Network Working Group T. Ts'o, Editor Internet-Draft VA Linux Systems draft-tso-telnet-enc-des-ofb-03.txt August 1999 Telnet Encryption: DES 64 bit Output Feedback 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. 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 RFC 2119. 0. Abstract This document specifies how to use the DES encryption algorithm in output feedback mode with the telnet encryption option. 1. Command Names and Codes Encryption Type DES_OFB64 2 Suboption Commands OFB64_IV 1 OFB64_IV_OK 2 OFB64_IV_BAD 3 Expires February 19100 [Page 1] Internet-Draft DES 64 bit Output Feedback August 1999 2. Command Meanings IAC SB ENCRYPT IS DES_OFB64 OFB64_IV IAC SE The sender of this command generates a random 8 byte initial vec- tor, and sends it to the other side of the connection using the OFB64_IV command. The initial vector is sent in clear text. Only the side of the connection that is WILL ENCRYPT may send the OFB64_IV command IAC SB ENCRYPT REPLY DES_OFB64 OFB64_IV_OK IAC SE IAC SB ENCRYPT REPLY DES_OFB64 OFB64_IV_BAD IAC SE The sender of these commands either accepts or rejects the initial vector received in a OFB64_IV command. Only the side of the con- nection that is DO ENCRYPT may send the OFB64_IV_OK and OFB64_IV_BAD commands. The OFB64_IV_OK command MUST be sent for backwards compatibility with existing implementations; there real- ly isn't any reason why a sender would need to send the OFB64_IV_BAD command except in the case of a protocol violation where the IV sent was not of the correct length (i.e., 8 bytes). 3. Implementation Rules Once a OFB64_IV_OK command has been received, the WILL ENCRYPT side of the connection should do keyid negotiation using the ENC_KEYID command. Once the keyid negotiation has successfully identified a common keyid, then START and END commands may be sent by the side of the connection that is WILL ENCRYPT. Data will be encrypted using the DES 64 bit Output Feedback algorithm. If encryption (decryption) is turned off and back on again, and the same keyid is used when re-starting the encryption (decryption), the intervening clear text must not change the state of the encryption (decryption) machine. If a START command is sent (received) with a different keyid, the en- cryption (decryption) machine must be re-initialized immediately fol- lowing the end of the START command with the new key and the initial vector sent (received) in the last OFB64_IV command. If a new OFB64_IV command is sent (received), and encryption (decryp- tion) is enabled, the encryption (decryption) machine must be re-ini- tialized immediately following the end of the OFB64_IV command with the new initial vector, and the keyid sent (received) in the last START command. If encryption (decryption) is not enabled when a OFB64_IV command is Expires February 19100 [Page 2] Internet-Draft DES 64 bit Output Feedback August 1999 sent (received), the encryption (decryption) machine must be re-ini- tialized after the next START command, with the keyid sent (received) in that START command, and the initial vector sent (received) in this OFB64_IV command. 4. Algorithm Given that V[i] is the initial 64 bit vector, V[n] is the nth 64 bit vector, D[n] is the nth chunk of 64 bits of data to encrypt (de- crypt), and O[n] is the nth chunk of 64 bits of encrypted (decrypted) data, then: V[0] = DES(V[i], key) V[n+1] = DES(V[n], key) O[n] = D[n] V[n] 5. Integration with the AUTHENTICATION telnet option As noted in the telnet ENCRYPTION option specifications, a keyid val- ue of zero indicates the default encryption key, as might be derived from the telnet AUTHENTICATION option. If the default encryption key negotiated as a result of the telnet AUTHENTICATION option contains less than 8 bytes, then the DES_OFB64 option may not be offered or used as a valid telnet encryption option. If the encryption key ne- gotiated as a result of the telnet AUTHENTICATION option is greater than 16 bytes the first 8 bytes of the key should be used as keyid 0 for data sent from the telnet server to the telnet client, and the second 8 bytes of the key should be used as keyid 0 for data sent by the telnet client to the telnet server. Otherwise, the first 8 bytes of the encryption key is used as keyid zero for the telnet ENCRYPTION option in both directions (with the client as WILL ENCRYPT and the server as WILL ENCRYPT). In all cases, if the key negotiated by the telnet AUTHENTICATION op- tion was not a DES key, the key used by the DES_CFB64 must have its parity corrected after it is detrmined using the above algorithm. Note that the above algorithm assumes that it is safe to use a non- DES key (or part of a non-DES key) as a DES key. This is not neces- sarily true of all cipher systems, but we specify this behaviour as the default since it is true for most authentication systems in popu- lar use today, and for compatibility with existing implementations. New telnet AUTHENTICATION mechanisms may specify althernative methods for determining the keys to be used for this cipher suite in their specification, if the session key negotiated by that authentication mechanism is not a DES key and and where this algorithm may not be safely used. 6. Security considerations Expires February 19100 [Page 3] Internet-Draft DES 64 bit Output Feedback August 1999 Encryption using Output Feedback does not ensure data integrity; an active attacker may be able to substitute text, if he can predict the clear-text that was being transmitted. This option was originally drafted back when CPU speeds where not necessarily fast enough to do allow use of CFB. Since then, CPU's have gotten much faster. Given the inherent weaknesses in Output Feedback mode, perhaps it should be deprecated in favor of CFB modes? 7. Acknowledgments This document was originally written by Dave Borman of Cray Research with the assistance of the IETF Telnet Working Group. Author's Address Theodore Ts'o, Editor VA Linux Systems 43 Pleasant St. Medford, MA 02155 Phone: (781) 391-3464 EMail: tytso@valinux.com Expires February 19100 [Page 4]