Network Working Group                                          R. Braden
Request for Comments: 1009                                     J. Postel
Obsoletes: 985                                                       ISI
                                                               June 1987

                   Requirements for Internet Gateways

Status of this Memo

   This document is a formal statement of the requirements to be met by
   gateways used in the Internet system.  As such, it is an official
   specification for the Internet community.  Distribution of this memo
   is unlimited.

   This RFC summarizes the requirements for gateways to be used between
   networks supporting the Internet protocols.  While it was written
   specifically to support National Science Foundation research
   programs, the requirements are stated in a general context and are
   applicable throughout the Internet community.

   The purpose of this document is to present guidance for vendors
   offering gateway products that might be used or adapted for use in an
   Internet application.  It enumerates the protocols required and gives
   references to RFCs and other documents describing the current
   specifications.  In a number of cases the specifications are evolving
   and may contain ambiguous or incomplete information.  In these cases
   further discussion giving specific guidance is included in this
   document.  Specific policy issues relevant to the NSF scientific
   networking community are summarized in an Appendix.  As other
   specifications are updated this document will be revised.  Vendors
   are encouraged to maintain contact with the Internet research

1.  Introduction

   The following material is intended as an introduction and background
   for those unfamiliar with the Internet architecture and the Internet
   gateway model.  General background and discussion on the Internet
   architecture and supporting protocol suite can be found in the DDN
   Protocol Handbook [25] and ARPANET Information Brochure [26], see
   also [19, 28, 30, 31].

   The Internet protocol architecture was originally developed under
   DARPA sponsorship to meet both military and civilian communication
   requirements [32].  The Internet system presently supports a variety
   of government and government-sponsored operational and research
   activities.  In particular, the National Science Foundation (NSF) is
   building a major extension to the Internet to provide user access to

Braden & Postel                                                 [Page 1]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   national supercomputer centers and other national scientific
   resources, and to provide a computer networking capability to a large
   number of universities and colleges.

   In this document there are many terms that may be obscure to one
   unfamiliar with the Internet protocols.  There is not much to be done
   about that but to learn, so dive in.  There are a few terms that are
   much abused in general discussion but are carefully and intentionally
   used in this document.  These few terms are defined here.

      Packet      A packet is the unit of transmission on a physical

      Datagram    A datagram is the unit of transmission in the IP
                  protocol.  To cross a particular network a datagram is
                  encapsulated inside a packet.

      Router      A router is a switch that receives data transmission
                  units from input interfaces and, depending on the
                  addresses in those units, routes them to the
                  appropriate output interfaces.  There can be routers
                  at different levels of protocol.  For example,
                  Interface Message Processors (IMPs) are packet-level

      Gateway     In the Internet documentation generally, and in this
                  document specifically, a gateway is an IP-level
                  router.  In the Internet community the term has a long
                  history of this usage [32].

   1.1.  The DARPA Internet Architecture

      1.1.1.  Internet Protocols

         The Internet system consists of a number of interconnected
         packet networks supporting communication among host computers
         using the Internet protocols.  These protocols include the
         Internet Protocol (IP), the Internet Control Message Protocol
         (ICMP), the Transmission Control Protocol (TCP), and
         application protocols depending upon them [22].

         All Internet protocols use IP as the basic data transport
         mechanism.  IP [1,31] is a datagram, or connectionless,
         internetwork service and includes provision for addressing,
         type-of-service specification, fragmentation and reassembly,
         and security information.  ICMP [2] is considered an integral

Braden & Postel                                                 [Page 2]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         part of IP, although it is architecturally layered upon IP.
         ICMP provides error reporting, flow control and first-hop
         gateway redirection.

         Reliable data delivery is provided in the Internet protocol
         suite by transport-level protocols such as the Transmission
         Control Protocol (TCP), which provides end-end retransmission,
         resequencing and connection control.  Transport-level
         connectionless service is provided by the User Datagram
         Protocol (UDP).

