate, the verifier SHOULD ignore the 2235 signature and return PERMFAIL (signature did not verify). 2237 5. Otherwise, the signature has correctly verified. 2239 INFORMATIVE IMPLEMENTER'S NOTE: Implementations might wish to 2240 initiate the public-key query in parallel with calculating the 2241 hash as the public key is not needed until the final decryption is 2242 calculated. Implementations may also verify the signature on the 2243 message header before validating that the message hash listed in 2244 the "bh=" tag in the DKIM-Signature header field matches that of 2245 the actual message body; however, if the body hash does not match, 2246 the entire signature must be considered to have failed. 2248 A body length specified in the "l=" tag of the signature limits the 2249 number of bytes of the body passed to the verification algorithm. 2250 All data beyond that limit is not validated by DKIM. Hence, 2251 verifiers might treat a message that contains bytes beyond the 2252 indicated body length with suspicion, such as by truncating the 2253 message at the indicated body length, declaring the signature invalid 2254 (e.g., by returning PERMFAIL (unsigned content)), or conveying the 2255 partial verification to the policy module. 2257 INFORMATIVE IMPLEMENTATION NOTE: Verifiers that truncate the body 2258 at the indicated body length might pass on a malformed MIME 2259 message if the signer used the "N-4" trick (omitting the final 2260 "--CRLF") described in the informative note in Section 3.4.5. 2261 Such verifiers may wish to check for this case and include a 2262 trailing "--CRLF" to avoid breaking the MIME structure. A simple 2263 way to achieve this might be to append "--CRLF" to any "multipart" 2264 message with a body length; if the MIME structure is already 2265 correctly formed, this will appear in the postlude and will not be 2266 displayed to the end user. 2268 6.2. Communicate Verification Results 2270 Verifiers wishing to communicate the results of verification to other 2271 parts of the mail system may do so in whatever manner they see fit. 2272 For example, implementations might choose to add an email header 2273 field to the message before passing it on. Any such header field 2274 SHOULD be inserted before any existing DKIM-Signature or preexisting 2275 authentication status header fields in the header field block. The 2276 Authentication-Results: header field ([RFC5451]) MAY be used for this 2277 purpose. 2279 INFORMATIVE ADVICE to MUA filter writers: Patterns intended to 2280 search for results header fields to visibly mark authenticated 2281 mail for end users should verify that such header field was added 2282 by the appropriate verifying domain and that the verified identity 2283 matches the author identity that will be displayed by the MUA. In 2284 particular, MUA filters should not be influenced by bogus results 2285 header fields added by attackers. To circumvent this attack, 2286 verifiers may wish to delete existing results header fields after 2287 verification and before adding a new header field. 2289 6.3. Interpret Results/Apply Local Policy 2291 It is beyond the scope of this specification to describe what actions 2292 an Identity Assessor can make, but mail carrying a validated SDID 2293 presents an opportunity to an Identity Assessor that unauthenticated 2294 email does not. Specifically, an authenticated email creates a 2295 predictable identifier by which other decisions can reliably be 2296 managed, such as trust and reputation. Conversely, unauthenticated 2297 email lacks a reliable identifier that can be used to assign trust 2298 and reputation. It is reasonable to treat unauthenticated email as 2299 lacking any trust and having no positive reputation. 2301 In general, verifiers SHOULD NOT reject messages solely on the basis 2302 of a lack of signature or an unverifiable signature; such rejection 2303 would cause severe interoperability problems. However, if the 2304 verifier does opt to reject such messages (for example, when 2305 communicating with a peer who, by prior agreement, agrees to only 2306 send signed messages), and the verifier runs synchronously with the 2307 SMTP session and a signature is missing or does not verify, the MTA 2308 SHOULD use a 550/5.7.x reply code. 2310 If it is not possible to fetch the public key, perhaps because the 2311 key server is not available, a temporary failure message MAY be 2312 generated using a 451/4.7.5 reply code, such as: 2313 451 4.7.5 Unable to verify signature - key server unavailable 2315 Temporary failures such as inability to access the key server or 2316 other external service are the only conditions that SHOULD use a 4xx 2317 SMTP reply code. In particular, cryptographic signature verification 2318 failures MUST NOT return 4xx SMTP replies. 2320 Once the signature has been verified, that information MUST be 2321 conveyed to the Identity Assessor (such as an explicit allow/ 2322 whitelist and reputation system) and/or to the end user. If the SDID 2323 is not the same as the address in the From: header field, the mail 2324 system SHOULD take pains to ensure that the actual SDID is clear to 2325 the reader. 2327 The verifier MAY treat unsigned header fields with extreme 2328 skepticism, including marking them as untrusted or even deleting them 2329 before display to the end user. 2331 While the symptoms of a failed verification are obvious -- the 2332 signature doesn't verify -- establishing the exact cause can be more 2333 difficult. If a selector cannot be found, is that because the 2334 selector has been removed, or was the value changed somehow in 2335 transit? If the signature line is missing, is that because it was 2336 never there, or was it removed by an overzealous filter? For 2337 diagnostic purposes, the exact reason why the verification fails 2338 SHOULD be made available to the policy module and possibly recorded 2339 in the system logs. If the email cannot be verified, then it SHOULD 2340 be rendered the same as all unverified email regardless of whether or 2341 not it looks like it was signed. 2343 7. IANA Considerations 2345 DKIM has registered namespaces with IANA. In all cases, new values 2346 are assigned only for values that have been documented in a published 2347 RFC that has IETF Consensus [RFC5226]. 2349 This memo updates these registries as described below. Of note is 2350 the addition of a new "status" column. All registrations into these 2351 namespaces MUST include the name being registered, the document in 2352 which it was registered or updated, and an indication of its current 2353 status which MUST be one of "active" (in current use) or "historic" 2354 (no longer in current use). 2356 7.1. DKIM-Signature Tag Specifications 2358 A DKIM-Signature provides for a list of tag specifications. IANA has 2359 established the DKIM-Signature Tag Specification Registry for tag 2360 specifications that can be used in DKIM-Signature fields. 