Network Working Group                                        R. Ullmann
Request for Comments: 1476                 Process Software Corporation
                                                             June 1993


                 RAP: Internet Route Access Protocol

Status of this Memo

  This memo defines an Experimental Protocol for the Internet
  community.  It does not specify an Internet standard.  Discussion and
  suggestions for improvement are requested.  Please refer to the
  current edition of the "IAB Official Protocol Standards" for the
  standardization state and status of this protocol.  Distribution of
  this memo is unlimited.

Abstract

  This RFC describes an open distance vector routing protocol for use
  at all levels of the internet, from isolated LANs to the major
  routers of an international commercial network provider.

Table of Contents

  1.       Introduction  . . . . . . . . . . . . . . . . . . . 2
  1.1       Link-State and Distance-Vector . . . . . . . . . . 3
  1.2       Terminology  . . . . . . . . . . . . . . . . . . . 3
  1.3       Philosophy . . . . . . . . . . . . . . . . . . . . 3
  2.       RAP Protocol  . . . . . . . . . . . . . . . . . . . 4
  2.1       Command Header Format  . . . . . . . . . . . . . . 4
  2.1.1     Length field . . . . . . . . . . . . . . . . . . . 4
  2.1.2     RAP version  . . . . . . . . . . . . . . . . . . . 5
  2.2       RAP Commands . . . . . . . . . . . . . . . . . . . 5
  2.2.1     No operation . . . . . . . . . . . . . . . . . . . 5
  2.2.2     Poll . . . . . . . . . . . . . . . . . . . . . . . 6
  2.2.3     Error  . . . . . . . . . . . . . . . . . . . . . . 7
  2.2.4     Add Route  . . . . . . . . . . . . . . . . . . . . 8
  2.2.5     Purge Route  . . . . . . . . . . . . . . . . . . . 9
  3.       Attributes of Routes  . . . . . . . . . . . . . . . 9
  3.1       Metric and Option Format . . . . . . . . . . . . .10
  3.1.1     Option Class . . . . . . . . . . . . . . . . . .  10
  3.1.2     Type . . . . . . . . . . . . . . . . . . . . . .  10
  3.1.3     Format . . . . . . . . . . . . . . . . . . . . .  11
  3.2       Metrics and Options  . . . . . . . . . . . . . .  11
  3.2.1     Distance . . . . . . . . . . . . . . . . . . . .  12
  3.2.2     Delay  . . . . . . . . . . . . . . . . . . . . .  12
  3.2.3     MTU  . . . . . . . . . . . . . . . . . . . . . .  12
  3.2.4     Bandwidth  . . . . . . . . . . . . . . . . . . .  12



Ullmann                                                         [Page 1]

RFC 1476                          RAP                          June 1993


  3.2.5     Origin . . . . . . . . . . . . . . . . . . . . .  12
  3.2.6     Target . . . . . . . . . . . . . . . . . . . . .  13
  3.2.7     Packet Cost  . . . . . . . . . . . . . . . . . .  13
  3.2.8     Time Cost  . . . . . . . . . . . . . . . . . . .  13
  3.2.9     Source Restriction . . . . . . . . . . . . . . .  14
  3.2.10    Destination Restriction  . . . . . . . . . . . .  14
  3.2.11    Trace  . . . . . . . . . . . . . . . . . . . . .  14
  3.2.12    AUP  . . . . . . . . . . . . . . . . . . . . . .  15
  3.2.13    Public . . . . . . . . . . . . . . . . . . . . .  15
  4.       Procedure   . . . . . . . . . . . . . . . . . . .  15
  4.1       Receiver filtering . . . . . . . . . . . . . . .  16
  4.2       Update of metrics and options  . . . . . . . . .  16
  4.3       Aggregation  . . . . . . . . . . . . . . . . . .  17
  4.4       Active route selection . . . . . . . . . . . . .  17
  4.5       Transmitter filtering  . . . . . . . . . . . . .  17
  4.6       Last resort loop prevention  . . . . . . . . . .  18
  5.       Conclusion  . . . . . . . . . . . . . . . . . . .  18
  6.       Appendix: Real Number Representation  . . . . . .  19
  7.       References  . . . . . . . . . . . . . . . . . . .  20
  8.       Security Considerations . . . . . . . . . . . . .  20
  9.       Author's Address  . . . . . . . . . . . . . . . .  20

1.  Introduction

  RAP is a general protocol for distributing routing information at all
  levels of the Internet, from private LANs to the widest-flung
  international carrier networks.  It does not distinguish between
  "interior" and "exterior" routing (except as restricted by specific
  policy), and therefore is not as restricted nor complex as those
  protocols that have strict level and area definitions in their
  models.

