Network Working Group                                          A. Farrel
Request for Comments: 5511                            Old Dog Consulting
Category: Standards Track                                     April 2009


         Routing Backus-Naur Form (RBNF): A Syntax Used to Form
       Encoding Rules in Various Routing Protocol Specifications

Status of This Memo

   This document specifies an Internet standards track protocol for the
   Internet community, and requests discussion and suggestions for
   improvements.  Please refer to the current edition of the "Internet
   Official Protocol Standards" (STD 1) for the standardization state
   and status of this protocol.  Distribution of this memo is unlimited.

Copyright Notice

   Copyright (c) 2009 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 in effect on the date of
   publication of this document (http://trustee.ietf.org/license-info).
   Please review these documents carefully, as they describe your rights
   and restrictions with respect to this document.

   This document may contain material from IETF Documents or IETF
   Contributions published or made publicly available before November
   10, 2008.  The person(s) controlling the copyright in some of this
   material may not have granted the IETF Trust the right to allow
   modifications of such material outside the IETF Standards Process.
   Without obtaining an adequate license from the person(s) controlling
   the copyright in such materials, this document may not be modified
   outside the IETF Standards Process, and derivative works of it may
   not be created outside the IETF Standards Process, except to format
   it for publication as an RFC or to translate it into languages other
   than English.

Abstract

   Several protocols have been specified in the Routing Area of the IETF
   using a common variant of the Backus-Naur Form (BNF) of representing
   message syntax.  However, there is no formal definition of this
   version of BNF.

   There is value in using the same variant of BNF for the set of
   protocols that are commonly used together.  This reduces confusion
   and simplifies implementation.

   Updating existing documents to use some other variant of BNF that is
   already formally documented would be a substantial piece of work.

   This document provides a formal definition of the variant of BNF that
   has been used (that we call Routing BNF) and makes it available for
   use by new protocols.

Table of Contents

   1. Introduction ....................................................3
      1.1. Terminology ................................................3
      1.2. Existing Uses ..............................................3
      1.3. Applicability Statement ....................................4
   2. Formal Definitions ..............................................4
      2.1. Rule Definitions ...........................................5
           2.1.1. Rule Name Delimitation ..............................5
           2.1.2. Objects .............................................5
           2.1.3. Constructs ..........................................6
           2.1.4. Messages ............................................6
      2.2. Operators ..................................................6
           2.2.1. Assignment ..........................................6
           2.2.2. Concatenation .......................................7
           2.2.3. Optional Presence ...................................7
           2.2.4. Alternatives ........................................8
           2.2.5. Repetition ..........................................9
           2.2.6. Grouping ...........................................10
      2.3. Editorial Conventions .....................................11
           2.3.1. White Space ........................................11
           2.3.2. Line Breaks ........................................11
           2.3.3. Ordering ...........................................11
      2.4. Precedence ................................................11
   3. Automated Validation ...........................................13
   4. Security Considerations ........................................13
   5. Acknowledgments ................................................13
   6. References .....................................................13
      6.1. Normative References ......................................13
      6.2. Informative References ....................................13

1.  Introduction

   Backus-Naur Form (BNF) has been used to specify the message formats
   of several protocols within the Routing Area of the IETF.
   Unfortunately, these specifications are not based on any specific
   formal definition of BNF, and they differ slightly from the
   definitions provided in other places.

   It is clearly valuable to have a formal definition of the syntax-
   defining language that is used.  It would be possible to convert all
   existing specifications to use an established specification of BNF
   (for example, Augmented BNF or ABNF [RFC5234]); however, this would
   require a lot of work.  It should be noted that in ABNF the terminals
   are integers (characters/bytes), while in the BNF form used to define
   message formats, the terminals are "objects" (some kind of message
   elements, but not individual bytes or characters) or entire
   "messages".  This means that converting existing specifications to
   use an established BNF specification would also require extensions to
   that BNF specification.

   On the other hand, the variant of BNF used by the specifications in
   question (which is similar to a subset of Extended BNF [EBNF]) is
   consistent and has only a small number of constructs.  It makes
   sense, therefore, to provide a definition of this variant of BNF to
   allow ease of interpretation of existing documents and to facilitate
   the development of new protocol specifications using the same variant
   of BNF.  A specification will also facilitate automated verification
   of the formal definitions used in future documents.

