Key Mgmt Guidelines: Guidelines for Cryptographic Key Management
When a protocol must use automated key management, and the narrow cases where manual keying is acceptable.
Bellovin, S. and R. Housley, "Guidelines for Cryptographic Key Management", BCP 107, RFC 4107, DOI 10.17487/RFC4107, June 2005, <https://www.rfc-editor.org/info/rfc4107>.
@misc{rfc4107,
series = {Request for Comments},
number = 4107,
howpublished = {RFC 4107},
publisher = {RFC Editor},
doi = {10.17487/RFC4107},
url = {https://www.rfc-editor.org/info/rfc4107},
author = {S. Bellovin and R. Housley},
title = {{Guidelines for Cryptographic Key Management}},
pagetotal = 7,
year = 2005,
month = jun,
} RFC text
Contents
Network Working Group S. Bellovin
Request for Comments: 4107 Columbia University
BCP: 107 R. Housley
Category: Best Current Practice Vigil Security
June 2005
Guidelines for Cryptographic Key Management
Status of This Memo
This document specifies an Internet Best Current Practices for the
Internet Community, and requests discussion and suggestions for
improvements. Distribution of this memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (2005).
Abstract
The question often arises of whether a given security system requires
some form of automated key management, or whether manual keying is
sufficient. This memo provides guidelines for making such decisions.
When symmetric cryptographic mechanisms are used in a protocol, the
presumption is that automated key management is generally but not
always needed. If manual keying is proposed, the burden of proving
that automated key management is not required falls to the proposer.
1. Introduction
The question often arises of whether or not a given security system
requires some form of automated key management, or whether manual
keying is sufficient.
There is not one answer to that question; circumstances differ. In
general, automated key management SHOULD be used. Occasionally,
relying on manual key management is reasonable; we propose some
guidelines for making that judgment.
On the other hand, relying on manual key management has significant
disadvantages, and we outline the security concerns that justify the
preference for automated key management. However, there are
situations in which manual key management is acceptable.
1.1. Terminology
The keywords MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD,
SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL, when they appear in this
document, are to be interpreted as described in RFC 2119 [B].
2. Guidelines
These guidelines are for use by IETF working groups and protocol
authors who are determining whether to mandate automated key
management and whether manual key management is acceptable. Informed
judgment is needed.
The term "key management" refers to the establishment of
cryptographic keying material for use with a cryptographic algorithm
to provide protocol security services, especially integrity,
authentication, and confidentiality. Automated key management
derives one or more short-term session keys. The key derivation
function may make use of long-term keys to incorporate authentication
into the process. The manner in which this long-term key is
distributed to the peers and the type of key used (pre-shared
symmetric secret value, RSA public key, DSA public key, and others)
is beyond the scope of this document. However, it is part of the
overall key management solution. Manual key management is used to
distribute such values. Manual key management can also be used to
distribute long-term session keys.
Automated key management and manual key management provide very
different features. In particular, the protocol associated with an
automated key management technique will confirm the liveness of the
peer, protect against replay, authenticate the source of the short-
term session key, associate protocol state information with the
short-term session key, and ensure that a fresh short-term session
key is generated. Further, an automated key management protocol can
improve interoperability by including negotiation mechanisms for
cryptographic algorithms. These valuable features are impossible or
extremely cumbersome to accomplish with manual key management.
For some symmetric cryptographic algorithms, implementations must
prevent overuse of a given key. An implementation of such algorithms
can make use of automated key management when the usage limits are
nearly exhausted, in order to establish replacement keys before the
limits are reached, thereby maintaining secure communications.
Examples of automated key management systems include IPsec IKE and
Kerberos. S/MIME and TLS also include automated key management
functions.
Key management schemes should not be designed by amateurs; it is
almost certainly inappropriate for working groups to design their
own. To put it in concrete terms, the very first key management
protocol in the open literature was published in 1978 [NS]. A flaw
and a fix were published in 1981 [DS], and the fix was cracked in
1994 [AN]. In 1995 [L], a new flaw was found in the original 1978
version, in an area not affected by the 1981/1994 issue. All of
these flaws were obvious once described -- yet no one spotted them
earlier. Note that the original protocol (translated to employ
certificates, which had not been invented at that time) was only
three messages.
