Internet Draft Editor: Peter Gutmann draft-ietf-smime-password-02.txt University of Auckland March 2, 2000 Expires September 2000 Password-based Encryption for S/MIME 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. Abstract The Cryptographic Message Syntax data format doesn't currently contain any provisions for password-based data encryption. This document provides a method of encrypting data using user-supplied passwords and, by extension, any form of variable-length keying material which isn't necessarily an algorithm-specific fixed-format key. This draft is being discussed on the "ietf-smime" mailing list. To join the list, send a message to with the single word "subscribe" in the body of the message. Also, there is a Web site for the mailing list at . 1. Introduction This document describes a password-based content encryption mechanism for S/MIME. This is implemented as a new RecipientInfo type and is an extension to the RecipientInfo types currently defined in RFC 2640 [RFC2640]. The format of the messages are described in ASN.1:1994 [ASN1]. The key words "MUST", "MUST NOT", "REQUIRED", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119]. 1.1 Password-based Content Encryption CMS currently defined three recipient information types for public-key key wrapping (KeyTransRecipientInfo), conventional key wrapping (KEKRecipientInfo), and key agreement (KeyAgreeRecipientInfo). The recipient information described here adds a fourth type, PasswordRecipientInfo, which provides for password-based key wrapping. 1.2 RecipientInfo Types The new recipient information type is an extension to the RecipientInfo type defined in section 6.2 of CMS, extending the types to: RecipientInfo ::= CHOICE { ktri KeyTransRecipientInfo, kari [1] KeyAgreeRecipientInfo, kekri [2] KEKRecipientInfo, pwri [3] PasswordRecipientinfo -- New RecipientInfo type } Although the recipient information generation process is described in terms of a password-based operation (since this will be its most common use), the transformation employed is a general-purpose key derivation one which allows any type of keying material to be converted into a key specific to a particular content-encryption algorithm. Since the most common use for password-based encryption is to encrypt files which are stored locally (rather than being transmitted across a network), the term "recipient" is somewhat misleading, but is used here because the other key transport mechanisms have always been described in similar terms. 1.2.1 PasswordRecipientInfo Type Recipient information using a user-supplied password or previously agreed-upon key is represented in the type PasswordRecipientInfo. Each instance of PasswordRecipientInfo will transfer the content-encryption key (CEK) to one or more recipients who have the previously agreed-upon password or key-encryption key (KEK). PasswordRecipientInfo ::= SEQUENCE { version CMSVersion, -- Always set to 0 keyDerivationAlgorithm [0] KeyDerivationAlgorithmIdentifier OPTIONAL, keyEncryptionAlgorithm KeyEncryptionAlgorithmIdentifier, encryptedKey EncryptedKey } The fields of type PasswordRecipientInfo have the following meanings: version is the syntax version number. It shall always be 0. keyDerivationAlgorithm identifies the key-derivation algorithm, and any associated parameters, used to derive the KEK from the user-supplied password. If this field is absent, the KEK is supplied from an external source, for example a crypto token such as a smart card. keyEncryptionAlgorithm identifies the content-encryption algorithm, and any associated parameters, used to encrypt the CEK with the KEK. encryptedKey is the result of encrypting the content-encryption key with the KEK. 1.2.2 Rationale Password-based key wrapping is a two-stage process, a first stage in which a user-supplied password is converted into a KEK if required, and a second stage in which the KEK is used to encrypt a CEK. These two stages are identified by the two algorithm identifiers. Although the PKCS #5 standard goes one step further to wrap these up into a single algorithm identifier, this design is particular to that standard and may not be applicable for other key wrapping mechanisms. For this reason the two steps are specified separately. 2 Supported Algorithms This section lists the algorithms that must be implemented. Additional algorithms that should be implemented are also included. 