  The protocol encourages the widest possible dissemination of topology
  information, aggregating it only when limits of thrust, bandwidth, or
  administrative policy require it.  Thus RAP permits aggressive use of
  resources to optimize routes where desired, without the restrictions
  inherent in the simplifications of other models.

  While RAP uses IPv7 [RFC1475] addressing internally, it is run over
  both IPv4 and IPv7 networks, and shares routing information between
  them.  A IPv4 router will only be able to activate and propagate
  routes that are defined within the local Administrative Domain (AD),
  loading the version 4 subset of the address into the local IP
  forwarding database.







Ullmann                                                         [Page 2]

RFC 1476                          RAP                          June 1993


1.1  Link-State and Distance-Vector

  Of the two major classes of routing algorithm, link-state and
  distance vector, only distance vector seems to scale from the local
  network (where RIP is existence-proof of its validity) to large scale
  inter-domain policy routing, where the number of links and policies
  exceeds the ability of each router to map the entire network.

  In between, we have OSPF, an open link state (specifically, using
  shortest-path-first analysis of the graph, hence the acronym)
  protocol, with extensive development in intra-area routing.

  Since distance vector has proven useful at both ends of the range, it
  seems reasonable to apply it to the entire range of scales, creating
  a protocol that works automatically on small groups of LANs, but can
  apply fairly arbitrary policy in the largest networks.

  This helps model the real world, where networks are not clearly
  divided into hierarchical domains with identifiable "border" routers,
  but have many links across organizational structure and over back
  fences.

1.2  Terminology

  The RAP protocol propagates routes in the opposite direction to the
  travel of datagrams using the routes.  To avoid confusion explaining
  the routing protocol, several terms are distinguished:

  source          where datagrams come from, the source of the
                  datagrams

  destination     where datagrams go to, the destination of the
                  datagrams

  origin          where routing information originates, the router
                  initially inserting route information into the
                  RAP domain

  target          where routing information goes, the target uses the
                  information to send datagrams

1.3  Philosophy

  Protocols should become simpler as they evolve.







Ullmann                                                         [Page 3]

RFC 1476                          RAP                          June 1993


2.  RAP Protocol

  The RAP protocol operates on TCP port 38, with peers opening a
  symmetric TCP connection between the RAP ports on each system.  Thus
  only one RAP connection exists between any pair of peers.

  RAP is also used on UDP port 38, as a peer discovery method.  Hosts
  (i.e., non-routing systems) may listen to RAP datagrams on this port
  to discover local gateways.  This is in addition to, or in lieu of,
  an Internet Standard gateway discovery protocol, which does not exist
  at this writing.

  The peers then use RAP commands to send each other all routes
  available though the sending peer.  This occurs as a full-duplex
  (i.e., simultaneous) exchange of information, with no acknowledgement
  of individual commands.

  Once the initial exchange has been completed, the peers send only
  updates to routes, new routes, and purge commands to delete routes
  previously offered.

  When the connection is broken, each system purges all routes that had
  been offered by the peer.

2.1  Command Header Format

  Each RAP command starts with a header.  The header contains a length
  field to identify the start of the next packet in the TCP stream, a
  version number, and the code for the command.  On UDP, the length
  field does not appear:  each UDP datagram must contain exactly one
  RAP command and not contain data or padding after the end of the
  command.

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        length                                                 |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        RAP version            |       command code            |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

2.1.1  Length field

  The length is a 32 bit unsigned number specifying the offset in bytes
  from the first byte of the length field of this command packet to the
  start of the length field of the next.  The minimum value is 8.
  There is no specific limit to the length of a command packet;
  implementations MUST be able to at least count and skip over a packet



Ullmann                                                         [Page 4]

RFC 1476                          RAP                          June 1993


  that is too large and then MAY send an error indication.

  Each version of the protocol will profile what size should be
  considered the limit for senders, and what (larger) size should be
  considered by receivers to mean that the connection is insane:
  either unsynchronized or worse.

  For version 1 of the protocol, senders MUST NOT send command packets
  greater than 16384 bytes.  Receivers SHOULD consider packets that
  appear to be greater than 162144 bytes in length to be an indication
  of an unrecoverable error.