      1.1.2.  Networks and Gateways

         The constituent networks of the Internet system are required
         only to provide packet (connectionless) transport.  This
         requires only delivery of individual packets.  According to the
         IP service specification, datagrams can be delivered out of
         order, be lost or duplicated and/or contain errors.  Reasonable
         performance of the protocols that use IP (e.g., TCP) requires
         an IP datagram loss rate of less than 5%.  In those networks
         providing connection-oriented service, the extra reliability
         provided by virtual circuits enhances the end-end robustness of
         the system, but is not necessary for Internet operation.

         Constituent networks may generally be divided into two classes:

         *  Local-Area Networks (LANs)

            LANs may have a variety of designs, typically based upon
            buss, ring, or star topologies.  In general, a LAN will
            cover a small geographical area (e.g., a single building or
            plant site) and provide high bandwidth with low delays.

         *  Wide-Area Networks (WANs)

            Geographically-dispersed hosts and LANs are interconnected
            by wide-area networks, also called long-haul networks.
            These networks may have a complex internal structure of
            lines and packet-routers (typified by ARPANET), or they may
            be as simple as point-to-point lines.

         In the Internet model, constituent networks are connected
         together by IP datagram forwarders which are called "gateways"
         or "IP routers".  In this document, every use of the term
         "gateway" is equivalent to "IP router".  In current practice,
         gateways are normally realized with packet-switching software

Braden & Postel                                                 [Page 3]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         executing on a general-purpose CPU, but special-purpose
         hardware may also be used (and may be required for future
         higher-throughput gateways).

         A gateway is connected to two or more networks, appearing to
         each of these networks as a connected host.  Thus, it has a
         physical interface and an IP address on each of the connected
         networks.  Forwarding an IP datagram generally requires the
         gateway to choose the address of the next-hop gateway or (for
         the final hop) the destination host.  This choice, called
         "routing", depends upon a routing data-base within the gateway.
         This routing data-base should be maintained dynamically to
         reflect the current topology of the Internet system; a gateway
         normally accomplishes this by participating in distributed
         routing and reachability algorithms with other gateways.
         Gateways provide datagram transport only, and they seek to
         minimize the state information necessary to sustain this
         service in the interest of routing flexibility and robustness.

         Routing devices may also operate at the network level; in this
         memo we will call such devices MAC routers (informally called
         "level-2 routers", and also called "bridges").  The name
         derives from the fact that MAC routers base their routing
         decision on the addresses in the MAC headers; e.g., in IEEE
         802.3 networks, a MAC router bases its decision on the 48-bit
         addresses in the MAC header.  Network segments which are
         connected by MAC routers share the same IP network number,
         i.e., they logically form a single IP network.

         Another variation on the simple model of networks connected
         with gateways sometimes occurs: a set of gateways may be
         interconnected with only serial lines, to effectively form a
         network in which the routing is performed at the internetwork
         (IP) level rather than the network level.

      1.1.3.  Autonomous Systems

         For technical, managerial, and sometimes political reasons, the
         gateways of the Internet system are grouped into collections
         called "autonomous systems" [35].  The gateways included in a
         single autonomous system (AS) are expected to:

            *  Be under the control of a single operations and
               maintenance (O&M) organization;

            *  Employ common routing protocols among themselves, to
               maintain their routing data-bases dynamically.

Braden & Postel                                                 [Page 4]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         A number of different dynamic routing protocols have been
         developed (see Section 4.1); the particular choice of routing
         protocol within a single AS is generically called an interior
         gateway protocol or IGP.

         An IP datagram may have to traverse the gateways of two or more
         ASs to reach its destination, and the ASs must provide each
         other with topology information to allow such forwarding.  The
         Exterior Gateway Protocol (EGP) is used for this purpose,
         between gateways of different autonomous systems.

      1.1.4.  Addresses and Subnets

         An IP datagram carries 32-bit source and destination addresses,
         each of which is partitioned into two parts -- a constituent
         network number and a host number on that network.

            IP-address ::=  { <Network-number>,  <Host-number> }

         To finally deliver the datagram, the last gateway in its path
         must map the host-number (or "rest") part of an IP address into
         the physical address of a host connection to the constituent

         This simple notion has been extended by the concept of
         "subnets", which were introduced in order to allow arbitrary
         complexity of interconnected LAN structures within an
         organization, while insulating the Internet system against
         explosive growth in network numbers and routing complexity.
         Subnets essentially provide a two-level hierarchical routing
         structure for the Internet system.  The subnet extension,
         described in RFC-950 [21], is now a required part of the
         Internet architecture.  The basic idea is to partition the
         <host number> field into two parts: a subnet number, and a true
         host number on that subnet.