Key management software is not always large or bloated. Even IKEv1
[HC] can be done in less than 200 Kbytes of object code, and TLS [DA]
in half that space. Note that this TLS estimate includes other
functionality as well.
A session key is used to protect a payload. The nature of the
payload depends on the layer where the symmetric cryptography is
applied.
In general, automated key management SHOULD be used to establish
session keys. Strong justification is needed in the security
considerations section of a proposal that makes use of manual key
management.
2.1. Automated Key Management
Automated key management MUST be used if any of these conditions
hold:
A party will have to manage n^2 static keys, where n may become
large.
Any stream cipher (such as RC4 [TK], AES-CTR [NIST], or AES-CCM
[WHF]) is used.
An initialization vector (IV) might be reused, especially an
implicit IV. Note that random or pseudo-random explicit IVs are
not a problem unless the probability of repetition is high.
Large amounts of data might need to be encrypted in a short time,
causing frequent change of the short-term session key.
Long-term session keys are used by more than two parties.
Multicast is a necessary exception, but multicast key management
standards are emerging in order to avoid this in the future.
Sharing long-term session keys should generally be discouraged.
The likely operational environment is one where personnel (or
device) turnover is frequent, causing frequent change of the
short-term session key.
2.2. Manual Key Management
Manual key management may be a reasonable approach in any of these
situations:
The environment has very limited available bandwidth or very high
round-trip times. Public key systems tend to require long
messages and lots of computation; symmetric key alternatives, such
as Kerberos, often require several round trips and interaction
with third parties.
The information being protected has low value.
The total volume of traffic over the entire lifetime of the long-
term session key will be very low.
The scale of each deployment is very limited.
Note that assertions about such things should often be viewed with
skepticism. The burden of demonstrating that manual key management
is appropriate falls to the proponents -- and it is a fairly high
hurdle.
Systems that employ manual key management need provisions for key
changes. There MUST be some way to indicate which key is in use to
avoid problems during transition. Designs SHOULD sketch plausible
mechanisms for deploying new keys and replacing old ones that might
have been compromised. If done well, such mechanisms can later be
used by an add-on key management scheme.
Lack of clarity about the parties involved in authentication is not a
valid reason for avoiding key management. Rather, it tends to
indicate a deeper problem with the underlying security model.
2.3. Key Size and Random Values
Guidance on cryptographic key size for public keys that are used for
exchanging symmetric keys can be found in BCP 86 [OH].
When manual key management is used, long-term shared secret values
SHOULD be at least 128 bits.
Guidance on random number generation can be found in BCP 106 [ESC].
When manual key management is used, long-term shared secrets MUST be
unpredictable "random" values, ensuring that an adversary will have
no greater expectation than 50% of finding the value after searching
half the key search space.
3. Security Considerations
This document provides guidance to working groups and protocol
designers. The security of the Internet is improved when automated
key management is employed.
The inclusion of automated key management does not mean that an
interface for manual key management is prohibited. In fact, manual
key management is very helpful for debugging. Therefore,
implementations ought to provide a manual key management interface
for such purposes, even if it is not specified by the protocol.
4. References
This section contains normative and informative references.
4.1. Normative References
[B] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997.
[ESC] Eastlake, D., 3rd, Schiller, J., and S. Crocker, "Randomness
Requirements for Security", BCP 106, RFC 4086, June 2005.
[OH] Orman, H. and P. Hoffman, "Determining Strengths For Public
Keys Used For Exchanging Symmetric Keys", BCP 86, RFC 3766,
April 2004
4.2. Informative References
[AN] M. Abadi and R. Needham, "Prudent Engineering Practice for
Cryptographic Protocols", Proc. IEEE Computer Society
Symposium on Research in Security and Privacy, May 1994.
