1 Network Working Group M. Horowitz
2 <draft-horowitz-key-derivation-01.txt> Cygnus Solutions
3 Internet-Draft March, 1997
6 Key Derivation for Authentication, Integrity, and Privacy
10 This document is an Internet-Draft. Internet-Drafts are working
11 documents of the Internet Engineering Task Force (IETF), its areas,
12 and its working groups. Note that other groups may also distribute
13 working documents as Internet-Drafts.
15 Internet-Drafts are draft documents valid for a maximum of six months
16 and may be updated, replaced, or obsoleted by other documents at any
17 time. It is inappropriate to use Internet-Drafts as reference
18 material or to cite them other than as ``work in progress.''
20 To learn the current status of any Internet-Draft, please check the
21 ``1id-abstracts.txt'' listing contained in the Internet-Drafts Shadow
22 Directories on ds.internic.net (US East Coast), nic.nordu.net
23 (Europe), ftp.isi.edu (US West Coast), or munnari.oz.au (Pacific
26 Distribution of this memo is unlimited. Please send comments to the
31 Recent advances in cryptography have made it desirable to use longer
32 cryptographic keys, and to make more careful use of these keys. In
33 particular, it is considered unwise by some cryptographers to use the
34 same key for multiple purposes. Since most cryptographic-based
35 systems perform a range of functions, such as authentication, key
36 exchange, integrity, and encryption, it is desirable to use different
37 cryptographic keys for these purposes.
39 This RFC does not define a particular protocol, but defines a set of
40 cryptographic transformations for use with arbitrary network
41 protocols and block cryptographic algorithm.
46 In order to use multiple keys for different functions, there are two
49 - Each protocol ``key'' contains multiple cryptographic keys. The
50 implementation would know how to break up the protocol ``key'' for
51 use by the underlying cryptographic routines.
53 - The protocol ``key'' is used to derive the cryptographic keys.
54 The implementation would perform this derivation before calling
60 Internet Draft Key Derivation March, 1997
63 the underlying cryptographic routines.
65 In the first solution, the system has the opportunity to provide
66 separate keys for different functions. This has the advantage that
67 if one of these keys is broken, the others remain secret. However,
68 this comes at the cost of larger ``keys'' at the protocol layer. In
69 addition, since these ``keys'' may be encrypted, compromising the
70 cryptographic key which is used to encrypt them compromises all the
71 component keys. Also, the not all ``keys'' are used for all possible
72 functions. Some ``keys'', especially those derived from passwords,
73 are generated from limited amounts of entropy. Wasting some of this
74 entropy on cryptographic keys which are never used is unwise.
76 The second solution uses keys derived from a base key to perform
77 cryptographic operations. By carefully specifying how this key is
78 used, all of the advantages of the first solution can be kept, while
79 eliminating some disadvantages. In particular, the base key must be
80 used only for generating the derived keys, and this derivation must
81 be non-invertible and entropy-preserving. Given these restrictions,
82 compromise of one derived keys does not compromise the other subkeys.
83 Attack of the base key is limited, since it is only used for
84 derivation, and is not exposed to any user data.
86 Since the derived key has as much entropy as the base keys (if the
87 cryptosystem is good), password-derived keys have the full benefit of
88 all the entropy in the password.
90 To generate a derived key from a base key:
92 Derived Key = DK(Base Key, Well-Known Constant)
96 DK(Key, Constant) = n-truncate(E(Key, Constant))
98 In this construction, E(Key, Plaintext) is a block cipher, Constant
99 is a well-known constant defined by the protocol, and n-truncate
100 truncates its argument by taking the first n bits; here, n is the key
103 If the output of E is is shorter than n bits, then some entropy in
104 the key will be lost. If the Constant is smaller than the block size
105 of E, then it must be padded so it may be encrypted. If the Constant
106 is larger than the block size, then it must be folded down to the
107 block size to avoid chaining, which affects the distribution of
110 In any of these situations, a variation of the above construction is
111 used, where the folded Constant is encrypted, and the resulting
112 output is fed back into the encryption as necessary (the | indicates
115 K1 = E(Key, n-fold(Constant))
122 Internet Draft Key Derivation March, 1997
128 DK(Key, Constant) = n-truncate(K1 | K2 | K3 | K4 ...)
130 n-fold is an algorithm which takes m input bits and ``stretches''
131 them to form n output bits with no loss of entropy, as described in
132 [Blumenthal96]. In this document, n-fold is always used to produce n
133 bits of output, where n is the key size of E.
135 If the size of the Constant is not equal to the block size of E, then
136 the Constant must be n-folded to the block size of E. This number is
137 used as input to E. If the block size of E is less than the key
138 size, then the output from E is taken as input to a second invocation
139 of E. This process is repeated until the number of bits accumulated
140 is greater than or equal to the key size of E. When enough bits have
141 been computed, the first n are taken as the derived key.
143 Since the derived key is the result of one or more encryptions in the
144 base key, deriving the base key from the derived key is equivalent to
145 determining the key from a very small number of plaintext/ciphertext
146 pairs. Thus, this construction is as strong as the cryptosystem
150 Deriving Keys from Passwords
152 When protecting information with a password or other user data, it is
153 necessary to convert an arbitrary bit string into an encryption key.
154 In addition, it is sometimes desirable that the transformation from
155 password to key be difficult to reverse. A simple variation on the
156 construction in the prior section can be used:
158 Key = DK(n-fold(Password), Well-Known Constant)
160 The n-fold algorithm is reversible, so recovery of the n-fold output
161 is equivalent to recovery of Password. However, recovering the n-
162 fold output is difficult for the same reason recovering the base key
163 from a derived key is difficult.
167 Traditionally, the transformation from plaintext to ciphertext, or
168 vice versa, is determined by the cryptographic algorithm and the key.
169 A simple way to think of derived keys is that the transformation is
170 determined by the cryptographic algorithm, the constant, and the key.
172 For interoperability, the constants used to derive keys for different
173 purposes must be specified in the protocol specification. The
174 constants must not be specified on the wire, or else an attacker who
175 determined one derived key could provide the associated constant and
176 spoof data using that derived key, rather than the one the protocol
184 Internet Draft Key Derivation March, 1997
187 Determining which parts of a protocol require their own constants is
188 an issue for the designer of protocol using derived keys.
191 Security Considerations
193 This entire document deals with security considerations relating to
194 the use of cryptography in network protocols.
199 I would like to thank Uri Blumenthal, Hugo Krawczyk, and Bill
200 Sommerfeld for their contributions to this document.
205 [Blumenthal96] Blumenthal, U., "A Better Key Schedule for DES-Like
206 Ciphers", Proceedings of PRAGOCRYPT '96, 1996.
213 955 Massachusetts Avenue
216 Phone: +1 617 354 7688
217 Email: marc@cygnus.com