            IP-address ::=
                    { <Network-number>, <Subnet-number>, <Host-number> }

         The interconnected LANs of an organization will be given the
         same network number but different subnet numbers.  The
         distinction between the subnets of such a subnetted network
         must not be visible outside that network.  Thus, wide-area
         routing in the rest of the Internet will be based only upon the
         <Network-number> part of the IP destination address; gateways
         outside the network will lump <Subnet-number> and <Host-number>

Braden & Postel                                                 [Page 5]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         together to form an uninterpreted "rest" part of the 32-bit IP
         address.  Within the subnetted network, the local gateways must
         route on the basis of an extended network number:

            { <Network-number>, <Subnet-number> }.

         The bit positions containing this extended network number are
         indicated by a 32-bit mask called the "subnet mask" [21]; it is
         recommended but not required that the <Subnet-number> bits be
         contiguous and fall between the <Network-number> and the
         <Host-number> fields.  No subnet should be assigned the value
         zero or -1 (all one bits).

         Flexible use of the available address space will be
         increasingly important in coping with the anticipated growth of
         the Internet.  Thus, we allow a particular subnetted network to
         use more than one subnet mask.  Several campuses with very
         large LAN configurations are also creating nested hierarchies
         of subnets, sub-subnets, etc.

         There are special considerations for the gateway when a
         connected network provides a broadcast or multicast capability;
         these will be discussed later.

   1.2.  The Internet Gateway Model

      There are two basic models for interconnecting local-area networks
      and wide-area (or long-haul) networks in the Internet.  In the
      first, the local-area network is assigned a network number and all
      gateways in the Internet must know how to route to that network.
      In the second, the local-area network shares (a small part of) the
      address space of the wide-area network.  Gateways that support
      this second model are called "address sharing gateways" or
      "transparent gateways".  The focus of this memo is on gateways
      that support the first model, but this is not intended to exclude
      the use of transparent gateways.

      1.2.1.  Internet Gateways

         An Internet gateway is an IP-level router that performs the
         following functions:

            1.  Conforms to specific Internet protocols specified in
                this document, including the Internet Protocol (IP),
                Internet Control Message Protocol (ICMP), and others as
                necessary.  See Section 2 (Protocols Required).

            2.  Interfaces to two or more packet networks.  For each

Braden & Postel                                                 [Page 6]

RFC 1009 - Requirements for Internet Gateways                  June 1987

                connected network the gateway must implement the
                functions required by that network.  These functions
                typically include:

               a.  encapsulating and decapsulating the IP datagrams with
                   the connected network framing (e.g., an Ethernet
                   header and checksum);

               b.  sending and receiving IP datagrams up to the maximum
                   size supported by that network, this size is the
                   network's "Maximum Transmission Unit" or "MTU";

               c.  translating the IP destination address into an
                   appropriate network-level address for the connected
                   network (e.g., an Ethernet hardware address);

               d.  responding to the network flow control and error
                   indication, if any.

               See Section 3 (Constituent Network Interface), for
               details on particular constituent network interfaces.

            3.  Receives and forwards Internet datagrams.  Important
                issues are buffer management, congestion control, and
                fairness.  See Section 4 (Gateway Algorithms).

               a.  Recognizes various error conditions and generates
                   ICMP error and information messages as required.

               b.  Drops datagrams whose time-to-live fields have
                   reached zero.

               c.  Fragments datagrams when necessary to fit into the
                   MTU of the next network.

            4.  Chooses a next-hop destination for each IP datagram,
                based on the information in its routing data-base.  See
                Section 4 (Gateway Algorithms).

            5.  Supports an interior gateway protocol (IGP) to carry out
                distributed routing and reachability algorithms with the
                other gateways in the same autonomous system.  In
                addition, some gateways will need to support the
                Exterior Gateway Protocol (EGP) to exchange topological
                information with other autonomous systems.  See
                Section 4 (Gateway Algorithms).