   This document provides such a specification and names the BNF variant
   Routing BNF (RBNF).

1.1.  Terminology

   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 [RFC2119].

1.2.  Existing Uses

   The first notable use of the variant of BNF that concerns us is in
   the specification of the Resource Reservation Protocol (RSVP)
   [RFC2205].  RSVP has been extended for use in Multiprotocol Label
   Switching (MPLS) networks to provide signaling for Traffic
   Engineering (TE) [RFC3209], and this has been developed for use as
   the signaling protocol in Generalized MPLS (GMPLS) networks
   [RFC3473].

   Each of these three uses of RSVP has given rise to a large number of
   specifications of protocol extensions to provide additional features
   over and above those in the base documents.  Each new feature is
   defined in its own document using the common variant of BNF.

   New protocols have also been specified using the same variant of BNF.
   This has arisen partly because the developers were familiar with the
   BNF used in [RFC2205], etc., but also because of the overlap between
   the protocols, especially with respect to the network objects
   controlled and operated.

   Notable among these additional protocols are the Link Management
   Protocol (LMP) [RFC4204] and the Path Computation Element Protocol
   (PCEP) [RFC5440].  In both cases, further documents that specify
   protocol extensions also use the same variant of BNF.

1.3.  Applicability Statement

   RBNF as defined in this document is primarily applicable for the
   protocols listed in the previous section.  The specification may be
   used to facilitate the interpretation of the pre-existing RFCs that
   are referenced.  It should also be used in the specification of
   extensions to those protocols.

   RBNF could also be used for the specification of new protocols.  This
   is most appropriate for the development of new protocols that are
   closely related to those that already use RBNF.  For example, PCEP is
   closely related to RSVP-TE, and when it was developed, the PCE
   working group chose to use the same form of BNF as was already used
   in the RSVP-TE specifications.

   If a wholly new protocol is being developed and is not related to a
   protocol that already uses RBNF, the working group should consider
   carefully whether to use RBNF or to use a more formally specified and
   broader form of BNF such as ABNF [RFC5234].

   The use of RBNF to specify extensions to protocols that do not
   already use RBNF (i.e., that use some other form of BNF) is not
   recommended.

2.  Formal Definitions

   The basic building blocks of BNF are rules and operators.  At its
   simplest form, a rule in the context we are defining is a protocol
   object that is traditionally defined by a bit diagram in the protocol
   specification.  Further and more complex rules are constructed by

   combining other rules using operators.  The most complex rule is the
   message that is constructed from an organization of protocol objects
   as specified by the operators.

   An RBNF specification consists of a sequence of rule definitions
   using the operators defined in Section 2.2.  One rule may be
   constructed from a set of other rules using operators.  The order of
   definition of rules does not matter.  That is, the subordinate rules
   MAY be defined first and then used in subsequent definitions of
   further rules, or the top-level rules MAY be defined first followed
   by a set of definitions of the subordinate rules.

   Rule definitions are read left-to-right on any line, and the lines
   are read top-to-bottom on the page.  This becomes particularly
   important when considering sequences of rules and operators.

2.1.  Rule Definitions

   No semantics should be assumed from special characters used in rule
   names.  For example, it would be wrong to assume that a rule carries
   a decimal number because the rule name begins or ends with the letter
   "d".  However, individual specifications MAY choose to assign rule
   names in any way that makes the human interpretation of the rule
   easier.

2.1.1.  Rule Name Delimitation

   All rule names are enclosed by angle brackets ("<" and ">").  Rule
   names MAY include any printable characters, but MUST NOT include tabs
   or line feeds/breaks.

   Example:
     <Path Message>

2.1.2.  Objects

   The most basic (indivisible) rule is termed an object.  The
   definition of an object is derived from its context.

   Objects are typically named in uppercase.  They do not usually use
   spaces within the name, favoring underbars ("_").

   Example:
     <SENDER_TEMPLATE>

2.1.3.  Constructs

   Rules that are constructed from other rules using operators are
   termed constructs.

   Constructs are named in lowercase, although capitals are commonly
   used to indicate acronyms.  Spaces and hyphens are used between words
   within names.