2362 The updated entries in the registry comprise: 2364 +------+-----------------+--------+ 2365 | TYPE | REFERENCE | STATUS | 2366 +------+-----------------+--------+ 2367 | v | (this document) | active | 2368 | a | (this document) | active | 2369 | b | (this document) | active | 2370 | bh | (this document) | active | 2371 | c | (this document) | active | 2372 | d | (this document) | active | 2373 | h | (this document) | active | 2374 | i | (this document) | active | 2375 | l | (this document) | active | 2376 | q | (this document) | active | 2377 | s | (this document) | active | 2378 | t | (this document) | active | 2379 | x | (this document) | active | 2380 | z | (this document) | active | 2381 +------+-----------------+--------+ 2383 Table 1: DKIM-Signature Tag Specification Registry Updated Values 2385 7.2. DKIM-Signature Query Method Registry 2387 The "q=" tag-spec (specified in Section 3.5) provides for a list of 2388 query methods. 2390 IANA has established the DKIM-Signature Query Method Registry for 2391 mechanisms that can be used to retrieve the key that will permit 2392 validation processing of a message signed using DKIM. 2394 The updated entry in the registry comprises: 2396 +------+--------+-----------------+--------+ 2397 | TYPE | OPTION | REFERENCE | STATUS | 2398 +------+--------+-----------------+--------+ 2399 | dns | txt | (this document) | active | 2400 +------+--------+-----------------+--------+ 2402 DKIM-Signature Query Method Registry Updated Values 2404 7.3. DKIM-Signature Canonicalization Registry 2406 The "c=" tag-spec (specified in Section 3.5) provides for a specifier 2407 for canonicalization algorithms for the header and body of the 2408 message. 2410 IANA has established the DKIM-Signature Canonicalization Algorithm 2411 Registry for algorithms for converting a message into a canonical 2412 form before signing or verifying using DKIM. 2414 The updated entries in the header registry comprise: 2416 +---------+-----------------+--------+ 2417 | TYPE | REFERENCE | STATUS | 2418 +---------+-----------------+--------+ 2419 | simple | (this document) | active | 2420 | relaxed | (this document) | active | 2421 +---------+-----------------+--------+ 2423 DKIM-Signature Header Canonicalization Algorithm Registry 2424 Updated Values 2426 The updated entries in the body registry comprise: 2428 +---------+-----------------+--------+ 2429 | TYPE | REFERENCE | STATUS | 2430 +---------+-----------------+--------+ 2431 | simple | (this document) | active | 2432 | relaxed | (this document) | active | 2433 +---------+-----------------+--------+ 2435 DKIM-Signature Body Canonicalization Algorithm Registry 2436 Updated Values 2438 7.4. _domainkey DNS TXT Record Tag Specifications 2440 A _domainkey DNS TXT record provides for a list of tag 2441 specifications. IANA has established the DKIM _domainkey DNS TXT Tag 2442 Specification Registry for tag specifications that can be used in DNS 2443 TXT Records. 2445 The updated entries in the registry comprise: 2447 +------+-----------------+----------+ 2448 | TYPE | REFERENCE | STATUS | 2449 +------+-----------------+----------+ 2450 | v | (this document) | active | 2451 | g | [RFC4871] | historic | 2452 | h | (this document) | active | 2453 | k | (this document) | active | 2454 | n | (this document) | active | 2455 | p | (this document) | active | 2456 | s | (this document) | active | 2457 | t | (this document) | active | 2458 +------+-----------------+----------+ 2460 DKIM _domainkey DNS TXT Record Tag Specification Registry 2461 Updated Values 2463 7.5. DKIM Key Type Registry 2465 The "k=" (specified in Section 3.6.1) and the "a=" (specified in Section 3.5) tags provide for a list of 2467 mechanisms that can be used to decode a DKIM signature. 2469 IANA has established the DKIM Key Type Registry for such mechanisms. 2471 The updated entry in the registry comprises: 2473 +------+-----------+--------+ 2474 | TYPE | REFERENCE | STATUS | 2475 +------+-----------+--------+ 2476 | rsa | [RFC3447] | active | 2477 +------+-----------+--------+ 2479 DKIM Key Type Updated Values 2481 7.6. DKIM Hash Algorithms Registry 2483 The "h=" (specified in Section 3.6.1) and the "a=" (specified in Section 3.5) tags provide for a list of 2485 mechanisms that can be used to produce a digest of message data. 2487 IANA has established the DKIM Hash Algorithms Registry for such 2488 mechanisms. 2490 The updated entries in the registry comprise: 2492 +--------+-------------------+--------+ 2493 | TYPE | REFERENCE | STATUS | 2494 +--------+-------------------+--------+ 2495 | sha1 | [FIPS-180-2-2002] | active | 2496 | sha256 | [FIPS-180-2-2002] | active | 2497 +--------+-------------------+--------+ 2499 DKIM Hash Algorithms Updated Values 2501 7.7. DKIM Service Types Registry 2503 The "s=" tag (specified in Section 3.6.1) provides for a 2504 list of service types to which this selector may apply. 2506 IANA has established the DKIM Service Types Registry for service 2507 types. 2509 The updated entries in the registry comprise: 2511 +-------+-----------------+--------+ 2512 | TYPE | REFERENCE | STATUS | 2513 +-------+-----------------+--------+ 2514 | email | (this document) | active | 2515 | * | (this document) | active | 2516 +-------+-----------------+--------+ 2518 DKIM Service Types Registry Updated Values 2520 7.8. DKIM Selector Flags Registry 2522 The "t=" tag (specified in Section 3.6.1) provides for a 2523 list of flags to modify interpretation of the selector. 2525 IANA has established the DKIM Selector Flags Registry for additional 2526 flags. 2528 The updated entries in the registry comprise: 2530 +------+-----------------+--------+ 2531 | TYPE | REFERENCE | STATUS | 2532 +------+-----------------+--------+ 2533 | y | (this document) | active | 2534 | s | (this document) | active | 2535 +------+-----------------+--------+ 2537 DKIM Selector Flags Registry Updated Values 2539 7.9. DKIM-Signature Header Field 2541 IANA has added DKIM-Signature to the "Permanent Message Header 2542 Fields" registry (see [RFC3864]) for the "mail" protocol, using this 2543 document as the reference. 2545 8. Security Considerations 2547 It has been observed that any mechanism that is introduced that 2548 attempts to stem the flow of spam is subject to intensive attack. 2549 DKIM needs to be carefully scrutinized to identify potential attack 2550 vectors and the vulnerability to each. See also [RFC4686]. 2552 8.1. Misuse of Body Length Limits ("l=" Tag) 2554 Body length limits (in the form of the "l=" tag) are subject to 2555 several potential attacks. 2557 8.1.1. Addition of New MIME Parts to Multipart/* 2559 If the body length limit does not cover a closing MIME multipart 2560 section (including the trailing "--CRLF" portion), then it is 2561 possible for an attacker to intercept a properly signed multipart 2562 message and add a new body part. Depending on the details of the 2563 MIME type and the implementation of the verifying MTA and the 2564 receiving MUA, this could allow an attacker to change the information 2565 displayed to an end user from an apparently trusted source. 2567 For example, if attackers can append information to a "text/html" 2568 body part, they may be able to exploit a bug in some MUAs that 2569 continue to read after a "" marker, and thus display HTML text 2570 on top of already displayed text. If a message has a "multipart/ 2571 alternative" body part, they might be able to add a new body part 2572 that is preferred by the displaying MUA. 2574 8.1.2. Addition of new HTML content to existing content 2576 Several receiving MUA implementations do not cease display after a 2577 """" tag. In particular, this allows attacks involving 2578 overlaying images on top of existing text. 2580 INFORMATIVE EXAMPLE: Appending the following text to an existing, 2581 properly closed message will in many MUAs result in inappropriate 2582 data being rendered on top of existing, correct data: 2584