Braden & Postel                                                 [Page 7]

RFC 1009 - Requirements for Internet Gateways                  June 1987

            6.  Provides system support facilities, including loading,
                debugging, status reporting, exception reporting and
                control.  See Section 5 (Operation and Maintenance).

      1.2.2.  Embedded Gateways

         A gateway may be a stand-alone computer system, dedicated to
         its IP router functions.  Alternatively, it is possible to
         embed gateway functionality within a host operating system
         which supports connections to two or more networks.  The
         best-known example of an operating system with embedded gateway
         code is the Berkeley BSD system.  The embedded gateway feature
         seems to make internetting easy, but it has a number of hidden

            1.  If a host has only a single constituent-network
                interface, it should not act as a gateway.

                For example, hosts with embedded gateway code that
                gratuitously forward broadcast packets or datagrams on
                the same net often cause packet avalanches.

            2.  If a (multihomed) host acts as a gateway, it must
                implement ALL the relevant gateway requirements
                contained in this document.

                For example, the routing protocol issues (see Sections
                2.6 and 4.1) and the control and monitoring problems are
                as hard and important for embedded gateways as for
                stand-alone gateways.

                   Since Internet gateway requirements and
                   specifications may change independently of operating
                   system changes, an administration that operates an
                   embedded gateway in the Internet is strongly advised
                   to have an ability to maintain and update the gateway
                   code (e.g., this might require gateway code source).

            3.  Once a host runs embedded gateway code, it becomes part
                of the Internet system.  Thus, errors in software or
                configuration of such a host can hinder communication
                between other hosts.  As a consequence, the host
                administrator must lose some autonomy.

                In many circumstances, a host administrator will need to
                disable gateway coded embedded in the operating system,
                and any embedded gateway code must be organized so it
                can be easily disabled.

Braden & Postel                                                 [Page 8]

RFC 1009 - Requirements for Internet Gateways                  June 1987

            4.  If a host running embedded gateway code is concurrently
                used for other services, the O&M (operation and
                maintenance) requirements for the two modes of use may
                be in serious conflict.

                For example, gateway O&M will in many cases be performed
                remotely by an operations center; this may require
                privileged system access which the host administrator
                would not normally want to distribute.

      1.2.3.  Transparent Gateways

         The basic idea of a transparent gateway is that the hosts on
         the local-area network behind such a gateway share the address
         space of the wide-area network in front of the gateway.  In
         certain situations this is a very useful approach and the
         limitations do not present significant drawbacks.

         The words "in front" and "behind" indicate one of the
         limitations of this approach: this model of interconnection is
         suitable only for a geographically (and topologically) limited
         stub environment.  It requires that there be some form of
         logical addressing in the network level addressing of the
         wide-area network (that is, all the IP addresses in the local
         environment map to a few (usually one) physical address in the
         wide-area network, in a way consistent with the { IP address
         <-> network address } mapping used throughout the wide-area

         Multihoming is possible on one wide-area network, but may
         present routing problems if the interfaces are geographically
         or topologically separated.  Multihoming on two (or more)
         wide-area networks is a problem due to the confusion of

         The behavior that hosts see from other hosts in what is
         apparently the same network may differ if the transparent
         gateway cannot fully emulate the normal wide-area network
         service.  For example, if there were a transparent gateway
         between the ARPANET and an Ethernet, a remote host would not
         receive a Destination Dead message [3] if it sent a datagram to
         an Ethernet host that was powered off.

Braden & Postel                                                 [Page 9]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   1.3.  Gateway Characteristics

      Every Internet gateway must perform the functions listed above.
      However, a vendor will have many choices on power, complexity, and
      features for a particular gateway product.  It may be helpful to
      observe that the Internet system is neither homogeneous nor
      fully-connected.  For reasons of technology and geography, it is
      growing into a global-interconnect system plus a "fringe" of LANs
      around the "edge".

