Network Working Group                                       J. Klensin
Request for Comments: 2195                                    R. Catoe
Category: Standards Track                                 P. Krumviede
Obsoletes: 2095                                                    MCI
                                                       September 1997


      IMAP/POP AUTHorize Extension for Simple Challenge/Response

Status of this Memo

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

Abstract

  While IMAP4 supports a number of strong authentication mechanisms as
  described in RFC 1731, it lacks any mechanism that neither passes
  cleartext, reusable passwords across the network nor requires either
  a significant security infrastructure or that the mail server update
  a mail-system-wide user authentication file on each mail access.
  This specification provides a simple challenge-response
  authentication protocol that is suitable for use with IMAP4.  Since
  it utilizes Keyed-MD5 digests and does not require that the secret be
  stored in the clear on the server, it may also constitute an
  improvement on APOP for POP3 use as specified in RFC 1734.

1. Introduction

  Existing Proposed Standards specify an AUTHENTICATE mechanism for the
  IMAP4 protocol [IMAP, IMAP-AUTH] and a parallel AUTH mechanism for
  the POP3 protocol [POP3-AUTH].  The AUTHENTICATE mechanism is
  intended to be extensible; the four methods specified in [IMAP-AUTH]
  are all fairly powerful and require some security infrastructure to
  support.  The base POP3 specification [POP3] also contains a
  lightweight challenge-response mechanism called APOP.  APOP is
  associated with most of the risks associated with such protocols: in
  particular, it requires that both the client and server machines have
  access to the shared secret in cleartext form. CRAM offers a method
  for avoiding such cleartext storage while retaining the algorithmic
  simplicity of APOP in using only MD5, though in a "keyed" method.







Klensin, Catoe & Krumviede  Standards Track                     [Page 1]

RFC 2195              IMAP/POP AUTHorize Extension        September 1997


  At present, IMAP [IMAP] lacks any facility corresponding to APOP.
  The only alternative to the strong mechanisms identified in [IMAP-
  AUTH] is a presumably cleartext username and password, supported
  through the LOGIN command in [IMAP].  This document describes a
  simple challenge-response mechanism, similar to APOP and PPP CHAP
  [PPP], that can be used with IMAP (and, in principle, with POP3).

  This mechanism also has the advantage over some possible alternatives
  of not requiring that the server maintain information about email
  "logins" on a per-login basis.  While mechanisms that do require such
  per-login history records may offer enhanced security, protocols such
  as IMAP, which may have several connections between a given client
  and server open more or less simultaneous, may make their
  implementation particularly challenging.

2. Challenge-Response Authentication Mechanism (CRAM)

  The authentication type associated with CRAM is "CRAM-MD5".

  The data encoded in the first ready response contains an
  presumptively arbitrary string of random digits, a timestamp, and the
  fully-qualified primary host name of the server.  The syntax of the
  unencoded form must correspond to that of an RFC 822 'msg-id'
  [RFC822] as described in [POP3].

  The client makes note of the data and then responds with a string
  consisting of the user name, a space, and a 'digest'.  The latter is
  computed by applying the keyed MD5 algorithm from [KEYED-MD5] where
  the key is a shared secret and the digested text is the timestamp
  (including angle-brackets).

  This shared secret is a string known only to the client and server.
  The `digest' parameter itself is a 16-octet value which is sent in
  hexadecimal format, using lower-case ASCII characters.

  When the server receives this client response, it verifies the digest
  provided.  If the digest is correct, the server should consider the
  client authenticated and respond appropriately.

  Keyed MD5 is chosen for this application because of the greater
  security imparted to authentication of short messages. In addition,
  the use of the techniques described in [KEYED-MD5] for precomputation
  of intermediate results make it possible to avoid explicit cleartext
  storage of the shared secret on the server system by instead storing
  the intermediate results which are known as "contexts".






Klensin, Catoe & Krumviede  Standards Track                     [Page 2]

RFC 2195              IMAP/POP AUTHorize Extension        September 1997


  CRAM does not support a protection mechanism.

  Example:

  The examples in this document show the use of the CRAM mechanism with
  the IMAP4 AUTHENTICATE command [IMAP-AUTH].  The base64 encoding of
  the challenges and responses is part of the IMAP4 AUTHENTICATE
  command, not part of the CRAM specification itself.