  Note that these limits probably will not be approached in normal
  operation of version 1 of the protocol; receivers may reasonably
  decline to use routes described by 16K bytes of metrics and policy.
  But even the most memory-restricted implementation MUST be able to
  skip such a command packet.

2.1.2  RAP version

  The version field is a 16 bit unsigned number.  It identifies the
  version of RAP used for that command.  Note that commands with
  different versions may be mixed on the same connection, although the
  usual procedure will be to do the serious protocol (exchanging route
  updates) only at the highest version common to both ends of the
  connection.

  Each side starts the connection by sending a poll command, using the
  highest version supported and continues by using the highest version
  received in any command from the remote.  The response to the poll
  will either be a no-operation packet at that version or an error
  packet at the highest version supported by the remote.

  This document describes version 1 of the RAP protocol.

2.2  RAP Commands

  There five simple RAP commands, described in the following sections.

2.2.1  No operation

  The no operation command serves to reset the poll timer (see next
  section) of the receiver, or (as a side effect) to tell the receiver
  that a particular version is supported.  It never contains option
  specific data and its length is always 8.

  The no operation command is also used in a UDP broadcast to inform
  other systems that the sender is running RAP actively on the network



Ullmann                                                         [Page 5]

RFC 1476                          RAP                          June 1993


  and is both a possible gateway and a candidate peer.  If this command
  is being sent in response to a broadcast poll, it should be sent only
  to the poller.

  A RAP process may send such broadcasts in a startup sequence, or it
  may persist indefinitely to inform other systems coming on line.  If
  it persists, it must not send them more than once every 10 minutes
  (after the initial startup sequence).  If the RAP process sends polls
  as part of its startup, it must not persist in sending them after the
  startup sequence.

  The command code for no-operation is always 0, regardless of RAP
  version.

2.2.2  Poll

  A poll command packet requests that the other side transmit either a
  no-operation packet, or some other packet if sent without delay.
  (i.e., receivers MUST NOT delay a response to a poll by waiting for
  some other packet expected to be queued soon.)

  The poll command code is always 1, regardless of version, and the
  length is always 8.

  Each RAP implementation runs a timer for each connection, to ensure
  that if the other system becomes unreachable, the connection will be
  closed or reset.  The timers run at each end of the connection are
  independent; each system is responsible for sending polls in time to
  reset its own timer.

  The timer MUST be reset (restarted) on the receipt of any RAP packet,
  regardless of whether the version or command code is known.

  In normal operation, if route updates are being sent in both
  directions, polls may not be necessary for long periods of time as
  the timers are continually reset.  When the connection is quiescent,
  both timers will typically get reset as a result of the side with the
  shorter timer doing a poll, and then getting a no-operation in
  response.  RAP implementations MUST NOT be dependent in any way on
  the size or existence of the remote timer.

  An implementation that has access to information from the TCP layer,
  such as the results of TCP layer keepalives, MAY use this instead of
  or in addition to a timer.  However, the use of TCP keepalives is
  discouraged, and this procedure does not ensure that the remote RAP
  process is alive, only that its TCP is accepting data.  Thus a
  failure mode exists that would not exist for active RAP layer polls.




Ullmann                                                         [Page 6]

RFC 1476                          RAP                          June 1993


  The timer MUST be implemented, SHOULD be configurable in at least the
  range 1 to 10 minutes on a per-peer basis, and MAY be infinite
  (disabled) by explicit configuration.

  On UDP, a system (router or non-routing host) may send RAP polls to
  attempt to locate candidate peers or possible gateways.  Such a
  system must not persist in sending polls after its startup sequence,
  except that a system which actually has offered traffic for non-local
  destinations, and has no available gateways, may continue to send
  periodic polls to attempt to acquire a gateway.

2.2.3  Error

  The error packet is used to report an error, whether fatal, serious
  or informational.  It includes a null terminated text description in
  ISO-10646-UTF-1 of the condition, which may be useful to a human
  administrator, and SHOULD be written to a log file.  (The machine is
  not expected to understand the text.)

  Errors are actual failures (in the interpretation of the receiver) to
  use the correct syntax and semantics of the RAP protocol itself, or
  "failure" of the receiver to implement a version of the protocol.
  Other conditions that may require action on the part of the peer
  (such as purging a route) are given their own command codes.

