Network Working Group P. Eronen
Internet-Draft Nokia
Expires: August 6, 2004 H. Tschofenig
Siemens
February 6, 2004
Pre-Shared Key Ciphersuites for Transport Layer Security (TLS)
draft-eronen-tls-psk-00.txt
Status of this Memo
This document is an Internet-Draft and is in full conformance with
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/ietf/1id-abstracts.txt.
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.
This Internet-Draft will expire on August 6, 2004.
Copyright Notice
Copyright (C) The Internet Society (2004). All Rights Reserved.
Abstract
This document specifies new ciphersuites for the Transport Layer
Security (TLS) protocol to support authentication based on pre-shared
keys. These pre-shared keys are symmetric keys, shared in advance
among the communicating parties, and do not require any public key
operations.
Conventions used in this document
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [KEYWORDS].
Eronen & Tschofenig Expires August 6, 2004 [Page 1]
Internet-Draft PSK Ciphersuites for TLS February 2004
1. Introduction
Usually TLS uses public key certificates [TLS] or Kerberos [TLS-KRB]
for authentication. This document describes how to use symmetric keys
(later called pre-shared keys or PSKs), shared in advance among the
communicating parties, to establish a TLS connection.
There are basically two reasons why one might want to do this:
o First, TLS may be used in performance-constrained environments
where the CPU power needed for public key operations is not
available.
o Second, pre-shared keys may be more convenient from a key
management point of view. For instance, in closed environments
where the connections are mostly configured manually in advance,
it may be easier to configure a PSK than to use certificates.
Another case is when the parties already have a mechanism for
setting up a shared secret key, and that mechanism could be used
to "bootstrap" a key for authenticating a TLS connection.
This document specifies a number of new ciphersuites for TLS. These
ciphersuites use a new authentication and key exchange algorithm for
PSKs, and re-use existing cipher and MAC algorithms from [TLS] and
[TLS-AES].
1.1 Applicability statement
The ciphersuites defined in this document are intended for a rather
limited set of applications, usually involving only a very small
number of clients and servers. Even in such environments, other
alternatives may be more appropriate.
If the main goal is to avoid PKIs, another possibility worth
considering is to use self-signed certificates with public key
fingerprints. Instead of manually configuring a shared secret in,
for instance, some configuration file, a fingerprint (hash) of the
other party's public key (or certificate) could be placed there
instead.
It is also possible to use SRP for shared secret authentication
[TLS-SRP]. However, SRP requires more computational resources and may
have some IPR issues. However, it does provide protection against
dictionary attacks.
Eronen & Tschofenig Expires August 6, 2004 [Page 2]
Internet-Draft PSK Ciphersuites for TLS February 2004
2. Protocol
It is assumed that the reader is familiar with ordinary TLS
handshake, shown below. The elements in parenthesis are not included
in PSK-based ciphersuites.
Client Server
------ ------
ClientHello -------->
ServerHello
(Certificate)
ServerKeyExchange
(CertificateRequest)
<-------- ServerHelloDone
(Certificate)
ClientKeyExchange
(CertificateVerify)
ChangeCipherSpec
Finished -------->
ChangeCipherSpec
<-------- Finished
Application Data <-------> Application Data
The client indicates its willingness to use pre-shared key
authentication by including one or more PSK-based ciphersuites in the
ClientHello message. The following ciphersuites are defined in this
document:
CipherSuite TLS_PSK_WITH_RC4_128_SHA = { 0x00, 0xTBD };
CipherSuite TLS_PSK_WITH_3DES_EDE_CBC_SHA = { 0x00, 0xTBD };
CipherSuite TLS_PSK_WITH_AES_128_CBC_SHA = { 0x00, 0xTBD };
CipherSuite TLS_PSK_WITH_AES_256_CBC_SHA = { 0x00, 0xTBD };
Note that this document defines only a new authentication and key
exchange algorithm; see [TLS] and [TLS-AES] for description of the
cipher and MAC algorithms.
If the TLS server also wants to use pre-shared keys, it selects one
of the PSK ciphersuites, places the selected ciphersuite in the
ServerHello message, and includes an appropriate ServerKeyExchange
message (see below). The Certificate and CertificateRequest payloads
are omitted from the response.
Both clients and servers may have pre-shared keys with several
different parties. The client indicates which key to use by including
a "PSK identity" in the ClientKeyExchange message (note that unlike
Eronen & Tschofenig Expires August 6, 2004 [Page 3]
Internet-Draft PSK Ciphersuites for TLS February 2004
in [TLS-SHAREDKEYS], the session_id field in ClientHello message
keeps its usual meaning). To help the client in selecting which
identity to use, the server can provide a "PSK identity hint" in the
ServerKeyExchange message (note that if no hint is provided, a
ServerKeyExchange message is still sent).
This document does not specify the format of the PSK identity or PSK
identity hint; neither is specified how exactly the client uses the
hint (if it uses it at all). The parties have to agree on the
identities when the shared secret is configured (however, see Section
4 for related security considerations).