2.1 Key Derivation Algorithms These algorithms are used to convert the password into a KEK. The key derivation algorithms are: KeyDerivationAlgorithmIdentifer ALGORITHM-IDENTIFIER ::= { { SYNTAX PBKDF2-params IDENTIFIED BY id-PBKDF2 }, ... } CMS implementations MUST include PBKDF2 [PKCS5v2]. 2.2 Key Encryption Algorithms These algorithms are used to encrypt the content (the key) using the derived KEK. The content encryption algorithms are: KeyEncryptionAlgorithmIdentifer ALGORITHM-IDENTIFIER ::= PBES2-Encs CMS implementations MUST include Triple-DES in CBC mode, SHOULD include RC2 in CBC mode, and MAY include other algorithms such as CAST-128, RC5, IDEA, Skipjack, and encryption modes as required. CMS implementations SHOULD NOT include any KSG ciphers such as RC4 or a block cipher in OFB mode, and SHOULD NOT include a block cipher in ECB mode. The use of RC2 has special requirements, see section 2.4 for details. 2.3 Symmetric Key Encryption Algorithms The key wrap algorithm is used to wrap the CEK with the KEK. There is no requirement that the content-encryption algorithm match the KEK algorithm, although care should be taken to ensure that, if different algorithms are used, they offer an equivalent level of security (for example wrapping a Triple-DES key with an RC2/40 key leads to a severe impedance mismatch in encryption strength). The key wrap algorithm specified below is independent of the content-encryption or wrapping algorithms, relying only on the use of a block cipher to perform the wrapping. 2.3.1 Key Wrap The key wrap algorithm encrypts a CEK with a KEK in a manner which ensures that every bit of plaintext effects every bit of ciphertext. This makes it equivalent in function to the package transform [PACKAGE] without requiring additional mechanisms or resources such as hash functions or cryptographically strong random numbers. The key wrap algorithm is performed in two phases, a first phase which formats the CEK into a form suitable for encryption by the KEK, and a second phase which wraps the formatted CEK using the KEK. Key formatting: Create a formatted CEK block consisting of the following: 1. A one-byte count of the number of bytes in the CEK. 2. A check value containing the bitwise complement of the first three bytes of the CEK. 3. The CEK. 4. Enough random padding data to make the CEK data block at least two KEK cipher blocks long (the fact that 32 bits of count+check value are used means that even with a 40-bit CEK, the resulting data size will always be at least two (64-bit) cipher blocks long). The padding data does not have to be cryptographically strong, although unpredictability helps. The formatted CEK block then looks as follows: CEK byte count || check value || CEK || padding (if required) Key wrapping: 1. Encrypt the padded key using the KEK. 2. Without resetting the IV (that is, using the last ciphertext block as the IV), encrypt the encrypted padded key a second time. The resulting double-encrypted data is the EncryptedKey. 2.3.2 Key Unwrap Key unwrapping: 1. Using the n-1'th ciphertext block as the IV, decrypt the n'th ciphertext block. 2. Using the decrypted n'th ciphertext block as the IV, decrypt the 1st ... n-1'th ciphertext blocks. This strips the outer layer of encryption. 3. Decrypt the inner layer of encryption using the KEK. Key format verification: 1a.If the CEK byte count is less than the minimum allowed key size (usually 5 bytes for 40-bit keys) or greater than the wrapped CEK length or not valid for the CEK algorithm (eg not 16 or 24 bytes for triple DES), the KEK was invalid. 1b.If the bitwise complement of the key check value doesn't match the first three bytes of the key, the KEK was invalid. 2.3.3 Example Given a content-encryption algorithm of Skipjack and a KEK algorithm of Triple-DES, the wrap steps are as follows: 1. Set the first 4 bytes of the CEK block to the Skipjack key size (10 bytes) and the bitwise complement of the first three bytes of the CEK. 2. Append the 80-bit (10-byte) Skipjack CEK and pad the total to 16 bytes (two triple-DES blocks) using 2 bytes of random data. 2. Using the IV given in the KeyEncryptionAlgorithmIdentifer, encrypted the padded Skipjack key. 3. Without resetting the IV, encrypt the encrypted padded key a second time. The unwrap steps are as follows: 1. Using the first 8 bytes of the double-encrypted key as the IV, decrypt the second 8 bytes. 