2585
2587 8.2. Misappropriated Private Key 2589 If the private key for a user is resident on their computer and is 2590 not protected by an appropriately secure mechanism, it is possible 2591 for malware to send mail as that user and any other user sharing the 2592 same private key. The malware would not, however, be able to 2593 generate signed spoofs of other signers' addresses, which would aid 2594 in identification of the infected user and would limit the 2595 possibilities for certain types of attacks involving socially 2596 engineered messages. This threat applies mainly to MUA-based 2597 implementations; protection of private keys on servers can be easily 2598 achieved through the use of specialized cryptographic hardware. 2600 A larger problem occurs if malware on many users' computers obtains 2601 the private keys for those users and transmits them via a covert 2602 channel to a site where they can be shared. The compromised users 2603 would likely not know of the misappropriation until they receive 2604 "bounce" messages from messages they are purported to have sent. 2605 Many users might not understand the significance of these bounce 2606 messages and would not take action. 2608 One countermeasure is to use a user-entered passphrase to encrypt the 2609 private key, although users tend to choose weak passphrases and often 2610 reuse them for different purposes, possibly allowing an attack 2611 against DKIM to be extended into other domains. Nevertheless, the 2612 decoded private key might be briefly available to compromise by 2613 malware when it is entered, or might be discovered via keystroke 2614 logging. The added complexity of entering a passphrase each time one 2615 sends a message would also tend to discourage the use of a secure 2616 passphrase. 2618 A somewhat more effective countermeasure is to send messages through 2619 an outgoing MTA that can authenticate the submitter using existing 2620 techniques (e.g., SMTP Authentication), possibly validate the message 2621 itself (e.g., verify that the header is legitimate and that the 2622 content passes a spam content check), and sign the message using a 2623 key appropriate for the submitter address. Such an MTA can also 2624 apply controls on the volume of outgoing mail each user is permitted 2625 to originate in order to further limit the ability of malware to 2626 generate bulk email. 2628 8.3. Key Server Denial-of-Service Attacks 2630 Since the key servers are distributed (potentially separate for each 2631 domain), the number of servers that would need to be attacked to 2632 defeat this mechanism on an Internet-wide basis is very large. 2633 Nevertheless, key servers for individual domains could be attacked, 2634 impeding the verification of messages from that domain. This is not 2635 significantly different from the ability of an attacker to deny 2636 service to the mail exchangers for a given domain, although it 2637 affects outgoing, not incoming, mail. 2639 A variation on this attack is that if a very large amount of mail 2640 were to be sent using spoofed addresses from a given domain, the key 2641 servers for that domain could be overwhelmed with requests. However, 2642 given the low overhead of verification compared with handling of the 2643 email message itself, such an attack would be difficult to mount. 2645 8.4. Attacks Against the DNS 2647 Since the DNS is a required binding for key services, specific 2648 attacks against the DNS must be considered. 2650 While the DNS is currently insecure [RFC3833], these security 2651 problems are the motivation behind DNS Security (DNSSEC) [RFC4033], 2652 and all users of the DNS will reap the benefit of that work. 2654 DKIM is only intended as a "sufficient" method of proving 2655 authenticity. It is not intended to provide strong cryptographic 2656 proof about authorship or contents. Other technologies such as 2657 OpenPGP [RFC4880] and S/MIME [RFC5751] address those requirements. 2659 A second security issue related to the DNS revolves around the 2660 increased DNS traffic as a consequence of fetching selector-based 2661 data as well as fetching signing domain policy. Widespread 2662 deployment of DKIM will result in a significant increase in DNS 2663 queries to the claimed signing domain. In the case of forgeries on a 2664 large scale, DNS servers could see a substantial increase in queries. 2666 A specific DNS security issue that should be considered by DKIM 2667 verifiers is the name chaining attack described in Section 2.3 of 2668 [RFC3833]. A DKIM verifier, while verifying a DKIM-Signature header 2669 field, could be prompted to retrieve a key record of an attacker's 2670 choosing. This threat can be minimized by ensuring that name 2671 servers, including recursive name servers, used by the verifier 2672 enforce strict checking of "glue" and other additional information in 2673 DNS responses and are therefore not vulnerable to this attack. 2675 8.5. Replay Attacks 2677 In this attack, a spammer sends a message to be spammed to an 2678 accomplice, which results in the message being signed by the 2679 originating MTA. The accomplice resends the message, including the 2680 original signature, to a large number of recipients, possibly by 2681 sending the message to many compromised machines that act as MTAs. 2682 The messages, not having been modified by the accomplice, have valid 2683 signatures. 2685 Partial solutions to this problem involve the use of reputation 2686 services to convey the fact that the specific email address is being 2687 used for spam and that messages from that signer are likely to be 2688 spam. This requires a real-time detection mechanism in order to 2689 react quickly enough. However, such measures might be prone to 2690 abuse, if for example an attacker resent a large number of messages 2691 received from a victim in order to make them appear to be a spammer. 2693 Large verifiers might be able to detect unusually large volumes of 2694 mails with the same signature in a short time period. Smaller 2695 verifiers can get substantially the same volume of information via 2696 existing collaborative systems. 