The format of the ServerKeyExchange and ClientKeyExchange messages is
shown below.
struct {
select (KeyExchangeAlgorithm) {
case diffie_hellman:
ServerDHParams params;
Signature signed_params;
case rsa:
ServerRSAParams params;
Signature signed_params;
case psk: /* NEW */
opaque psk_identity_hint<0..2^16-1>;
};
} ServerKeyExchange;
struct {
select (KeyExchangeAlgorithm) {
case rsa: EncryptedPreMasterSecret;
case diffie_hellman: ClientDiffieHellmanPublic;
case psk: opaque psk_identity<0..2^16-1>; /* NEW */
} exchange_keys;
} ClientKeyExchange;
The premaster secret is formed as follows: concatenate 24 zero
octets, followed by SHA-1 hash [FIPS180-2] of the PSK itself,
followed by 4 zero octets.
Note: This effectively means that only the HMAC-SHA1 part of the
TLS PRF is used, and the HMAC-MD5 part is not used. See
[Krawczyk20040113] for a more detailed rationale. The PSK is first
hashed so that PSKs longer than 24 octets can be used; this is
similar to what is done in [HMAC] if the key length is longer than
the hash block size.
If the server does not recognize the PSK identity, it SHOULD respond
Eronen & Tschofenig Expires August 6, 2004 [Page 4]
Internet-Draft PSK Ciphersuites for TLS February 2004
with a decrypt_error alert message. This alert is also sent if
validating the Finished message fails. The use of the same alert
message makes it more difficult to find out which PSK identities are
known to the server.
3. IANA considerations
This document does not define any new namespaces to be managed by
IANA. It does require assignment of several new ciphersuite numbers,
but it is unclear how this is done, since the TLS spec does not say
who is responsible for assigning them :-)
4. Security Considerations
As with all schemes involving shared keys, special care should be
taken to protect the shared values and to limit their exposure over
time.
The ciphersuites defined in this document do not provide Perfect
Forward Secrecy (PFS). That is, if the shared secret key is somehow
compromised, an attacker can decrypt old conversations. (Note that
the most popular TLS key exchange algorithm, RSA, does not provide
PFS either.)
Use of a fixed shared secret of limited entropy (such as a password)
allows an attacker to perform a brute-force or dictionary attack to
recover the secret. This may be either an off-line attack (against a
captured TLS conversation), or an on-line attack where the attacker
tries to connect to the server and tries different keys. Note that
the protocol requires the client to prove it knows the key first, so
just attempting to connect to a server does not reveal information
required for an off-line attack. It is RECOMMENDED that
implementations that allow the administrator to manually configure
the PSK also provide a functionality for generating a new random PSK,
taking [RANDOMNESS] into account.
The PSK identity is sent in cleartext. While using a user name or
other similar string as the PSK identity is the most straightforward
option, it may lead to problems in some environments since an
eavesdropper is able to identify the communicating parties. Even when
the identity does not reveal any information itself, reusing the same
identity over time may eventually allow an attacker to use traffic
analysis to the identify parties. It should be noted that this is no
worse than client certificates, since they are also sent in
cleartext.
Eronen & Tschofenig Expires August 6, 2004 [Page 5]
Internet-Draft PSK Ciphersuites for TLS February 2004
5. Acknowledgments
The protocol defined in this document is heavily based on work by Tim
Dierks and Peter Gutmann, and borrows some text from [TLS-SHAREDKEYS]
and [TLS-AES]. Valuable feedback was also provided by Peter Gutmann
and Mika Tervonen.
When the first version of this draft was almost ready, the authors
learned that something similar had been proposed already in 1996
[TLS-PASSAUTH]. However, this draft is not intended for web password
authentication, but rather other uses of TLS.
Normative References
[KEYWORDS]
Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", RFC 2119, March 1997.
[TLS-AES] Chown, P., "Advanced Encryption Standard (AES)
Ciphersuites for Transport Layer Security (TLS)", RFC
3268, June 2002.
[TLS] Dierks, T. and C. Allen, "The TLS Protocol Version 1.0",
RFC 2246, January 1999.
[RANDOMNESS]
Eastlake, D., Crocker, S. and J. Schiller, "Randomness
Recommendations for Security", RFC 1750, December 1994.
[FIPS180-2]
National Institute of Standards and Technology,
"Specifications for the Secure Hash Standard", Federal
Information Processing Standard (FIPS) Publication 180-2,
August 2002.
Informative References
[TLS-SHAREDKEYS]
Gutmann, P., "Use of Shared Keys in the TLS Protocol",
draft-ietf-tls-sharedkeys-02 (work in progress), October
2003.
[HMAC] Krawczyk, H., Bellare, M. and R. Canetti, "HMAC:
Keyed-Hashing for Message Authentication", RFC 2104,
February 1997.
[Krawczyk20040113]
Krawczyk, H., "Re: TLS shared keys PRF", message on
Eronen & Tschofenig Expires August 6, 2004 [Page 6]
Internet-Draft PSK Ciphersuites for TLS February 2004
ietf-tls@lists.certicom.com mailing list 2004-01-13,
http://www.imc.org/ietf-tls/mail-archive/msg04098.html.