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        length                                                 |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        RAP version (1)        |       command code (2)        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        error code (0)  [reserved]                             |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        description                                            |
   +                                                               +
   |                       ...                                     |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

  The RAP system receiving an Error packet MUST NOT regard it as fatal,
  and close the connection or discard routes.  If the sending system
  desires the condition to be fatal (unrecoverable), its proper action
  is to close the connection.  This requirement is to prevent the kind
  of failure mode demonstrated by hosts that killed off TCP connections
  on the receipt of ICMP Host-Unreachable notifications, even when the
  condition is transient.  We do not want to discourage the reporting
  of errors, in the way that some implementations avoided sending ICMP
  datagrams to deal with overly sensitive hosts.



Ullmann                                                         [Page 7]

RFC 1476                          RAP                          June 1993


  An error packet MUST NOT be sent in response to something that is (or
  might be) an error packet itself.  Subsequent versions of RAP should
  keep the command code point (2) of the error packet.

2.2.4  Add Route

  The add route command offers a route to the receiving peer.  As noted
  later, it MUST be a route actually loaded into the forwarding
  database of the offering peer at the time the add route is sent.

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        length                                                 |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        RAP version (1)        |       command code (3)        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        distance               |     (MBZ)     |     mask      |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        destination network                                    |
   +                                                               +
   |                    ...                                        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        route identifier                                       |
   +                                                               +
   |                    ...                                        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        metrics and options    ....                            |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

  The add route command describes a single offered route, with the
  metrics and other options (such as policies) associated with the
  route.

  Distance is a simple count of the hops to the RAP process (or other
  routing process) that originated the route, incremented every time
  the route is forwarded.  Its initial value may be greater than 1,
  particularily for a route that is administratively configured to
  aggregate routes for a large network or AD.  It may also enter the
  RAP routing domain for the first time with a non-zero distance
  because the route originated in RIP, OSPF, or BGP; if so, the
  distance carried in that protocol is copied into the RAP route.

  The mask is a count of the number of bits of prefix ones in the
  binary representation of the network mask.  Non-contiguous masks are
  not supported directly.  (The destination restriction option may be
  used to give another, non-contiguous, mask; the header mask would
  then describes the number of contiguous ones.)



Ullmann                                                         [Page 8]

RFC 1476                          RAP                          June 1993


  The route identifier is a 64 bit value that the IP forwarding module
  on the sending host can use to rapidly identify the route and the
  next hop for each incoming datagram.  The host receiving the route
  places this identifier into the forward route ID field of the
  datagrams being sent to this host.

  The route ID is also used to uniquely identify the route in the purge
  route operation.

2.2.5  Purge Route

  The purge route command requires that the receiving peer delete a
  route from its database if in use, and requires that it revoke that
  route from any of its peers to whom it has offered the route.  This
  command should preferably be sent before the route is deleted from
  the sending peer's forwarding database, but this is not (cannot be)
  required; it should be sent without delay when the route is removed.

  The command code is 4.  The format is the same as add route without
  any added metrics or options.

  If the route identifier in a purge route command is zero, the command
  requires the deletion of all routes to the destination previously
  offered by this peer.

3.  Attributes of Routes

  There are a rather large number of possible attributes.
  Possibilities include both metrics, and other options describing for
  example policy restrictions and alterations of proximity.  Any
  particular route will usefully carry only a few attributes or none at
  all, particularily on an infrastructure backbone.  A reasonable
  policy for the routers that make up a backbone might be to strip all
  attributes before propagating routes (discarding routes that carry
  attributes with class indications prohibiting this), and then adding
  (for example) an AUP attribute to all routes propagated off of the
  backbone.  A less drastic method would be to simply prefer routes
  with no restrictions, but still propagate a route with restrictions
  if no other is available.

  Most options can occur more than once in a route if there is any
  sensible reason to do so.









Ullmann                                                         [Page 9]

RFC 1476                          RAP                          June 1993


3.1  Metric and Option Format

  Each metric or option for a route begins with a 32 bit header:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |   length      | C |  format   |           type                |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |        option data                 ...        |   padding     |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                  RAP Option/Metric Header Format

A description of each field:

  length       length of the option or metric
  C            option class, see below
  format       data format
  type         option type identifier
  data         variable length

3.1.1  Option Class

  This field tells implementations what to do with routes containing
  options or metrics they do not understand.  No implementation is
  required to implement (i.e., understand) any given option or metric
  by the RAP specification itself, except for the distance metric in
  the RAP header.