2. Without resetting the IV, decrypt the first 8 bytes. 3. Decrypt the inner layer of encryption using the the IV given in the KeyEncryptionAlgorithmIdentifer to recover the padded Skipjack key. 4. If the length byte isn't equal to the Skipjack key size (80 bits or 10 bytes) or the bitwise complement of the check bytes doesn't match the first three bytes of the CEK, the KEK was invalid. 2.3.4 Rationale for the Double Wrapping If many CEK's are encrypted in a standard way with the same KEK and the KEK has a 64-bit block size then after about 2^32 encryptions there is a high probability of a collision between different blocks of encrypted CEK's. If an opponent manages to obtain a CEK, they may be able to solve for other CEK's. The double-encryption wrapping process, which makes every bit of ciphertext dependent on every bit of the CEK, eliminates this collision problem (as well as preventing other potential problems such as bit-flipping attacks). Since the IV is applied to the inner layer of encryption, even wrapping the same CEK with the same KEK will result in a completely different wrapped key each time. 2.4 Special Handling for RC2 Keys For a variety of historical, political, and software-peculiarity reasons which are beyond the scope of this document, the handling of keys for the RC2 algorithm [RFC2268] by different implementations is somewhat arbitrary. In particular, the choice of actual vs effective key bits used in the algorithm is often unclear. The standard RC2 AlgorithmIdentifier only allows the effective key bits to be specified, leaving the actual key bits to be communicated via out-of-band means, which in some cases means hardcoding them into applications. Solving this problem requires two things, a precise definition of how keys represented with the standard RC2 AlgorithmIdentifier are handled, and a new RC2 AlgorithmIdentifier which allows keys currently in use by different applications to be handled. 2.4.1 Handling of RC2 with RFC 2268 AlgorithmIdentifier RFC 2268 defines the following AlgorithmIdentifier for RC2: rc2CBC OBJECT IDENTIFIER ::= {iso(1) member-body(2) US(840) rsadsi(113549) encryptionAlgorithm(3) 2} RC2-CBCParameter ::= CHOICE { iv IV, params SEQUENCE { version INTEGER, iv OCTET STRING } } where the version field encodes the effective key size in a complex manner specified in the RFC. Where this algorithm identifier is used, the actual key size shall be the same size as the effective key size as given by the version field. When RC2 is to be used, implementations should use this AlgorithmIdentifier and parameters, and when this AlgorithmIdentifier is used the actual key size MUST NOT be a value other than the effective key size (to use a different size, see section 2.4.2). 2.4.2 Handling of RC2 with Other Key Sizes If the use of an actual key size of other than the effective key size is required, implementations MUST use the following AlgorithmIdentifier: rc2CBC OBJECT IDENTIFIER ::= {1 3 6 1 4 1 3029 666 13} RC2-CBCParameter ::= SEQUENCE { actualKeySize INTEGER, -- Actual key size in bits effectiveKeySize INTEGER, -- Effective key size in bits iv OCTET STRING } This allows arbitrary actual and effective key sizes to be specified for compatibility with existing usage. Although implementations SHOULD NOT use this alternative (using instead the one in section 2.4.1) experience has shown that implementors will continue to use oddball RC2 parameters anyway, so new implementations should be prepared to encounter and handle actual and effective key sizes ranging from 40 up to around 200 bits. 2.4.3 Rationale The reason for providing for the handling of oddball key sizes is compatibility with existing applications, for example a mailing-list exploder or mail gateway may take an RSA-wrapped CEK generated by a current application and repackage it with a KEK, so we need a mechanism for handling strange key lengths in a manner which is compatible with existing usage. The alternative RC2 AlgorithmIdentifier, although not recommended, provides a means of ensuring this compatibility. 