   Example:
     <sender descriptor>

2.1.4.  Messages

   The final objective is the definition of messages.  These are rules
   that are constructed from objects and constructs using operators.
   The only syntactic difference between a message and a construct is
   that no other rule is typically constructed from a message.

   Messages are typically named in title case.

   Example:
     <Path Message>

2.2.  Operators

   Operators are used to build constructs and messages from objects and
   constructs.

2.2.1.  Assignment

   Assignment is used to form constructs and messages.

   Meaning:
     The named construct or message on the left-hand side is defined to
     be set equal to the right-hand side of the assignment.

   Encoding:
     colon, colon, equal sign ("::=")

   Example:
     <WF flow descriptor> ::= <FLOWSPEC>

   Note:
     The left-hand side of the assignment and the assignment operator
     MUST be present on the same line.

2.2.2.  Concatenation

   Objects and constructs can be combined as a sequence to form a new
   construct or a message.

   Meaning:
     The objects or constructs MUST be present in the order specified.
     The order of reading RBNF is stated in Section 2.

   Encoding:
     A sequence of objects and constructs usually separated by spaces.
     The objects in a sequence MAY be separated by line breaks.

   Example:
     <SE flow descriptor> ::= <FLOWSPEC> <filter spec list>

   Note:
     See Section 2.3.3 for further comments on the ordering of objects
     and constructs.

2.2.3.  Optional Presence

   Objects and constructs can be marked as optionally present.

   Meaning:
     The optional objects or constructs MAY be present or absent within
     the assignment.  Unless indicated as optional, objects and
     constructs are mandatory and MUST be present.  The optional
     operator can also be nested to give a hierarchical dependency of
     presence as shown in the example below.

   Encoding:
     Contained in square brackets ("[" and "]").

   Example:
     <PathTear Message> ::= <Common Header> [ <INTEGRITY> ]
                            <SESSION> <RSVP_HOP>
                            [ <sender descriptor> ]

   Example of nesting:
     The optional operator can be nested.  For example,

       <construct> ::= <MAND> [ <OPT_1> [ <OPT_2> ] ]

     In this construction, the object OPT_2 can only be present if OPT_1
     is also present.

   Note:
     The set of objects and constructs within the same pair of square
     brackets is treated as a unit (an unnamed construct).  This means
     that when multiple objects and constructs are included within the
     same pair of square brackets, all MUST be included when one is
     included, unless nested square brackets are used as in the previous
     example.

2.2.4.  Alternatives

   Choices can be indicated within assignments.

   Meaning:
     Either one rule or the other MUST be present.

   Encoding:
     The pipe symbol ("|") is used between the objects or constructs
     that are alternatives.

   Example:
     <flow descriptor list> ::= <FF flow descriptor list>
                                | <SE flow descriptor>

   Notes:
     1. Use of explicit grouping (Section 2.2.6) is RECOMMENDED to avoid
        confusion.  Implicit grouping using line breaks (Section 2.3.2)
        is often used, but gives rise to potential misinterpretation and
        SHOULD be avoided in new definitions.

     2. Multiple members of alternate sets can give rise to confusion.
        For example:

        <flow descriptor list> ::=  <empty> |
                             <flow descriptor list> <flow descriptor>

        could be read to mean that an instance of <flow descriptor> must
        be present or that it is optional.

        To avoid this type of issue, explicit grouping (see Section
        2.2.6), or an intermediary MUST be used in all new documents
        (existing uses are not deprecated, and automatic parsers need to
        handle existing RFCs).  See also Section 2.4 for a description
        of precedence rules.

        Thus:

          <construct> ::= <ALT_A> <ALT_B> | <ALT_C> <ALT_D>

        is not allowed in new documents and MUST be presented using
        grouping or using an intermediary construct.  For example, and
        depending on intended meaning:

          <construct> ::= ( <ALT_A> <ALT_B> ) | ( <ALT_C> <ALT_D> )

          or

          <construct> ::= <ALT_A> ( <ALT_B> | <ALT_C> ) <ALT_D>

        or

          <intermediary X> ::= <ALT_A> <ALT_B>
          <intermediary Y> ::= <ALT_C> <ALT_D>
          <construct> ::= <intermediary X> | <intermediary Y>

          or

          <intermediary Z> ::= <ALT_B> | <ALT_C>
          <construct> ::= <ALT_A> <intermediary Z> <ALT_D>

2.2.5.  Repetition

   It could be the case that a sequence of identical objects or
   constructs is required within an assignment.

