TCPM Working Group J. Touch Internet Draft USC/ISI Intended status: Proposed Standard November 28, 2012 Expires: May 2013 Shared Use of Experimental TCP Options draft-ietf-tcpm-experimental-options-03.txt Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. 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 May 28, 2013. Copyright Notice Copyright (c) 2012 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Touch Expires May 28, 2013 [Page 1] Internet-Draft Shared Use of Experimental TCP Options November 2012 Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Abstract This document describes how the experimental TCP option codepoints can support concurrent use through the use of a magic number. This mechanism avoids the need for a coordinated registry and is backward-compatible with currently known uses. It is recommended for all new TCP options that use these codepoints. Table of Contents 1. Introduction...................................................2 2. Conventions used in this document..............................4 3. TCP Experimental Option Structure..............................4 3.1. Selecting a Magic Number..................................5 3.2. Impact on TCP Option Processing...........................5 4. Reducing the Impact of False Positives.........................6 5. Migration to Assigned Options..................................7 6. Security Considerations........................................7 7. IANA Considerations............................................7 8. References.....................................................8 8.1. Normative References......................................8 8.2. Informative References....................................8 9. Acknowledgments................................................9 1. Introduction TCP includes options to enable new protocol capabilities that can be activated only where needed and supported [RFC793]. The space for identifying such options is small - 256 values, of which 30 are assigned at the time this document was published [IANA]. Two of these codepoints are allocated to support experiments (253, 254) [RFC4727]. These values are intended for testing purposes or anytime an assigned codepoint is either not warranted or available, e.g., based on the maturity status of the defined capability (i.e., Experimental or Informational, rather than Standards Track). The term "experimental TCP options" refers here to options that use the experimental TCP option codepoints [RFC4727]. Such experiments can be described in any type of RFC - Experimental, Informational, etc., and are intended to be used both in controlled environments and in are allowed in public deployments (when not enabled as default) [RFC3962]. Nothing prohibits deploying multiple experiments Touch, (TBD) Expires May 28, 2013 [Page 2] Internet-Draft Shared Use of Experimental TCP Options November 2012 in the same environment - controlled or public. Further, some protocols are specified in Experimental or Informational RFCs, which either include parameters or design choices not yet understood or which might not be widely deployed [RFC2026]. TCP options in such RFCs are typically not eligible for assigned TCP option codepoints [RFC2780], and so there is a need to share use of the experimental option codepoints. There is currently no mechanism to support shared use of the experimental TCP option codepoints. Experimental options 253 and 254 are already deployed in operational code to support an early version of TCP authentication. Option 253 is also documented for the experimental TCP Cookie Transaction option [RFC6013]. This shared use results in collisions in which a single codepoint can appear multiple times in a single TCP segment and for which each use is ambiguous. Other codepoints have been used without assignment (known as "squatting"), notably 31-32 (TCP cookie transactions, as originally distributed and in its API doc) and 76-78 (tcpcrypt) [Bi11][Si11]. Commercial products reportedly also use unassigned options 33, 69- 70, and 76-78 as well. Even though these uses are unauthorized, they currently impact legitimate assignees. Both such misuses (squatting on both experimental and assigned codepoints) are expected to continue, but there are several approaches which can alleviate the impact on cooperating protocol designers. One proposal relaxes the requirements for assignment of TCP options, allowing them to be assigned more readily for protocols that have not been standardized through the IETF process [RFC5226]. Another proposal assigns a larger pool to options and manages their sharing through IANA coordination [Ed11]. The approach proposed in this document does not require additional codepoints and also avoids IANA involvement. The solution adds a field to the structure of the experimental TCP option. This field is populated with a fixed "magic number" defined as part of a specific option experiment. The magic number helps reduce the probability of a collision of independent experimental uses of the same option codepoint, both for those who follow this document (using other magic numbers) and those who do not (squatters). The solution proposed in this document is recommended for all new protocols that use experimental TCP option codepoints. The techniques used here may also help share other experimental codepoints, but that issue is out of scope for this document. Touch, (TBD) Expires May 28, 2013 [Page 3] Internet-Draft Shared Use of Experimental TCP Options November 2012 2. 