2698 8.6. Limits on Revoking Keys 2700 When a large domain detects undesirable behavior on the part of one 2701 of its users, it might wish to revoke the key used to sign that 2702 user's messages in order to disavow responsibility for messages that 2703 have not yet been verified or that are the subject of a replay 2704 attack. However, the ability of the domain to do so can be limited 2705 if the same key, for scalability reasons, is used to sign messages 2706 for many other users. Mechanisms for explicitly revoking keys on a 2707 per-address basis have been proposed but require further study as to 2708 their utility and the DNS load they represent. 2710 8.7. Intentionally Malformed Key Records 2712 It is possible for an attacker to publish key records in DNS that are 2713 intentionally malformed, with the intent of causing a denial-of- 2714 service attack on a non-robust verifier implementation. The attacker 2715 could then cause a verifier to read the malformed key record by 2716 sending a message to one of its users referencing the malformed 2717 record in a (not necessarily valid) signature. Verifiers MUST 2718 thoroughly verify all key records retrieved from the DNS and be 2719 robust against intentionally as well as unintentionally malformed key 2720 records. 2722 8.8. Intentionally Malformed DKIM-Signature Header Fields 2724 Verifiers MUST be prepared to receive messages with malformed DKIM- 2725 Signature header fields, and thoroughly verify the header field 2726 before depending on any of its contents. 2728 8.9. Information Leakage 2730 An attacker could determine when a particular signature was verified 2731 by using a per-message selector and then monitoring their DNS traffic 2732 for the key lookup. This would act as the equivalent of a "web bug" 2733 for verification time rather than when the message was read. 2735 8.10. Remote Timing Attacks 2737 In some cases it may be possible to extract private keys using a 2738 remote timing attack [BONEH03]. Implementations should consider 2739 obfuscating the timing to prevent such attacks. 2741 8.11. Reordered Header Fields 2743 Existing standards allow intermediate MTAs to reorder header fields. 2744 If a signer signs two or more header fields of the same name, this 2745 can cause spurious verification errors on otherwise legitimate 2746 messages. In particular, signers that sign any existing DKIM- 2747 Signature fields run the risk of having messages incorrectly fail to 2748 verify. 2750 8.12. RSA Attacks 2752 An attacker could create a large RSA signing key with a small 2753 exponent, thus requiring that the verification key have a large 2754 exponent. This will force verifiers to use considerable computing 2755 resources to verify the signature. Verifiers might avoid this attack 2756 by refusing to verify signatures that reference selectors with public 2757 keys having unreasonable exponents. 2759 In general, an attacker might try to overwhelm a verifier by flooding 2760 it with messages requiring verification. This is similar to other 2761 MTA denial-of-service attacks and should be dealt with in a similar 2762 fashion. 2764 8.13. Inappropriate Signing by Parent Domains 2766 The trust relationship described in Section 3.9 could conceivably be 2767 used by a parent domain to sign messages with identities in a 2768 subdomain not administratively related to the parent. For example, 2769 the ".com" registry could create messages with signatures using an 2770 "i=" value in the example.com domain. There is no general solution 2771 to this problem, since the administrative cut could occur anywhere in 2772 the domain name. For example, in the domain "example.podunk.ca.us" 2773 there are three administrative cuts (podunk.ca.us, ca.us, and us), 2774 any of which could create messages with an identity in the full 2775 domain. 2777 INFORMATIVE NOTE: This is considered an acceptable risk for the 2778 same reason that it is acceptable for domain delegation. For 2779 example, in the example above any of the domains could potentially 2780 simply delegate "example.podunk.ca.us" to a server of their choice 2781 and completely replace all DNS-served information. Note that a 2782 verifier MAY ignore signatures that come from an unlikely domain 2783 such as ".com", as discussed in Section 6.1.1. 2785 8.14. Attacks Involving Addition of Header Fields 2787 Many email implementations do not enforce [RFC5322] with strictness. 2788 As discussed in Section 5.3 DKIM processing is predicated on a valid 2789 mail message as its input. However, DKIM implementers should be 2790 aware of the potential effect of having loose enforcement by email 2791 components interacting with DKIM modules. 2793 For example, a message with multiple From: header fields violates 2794 Section 3.6 of [RFC5322]. With the intent of providing a better user 2795 experience, many agents tolerate these violations and deliver the 2796 message anyway. An MUA then might elect to render to the user the 2797 value of the last, or "top", From: field. This may also be done 2798 simply out of the expectation that there is only one, where a "find 2799 first" algorithm would have the same result. Such code in an MUA can 2800 be exploited to fool the user if it is also known that the other 2801 From: field is the one checked by arriving message filters. Such is 2802 the case with DKIM; although the From: field must be signed, a 2803 malformed message bearing more than one From: field might only have 2804 the first ("bottom") one signed, in an attempt to show the message 2805 with some "DKIM passed" annotation while also rendering the From: 2806 field that was not authenticated. (This can also be taken as a 2807 demonstration that DKIM is not designed to support author 2808 validation.) 2810 To resist this specific attack the signed header field list can 2811 include an additional reference for each field that was present at 2812 signing. For example, a proper message with one From: field could be 2813 signed using "h=From:From:..." Due to the way header fields are 2814 canonicalized for input to the hash function, the extra field 2815 references will prevent instances of the cited fields from being 2816 added after signing, as doing so would render the signature invalid. 2818 The From: field is used above to illustrate this issue, but it is 2819 only one of several fields that Section 3.6 of [RFC5322] constrains 2820 in this way. In reality any agent that forgives malformations, or is 2821 careless about identifying which parts of a message were 2822 authenticated, is open to exploitation. 2824 8.15. Malformed Inputs 2826 DKIM allows additional header fields to be added to a signed message 2827 without breaking the signature. This tolerance can be abused, for 2828 example in a replay attack. The attack is accomplished by creating 2829 additional instances of header fields to an already signed message, 2830 without breaking the signature. These then might be displayed to the 2831 end user or are used as filtering input. Salient fields could 2832 include From: and Subject:, 2834 The resulting message violates section 3.6 of [RFC5322]. The way 2835 such input will be handled and displayed by an MUA is unpredictable, 2836 but in some cases it might display the newly added header fields 2837 rather than those that are part of the originally signed message 2838 alongside some "valid DKIM signature" annotation. This might allow 2839 an attacker to replay a previously sent, signed message with a 2840 different Subject:, From: or To: field. 