INTERNET DRAFT M. S. Kucherawy Title: draft-kucherawy-sender-auth-header-00 Sendmail, Inc. Expires: March 2005 September 2004 Message Header for Indicating Sender Authentication Status Status of this Memo By submitting this Internet-Draft, I certify that any applicable patent or other IPR claims of which I am aware have been disclosed, or will be disclosed, and any of which I become aware will be disclosed, in accordance with RFC 3668. This document may not be modified, and derivative works of it may not be created, except to publish it as an RFC and to translate it into languages other than English. This document is an Internet-Draft and is subject to all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet- Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/1id-abstracts.html The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html Abstract This memo defines a new message header for use with [MAIL] messages to indicate the results of sender authentication efforts to mail user agents (MUAs) in order to equip them to relay that information in a convenient way to users. Table of Contents Kucherawy Expires February 2005 [Page 1] Internet-Draft Message Sender Authentication Header September 2004 1. Introduction This memo defines a new message header for [MAIL] messages which presents the results of a sender authentication effort in a machine- readable format. The intent is to create a place to collect such data when sender authentication mechanisms are in use so that an MUA can provide a recommendation to the user as to the trustworthiness of the message's origin and content. This memo defines both the format of this new header, and discusses the implications of its presence or absence. At the time of publication of this draft, only [AUTH] is a published sender authentication standard. However, several more are in the Internet Draft stage. As various methods emerge, it is necessary to prepare their appearance and encourage convergence in the area of interfacing these filters to MUAs. 1.1. Purposes The header defined in this memo is expected to serve several pur- poses: (1) Convey to MUAs from filters and MTAs the results of various sender authentication checks being applied; (2) Provide a common location for the presentation of this data; (3) Create an extensible framework for specifying new authentication methods as such emerge. (4) Convey the results of sender authentication tests to later filter- ing agents within the same "trust domain", as such agents might apply more or less stringent checks based on sender authentication results. 1.2. Requirements This memo establishes no new requirements on existing protocols or servers, as there is currently not a standard place for the described data to be collected or presented. Kucherawy Expires February 2005 [Page 2] Internet-Draft Message Sender Authentication Header September 2004 1.3. Terminology This document occasionally uses terms that appear in capital letters. When the terms "MUST", "SHOULD", "RECOMMENDED", "MUST NOT", "SHOULD NOT", and "MAY" appear capitalized, they are being used to indicate particular requirements of this specification. A discussion of the meanings of these terms appears in RFC2119. Generally it is assumed that the work of applying sender authentica- tion schemes takes place at a border MTA, that is, an MTA which acts as a gateway between the general Internet and the users within an organizational boundary. This specification is written with that assumption in mind. However, there are some sites at which the entire mail infrastructure consists of a single host. In such cases, such terms as "border MTA" and "delivery MTA" may well apply to the same machine or even the very same agent. 2. Definition and Format of the Header The new header being defined here is called "Authentication-Results". It qualifies in [MAIL] terms as a Structured Header Field, and thus all of the related definitions in that document apply. The decommented value of the header consists of a hostname, some whitespace, a "header=mailbox" statement indicating which header was selected to determine who sent the message and what address was extracted from that header, followed by zero or more authentication method names and a result associated with each, returned by the code that implements the method. As it is currently a matter of some debate, the header MAY appear more than once in a single message, or more than one result MAY be represented in a single header, or a combination of these MAY be applied. Formally, the header is specified as follows: header := "Authentication-Results" ":" hostname 1*LWSP headerspec *(";" name "=" value) hostname := Domain ; as defined in [SMTP] Kucherawy Expires February 2005 [Page 3] Internet-Draft Message Sender Authentication Header September 2004 name := method ; a method indicates which method's result is ; is represented by "value", and is one of the methods ; explicitly defined as valid in this document ; or is an extension method as defined below method := token ; authentication method whose results are being relayed value := result result := "pass" / "fail" / "softfail" / "neutral" / "temperror" / "permerror" ; an indication of the results of the attempt to ; authenticate the sender token := ; As defined in Appendix A of [MIME] LWSP := 0x20 (linear space) / 0x09 (horizontal tab) addrspec := header "=" mailbox ; an indication of which header was used to ; determine the "true" sender of the message, ; against which authentication tests were done header := token ; which header was used to determine the ; "true" sender of the message being analyzed mailbox := As defined in section 3.4 of [MAIL] The six possible values of the "result" are: pass sending domain publishes an authentication policy of some kind, and the message passed the authentication tests fail sending domain publishes an authentication policy of some kind, and the message failed the authentication tests softfail sending domain publishes an authentication policy which doesn't require authentication of all messages from that domain, and the message failed the authentication tests Kucherawy Expires February 2005 [Page 4] Internet-Draft Message Sender Authentication Header September 2004 neutral sending domain does not publish any sender authentication policy temperror a temporary (recoverable) error occurred attempting to authenti- cate the sender; either the process couldn't be completed locally because of some transient condition, or there was a tem- porary failure retrieving the sending domain's policy; a later attempt to re-authenticate this message might produce a more final result permerror a permanent (unrecoverable) error occurred attempting to authen- ticate the sender; either the process couldn't be completed locally, or there was a permanent failure retrieving the sending domain's policy 3. Definition Of The "auth" Method As it is currently an existing standard for sender authentication, it is appropriate to define an authentication method identifier for [AUTH]. The authentication method identifier "auth" is thus defined for MTAs applying that standard for sender authentication. 4. Adding The Header To A Message This specification makes no attempt to evaluate the relative strengths of various sender authentication methods that may become available. As such, the order of the presented authentication meth- ods and results are not relevant since ultimately the importance on any given method over another is the decision of the MUA that is interpreting the value of the header. The "method" MUST refer to an authentication method declared in this memo, or in a subsequent one, or to an authentication method name assigned by IANA. The three values of the "result" should be fairly self-evident; "yes" indicates the sender was successfully authenticated, "no" indicates the opposite, and "unknown" indicates no conclusion was possible. If an MTA applies any authentication test, it MUST add this header to Kucherawy Expires February 2005 [Page 5] Internet-Draft Message Sender Authentication Header September 2004 indicate at which host the test was done, which test got applied, and what the result was. If an MTA applies more than one such test, it MUST either add this header once per test, or one header indicating all of the results. An MTA MUST NOT add a result to an existing header. An MTA adding this header in either form MUST use its own hostname only. It MUST be a fully-qualified domain name. For security reasons, an MTA SHOULD remove any discovered instance of this header for which the "hostname" is its own, i.e. headers which claim to be from the MTA but were added before the mail arrived at the MTA for processing. A border MTA MAY also delete any discovered instance of this header which claims to have been added within its trust boundary. For example, a border MTA at mx.example.com SHOULD delete any instance of this header claiming to come from mx.exam- ple.com and MAY delete any instance of this header claiming to come from any host in example.com prior to adding its own headers. This applies in both directions so that hosts outside the domain cannot claim results MUAs inside the domain might trust. An MTA compliant with this specification MUST add this header after performing one or more sender authentication tests. An MTA MAY add this header containing only the "hostname" portion to explicitly indicate that no sender authentication schemes were applied prior to delivery of this message. 