  Classes:

  0        use, propagate, and include this option unmodified
  1        use, propagate, but do not include this option
  2        use this route, but do not propagate it
  3        discard this route

  Note that class 0 is an imperative:  if the route is propagated, the
  option must be included.

  Class and type are entirely orthogonal, different implementations
  might use different classes for the same option or metric.

3.1.2  Type

  The type code identifies the specific option or metric.  The codes
  are part of the option descriptions following.




Ullmann                                                        [Page 10]

RFC 1476                          RAP                          June 1993


  Type 0 indicates a null (no-operation) option.  It should be class
  zero, but an implementation that "understands" the null option may
  decline to propagate it.

  Note that since an implementation may delete an option of class 1 by
  simply setting its type to 0 and forwarding the route description,
  class 1 does not provide any confidentiality of the content of an
  option.

3.1.3  Format

  The format field specifies the format of the data included after the
  option header.  Formats:

  0        none, no data present.
  1        one or more 32-bit signed integers
  2        a character string, null terminated
  3        one or more real numbers
  4        an octet string
  5        one real, followed by a character string

  Format is also orthogonal to type, but a particular type is usually
  only reasonably represented by one format.  This allows decoding of
  all option values for logging and other troubleshooting, even when
  the option type is unknown.  (A new unknown format will still present
  a problem.)

  Format 4, octet string, is to be represented in dotted-decimal byte
  form when printed; it is normally an internet address.

  Format 5 is intended for dimensioned parameters with the character
  string giving the dimension or scale.

3.2  Metrics and Options

  As much as possible, metrics are kept in the base units of bytes and
  seconds, by analogy to the physics systems of MKS (meter-kilogram-
  second) and CGS (centimeter-gram-second) of base units.

  Bytes aren't the real primitive, the bit is.  We are thus using a
  multiple of 8 that isn't part of what one would come to expect from a
  decimal metric system that uses the other prefixes.  However, since K
  (kilo) is often taken to be 1024, and M (mega) to be 1,048,576 (or
  even 1,024,000) we allow this liberty.

  Distance is measured in units also unique to the field.  It is the
  integer number of times that a datagram must be forwarded to reach
  the destination.  (Hop count.)



Ullmann                                                        [Page 11]

RFC 1476                          RAP                          June 1993


3.2.1  Distance

  The Distance metric counts the number of hops on a route; this is
  included in the RAP route command header.

  The initial distance at insertion into the RAP domain by the origin
  of the route MUST be less than or equal to 2z, where z is the number
  of zero bits in the route mask.

  If the origin derives the route from RIP or OSPF, and the distance
  exceeds 2z, the route must not be used.

  When a router originates a route designed to permit aggregation, the
  distance is usefully set to more than 0; this allows simple subset
  aggregation without propagating small distance changes repeatedly as
  the internal diameter of the described network changes.

  For example, for routers designated to announce a default route for
  an AD, with a 24/48 mask, the maximum initial distance is 96.

3.2.2  Delay

  The Delay metric (Type = 2) measures the one-way path delay.  It is
  usually the sum of delays configured for the gateways and interfaces,
  but might also include path segments that are actually measured.

  Format is real (3), with one value.  The units are seconds.

3.2.3  MTU

  The MTU metric (Type = 3) measures the minimum value over the route
  of the Maximum Transmission Unit, i.e., the largest IP datagram that
  can be routed without resulting in fragmentation.

  Format is one integer, measuring the MTU in bytes.

3.2.4  Bandwidth

  The Bandwidth metric (Type = 4) measures the minimum bandwidth of the
  path segments that make up the route.

  Format is one real, representing bandwidth in bytes/second.

3.2.5  Origin

  The origin attribute (type = 5) identifies the router that originally
  inserted the route into the RAP domain.  It is one of the IP
  addresses of the router, format is 4.



Ullmann                                                        [Page 12]

RFC 1476                          RAP                          June 1993


3.2.6  Target

  The target attribute (type = 6) identifies a host or network toward
  which the route should be propagated, regardless of proximity
  filtering that would otherwise occur.  This aids in the establishment
  of tunnels for hosts or subnets "away from home." It can be used to
  force the route to propagate all the way to the home network, or to
  try to propagate a better route to a host that the origin has
  established a connection (e.g., TCP) with.  Note that a router can
  distinguish these two cases during proximity filtering by comparing
  the route described with the host or network identified by the target
  option.