2842 However, [RFC5322] also tolerates obsolete message syntax, which does 2843 allow things like multiple From: fields on messages. The 2844 implementation of DKIM thus potentially creates a more stringent 2845 layer of expectation regarding the meaning of an identity, while that 2846 additional meaning is either obscured from or incorrectly presented 2847 to an end user in this context. 2849 Implementers need to consider this possibility when designing their 2850 input handling functions. Outright rejection of messages that 2851 violate Section 3.6 of [RFC5322] will interfere with delivery of 2852 legacy formats. Instead, given such input, a signing module could 2853 return an error rather than generate a signature; a verifying module 2854 might return a syntax error code or arrange not to return a positive 2855 result even if the signature technically validates. 2857 Senders concerned that their messages might be particularly 2858 vulnerable to this sort of attack and who do not wish to rely on 2859 receiver filtering of invalid messages can ensure that adding 2860 additional header fields will break the DKIM signature by including 2861 two copies of the header fields about which they are concerned in the 2862 signature (e.g. "h= ... from:from:to:to:subject:subject ..."). See 2863 Sections 3.5 and 5.4 for further discussion of this mechanism. 2865 Specific validity rules for all known header fields can be gleaned 2866 from the IANA "Permanent Header Field Registry" and the reference 2867 documents it identifies. 2869 9. References 2871 9.1. Normative References 2873 [FIPS-180-2-2002] 2874 U.S. Department of Commerce, "Secure Hash Standard", FIPS 2875 PUB 180-2, August 2002. 2877 [ITU-X660-1997] 2878 "Information Technology - ASN.1 encoding rules: 2879 Specification of Basic Encoding Rules (BER), Canonical 2880 Encoding Rules (CER) and Distinguished Encoding Rules 2881 (DER)", 1997. 2883 [RFC1034] Mockapetris, P., "DOMAIN NAMES - CONCEPTS AND FACILITIES", 2884 RFC 1034, November 1987. 2886 [RFC2045] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2887 Extensions (MIME) Part One: Format of Internet Message 2888 Bodies", RFC 2045, November 1996. 2890 [RFC2047] Moore, K., "MIME (Multipurpose Internet Mail Extensions) 2891 Part Three: Message Header Extensions for Non-ASCII Text", 2892 RFC 2047, November 1996. 2894 [RFC2049] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2895 Extensions (MIME) Part Five: Conformance Criteria and 2896 Examples", RFC 2049, November 1996. 2898 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2899 Requirement Levels", BCP 14, RFC 2119, March 1997. 2901 [RFC3447] Jonsson, J. and B. Kaliski, "Public-Key Cryptography 2902 Standards (PKCS) #1: RSA Cryptography Specifications 2903 Version 2.1", RFC 3447, February 2003. 2905 [RFC3490] Faltstrom, P., Hoffman, P., and A. Costello, 2906 "Internationalizing Domain Names in Applications (IDNA)", 2907 RFC 3490, March 2003. 2909 [RFC5234] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax 2910 Specifications: ABNF", RFC 4234, January 2008. 2912 [RFC5321] Klensin, J., "Simple Mail Transfer Protocol", RFC 5321, 2913 October 2008. 2915 [RFC5322] Resnick, P., "Internet Message Format", RFC 5322, 2916 October 2008. 2918 [RFC5598] Crocker, D., "Internet Mail Architecture", RFC 5598, 2919 July 2009. 2921 9.2. Informative References 2923 [BONEH03] "Remote Timing Attacks are Practical", Proceedings 12th 2924 USENIX Security Symposium, 2003. 2926 [RFC1847] Galvin, J., Murphy, S., Crocker, S., and N. Freed, 2927 "Security Multiparts for MIME: Multipart/Signed and 2928 Multipart/Encrypted", RFC 1847, October 1995. 2930 [RFC3766] Orman, H. and P. Hoffman, "Determining Strengths For 2931 Public Keys Used For Exchanging Symmetric Keys", BCP 86, 2932 RFC 3766, April 2004. 2934 [RFC3833] Atkins, D. and R. Austein, "Threat Analysis of the Domain 2935 Name System (DNS)", RFC 3833, August 2004. 2937 [RFC3864] Klyne, G., Nottingham, M., and J. Mogul, "Registration 2938 Procedures for Message Header Fields", BCP 90, RFC 3864, 2939 September 2004. 2941 [RFC4033] Arends, R., Austein, R., Larson, M., Massey, D., and S. 2942 Rose, "DNS Security Introduction and Requirements", 2943 RFC 4033, March 2005. 2945 [RFC4409] Gellens, R. and J. Klensin, "Message Submission for Mail", 2946 RFC 4409, April 2006. 2948 [RFC4686] Fenton, J., "Analysis of Threats Motivating DomainKeys 2949 Identified Mail (DKIM)", RFC 4686, September 2006. 2951 [RFC4870] Delany, M., "Domain-Based Email Authentication Using 2952 Public Keys Advertised in the DNS (DomainKeys)", RFC 4870, 2953 May 2007. 2955 [RFC4871] Allman, E., Callas, J., Delany, M., Libbey, M., Fenton, 2956 J., and M. Thomas, "DomainKeys Identified Mail (DKIM) 2957 Signatures", RFC 4871, May 2007. 2959 [RFC4880] Callas, J., Donnerhacke, L., Finney, H., and R. Thayer, 2960 "OpenPGP Message Format", RFC 4880, November 2007. 2962 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 2963 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 2964 May 2008. 2966 [RFC5451] Kucherawy, M., "Message Header Field for Indicating 2967 Message Authentication Status", RFC 5451, April 2009. 2969 [RFC5751] Ramsdell, B., "Secure/Multipurpose Internet Mail 2970 Extensions (S/MIME) Version 3.1 Message Specification", 2971 RFC 5751, January 2010. 2973 Appendix A. Example of Use (INFORMATIVE) 2975 This section shows the complete flow of an email from submission to 2976 final delivery, demonstrating how the various components fit 2977 together. The key used in this example is shown in Appendix C. 2979 A.1. The User Composes an Email 2980 From: Joe SixPack 2981 To: Suzie Q 2982 Subject: Is dinner ready? 2983 Date: Fri, 11 Jul 2003 21:00:37 -0700 (PDT) 2984 Message-ID: <;20030712040037.46341.5F8J@football.example.com> 2986 Hi. 2988 We lost the game. Are you hungry yet? 2990 Joe. 2992 Figure 1: The User Composes an Email 2994 A.2. The Email is Signed 2996 This email is signed by the example.com outbound email server and now 2997 looks like this: 2998 DKIM-Signature: v=1; a=rsa-sha256; s=brisbane; d=example.com; 2999 c=simple/simple; q=dns/txt; i=joe@football.example.com; 3000 h=Received : From : To : Subject : Date : Message-ID; 3001 bh=2jUSOH9NhtVGCQWNr9BrIAPreKQjO6Sn7XIkfJVOzv8=; 3002 b=AuUoFEfDxTDkHlLXSZEpZj79LICEps6eda7W3deTVFOk4yAUoqOB 3003 4nujc7YopdG5dWLSdNg6xNAZpOPr+kHxt1IrE+NahM6L/LbvaHut 3004 KVdkLLkpVaVVQPzeRDI009SO2Il5Lu7rDNH6mZckBdrIx0orEtZV 3005 4bmp/YzhwvcubU4=; 3006 Received: from client1.football.example.com [192.0.2.1] 3007 by submitserver.example.com with SUBMISSION; 3008 Fri, 11 Jul 2003 21:01:54 -0700 (PDT) 3009 From: Joe SixPack 3010 To: Suzie Q 3011 Subject: Is dinner ready? 