4.1. Header Position and Interpretation In order to ensure non-ambiguous results and avoid the impact of false headers, an MUA SHOULD NOT interpret this header unless specif- ically instructed to do so by the user. That is, this should not be "on by default". Naturally then, users would not activate such a feature unless they are certain the header will be added by the receiving MTA that accepts the mail which is ultimately read by the MUA, and instances of the header added by foreign MTAs will be removed before delivery. Furthermore, an MUA SHOULD NOT interpret this header unless the host- name it bears appears to be one within its own trust domain as con- figured by the user. An MTA adding a header MUST add the header at the top of the message so that there is generally some indication upon delivery of where in the chain of handling MTAs the sender authentcation was done. Kucherawy Expires February 2005 [Page 6] Internet-Draft Message Sender Authentication Header September 2004 Further discussion of this can be found in the Security Considera- tions section below. 4.2. Extension Fields Additional authentication method identifiers may be defined in the future by later revisions or extensions to this specification. Extension identifiers beginning with "x-" will never be defined as standard fields; such names are reserved for experimental use. Method identifiers NOT beginning with "x-" MUST be registered with the Internet Assigned Numbers Authority (IANA) and published in an RFC. Extension identifiers may be defined for the following reasons: (a) To allow additional information from emergent authentication sys- tems to be communicated to MUAs. The names of such identifiers should reflect the name of the method being defined, but should not be needlessly long. (b) To allow the creation of "sub-identifiers" which indicate different levels of authentication and differentiate between their relative strengths, e.g. "auth1-weak" and "auth1-strong". Authentication method implementors are encouraged to provide adequate information, via [MAIL] comments if necessary, to allow an MUA devel- oper to understand or relay ancilliary details of authentication results. For example, if it might be of interest to relay what data was used to perform an evaluation, such information could be relayed as a comment in the header, such as: Authentication-Results: mx.example.com; foo=yes (2 out of 3 tests passed) 5. Discussion This section discusses various implementation issues not specifically related to security. Security issues are discussed in a later sec- tion. Kucherawy Expires February 2005 [Page 7] Internet-Draft Message Sender Authentication Header September 2004 5.1. Legacy MUAs Implementors of this proposal should be aware that many MUAs are unlikely to be retrofit to support the new header and its semantics. In the interests of convenience and quicker adaptation, a delivery MTA might want to consider adding things that are processed by exist- ing MUAs as well as the header defined by this specification. One suggestion is to provide a Priority: header with a value that reflects the strength of the authentication that was accomplished, e.g. "low" for weak or no authentication, "normal" or "high" for good authentication. Certainly some modern MUAs can filter based on the content of this header, but as there is keen interest in having MUAs make some kind of graphical representation of this header's meaning, other interim means of doing so may be necessary while this proposal is adopted. 5.2. Trusting The Header Rather than the coding rigor of doing a digital signature on the con- tent of this header, it may be sufficient to establish a maximum SMTP path length between the addition of the header and final delivery. The path length is defined as the count of Received headers above the Authentication-Results header. If more than that maximum path length is traversed between insertion of the header and delivery, the value of the header should no longer be trusted to be valid. This requires a configurable MUA setting to define this maximum path length. The setting would need to take into account possible addi- tional hops should the final delivery server be unavailable. Although this approach likely bears a much shorter time to implement, it is a less general solution than the proposed one and would proba- bly require additional user education above the norm and thus lead to more confusion on deployment. The MUA SHOULD trust this header if it is trusting such headers and the path length after the addition of the header is less than three. 6. Conformance and Usage Requirements An MTA or gateway conforms to this specification if it applies one or Kucherawy Expires February 2005 [Page 8] Internet-Draft Message Sender Authentication Header September 2004 more sender authentication mechanisms and inserts a header corre- sponding to this standard after doing so and prior to delivery. MTAs that are relaying mail rather than delivering it MAY perform sender authentication or even take actions based on the results found, but MUST NOT add a "Authentication-Results" header if relaying rather than rejecting or discarding at the gateway. Conversely, an MTA doing local delivery MUST add this header prior to delivery the message in order to be compliant. A minimal implementation which does at least one sender authentica- tion check will add the header defined by this memo prior to invoking local delivery procedures. This specification places no restrictions on the processing of the header's contents by user agents or distribution lists. It is pre- sented to those packages solely for their own information. 