  Format is 4.

3.2.7  Packet Cost

  The packet cost metric (type = 7) measures the actual cost (to
  someone) of sending data over the route.  It is probably either class
  3 or 0.  Format is 5.

  The real number is the cost in currency units/byte.  Tariffs set in
  packets or "segments" should be converted using the nominal (or
  actual path) size.  For example, Sprintnet charges for DAF
  connections within its network are US$1.40/Ksegment thus for segments
  of 64 bytes, the cost is 0.000021875 USD.

  The string is the 3 capital letter ISO code [ISO4217] for the
  currency used.  Funds codes and codes XAU, XBA, XBB, XBC, XBD, and
  XXX are not used.

  If a route already has a packet cost in a different currency
  associated with it, another instance of this option should be added.
  RAP implementations MUST NOT attempt to convert the currency units
  except when actually making a route selection decision.  That is, the
  effects of a currency conversion should never be propagated, except
  for the proper effect of such a selection decision.

3.2.8  Time Cost

  The time cost metric (type = 8) measures the actual cost of holding
  one or more paths in the route open to send data.  It is probably
  either class 3 or 0.  Format is 5.

  The real number is the cost in currency units/second.  For example,
  Sprintnet charges for international connections (to typical
  destinations) are US$10/hour so the cost is 0.002777778 USD.




Ullmann                                                        [Page 13]

RFC 1476                          RAP                          June 1993


  The other notes re codes used and conversions in the previous section
  also apply.

3.2.9  Source Restriction

  A source restriction option (type 9, format 4, class 2 or 3)
  indicates that a route may only be used by datagrams from a
  particular source or set of sources.  The data consists of a network
  or host number, and a mask to qualify it.  If multiple source
  restriction options are included, the restriction is the logical
  union of the sources specified; i.e., any are permitted.

  Source restrictions must be added to routes when the RAP system has
  security filters set in the IP forwarding layer.  This is necessary
  to prevent datagrams from taking "better" routes that end in the
  datagram being silently discarded at the filter.  Note that this
  propagates confidential information about the security configuration,
  but only toward the net authorized to use the route if the RAP
  implementation is careful about where it is propagated.

3.2.10  Destination Restriction

  A destination restriction option (type 10, format 4, class 3) serves
  only to provide a non-contiguous mask, the destination already having
  been specified in the command header.  Data is the destination
  network and mask.

3.2.11  Trace

  Trace (type 11, format 4, class 0) provides an indication that the
  route has propagated through a particular system.  This can be used
  for loop detection, as well as various methods of troubleshooting.
  The data is one internet address, one of the addresses of the system.
  If an arriving route already carries a trace identifying this system,
  and is not an update, it is discarded.  If it is an update, the route
  is purged.

  Trace SHOULD NOT be simply added to every route traversing a system.
  Rather, it should be added (if being used for loop detection) when
  there is a suspicion that a loop has formed.

  When the distance to a destination has increased twice in a row in a
  fairly short period of time, and the number of trace options present
  in the route did not increase as a result of the last update, the RAP
  process should add a trace option identifying itself to the route.
  Effectively, when a loop forms, one router will select itself to be a
  tracer, adding itself and breaking the loop after one more turn.  If
  that fails for some reason, another router will add its trace.  Each



Ullmann                                                        [Page 14]

RFC 1476                          RAP                          June 1993


  router thus depends in the end only on its own trace and will break
  the loop, even if the other routers are using other methods, or
  simply counting-out the route.

3.2.12  AUP

  The AUP (Acceptable Use Policy) option (type 12, format 2, class
  any), tags a route as being useable only according to the policy of a
  network.  This may be used to avoid traversal of the net by (for
  example) commercial traffic, or to prevent un-intentional use of an
  organization's internal net.  (It does not provide a security barrier
  in the sense of forwarding filters, but does provide cooperative
  exchange of information on the useability of a net.)

  The data is a domain name, probably the name of the network, although
  it may be the name of another organization.  E.g., the routers that
  are subject to the NSF AUP might add NSF.NET as the descriptor of
  that policy.

3.2.13  Public

  Public (type 13, format 0, class 2 or 3) marks the route as
  consisting in part of a public broadcast medium.  Examples of a
  public medium are direct radio broadcast or a multi-drop cable in
  which other receivers, not associated with the destination may read
  the traffic.  I.e., a TV cable is a public medium, a LAN within an
  organization is not, even if it can be easily wiretapped.