3012 Date: Fri, 11 Jul 2003 21:00:37 -0700 (PDT) 3013 Message-ID: <20030712040037.46341.5F8J@football.example.com> 3015 Hi. 3017 We lost the game. Are you hungry yet? 3019 Joe. 3021 The Email is Signed 3023 The signing email server requires access to the private key 3024 associated with the "brisbane" selector to generate this signature. 3026 A.3. The Email Signature is Verified 3028 The signature is normally verified by an inbound SMTP server or 3029 possibly the final delivery agent. However, intervening MTAs can 3030 also perform this verification if they choose to do so. The 3031 verification process uses the domain "example.com" extracted from the 3032 "d=" tag and the selector "brisbane" from the "s=" tag in the DKIM- 3033 Signature header field to form the DNS DKIM query for: 3034 brisbane._domainkey.example.com 3036 Signature verification starts with the physically last Received 3037 header field, the From header field, and so forth, in the order 3038 listed in the "h=" tag. Verification follows with a single CRLF 3039 followed by the body (starting with "Hi."). The email is canonically 3040 prepared for verifying with the "simple" method. The result of the 3041 query and subsequent verification of the signature is stored (in this 3042 example) in the X-Authentication-Results header field line. After 3043 successful verification, the email looks like this: 3044 X-Authentication-Results: shopping.example.net 3045 header.from=joe@football.example.com; dkim=pass 3046 Received: from mout23.football.example.com (192.168.1.1) 3047 by shopping.example.net with SMTP; 3048 Fri, 11 Jul 2003 21:01:59 -0700 (PDT) 3049 DKIM-Signature: v=1; a=rsa-sha256; s=brisbane; d=example.com; 3050 c=simple/simple; q=dns/txt; i=joe@football.example.com; 3051 h=Received : From : To : Subject : Date : Message-ID; 3052 bh=2jUSOH9NhtVGCQWNr9BrIAPreKQjO6Sn7XIkfJVOzv8=; 3053 b=AuUoFEfDxTDkHlLXSZEpZj79LICEps6eda7W3deTVFOk4yAUoqOB 3054 4nujc7YopdG5dWLSdNg6xNAZpOPr+kHxt1IrE+NahM6L/LbvaHut 3055 KVdkLLkpVaVVQPzeRDI009SO2Il5Lu7rDNH6mZckBdrIx0orEtZV 3056 4bmp/YzhwvcubU4=; 3057 Received: from client1.football.example.com [192.0.2.1] 3058 by submitserver.example.com with SUBMISSION; 3059 Fri, 11 Jul 2003 21:01:54 -0700 (PDT) 3060 From: Joe SixPack 3061 To: Suzie Q 3062 Subject: Is dinner ready? 3063 Date: Fri, 11 Jul 2003 21:00:37 -0700 (PDT) 3064 Message-ID: <20030712040037.46341.5F8J@football.example.com> 3066 Hi. 3068 We lost the game. Are you hungry yet? 3070 Joe. 3072 Successful verification 3074 Appendix B. Usage Examples (INFORMATIVE) 3076 DKIM signing and validating can be used in different ways, for 3077 different operational scenarios. This Appendix discusses some common 3078 examples. 3080 NOTE: Descriptions in this Appendix are for informational purposes 3081 only. They describe various ways that DKIM can be used, given 3082 particular constraints and needs. In no case are these examples 3083 intended to be taken as providing explanation or guidance 3084 concerning DKIM specification details, when creating an 3085 implementation. 3087 B.1. Alternate Submission Scenarios 3089 In the most simple scenario, a user's MUA, MSA, and Internet 3090 (boundary) MTA are all within the same administrative environment, 3091 using the same domain name. Therefore, all of the components 3092 involved in submission and initial transfer are related. However, it 3093 is common for two or more of the components to be under independent 3094 administrative control. This creates challenges for choosing and 3095 administering the domain name to use for signing, and for its 3096 relationship to common email identity header fields. 3098 B.1.1. Delegated Business Functions 3100 Some organizations assign specific business functions to discrete 3101 groups, inside or outside the organization. The goal, then, is to 3102 authorize that group to sign some mail, but to constrain what 3103 signatures they can generate. DKIM selectors (the "s=" signature 3104 tag) facilitate this kind of restricted authorization. Examples of 3105 these outsourced business functions are legitimate email marketing 3106 providers and corporate benefits providers. 3108 Here, the delegated group needs to be able to send messages that are 3109 signed, using the email domain of the client company. At the same 3110 time, the client often is reluctant to register a key for the 3111 provider that grants the ability to send messages for arbitrary 3112 addresses in the domain. 3114 There are multiple ways to administer these usage scenarios. In one 3115 case, the client organization provides all of the public query 3116 service (for example, DNS) administration, and in another it uses DNS 3117 delegation to enable all ongoing administration of the DKIM key 3118 record by the delegated group. 3120 If the client organization retains responsibility for all of the DNS 3121 administration, the outsourcing company can generate a key pair, 3122 supplying the public key to the client company, which then registers 3123 it in the query service, using a unique selector. The client company 3124 retains control over the use of the delegated key because it retains 3125 the ability to revoke the key at any time. 3127 If the client wants the delegated group to do the DNS administration, 3128 it can have the domain name that is specified with the selector point 3129 to the provider's DNS server. The provider then creates and 3130 maintains all of the DKIM signature information for that selector. 3131 Hence, the client cannot provide constraints on the Local-part of 3132 addresses that get signed, but it can revoke the provider's signing 3133 rights by removing the DNS delegation record. 3135 B.1.2. PDAs and Similar Devices 3137 PDAs demonstrate the need for using multiple keys per domain. 3138 Suppose that John Doe wanted to be able to send messages using his 3139 corporate email address, jdoe@example.com, and his email device did 3140 not have the ability to make a Virtual Private Network (VPN) 3141 connection to the corporate network, either because the device is 3142 limited or because there are restrictions enforced by his Internet 3143 access provider. If the device was equipped with a private key 3144 registered for jdoe@example.com by the administrator of the 3145 example.com domain, and appropriate software to sign messages, John 3146 could sign the message on the device itself before transmission 3147 through the outgoing network of the access service provider. 3149 B.1.3. Roaming Users 3151 Roaming users often find themselves in circumstances where it is 3152 convenient or necessary to use an SMTP server other than their home 3153 server; examples are conferences and many hotels. In such 3154 circumstances, a signature that is added by the submission service 3155 will use an identity that is different from the user's home system. 