   Meaning:
     MAY repeat the preceding object, intermediate construct, or
     construct.

   Encoding:
     Three dots ("...").

   Example:
     <Path Message> ::= <Common Header> [ <INTEGRITY> ]
                        <SESSION> <RSVP_HOP>
                        <TIME_VALUES>
                        [ <POLICY_DATA> ... ]
                        [ <sender descriptor> ]

   Notes:
     1. A set of zero or more objects or constructs can be achieved by
        combining with the Optional concept as shown in the example
        above.

     2. Sequences can also be encoded by building a recursive construct
        using the Alternative operator.  For example:

          <sequence> ::= <OBJECT> |
                         ( <OBJECT> <sequence> )

     3. Repetition can also be applied to a component of an assignment
        to indicate the optional repetition of that component.  For
        example, the Notify message in [RFC3473] is defined as follows:

         <Notify message> ::=
                          <Common Header> [<INTEGRITY>]
                          [ [<MESSAGE_ID_ACK> | <MESSAGE_ID_NACK>] ... ]
                          [ <MESSAGE_ID> ]
                          <ERROR_SPEC> <notify session list>

        In this example, there is a sequence of zero or more instances
        of [<MESSAGE_ID_ACK> | <MESSAGE_ID_NACK>].  One could argue that
        the use of grouping (see Section 2.2.6) or a recursive construct
        (see Note 2, above) would be more clear.

2.2.6.  Grouping

   Meaning:
     A group of objects or constructs to be treated together.  This
     notation is not mandatory but is RECOMMENDED for clarity.  See
     Section 2.4 on Precedence.

   Encoding:
     Round brackets ("(" and ")") enclosing a set of objects,
     constructs, and operators.

   Example:
     <group> ::= ( <this> <that> )

   Notes:
     1. The precedence rule in Section 2.4 means that the use of
        grouping is not necessary for the formal interpretation of the
        BNF representation.  However, grouping can make the BNF easier
        to parse unambiguously.  Either grouping or an intermediate
        construct MUST be used for multi-alternates (Section 2.2.4).

     2. Line breaks (Section 2.3.2) are often used to clarify grouping
        as can be seen in the definition of <sequence> in Section 2.2.5,
        but these are open to misinterpretation, and explicit grouping
        is RECOMMENDED.

     3. A practical alternative to grouping is the definition of
        intermediate constructs as illustrated in Note 2 of Section
        2.2.4.

2.3.  Editorial Conventions

2.3.1.  White Space

   White space (that is space characters) between operators, objects,
   and constructs is ignored but SHOULD be used for readability.

2.3.2.  Line Breaks

   Line breaks within an assignment are ignored but SHOULD be used for
   readability.

   Line breaks are often used to imply grouping within the precedence
   rules set out in Section 2.4, but explicit grouping (Section 2.2.6)
   or intermediary constructs (Section 2.2.4) SHOULD be used in new
   definitions.

   A line break MUST NOT be present between the left-hand side of an
   assignment and the assignment operator (see Section 2.2.1).

   New assignments (i.e., new construct or message definitions) MUST
   begin on a new line.

2.3.3.  Ordering

   The ordering of objects and constructs in an assignment is explicit.

   Protocol specifications MAY opt to state that ordering is only
   RECOMMENDED.  In this case, elements of a list of objects and
   constructs MAY be received in any order.

2.4.  Precedence

   Precedence is the main opportunity for confusion in the use of this
   BNF.  In particular, the use of alternatives mixed with
   concatenations can give rise to different interpretations of the BNF.
   Although precedence can be deduced from a "proper" reading of the BNF
   using the rules defined above and the precedence ordering shown
   below, authors are strongly RECOMMENDED to use grouping (Section
   2.2.6) and ordering (Section 2.3.3) to avoid cases where the reader
   would otherwise be required to understand the precedence rules.

   Automated readers are REQUIRED to parse rules correctly with or
   without this use of grouping.