7. IANA Considerations Following the policies outlined in [IANA-CONSIDERATIONS], names of sender authentication methods supported by this specification must be registered with IANA under the IETF Consensus method. 8. Security Considerations The following security considerations apply when applying or process- ing the "Authentication-Results" header: 8.1. Non-conformant MTAs An MUA that is aware of this specification which accesses a mailbox whose mail is handled by a non-conformant MTA is in a position to make false conclusions based on forged headers. A malicious user or agent could forge a header using the destination MX for a receiving domain as the hostname token in the value of the header, and with the rest of the value claim that the sender was properly authenticated. The non-conformant MTA would fail to strip the forged header, and the MUA could trust it. It is for this reason an MUA SHOULD NOT have processing of the Kucherawy Expires February 2005 [Page 9] Internet-Draft Message Sender Authentication Header September 2004 "Authentication-Results" header enabled by default; instead it must be ignored, at least for the purposes of enacting filtering deci- sions, unless specifically enabled by the user after verifying that the MTA is compliant. It is acceptable to have an MUA aware of this standard, but have an explicit list of hostnames whose "Authentica- tion-Results" headers are trustworthy, but this list SHOULD initially be empty. Proposed alternate solutions to this problem are nascent. Possibly the simplest is a digital signature on the header which can be veri- fied by a posted public key. Another would be a means to interrogate the MTA that added the header to see if it is actually providing any sender authentication services and saw the message in question. In either case, a method needs to exist to verify that the host which appears to have added the header (a) actually did so, and (b) is legitimately adding that header for this delivery. 8.2. Header Position Headers can sometimes be reordered enroute by intermediate MTAs. The goal of requiring header addition only at the top of a message is an acknowledgement that some MTAs do reorder headers, but most do not. Thus, in the general case, there will be some indication of which MTAs (if any) handled the message after the addition of the header defined here. However, this is not universally true. In an environment involving an MTA that does reorder headers, the above-mentioned hop count test for header validity may not be possible. 9. Appendix - Authentication-Results Examples 9.1. Trivial case; header not present The trivial case: From: sender@example.com Received: from mail-router.example.com (mail-router.example.com [1.2.3.4]) Kucherawy Expires February 2005 [Page 10] Internet-Draft Message Sender Authentication Header September 2004 by server.sendmail.com (8.11.6/8.11.6) with ESMTP id g1G0r1kA003489; Fri, Feb 15 2002 17:19:07 -0800 Date: Fri, Feb 15 2002 16:54:30 -0800 To: receiver@sendmail.com Message-Id: <12345.abc@example.com> Subject: here's a sample Hello! Goodbye! The "Authentication-Results" header is completely absent. The MUA may make no conclusion about the validity of the message. This could be the case because the sender authentication services were not available at the time of delivery, or no service is provided, or the MTA is not in compliance with this specification. 9.2. Nearly-trivial case; service provided, but no authentication done A message that was delivered by an MTA which conforms to this stan- dard but provides no actual sender authentication service: Authentication-Results: mail-router.example.com from=sender@example.com From: sender@example.com Received: from mail-router.example.com (mail-router.example.com [1.2.3.4]) by server.sendmail.com (8.11.6/8.11.6) with ESMTP id g1G0r1kA003489; Fri, Feb 15 2002 17:19:07 -0800 Date: Fri, Feb 15 2002 16:54:30 -0800 To: receiver@sendmail.com Message-Id: <12345.abc@example.com> Subject: here's a sample Hello! Goodbye! The "Authentication-Results" header is present, indicating the deliv- ering MTA (which is named in the value of the header) conforms to this specification. The absence of any method and result tokens indicates no sender authentication was done. 9.3. Service provided, authentication done A message that was delivered by an MTA which conforms to this Kucherawy Expires February 2005 [Page 11] Internet-Draft Message Sender Authentication Header September 2004 standard and applied some sender authentication: Authentication-Results: mail-router.example.com from=sender@example.com; auth=pass (cram-md5) From: sender@example.com Received: from dialup-1-2-3-4.example-isp.com (dialup-1-2-3-4.example-isp.com [1.2.3.4]) by mail-router.example.com (8.11.6/8.11.6) with ESMTP id g1G0r1kA003489; Fri, Feb 15 2002 17:19:07 -0800 Date: Fri, Feb 15 2002 16:54:30 -0800 To: receiver@sendmail.com Message-Id: <12345.abc@example.com> Subject: here's a sample Hello! Goodbye! The "Authentication-Results" header is present, indicating the deliv- ering MTA (which is named in the value of the header) conforms to this specification. Furthermore, the sender authenticated her- self/himself to the MTA via a method specified in [AUTH]. The actual method is identified in a header comment after the method's result is indicated. The MUA could extract and relay this extra information if desired. 9.4. Service provided, several authentications done, single MTA A message that was relayed inbound via a single MTA which conforms to this standard and applied two different sender authentication checks: Authentication-Results: mail-router.example.com from=sender@example.com; auth=pass (cram-md5); spf=pass; sender-id=pass From: sender@example.com Received: from mail-router.example.com (mail-router.example.com [1.2.3.4]) by dialup-1-2-3-4.example-isp.com (8.11.6/8.11.6) with ESMTP id g1G0r1kA003489; Fri, Feb 15 2002 17:19:07 -0800 Date: Fri, Feb 15 2002 16:54:30 -0800 To: receiver@sendmail.com Message-Id: <12345.abc@example.com> Subject: here's a sample Hello! Goodbye! Kucherawy Expires February 2005 [Page 12] Internet-Draft Message Sender Authentication Header September 2004 The "Authentication-Results" header is present, indicating the deliv- ering MTA (which is named in the value of the header) conforms to this specification. Furthermore, the sender authenticated her- self/himself to the MTA via a method specified in [AUTH], and both SPF and Sender-ID checks were done and passed. The MUA could extract and relay this extra information if desired. 9.5. Service provided, several authentications done, same MTA A message that was relayed inbound by a single MTA which conforms to this standard and applied multiple sender authentication checks: Authentication-Results: auth-checker.example.com from=sender@example.com; sender-id=pass; domainkeys=fail (signature) Received: from mail-router.example.com (mail-router.example.com [10.11.12.13]) by auth-checker.example.com (8.11.6/8.11.6) with ESMTP id i7PK0sH7021929; Fri, Feb 15 2002 17:19:22 -0800 Authentication-Results: mail-router.example.com from=sender@example.com; auth=yes (cram-md5); spf=yes Received: from dialup-1-2-3-4.example-isp.com (dialup-1-2-3-4.example-isp.com [1.2.3.4]) by mail-router.example.com (8.11.6/8.11.6) with ESMTP id g1G0r1kA003489; Fri, Feb 15 2002 17:19:07 -0800 DomainKey-Signature: a=rsa-sha1; s=gatsby; d=sendmail.net; c=simple; q=dns; b=EToRSuvUfQVP3Bkz ... rTB0t0gYnBVCM= From: sender@example.com Date: Fri, Feb 15 2002 16:54:30 -0800 To: receiver@sendmail.com Message-Id: <12345.abc@example.com> Subject: here's a sample Hello! Goodbye! The "Authentication-Results" header is present, indicating confor- mance to this specification. It is present twice because two differ- ent MTAs in the chain of delivery did authentication tests. The first, "mail-router.example.com" reports that [AUTH] and SPF were both used and passed. In the [AUTH] case, additional data is pro- vided in the comment field, which the MUA can choose to render if desired. The second MTA, "auth-checker.example.com", reports that it did a Sender-ID test which passed, but a DomainKeys test failed Kucherawy Expires February 2005 [Page 13] Internet-Draft Message Sender Authentication Header September 2004 because the signature verification was not successful. 10. Acknowledgements The author wishes to acknowledge the following for their review and constructive criticism of this proposal: Mark Delany and Miles Libbey of Yahoo! Inc., Jim Fenton of Cisco, and Rand Wacker of Sendmail, Inc. 11. References [AUTH] Myers, J., "SMTP Service Extention for Authentication", RFC 2554, Netscape Communications, March 1999. [MAIL] Resnick, P. (editor), "Internet Message Format", RFC 2822, Qual- comm, Inc., April 2001. [MIME] N. Freed, N. Borenstein, "Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies", RFC 2045, November 1996 [SMTP] J. Klensin, (editor), "Simple Mail Transfer Protocol", RFC 2821, April 2001 12. Author's Address: Murray S. Kucherawy Sendmail, Inc. 6425 Christie Ave., 4th Floor Emeryville, CA 94608 USA Comments may be sent to msk+ietf@sendmail.com. Kucherawy Expires February 2005 [Page 14] Internet-Draft Message Sender Authentication Header September 2004 13. Copyright Notice Copyright (C) The Internet Society (year). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFOR- MATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Kucherawy Expires February 2005 [Page 15]