3157 Ideally, roaming users would connect back to their home server using 3158 either a VPN or a SUBMISSION server running with SMTP AUTHentication 3159 on port 587. If the signing can be performed on the roaming user's 3160 laptop, then they can sign before submission, although the risk of 3161 further modification is high. If neither of these are possible, 3162 these roaming users will not be able to send mail signed using their 3163 own domain key. 3165 B.1.4. Independent (Kiosk) Message Submission 3167 Stand-alone services, such as walk-up kiosks and web-based 3168 information services, have no enduring email service relationship 3169 with the user, but users occasionally request that mail be sent on 3170 their behalf. For example, a website providing news often allows the 3171 reader to forward a copy of the article to a friend. This is 3172 typically done using the reader's own email address, to indicate who 3173 the author is. This is sometimes referred to as the "Evite problem", 3174 named after the website of the same name that allows a user to send 3175 invitations to friends. 3177 A common way this is handled is to continue to put the reader's email 3178 address in the From header field of the message, but put an address 3179 owned by the email posting site into the Sender header field. The 3180 posting site can then sign the message, using the domain that is in 3181 the Sender field. This provides useful information to the receiving 3182 email site, which is able to correlate the signing domain with the 3183 initial submission email role. 3185 Receiving sites often wish to provide their end users with 3186 information about mail that is mediated in this fashion. Although 3187 the real efficacy of different approaches is a subject for human 3188 factors usability research, one technique that is used is for the 3189 verifying system to rewrite the From header field, to indicate the 3190 address that was verified. For example: From: John Doe via 3191 news@news-site.com . (Note that such rewriting 3192 will break a signature, unless it is done after the verification pass 3193 is complete.) 3195 B.2. Alternate Delivery Scenarios 3197 Email is often received at a mailbox that has an address different 3198 from the one used during initial submission. In these cases, an 3199 intermediary mechanism operates at the address originally used and it 3200 then passes the message on to the final destination. This mediation 3201 process presents some challenges for DKIM signatures. 3203 B.2.1. Affinity Addresses 3205 "Affinity addresses" allow a user to have an email address that 3206 remains stable, even as the user moves among different email 3207 providers. They are typically associated with college alumni 3208 associations, professional organizations, and recreational 3209 organizations with which they expect to have a long-term 3210 relationship. These domains usually provide forwarding of incoming 3211 email, and they often have an associated Web application that 3212 authenticates the user and allows the forwarding address to be 3213 changed. However, these services usually depend on users sending 3214 outgoing messages through their own service providers' MTAs. Hence, 3215 mail that is signed with the domain of the affinity address is not 3216 signed by an entity that is administered by the organization owning 3217 that domain. 3219 With DKIM, affinity domains could use the Web application to allow 3220 users to register per-user keys to be used to sign messages on behalf 3221 of their affinity address. The user would take away the secret half 3222 of the key pair for signing, and the affinity domain would publish 3223 the public half in DNS for access by verifiers. 3225 This is another application that takes advantage of user-level 3226 keying, and domains used for affinity addresses would typically have 3227 a very large number of user-level keys. Alternatively, the affinity 3228 domain could handle outgoing mail, operating a mail submission agent 3229 that authenticates users before accepting and signing messages for 3230 them. This is of course dependent on the user's service provider not 3231 blocking the relevant TCP ports used for mail submission. 3233 B.2.2. Simple Address Aliasing (.forward) 3235 In some cases a recipient is allowed to configure an email address to 3236 cause automatic redirection of email messages from the original 3237 address to another, such as through the use of a Unix .forward file. 3238 In this case, messages are typically redirected by the mail handling 3239 service of the recipient's domain, without modification, except for 3240 the addition of a Received header field to the message and a change 3241 in the envelope recipient address. In this case, the recipient at 3242 the final address' mailbox is likely to be able to verify the 3243 original signature since the signed content has not changed, and DKIM 3244 is able to validate the message signature. 3246 B.2.3. Mailing Lists and Re-Posters 3248 There is a wide range of behaviors in services that take delivery of 3249 a message and then resubmit it. A primary example is with mailing 3250 lists (collectively called "forwarders" below), ranging from those 3251 that make no modification to the message itself, other than to add a 3252 Received header field and change the envelope information, to those 3253 that add header fields, change the Subject header field, add content 3254 to the body (typically at the end), or reformat the body in some 3255 manner. The simple ones produce messages that are quite similar to 3256 the automated alias services. More elaborate systems essentially 3257 create a new message. 3259 A Forwarder that does not modify the body or signed header fields of 3260 a message is likely to maintain the validity of the existing 3261 signature. It also could choose to add its own signature to the 3262 message. 3264 Forwarders which modify a message in a way that could make an 3265 existing signature invalid are particularly good candidates for 3266 adding their own signatures (e.g., mailing-list-name@example.net). 3268 Since (re-)signing is taking responsibility for the content of the 3269 message, these signing forwarders are likely to be selective, and 3270 forward or re-sign a message only if it is received with a valid 3271 signature or if they have some other basis for knowing that the 3272 message is not spoofed. 3274 A common practice among systems that are primarily redistributors of 3275 mail is to add a Sender header field to the message, to identify the 3276 address being used to sign the message. This practice will remove 3277 any preexisting Sender header field as required by [RFC5322]. The 3278 forwarder applies a new DKIM-Signature header field with the 3279 signature, public key, and related information of the forwarder. 