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 RFC-2119 [RFC2119]. In this document, these words will appear with that interpretation only when in ALL CAPS. Lower case uses of these words are not to be interpreted as carrying RFC-2119 significance. In this document, the characters ">>" preceding an indented line(s) indicates a compliance requirement statement using the key words listed above. This convention aids reviewers in quickly identifying or finding the explicit compliance requirements of this RFC. 3. TCP Experimental Option Structure TCP options have the current common structure [RFC793], in which the first byte is the codepoint (Kind) and the second byte is the length of the option in bytes (Length): 0 1 2 3 01234567 89012345 67890123 45678901 +--------+--------+--------+--------+ | Kind | Length | ... | +--------+--------+--------+--------+ | ... +-------- Figure 1 TCP Option Structure [RFC793] This document extends the option structure for experimental codepoints (253, 254) with a magic number, which is typically 4 bytes in length. The magic number is used to differentiate different experiments, and is the first field after the Kind and Length, as follows: 0 1 2 3 01234567 89012345 67890123 45678901 +--------+--------+--------+--------+ | Kind | Length | Magic Number | +--------+--------+--------+--------+ | Magic Number | ... +--------+--------+--------+--- Figure 2 TCP Experimental Option with a Magic Number Touch, (TBD) Expires May 28, 2013 [Page 4] Internet-Draft Shared Use of Experimental TCP Options November 2012 >> Protocols requiring new TCP option codepoints that are not eligible for assigned values SHOULD use the existing TCP experimental option codepoints (253, 254) with magic numbers as described in this document. >> All protocols using the TCP experimental option codepoints (253, 254) SHOULD use magic numbers as described in this document. Magic numbers are used in other protocols, e.g., BOOTP [RFC951] and DHCP [RFC2131]. In the use proposed in this document they help ensure that concurrent experiments that share the same TCP option codepoint do not interfere. 3.1. Selecting a Magic Number The magic number is selected by the protocol designer when an experimental option is defined, i.e., when the specification for that option is authored. The magic number is selected any of a variety of ways, e.g., using the Unix time() command or bits selected by an arbitrary function (such as a hash) of an arbitrary string (e.g., the document title). This operation occurs only when the option is specified, and is not implemented as part of the design of an option. This document does not proscribe a minimum magic number size. Larger magic numbers reduce the probability of a collision with other options sharing the Kind codepoint, but also increase the option size. A suggested size is 32 bits, in network standard byte order: >> The magic number size and value SHOULD be selected to reduce the probability of collision. >> The magic number SHOULD be 32 bits, but MAY be either longer or shorter. 3.2. Impact on TCP Option Processing The magic number is considered part of the TCP option, not the TCP option header. The presence of the magic number increases the effective option Length field by the size of the magic number. The presence of this magic number is thus transparent to implementations that do not support TCP options where it is used. During TCP processing, experimental options are matched against both the experimental codepoints and the magic number value for each implemented protocol. Touch, (TBD) Expires May 28, 2013 [Page 5] Internet-Draft Shared Use of Experimental TCP Options November 2012 >> Experimental options that have magic numbers that do not match implemented protocols MUST be ignored. The remainder of the option is specified by the particular experimental protocol. This includes the possibility that the magic number could appear in only a subset of instances of the option. Because TCP option capabilities are negotiated during connection establishment, the magic number might be omitted afterwards (e.g., in non-SYN segments). >> TCP experimental option magic numbers, if used in any TCP segment of a connection, MUST be present in TCP SYN segments of that connection. The specification of an experimental option needs to describe whether the magic number appears in non-SYN segments. If the magic number does not appear in all segments, the experimental option may need to be rejected during connection negotiation because options for different experiments in non-SYN segments may not be distinguishable. As a result, this document recommends that: >> TCP experimental option magic numbers, if used in any TCP segment of a connection, SHOULD be used in all TCP segments of that connection in which any experimental option is present. Use of a magic number uses additional space in the TCP header and requires additional protocol processing by experimental protocols. Because these are experiments, neither consideration is a substantial impediment; a finalized protocol can avoid both issues with the assignment of a dedicated option codepoint later. 