  This is intended for use by cable TV providers to identify routes
  that should not be used for private communications, in spite of the
  attractively high bandwidth being offered.

4.  Procedure

  Routing information arrives in the RAP process from other peers, from
  (local) static route and interface configuration, and from other
  protocols (e.g., RIP).  The RAP process filters out routes that are
  of no interest (too detailed or too "far away" in the topology) and
  builds an internal database of available routes.

  From this database, it selects routes that are to be active and loads
  them into the IP forwarding database.

  It then advertises those routes to its peers, at a greater distance.







Ullmann                                                        [Page 15]

RFC 1476                          RAP                          June 1993


  -------------------------------------------------------------------

          [incoming routes]
                  |
                  v
          [proximity filtering/aggregation]       [static routes]
                  |                                  |
                  v                                  v
          [route database]  --->  [selected active routes]
                  ^                       |
                  |                       v
          [RIP, etc. routes]      [output filtering]
                                          |
                                          v
                                  [routes advertised]

  -------------------------------------------------------------------

4.1  Receiver filtering

  The first step is to filter out offered routes that are too "far
  away" or too specific.  The filter consists of a maximum distance at
  which a route is considered usable for each possible (contiguous)
  mask.

  Routers that need universal connectivity must either pass through the
  filter all routes regardless of distance (short of "infinity"), and
  use aggregation to reduce them, or have a default route to a router
  that does this.

  The filter may be adjusted dynamically to fit limited resources, but
  if the filter is opened, i.e., made less restrictive, there may be
  routes that have already been offered and discarded that will never
  become available.

4.2  Update of metrics and options

  The process then updates any metrics present on the route to reflect
  the path to the RAP peer.  MTU and bandwidth are minimized, delay and
  cost are added in.  Distance is incremented.  Any unknown options
  cause class-dependent processing:  discarding the option (class 2) or
  route (3), or marking the route as non-propagatable (1).

  Policy options that are known may cause the route to be discarded at
  this stage.






Ullmann                                                        [Page 16]

RFC 1476                          RAP                          June 1993


4.3  Aggregation

  The next step is to aggregate routes that are subsetted by other
  routes through the same peer.  This should not be done automatically
  in every possible case.  The more information that is propagated, the
  more effective the use of forward route identifiers is likely to be,
  particularily in the case of aggregating into a default route.

  In general, a route can be included in an aggregate, and not
  propagated further, if it is through the same peer (next hop) and has
  a smaller distance metric than the containing route.  (Thus datagrams
  will always travel "downhill" as they take more specific routes.)

  The usual case of aggregation is that routes derived from interface
  configurations on the routers from which they originated are subsumed
  into routes offered by routers explicitly configured to route for an
  entire network, area, or AD.  If the larger area becomes partitioned,
  unaggregatable routes will appear (as routes outside the area become
  the shortest distance routes) and traffic will flow around the
  partition.

  Attributes of routes, particularily policy options, may prevent
  aggregation and may result in routes simply being discarded.

  Some information about aggregation also needs to be represented in
  the forwarding database, if the route is made active:  the router
  will need to make a decision as to which forward route identifier to
  use for each datagram arriving on the active route.

4.4  Active route selection

  The router selects those routes to be entered into the IP forwarding
  database and actively used to forward datagrams from the set of
  routes after aggregation, combined with routes derived from other
  protocols such as RIP.  This selection may be made on any combination
  of attributes and options desired by local policy.

4.5  Transmitter filtering

  Finally, the RAP process must decide which routes to offer to its
  peers.  These must be a subset of the active routes, and may in turn
  be a selected subset for each peer.  Arbitrary local policies may be
  used in deciding whether or not to offer any particular route to a
  given peer.

  However, the transmitter must ensure that any datagram filters are
  represented in the offered route, so that the peer (and its peers)
  will not route into a black hole.



Ullmann                                                        [Page 17]

RFC 1476                          RAP                          June 1993


4.6  Last resort loop prevention

  RAP is designed to support many different kinds of routing selection
  algorithms, and allow them to interact to varying extents.  Routes
  can be shared among administrations, and between systems managed with
  more or less sophistication.