3. Test Vectors The following values are obtained when wrapping a 256-bit Blowfish-CFB key using a triple DES-CBC key derived from the passphrase "All n-entities must communicate with other n-entities via n-1 entiteeheehees" with salt { 12 34 56 78 78 56 34 12 } using 500 iterations of PBKDF2. PKCS #5v2 values: input 41 6C 6C 20 6E 2D 65 6E 74 69 74 69 65 73 20 6D passphrase: 75 73 74 20 63 6F 6D 6D 75 6E 69 63 61 74 65 20 77 69 74 68 20 6F 74 68 65 72 20 6E 2d 65 6E 74 69 74 69 65 73 20 76 69 61 20 6E 2D 31 20 65 6E 74 69 74 65 65 68 65 65 68 65 65 73 "All n-entities must communicate with other " "n-entities via n-1 entiteeheehees" input salt: 12 34 56 78 78 56 34 12 iterations: 500 output 6A 89 70 BF 68 C9 2C AE A8 4A 8D F2 85 10 85 86 3DES key: 07 12 63 80 CC 47 AB 2D CEK formatting phase: length byte: 20 key check: 73 9C 82 CEK: 8C 63 7D 88 72 23 A2 F9 65 B5 66 EB 01 4B 0F A5 D5 23 00 A3 F7 EA 40 FF FC 57 72 03 C7 1B AF 3B padding: FA 06 0A 45 complete 20 93 9C 82 8C 63 7D 88 72 23 A2 F9 65 B5 66 EB CEK block: 01 4B 0F A5 D5 23 00 A3 F7 EA 40 FF FC 57 72 03 C7 1B AF 3B FA 06 0A 45 Key wrap phase (wrap CEK block using 3DES key): IV: BA F1 CA 79 31 21 3C 4E first encr. F8 3F 9E 16 78 51 41 10 64 27 65 A9 F5 D8 71 CD pass output: 27 DB AA 41 E7 BD 80 48 A9 08 20 FF 40 82 A2 80 96 9E 65 27 9E 12 6A EB second encr. C0 3C 51 4A BD B9 E2 C5 AA C0 38 57 2B 5E 24 55 pass output: 38 76 B3 77 AA FB 82 EC A5 A9 D7 3F 8A B1 43 D9 EC 74 E6 CA D7 DB 26 0C ASN.1 encoded PasswordRecipientInfo: 0 A3 96: [3] { 2 02 1: INTEGER 0 5 A0 27: [0] { 7 06 9: OBJECT IDENTIFIER id-PBKDF2 (1 2 840 113549 1 5 12) 18 30 14: SEQUENCE { 20 04 8: OCTET STRING : 12 34 56 78 78 56 34 12 30 02 2: INTEGER 500 : } : } 34 30 20: SEQUENCE { 36 06 8: OBJECT IDENTIFIER des-EDE3-CBC (1 2 840 113549 3 7) 46 04 8: OCTET STRING : BA F1 CA 79 31 21 3C 4E : } 56 04 40: OCTET STRING : C0 3C 51 4A BD B9 E2 C5 AA C0 38 57 2B 5E 24 55 : 38 76 B3 77 AA FB 82 EC A5 A9 D7 3F 8A B1 43 D9 : EC 74 E6 CA D7 DB 26 0C : } 4. Security Considerations The security of this recipient information type rests on the security of the underlying mechanisms employed, for which further information can be found in RFC 2640 and PKCS5v2. More importantly, however, when used with a password the security of this information type rests on the entropy of the user-selected password, which is typically quite low. Pass phrases (as opposed to simple passwords) are STRONGLY RECOMMENDED, although it should be recognized that even with pass phrases it will be difficult to use this recipient information type to derive a KEK with sufficient entropy to properly protect a 128-bit (or higher) CEK. 5. Changes Since the Last Version - Added length count due to concerns over odd-length RC2 keys - Added check value to allow better error reporting and balance out the single-byte length value - Included test vectors Author Address Peter Gutmann University of Auckland Private Bag 92019 Auckland, New Zealand pgut001@cs.auckland.ac.nz References ASN1 Recommendation X.680: Specification of Abstract Syntax Notation One (ASN.1), 1994. PKCS5v2 PKCS #5 v2.0: Password-Based Cryptography Standard, RSA Laboratories, 25 March 1999. RFC2119 Key Words for Use in RFC's to Indicate Requirement Levels, S.Bradner, March 1997. RFC2268 A Description of the RC2(r) Encryption Algorithm, R.Rivest, March 1998. RFC2640 Cryptographic Message Syntax, draft-ietf-smime-cms-11.txt, Russ Housley, April 1999. PACKAGE All-or-Nothing Encryption and the Package Transform, R.Rivest, Proceedings of Fast Software Encryption '97, Haifa, Israel, January 1997. Appendix A: ASN.1 Module PasswordRecipientInfo { iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-9(9) smime(16) modules(0) pwri(n+1) } DEFINITIONS IMPLICIT TAGS ::= BEGIN IMPORTS FROM PKCS5 { iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-5(5) } PBKDF2-params, PBES2-Encs; PasswordRecipientInfo ::= SEQUENCE { version CMSVersion, -- Always set to 0 keyDerivationAlgorithm [0] KeyDerivationAlgorithmIdentifier OPTIONAL, keyEncryptionAlgorithm KeyEncryptionAlgorithmIdentifier, encryptedKey EncryptedKey } KeyDerivationAlgorithmIdentifer ALGORITHM-IDENTIFIER ::= { { SYNTAX PBKDF2-params IDENTIFIED BY id-PBKDF2 }, ... } KeyEncryptionAlgorithmIdentifer ALGORITHM-IDENTIFIER ::= PBES2-Encs END Full Copyright Statement Copyright (C) The Internet Society 1999. 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 assigns. 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 HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.