    S: * OK IMAP4 Server
    C: A0001 AUTHENTICATE CRAM-MD5
    S: + PDE4OTYuNjk3MTcwOTUyQHBvc3RvZmZpY2UucmVzdG9uLm1jaS5uZXQ+
    C: dGltIGI5MTNhNjAyYzdlZGE3YTQ5NWI0ZTZlNzMzNGQzODkw
    S: A0001 OK CRAM authentication successful

     In this example, the shared secret is the string
     'tanstaaftanstaaf'.  Hence, the Keyed MD5 digest is produced by
     calculating

       MD5((tanstaaftanstaaf XOR opad),
           MD5((tanstaaftanstaaf XOR ipad),
           <[email protected]>))

     where ipad and opad are as defined in the keyed-MD5 Work in
     Progress [KEYED-MD5] and the string shown in the challenge is the
     base64 encoding of <[email protected]>. The
     shared secret is null-padded to a length of 64 bytes. If the
     shared secret is longer than 64 bytes, the MD5 digest of the
     shared secret is used as a 16 byte input to the keyed MD5
     calculation.

     This produces a digest value (in hexadecimal) of

          b913a602c7eda7a495b4e6e7334d3890

     The user name is then prepended to it, forming

          tim b913a602c7eda7a495b4e6e7334d3890

     Which is then base64 encoded to meet the requirements of the IMAP4
     AUTHENTICATE command (or the similar POP3 AUTH command), yielding

          dGltIGI5MTNhNjAyYzdlZGE3YTQ5NWI0ZTZlNzMzNGQzODkw








Klensin, Catoe & Krumviede  Standards Track                     [Page 3]

RFC 2195              IMAP/POP AUTHorize Extension        September 1997


3. References

  [CHAP]  Lloyd, B., and W. Simpson, "PPP Authentication Protocols",
      RFC 1334, October 1992.

  [IMAP] Crispin, M., "Internet Message Access Protocol - Version
      4rev1", RFC 2060, University of Washington, December 1996.

  [IMAP-AUTH] Myers, J., "IMAP4 Authentication Mechanisms",
      RFC 1731, Carnegie Mellon, December 1994.

  [KEYED-MD5] Krawczyk, Bellare, Canetti, "HMAC: Keyed-Hashing for
      Message Authentication", RFC 2104, February 1997.

  [MD5]  Rivest, R., "The MD5 Message Digest Algorithm",
      RFC 1321, MIT Laboratory for Computer Science, April 1992.

  [POP3] Myers, J., and M. Rose, "Post Office Protocol - Version 3",
      STD 53, RFC 1939, Carnegie Mellon, May 1996.

  [POP3-AUTH] Myers, J., "POP3 AUTHentication command", RFC 1734,
      Carnegie Mellon, December, 1994.

4. Security Considerations

  It is conjectured that use of the CRAM authentication mechanism
  provides origin identification and replay protection for a session.
  Accordingly, a server that implements both a cleartext password
  command and this authentication type should not allow both methods of
  access for a given user.

  While the saving, on the server, of "contexts" (see section 2) is
  marginally better than saving the shared secrets in cleartext as is
  required by CHAP [CHAP] and APOP [POP3], it is not sufficient to
  protect the secrets if the server itself is compromised.
  Consequently, servers that store the secrets or contexts must both be
  protected to a level appropriate to the potential information value
  in user mailboxes and identities.

  As the length of the shared secret increases, so does the difficulty
  of deriving it.

  While there are now suggestions in the literature that the use of MD5
  and keyed MD5 in authentication procedures probably has a limited
  effective lifetime, the technique is now widely deployed and widely
  understood.  It is believed that this general understanding may
  assist with the rapid replacement, by CRAM-MD5, of the current uses
  of permanent cleartext passwords in IMAP.   This document has been



Klensin, Catoe & Krumviede  Standards Track                     [Page 4]

RFC 2195              IMAP/POP AUTHorize Extension        September 1997


  deliberately written to permit easy upgrading to use SHA (or whatever
  alternatives emerge) when they are considered to be widely available
  and adequately safe.

  Even with the use of CRAM, users are still vulnerable to active
  attacks.  An example of an increasingly common active attack is 'TCP
  Session Hijacking' as described in CERT Advisory CA-95:01 [CERT95].

  See section 1 above for additional discussion.

5. Acknowledgements

  This memo borrows ideas and some text liberally from [POP3] and
  [RFC-1731] and thanks are due the authors of those documents.  Ran
  Atkinson made a number of valuable technical and editorial
  contributions to the document.

6. Authors' Addresses

  John C. Klensin
  MCI Telecommunications
  800 Boylston St, 7th floor
  Boston, MA 02199
  USA

  EMail: [email protected]
  Phone: +1 617 960 1011

  Randy Catoe
  MCI Telecommunications
  2100 Reston Parkway
  Reston, VA 22091
  USA

  EMail: [email protected]
  Phone: +1 703 715 7366

  Paul Krumviede
  MCI Telecommunications
  2100 Reston Parkway
  Reston, VA 22091
  USA

  EMail: [email protected]
  Phone: +1 703 715 7251






Klensin, Catoe & Krumviede  Standards Track                     [Page 5]


You are viewing proxied material from gopher.viste.fr. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.