3281 Appendix C. Creating a Public Key (INFORMATIVE) 3283 The default signature is an RSA signed SHA256 digest of the complete 3284 email. For ease of explanation, the openssl command is used to 3285 describe the mechanism by which keys and signatures are managed. One 3286 way to generate a 1024-bit, unencrypted private key suitable for DKIM 3287 is to use openssl like this: 3288 $ openssl genrsa -out rsa.private 1024 3289 For increased security, the "-passin" parameter can also be added to 3290 encrypt the private key. Use of this parameter will require entering 3291 a password for several of the following steps. Servers may prefer to 3292 use hardware cryptographic support. 3294 The "genrsa" step results in the file rsa.private containing the key 3295 information similar to this: 3296 -----BEGIN RSA PRIVATE KEY----- 3297 MIICXwIBAAKBgQDwIRP/UC3SBsEmGqZ9ZJW3/DkMoGeLnQg1fWn7/zYtIxN2SnFC 3298 jxOCKG9v3b4jYfcTNh5ijSsq631uBItLa7od+v/RtdC2UzJ1lWT947qR+Rcac2gb 3299 to/NMqJ0fzfVjH4OuKhitdY9tf6mcwGjaNBcWToIMmPSPDdQPNUYckcQ2QIDAQAB 3300 AoGBALmn+XwWk7akvkUlqb+dOxyLB9i5VBVfje89Teolwc9YJT36BGN/l4e0l6QX 3301 /1//6DWUTB3KI6wFcm7TWJcxbS0tcKZX7FsJvUz1SbQnkS54DJck1EZO/BLa5ckJ 3302 gAYIaqlA9C0ZwM6i58lLlPadX/rtHb7pWzeNcZHjKrjM461ZAkEA+itss2nRlmyO 3303 n1/5yDyCluST4dQfO8kAB3toSEVc7DeFeDhnC1mZdjASZNvdHS4gbLIA1hUGEF9m 3304 3hKsGUMMPwJBAPW5v/U+AWTADFCS22t72NUurgzeAbzb1HWMqO4y4+9Hpjk5wvL/ 3305 eVYizyuce3/fGke7aRYw/ADKygMJdW8H/OcCQQDz5OQb4j2QDpPZc0Nc4QlbvMsj 3306 7p7otWRO5xRa6SzXqqV3+F0VpqvDmshEBkoCydaYwc2o6WQ5EBmExeV8124XAkEA 3307 qZzGsIxVP+sEVRWZmW6KNFSdVUpk3qzK0Tz/WjQMe5z0UunY9Ax9/4PVhp/j61bf 3308 eAYXunajbBSOLlx4D+TunwJBANkPI5S9iylsbLs6NkaMHV6k5ioHBBmgCak95JGX 3309 GMot/L2x0IYyMLAz6oLWh2hm7zwtb0CgOrPo1ke44hFYnfc= 3310 -----END RSA PRIVATE KEY----- 3311 To extract the public-key component from the private key, use openssl 3312 like this: 3313 $ openssl rsa -in rsa.private -out rsa.public -pubout -outform PEM 3315 This results in the file rsa.public containing the key information 3316 similar to this: 3317 -----BEGIN PUBLIC KEY----- 3318 MIGfMA0GCSqGSIb3DQEBAQUAA4GNADCBiQKBgQDwIRP/UC3SBsEmGqZ9ZJW3/DkM 3319 oGeLnQg1fWn7/zYtIxN2SnFCjxOCKG9v3b4jYfcTNh5ijSsq631uBItLa7od+v/R 3320 tdC2UzJ1lWT947qR+Rcac2gbto/NMqJ0fzfVjH4OuKhitdY9tf6mcwGjaNBcWToI 3321 MmPSPDdQPNUYckcQ2QIDAQAB 3322 -----END PUBLIC KEY----- 3324 This public-key data (without the BEGIN and END tags) is placed in 3325 the DNS: 3326 brisbane IN TXT ("v=DKIM1; p=MIGfMA0GCSqGSIb3DQEBAQUAA4GNADCBiQ" 3327 "KBgQDwIRP/UC3SBsEmGqZ9ZJW3/DkMoGeLnQg1fWn7/zYt" 3328 "IxN2SnFCjxOCKG9v3b4jYfcTNh5ijSsq631uBItLa7od+v" 3329 "/RtdC2UzJ1lWT947qR+Rcac2gbto/NMqJ0fzfVjH4OuKhi" 3330 "tdY9tf6mcwGjaNBcWToIMmPSPDdQPNUYckcQ2QIDAQAB") 3332 C.1. Compatibility with DomainKeys Key Records 3334 DKIM key records were designed to be backwards-compatible in many 3335 cases with key records used by DomainKeys [RFC4870] (sometimes 3336 referred to as "selector records" in the DomainKeys context). One 3337 area of incompatibility warrants particular attention. The "g=" tag/ 3338 value may be used in DomainKeys and [RFC4871] key records to provide 3339 finer granularity of the validity of the key record to a specific 3340 local-part. A null "g=" value in DomainKeys is valid for all 3341 addresses in the domain. This differs from the usage in the original 3342 DKIM specification, where a null "g=" value is not valid for any 3343 address. In particular, the example public key record in Section 3344 3.2.3 of [RFC4870] with DKIM. 3346 Although the "g=" tag has been deprecated in this version of the DKIM 3347 specification, signers are advised not to include the "g=" tag in key 3348 records because some [RFC4871]-compliant verifiers will be in use for 3349 a considerable period to come. 3351 Appendix D. MUA Considerations 3353 When a DKIM signature is verified, the processing system sometimes 3354 makes the result available to the recipient user's MUA. How to 3355 present this information to the user in a way that helps them is a 3356 matter of continuing human factors usability research. The tendency 3357 is to have the MUA highlight the SDID, in an attempt to show the user 3358 the identity that is claiming responsibility for the message. An MUA 3359 might do this with visual cues such as graphics, or it might include 3360 the address in an alternate view, or it might even rewrite the 3361 original From address using the verified information. Some MUAs 3362 might indicate which header fields were protected by the validated 3363 DKIM signature. This could be done with a positive indication on the 3364 signed header fields, with a negative indication on the unsigned 3365 header fields, by visually hiding the unsigned header fields, or some 3366 combination of these. If an MUA uses visual indications for signed 3367 header fields, the MUA probably needs to be careful not to display 3368 unsigned header fields in a way that might be construed by the end 3369 user as having been signed. If the message has an l= tag whose value 3370 does not extend to the end of the message, the MUA might also hide or 3371 mark the portion of the message body that was not signed. 3373 The aforementioned information is not intended to be exhaustive. The 3374 MUA may choose to highlight, accentuate, hide, or otherwise display 3375 any other information that may, in the opinion of the MUA author, be 3376 deemed important to the end user. 3378 Appendix E. Acknowledgements 3380 The previous IETF version of DKIM [RFC4871] was edited by: Eric 3381 Allman, Jon Callas, Mark Delany, Miles Libbey, Jim Fenton and Michael 3382 Thomas. 3384 That specification was the result of an extended, collaborative 3385 effort, including participation by: Russ Allbery, Edwin Aoki, Claus 3386 Assmann, Steve Atkins, Rob Austein, Fred Baker, Mark Baugher, Steve 3387 Bellovin, Nathaniel Borenstein, Dave Crocker, Michael Cudahy, Dennis 3388 Dayman, Jutta Degener, Frank Ellermann, Patrik Faeltstroem, Mark 3389 Fanto, Stephen Farrell, Duncan Findlay, Elliot Gillum, Olafur 3390 Gu[eth]mundsson, Phillip Hallam-Baker, Tony Hansen, Sam Hartman, 3391 Arvel Hathcock, Amir Herzberg, Paul Hoffman, Russ Housley, Craig 3392 Hughes, Cullen Jennings, Don Johnsen, Harry Katz, Murray S. 3393 Kucherawy, Barry Leiba, John Levine, Charles Lindsey, Simon 3394 Longsdale, David Margrave, Justin Mason, David Mayne, Thierry Moreau, 3395 Steve Murphy, Russell Nelson, Dave Oran, Doug Otis, Shamim Pirzada, 3396 Juan Altmayer Pizzorno, Sanjay Pol, Blake Ramsdell, Christian Renaud, 3397 Scott Renfro, Neil Rerup, Eric Rescorla, Dave Rossetti, Hector 3398 Santos, Jim Schaad, the Spamhaus.org team, Malte S. Stretz, Robert 3399 Sanders, Rand Wacker, Sam Weiler, and Dan Wing. 3401 The earlier DomainKeys was a primary source from which DKIM was 3402 derived. Further information about DomainKeys is at [RFC4870]. 3404 Authors' Addresses 3406 D. Crocker (editor) 3407 Brandenburg InternetWorking 3408 675 Spruce Dr. 3409 Sunnyvale 3410 USA 3412 Phone: +1.408.246.8253 3413 Email: dcrocker@bbiw.net 3414 URI: http://bbiw.net 3416 Tony Hansen (editor) 3417 AT&T Laboratories 3418 200 Laurel Ave. South 3419 Middletown, NJ 07748 3420 USA 3422 Email: tony+dkimov@maillennium.att.com 3424 M. Kucherawy (editor) 3425 Cloudmark 3426 128 King St., 2nd Floor 3427 San Francisco, CA 94107 3428 USA 3430 Email: msk@cloudmark.com