   The various mechanisms described in the previous sections have the
   following precedence, from highest (binding tightest) at the top, to
   lowest (and loosest) at the bottom:

      objects, constructs
      repetition
      grouping, optional
      concatenation
      alternative

   Note:
     Precedence is the main opportunity for confusion in the use of BNF.
     Authors are strongly RECOMMENDED to use grouping (Section 2.2.6) in
     all places where there is any scope for misinterpretation even when
     the meaning is obvious to the authors.

   Example:

     An example of the confusion in precedence can be found in Section
     3.1.4 of [RFC2205] and is mentioned in Section 2.2.4.

     <flow descriptor list> ::=  <empty> |
                      <flow descriptor list> <flow descriptor>

     The implementer MUST decide which of the following is intended:

     a.  <flow descriptor list> ::= <empty> |
                            ( <flow descriptor list> <flow descriptor> )

     b.  <flow descriptor list> ::= ( <empty> | <flow descriptor list> )
                                    <flow descriptor>

     The line break MAY be interpreted as implying grouping, but that is
     not an explicit rule.  However, the precedence rules say that
     concatenation has higher precedence than the Alternative operator.
     Thus, the text in [RFC2205] SHOULD be interpreted as shown in
     formulation a.

     Similarly (from the same section of [RFC2205]):

       <flow descriptor list> ::=
                        <FLOWSPEC>  <FILTER_SPEC>  |
                        <flow descriptor list> <FF flow descriptor>

     SHALL be interpreted as:

       <flow descriptor list> ::=
                      ( <FLOWSPEC> <FILTER_SPEC> ) |
                      ( <flow descriptor list> <FF flow descriptor> )

     The use of explicit grouping or intermediary constructs is strongly
     RECOMMENDED in new text to avoid confusion.

3.  Automated Validation

   RBNF would be appropriate for verification using automated validation
   tools.  Validation tools need to be able to check for close
   conformance to the rules expressed in this document to be useful for
   verifying new documents, but should also be able to parse RBNF as
   used in existing RFCs.  No tools are known at this time.

4.  Security Considerations

   This document does not define any network behavior and does not
   introduce or seek to solve any security issues.

   It may be noted that clear and unambiguous protocol specifications
   reduce the likelihood of incompatible or defective implementations
   that might be exploited in security attacks.

5.  Acknowledgments

   Thanks to Magnus Westerlund, Nic Neate, Chris Newman, Alfred Hoenes,
   Lou Berger, Julien Meuric, Stuart Venters, Tom Petch, Sam Hartman,
   and Pasi Eronen for review and useful comments.

6.  References

6.1.  Normative References

   [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
             Requirement Levels", BCP 14, RFC 2119, March 1997.

6.2.  Informative References

   [RFC2205] Braden, R., Ed., Zhang, L., Berson, S., Herzog, S., and S.
             Jamin, "Resource ReSerVation Protocol (RSVP) -- Version 1
             Functional Specification", RFC 2205, September 1997.

   [RFC3209] Awduche, D., Berger, L., Gan, D., Li, T., Srinivasan, V.,
             and G. Swallow, "RSVP-TE: Extensions to RSVP for LSP
             Tunnels", RFC 3209, December 2001.

   [RFC3473] Berger, L., Ed., "Generalized Multi-Protocol Label
             Switching (GMPLS) Signaling Resource ReserVation Protocol-
             Traffic Engineering (RSVP-TE) Extensions", RFC 3473,
             January 2003.

   [RFC4204] Lang, J., Ed., "Link Management Protocol (LMP)", RFC 4204,
             October 2005.

   [RFC5234] Crocker, D., Ed., and P. Overell, "Augmented BNF for Syntax
             Specifications: ABNF", STD 68, RFC 5234, January 2008.

   [RFC5440] Vasseur, JP., Ed., and JL. Le Roux, Ed., "Path Computation
             Element (PCE) Communication Protocol (PCEP)", RFC 5440,
             March 2009.

   [EBNF]    ISO/IEC 14977, "Information technology -- Syntactic
             metalanguage -- Extended BNF", 1996.

Author's Address

   Adrian Farrel
   Old Dog Consulting

   EMail: adrian@olddog.co.uk