[DA] Dierks, T. and C. Allen, "The TLS Protocol Version 1.0", RFC
2246, January 1999.
[DS] D. Denning and G. Sacco. "Timestamps in key distributed
protocols", Communication of the ACM, 24(8):533--535, 1981.
[HC] Harkins, D. and D. Carrel, "The Internet Key Exchange (IKE)",
RFC 2409, November 1998.
[L] G. Lowe. "An attack on the Needham-Schroeder public key
authentication protocol", Information Processing Letters,
56(3):131--136, November 1995.
[NIST] National Institute of Standards and Technology.
"Recommendation for Block Cipher Modes of Operation -- Methods
and Techniques," NIST Special Publication SP 800-38A, December
2001.
[NS] R. Needham and M. Schroeder. "Using encryption for
authentication in large networks of computers", Communications
of the ACM, 21(12), December 1978.
[TK] Thayer, R. and K. Kaukonen. "A Stream Cipher Encryption
Algorithm", Work in Progress.
[WHF] Whiting, D., Housley, R., and N. Ferguson , "Counter with
CBC-MAC (CCM)", RFC 3610, September 2003.
Authors' Addresses
Steven M. Bellovin
Department of Computer Science
Columbia University
1214 Amsterdam Avenue, M.C. 0401
New York, NY 10027-7003
Phone: +1 212-939-7149
EMail: bellovin@acm.org
Russell Housley
Vigil Security, LLC
918 Spring Knoll Drive
Herndon, VA 20170
Phone: +1 703-435-1775
EMail: housley@vigilsec.com
Full Copyright Statement
Copyright (C) The Internet Society (2005).
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
INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Intellectual Property
The IETF takes no position regarding the validity or scope of any
Intellectual Property Rights 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; nor does it represent that it has
made any independent effort to identify any such rights. Information
on the procedures with respect to rights in RFC documents can be
found in BCP 78 and BCP 79.
Copies of IPR disclosures made to the IETF Secretariat 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 implementers or users of this
specification can be obtained from the IETF on-line IPR repository at
http://www.ietf.org/ipr.
The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights that may cover technology that may be required to implement
this standard. Please address the information to the IETF at ietf-
ipr@ietf.org.
Acknowledgement
Funding for the RFC Editor function is currently provided by the
Internet Society.
AI ELI5
On this page
§1The Question Being Answered
The question often arises whether a given security system needs some form of automated key management, or whether manual keying is sufficient. The RFC says there is no single answer because circumstances differ. In general, automated key management SHOULD be used, and relying on manual key management has significant disadvantages. Occasionally it is reasonable, and the document proposes guidelines for judging that.
The guidelines are for IETF working groups and protocol authors deciding whether to mandate automated key management and whether manual key management is acceptable. The abstract adds that when symmetric cryptography is used, the presumption is that automated key management is generally but not always needed, and that a proposer of manual keying carries the burden of proof.
When two computers want to keep secrets, they need keys. This RFC helps decide whether the computers should swap keys on their own or whether a person can set them up once.
§2What Key Management Is
Key management means establishing the keying material used with a cryptographic algorithm to provide protocol security services, especially integrity, authentication and confidentiality. Automated key management derives one or more short-term session keys, and the derivation may use long-term keys to bring in authentication. How those long-term keys reach the peers, and what type they are (such as a pre-shared symmetric secret, an RSA public key or a DSA public key), is outside the document's scope. Manual key management is used to distribute them, and it can also distribute long-term session keys.
The two approaches offer very different features. An automated protocol confirms the liveness of the peer, protects against replay, authenticates the source of the short-term session key, associates protocol state with it and ensures a fresh key is generated. It can also improve interoperability by negotiating cryptographic algorithms. The RFC says these are impossible or extremely cumbersome with manual key management.
Some symmetric algorithms require implementations to prevent overuse of a given key. Automated key management can establish replacement keys before the usage limits are reached. Examples named are IPsec IKE and Kerberos, and S/MIME and TLS include automated key management functions too.