4. Reducing the Impact of False Positives False positives occur where the magic number of one experiment matches the value of an option that does not use magic numbers or if two experiments select the same magic number. Such collisions can cause an option to be interpreted by the incorrect processing routine. >> Experiments that are not robust to magic number false positives SHOULD implement other detection measures, such as checksums or digital signatures. Use of checksums or signatures may help an experiment use a shorter magic number while reducing the corresponding increased potential for false positives. However this document recommends magic numbers are used together with such checksums/signatures, not as a Touch, (TBD) Expires May 28, 2013 [Page 6] Internet-Draft Shared Use of Experimental TCP Options November 2012 substitute thereof. Magic numbers are static and thus more easily identify the experiment using the experimental option; they can also be more efficiently interpreted at the TCP receiver. 5. Migration to Assigned Options Some experiments may transition from experiment, and become eligible for an assigned TCP option codepoint. This document does not recommend a specific migration plan to transition from use of the experimental TCP options/magic numbers to use of an assigned codepoint. However, once an assigned codepoint is allocated, use of a magic number represents unnecessary overhead. As a result: >> Once a TCP option codepoint is assigned to a protocol, that protocol SHOULD NOT continue to use a magic number as part of that assigned codepoint. This document does not recommend whether or how an implementation of an assigned codepoint should be backward-compatible with use of the experimental codepoint/magic number. However, some implementers may be tempted to include both the experimental and assigned codepoint in the same segment, e.g., in a SYN to support backward-compatibility during connection establishment. This is a poor use limited resources and so to ensure conservation of the TCP option space: >> A TCP segment MUST NOT contain both an assigned TCP option codepoint and an experimental TCP option codepoint/magic number for the same protocol. Instead, a TCP that intends backward compatibility might send multiple SYNs with alternates of the same option and discard all but the most desired successful connection. 6. Security Considerations The mechanism described in this document is not intended to provide (nor does it weaken existing) security for TCP option processing. 7. IANA Considerations This document has no IANA considerations. This section should be removed prior to publication. Touch, (TBD) Expires May 28, 2013 [Page 7] Internet-Draft Shared Use of Experimental TCP Options November 2012 8. References 8.1. Normative References [RFC793] Postel, J., "Transmission Control Protocol", STD 7, RFC 793, Sep. 1981. [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [RFC4727] Fenner, B., "Experimental Values in IPv4, IPv6, ICMPv4, ICMPv6, UDP, and TCP Headers", RFC 4727, Nov. 2006. 8.2. Informative References [Bi11] Bittau, A., D. Boneh, M. Hamburg, M. Handley, D. Mazieres, Q. Slack, "Cryptographic protection of TCP Streams (tcpcrypt)", work in progress, draft-bittau-tcp-crypt-03, Sep. 3, 2012. [Ed11] Eddy, W., "Additional TCP Experimental-Use Options", work in progress, draft-eddy-tcpm-addl-exp-options-00, Aug. 16, 2011. [IANA] IANA web pages, http://www.iana.org/ [RFC951] Croft, B., J. Gilmore, "BOOTSTRAP PROTOCOL (BOOTP)", RFC 951, Sept. 1985. [RFC2026] Bradner, S., "The Internet Standards Process -- Revision 3", BCP 9, RFC 2026, Oct. 1996. [RFC2131] Droms, R., "Dynamic Host Configuration Protocol", RFC 2131, Mar. 1997. [RFC2780] Bradner, S., V. Paxson, "IANA Allocation Guidelines For Values In the Internet Protocol and Related Headers", BCP 37, RFC 2780, Mar. 2000. [RFC3962] Narten, T., "Assigning Experimental and Testing Numbers Considered Useful", BCP 82, RFC 3962, Jan. 2004. [RFC5226] Narten, T., H. Alvestrand, "Guidelines for Writing an IANA Considerations Section in RFCs", BCP 26, RFC 5226, May 2008. Touch, (TBD) Expires May 28, 2013 [Page 8] Internet-Draft Shared Use of Experimental TCP Options November 2012 [RFC6013] Simpson, W., "TCP Cookie Transactions (TCPCT)", RFC 6013, Jan. 2011. [Si11] Simpson, W., "TCP Cookie Transactions (TCPCT) Sockets Application Program Interface (API)", work in progress, draft-simpson-tcpct-api-04, Apr. 7, 2011. 9. Acknowledgments This document was motivated by discussions on the IETF TCPM mailing list and by Wes Eddy's proposal [Ed11]. Yoshifumi Nishida, Pasi Sarolathi, and Michael Scharf provided detailed feedback. This document was prepared using 2-Word-v2.0.template.dot. Authors' Addresses Joe Touch USC/ISI 4676 Admiralty Way Marina del Rey, CA 90292-6695 U.S.A. Phone: +1 (310) 448-9151 Email: touch@isi.edu Touch, (TBD) Expires May 28, 2013 [Page 9]