  This leaves one absolute requirement:  routing loops must be self-
  healing, regardless of the algorithm used on each host.  There are
  two caveats:

    1.  A loop will not fix itself in the presence of an error that
        continually recurs (thus re-generating the loop)

    2.  The last resort algorithm does not provide rapid breaking of
        loops, only eventual breaking of them even in the absence of
        any intervention by (human) intelligence.

  The algorithm relies on the distance in the RAP route header.  This
  count must be updated (i.e., incremented by one) at each router
  forwarding the route.

  Routers must also impose some limit on the number of hops permitted
  in incoming routes, discarding any routes that exceed the limit.
  This limit is "infinity" in the classic algorithm.  In RIP, infinity
  is 15, much too low for general inter-domain routing.

  In RAP, infinity is defined as 2z + i, where z is the number of zero
  bits in the mask (as described previously) and i is a small number
  which MUST be configurable.

  Note that RAP depends on the last resort algorithm, "counting to
  infinity," much less than predecessors such as RIP.  Routes in the
  RAP domain will usually be purged from the net as the purge route
  command is flooded without the delays typical of periodic broadcast
  algorithms.  Only in some cases will loops form, and they will be
  counted out as fast as the routing processes can exchange the
  information.

5.  Conclusion

  Unlike prior routing protocols, RAP is designed to solve the entire
  problem, from hands-off autoconfiguration of LAN networks, to
  implementing the most complex policies of international carriers.  It
  provides a scaleable solution to carry the Internet forward to a
  future in which essentially all users of data transmission use IP as
  the fabric of their networks.




Ullmann                                                        [Page 18]

RFC 1476                          RAP                          June 1993


6.  Appendix:  Real Number Representation

  Real numbers are represented by a one byte exponent, e, in excess-128
  notation, and a fraction, f, in excess-8388608 notation, with the
  radix point at the right.  (I.e., the "fraction" is actually an
  integer.)

  e is thus in the range 0 to 255, representing exponents (powers of 2)
  in the range 2^-128 to 2^127.

  f is in the range 0 to 16777215, representing numbers from -8388608
  to 8388607

  The value is (f-8338608) x 2^(e-128)

  The real number is not necessarily normalized, but a normalized
  representation will, of course, provide more accuracy for numbers not
  exactly representable.

  Example code, in C:

  #include <math.h>

  typedef struct {
          unsigned e : 8;
          unsigned f : 24;
          } real;

  double a;          /* input value */
  real r;
  double b;          /* output value */
  double d;
  int e32;

  /* convert to real: */

  d = frexp(a, &e32);
  r.e = e32+105;
  r.f = (int)(d*8388608.0) + 8388608;

  /* convert back: */

  b = ldexp((double)r.f - 8388608.0, (int)r.e - 128);








Ullmann                                                        [Page 19]

RFC 1476                          RAP                          June 1993


7.  References

  [ISO3166]   International Organization for Standardization.  Codes
              for the Representation of Names of Countries.  ISO
              3166, ISO, 1988.

  [ISO4217]   International Organization for Standardization.  Codes
              for the representation of currencies and funds.  ISO
              4217, ISO, 1981.

  [RFC791]    Postel, J., "Internet Protocol - DARPA Internet Program
              Protocol Specification", STD 5, RFC 791, DARPA,
              September 1981.

  [RFC1058]   Hedrick, C., "Routing Information Protocol", STD 34,
              RFC 1058, Rutgers University, June 1988.

  [RFC1247]   Moy, J., "OSPF Version 2", RFC 1247, Proteon, Inc.,
              July 1991.

  [RFC1287]   Clark, D., Chapin, L., Cerf, V., Braden, R., and
              R. Hobby, "Towards the Future Internet Architecture",
              RFC 1287, MIT, BBN, CNRI, ISI, UCDavis, December 1991.

  [RFC1338]   Fuller, V., Li, T., Yu, J., and K. Varadhan,
              "Supernetting: an Address Assignment and Aggregation
              Strategy", RFC 1338, BARRNet, cicso, Merit, OARnet,
              June 1992.

  [RFC1475]   Ullmann, R., "TP/IX: The Next Internet", RFC 1475,
              Process Software Corporation, June 1993.

8.  Security Considerations

  Security issues are discussed in sections 3.2.9 and 3.2.12.

9.  Author's Address

  Robert Ullmann
  Process Software Corporation
  959 Concord Street
  Framingham, MA 01701
  USA

  Phone: +1 508 879 6994 x226
  Email: [email protected]





Ullmann                                                        [Page 20]