Automatic key swapping is like a security guard who checks that you are really there, gives you a brand-new key each visit, and makes sure nobody replays an old one. A key you set up by hand cannot do any of that.
| Feature | Automated | Manual |
|---|---|---|
| Peer liveness confirmed | Yes | Impossible or very cumbersome |
| Replay protection | Yes | Impossible or very cumbersome |
| Fresh short-term session key | Yes | Impossible or very cumbersome |
| Algorithm negotiation | Can be included | Impossible or very cumbersome |
§2.1When Automation Is Required
The RFC says key management schemes should not be designed by amateurs, and that it is almost certainly inappropriate for working groups to design their own. As evidence it recounts that the first key management protocol in the open literature was published in 1978, a flaw and a fix appeared in 1981, the fix was cracked in 1994, and a new flaw in the original was found in 1995, even though all these flaws were obvious once described. It also notes that key management software is not always large: IKEv1 can be done in less than 200 Kbytes of object code, and TLS in half that.
Automated key management MUST be used if any of these conditions hold:
- A party will have to manage n^2 static keys, where n may become large.
- Any stream cipher is used, such as RC4, AES-CTR or AES-CCM.
- An initialization vector (IV) might be reused, especially an implicit IV. Random or pseudo-random explicit IVs are not a problem unless the probability of repetition is high.
- Large amounts of data might need to be encrypted in a short time, causing frequent change of the short-term session key.
- Long-term session keys are used by more than two parties. Multicast is a necessary exception, but multicast key management standards are emerging to avoid it.
- The operational environment has frequent personnel or device turnover.
In general, automated key management SHOULD be used to establish session keys, and a proposal using manual key management needs strong justification in its security considerations section.
Some situations are too risky for hand-made keys: too many keys to keep track of, kinds of secret codes that break if a number repeats, lots of secret traffic, or people and devices that come and go often. Then the computer must do the swapping.
§2.2When Manual Keying May Be Reasonable
Manual key management may be a reasonable approach in any of these situations:
- The environment has very limited bandwidth or very high round-trip times. Public key systems tend to need long messages and a lot of computation, and symmetric alternatives such as Kerberos often need several round trips and third parties.
- The information being protected has low value.
- The total traffic over the lifetime of the long-term session key will be very low.
- The scale of each deployment is very limited.
The RFC warns that such assertions should often be viewed with skepticism. The burden of showing that manual keying is appropriate falls on its proponents, and it is a fairly high hurdle.
A system that uses manual keys needs provisions for changing them. There MUST be some way to indicate which key is in use, to avoid problems during a transition. Designs SHOULD sketch plausible ways to deploy new keys and replace old ones that might have been compromised, because a well-done mechanism can later be reused by an add-on key management scheme. Finally, lack of clarity about which parties are authenticating is not a valid reason to avoid key management. It tends to point to a deeper problem with the underlying security model.
Doing it by hand can be fine when the connection is very slow, the secret is not valuable, hardly anything is sent, or there are only a few users. Even then, you must be able to tell which key is current and have a plan for swapping a stolen one.
§2.3Key Size and Random Values
For public keys used to exchange symmetric keys, the RFC points to BCP 86 (RFC 3766) for key size guidance. For random number generation it points to BCP 106 (RFC 4086).
When manual key management is used, long-term shared secret values SHOULD be at least 128 bits. They MUST also be unpredictable "random" values, so that an adversary has no greater expectation than 50% of finding the value after searching half the key search space.
A hand-made secret must be long and truly unguessable, like a combination picked by rolling dice rather than by choosing a birthday.
§3Security Considerations
This document gives guidance to working groups and protocol designers, and it says the security of the Internet is improved when automated key management is employed.
Including automated key management does not prohibit an interface for manual key management. The RFC says manual key management is very helpful for debugging, so implementations ought to provide a manual interface for that purpose, even if the protocol does not specify one.
Even when keys are swapped automatically, it is handy to keep a manual setting for testing, like a spare handle to try the door by hand.