[TLS-KRB] Medvinsky, A. and M. Hur, "Addition of Kerberos Cipher
Suites to Transport Layer Security (TLS)", RFC 2712,
October 1999.
[TLS-PASSAUTH]
Simon, D., "Addition of Shared Key Authentication to
Transport Layer Security (TLS)",
draft-ietf-tls-passauth-00 (expired), November 1996.
[TLS-SRP] Taylor, D., Wu, T., Mavroyanopoulos, N. and T. Perrin,
"Using SRP for TLS Authentication", draft-ietf-tls-srp-06
(work in progress), January 2004.
Authors' Addresses
Pasi Eronen
Nokia Research Center
P.O. Box 407
FIN-00045 Nokia Group
Finland
EMail: pasi.eronen@nokia.com
Hannes Tschofenig
Siemens
Otto-Hahn-Ring 6
Munich, Bayern 81739
Germany
EMail: Hannes.Tschofenig@siemens.com
Appendix A. Comparison with draft-ietf-tls-sharedkeys-02 (informative)
[TLS-SHAREDKEYS] presents another way to use shared keys with TLS.
Instead of defining new ciphersuites, it re-uses the TLS session
cache and session resumption functionality.
The approach presented in this document is, in our opinion, more
elegant and better in line with the design of TLS. However, it does
probably require more changes to existing TLS implementations.
Nevertheless, these changes should be rather straightforward,
especially for implementations that already support multiple key
exchange algorithms and have a modular architecture.
Eronen & Tschofenig Expires August 6, 2004 [Page 7]
Internet-Draft PSK Ciphersuites for TLS February 2004
The changes required are roughly the following:
1. An API to pass psk_identities and keys around from the application
to the TLS library. Most likely, both push-style interface (use
this psk_identity and key) and callbacks (given a psk_identity,
fetch corresponding shared secret) would be useful.
2. An API to determine which psk_identity was used for a session.
3. PSK ciphersuite identifiers must be added to the list of supported
ciphersuites.
4. Processing of PSK messages in the handshake code.
The session-cache based approach probably requires the following
changes (depending on details of the TLS implementation):
1. Most TLS implementations do not expose an API that allows detailed
modification of the session cache, so some modifications are
required (especially if the implementation is done in some
reasonably type-safe language, the application cannot just use
some pointer tricks to access private data structures).
On the client side, we need an API to communicate session_id, key
and whatever is used to look up entries from the session cache
(for instance, some implementations use IP address and port
number) to the TLS implementation (and initialize other session
cache fields to some sensible values).
On the server side, we need to communicate session_id and key.
Most likely, both push-style interface (use this session_id and
key) and pull callbacks (given a session_id, fetch corresponding
shared secret) would be useful (but callbacks may require more
changes).
2. An API to determine which session_id was used (and to determine if
shared secret or normal RSA was used).
3. The session resumption code normally checks that resumed sessions
use the same ciphersuite as the original session. Unless a single
ciphersuite is hardcoded to the session cache, this code has to be
modified (and the session cache needs a flag indicating which
entries were created using ordinary handshake and which using
shared-secret API--unless the check is omitted for all sessions,
breaking TLS 1.0 rules).
Eronen & Tschofenig Expires August 6, 2004 [Page 8]
Internet-Draft PSK Ciphersuites for TLS February 2004
4. If the TLS implementation supports compression, resumed sessions
must use the same compression method as the original. Either
compressions has to be disabled or this code modified.
5. TLS implementation should also check that the resumed session uses
the same protocol version; this needs changes as well, unless a
single version number is hardcoded.
6. The session cache code may need modifications to ensure the stored
entries actually stay there long enough to be useful. Currently
implementations are free to discard entries whenever they want.
However, probably most implementations would not require any
changes.
Eronen & Tschofenig Expires August 6, 2004 [Page 9]
Internet-Draft PSK Ciphersuites for TLS February 2004
Intellectual Property Statement
The IETF takes no position regarding the validity or scope of any
intellectual property or other rights that might be claimed to
pertain to the implementation or use of the technology described in
this document or the extent to which any license under such rights
might or might not be available; neither does it represent that it
has made any effort to identify any such rights. Information on the
IETF's procedures with respect to rights in standards-track and
standards-related documentation can be found in BCP-11. Copies of
claims of rights made available for publication and any assurances of
licenses to be made available, or the result of an attempt made to
obtain a general license or permission for the use of such
proprietary rights by implementors or users of this specification can
be obtained from the IETF Secretariat.
The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights which may cover technology that may be required to practice
this standard. Please address the information to the IETF Executive
Director.
Full Copyright Statement
Copyright (C) The Internet Society (2004). All Rights Reserved.
This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph are
included on all such copies and derivative works. However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.
The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assignees.
This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
Eronen & Tschofenig Expires August 6, 2004 [Page 10]
Internet-Draft PSK Ciphersuites for TLS February 2004
HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Acknowledgment
Funding for the RFC Editor function is currently provided by the
Internet Society.
Eronen & Tschofenig Expires August 6, 2004 [Page 11]