         *  The global-interconnect system is comprised of a number of
            wide-area networks to which are attached gateways of several
            ASs; there are relatively few hosts connected directly to
            it.  The global-interconnect system includes the ARPANET,
            the NSFNET "backbone", the various NSF regional and
            consortium networks, other ARPA sponsored networks such as
            the SATNET and the WBNET, and the DCA sponsored MILNET.  It
            is anticipated that additional networks sponsored by these
            and other agencies (such as NASA and DOE) will join the
            global-interconnect system.

         *  Most hosts are connected to LANs, and many organizations
            have clusters of LANs interconnected by local gateways.
            Each such cluster is connected by gateways at one or more
            points into the global-interconnect system.  If it is
            connected at only one point, a LAN is known as a "stub"

      Gateways in the global-interconnect system generally require:

         *  Advanced routing and forwarding algorithms

            These gateways need routing algorithms which are highly
            dynamic and also offer type-of-service routing.  Congestion
            is still not a completely resolved issue [24].  Improvements
            to the current situation will be implemented soon, as the
            research community is actively working on these issues.

         *  High availability

            These gateways need to be highly reliable, providing 24 hour
            a day, 7 days a week service.  In case of failure, they must
            recover quickly.

         *  Advanced O&M features

            These gateways will typically be operated remotely from a
            regional or national monitoring center.  In their

Braden & Postel                                                [Page 10]

RFC 1009 - Requirements for Internet Gateways                  June 1987

            interconnect role, they will need to provide sophisticated
            means for monitoring and measuring traffic and other events
            and for diagnosing faults.

         *  High performance

            Although long-haul lines in the Internet today are most
            frequently 56 Kbps, DS1 lines (1.5 Mbps) are of increasing
            importance, and even higher speeds are likely in the future.
            Full-duplex operation is provided at any of these speeds.

            The average size of Internet datagrams is rather small, of
            the order of 100 bytes.  At DS1 line speeds, the
            per-datagram processing capability of the gateways, rather
            than the line speed, is likely to be the bottleneck.  To
            fill a DS1 line with average-sized Internet datagrams, a
            gateway would need to pass -- receive, route, and send --
            2,000 datagrams per second per interface.  That is, a
            gateway which supported 3 DS1 lines and and Ethernet
            interface would need to be able to pass a dazzling 2,000
            datagrams per second in each direction on each of the
            interfaces, or a aggregate throughput of 8,000 datagrams per
            second, in order to fully utilize DS1 lines.  This is beyond
            the capability of current gateways.

               Note: some vendors count input and output operations
               separately in datagrams per second figures; for these
               vendors, the above example would imply 16,000 datagrams
               per second !

      Gateways used in the "LAN fringe" (e.g., campus networks) will
      generally have to meet less stringent requirements for
      performance, availability, and maintenance.  These may be high or
      medium-performance devices, probably competitively procured from
      several different vendors and operated by an internal organization
      (e.g., a campus computing center).  The design of these gateways
      should emphasize low average delay and good burst performance,
      together with delay and type-of-service sensitive resource
      management.  In this environment, there will be less formal O&M,
      more hand-crafted static configurations for special cases, and
      more need for inter-operation with gateways of other vendors.  The
      routing mechanism will need to be very flexible, but need not be
      so highly dynamic as in the global-interconnect system.

      It is important to realize that Internet gateways normally operate
      in an unattended mode, but that equipment and software faults can
      have a wide-spread (sometimes global) effect.  In any environment,

Braden & Postel                                                [Page 11]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      a gateway must be highly robust and able to operate, possibly in a
      degraded state, under conditions of extreme congestion or failure
      of network resources.

      Even though the Internet system is not fully-interconnected, many
      parts of the system do need to have redundant connectivity.  A
      rich connectivity allows reliable service despite failures of
      communication lines and gateways, and it can also improve service
      by shortening Internet paths and by providing additional capacity.
      The engineering tradeoff between cost and reliability must be made
      for each component of the Internet system.

Braden & Postel                                                [Page 12]

RFC 1009 - Requirements for Internet Gateways                  June 1987

2.  Protocols Required in Gateways

   The Internet architecture uses datagram gateways to interconnect
   constituent networks.  This section describes the various protocols
   which a gateway needs to implement.

   2.1.  Internet Protocol (IP)

      IP is the basic datagram protocol used in the Internet system [19,
      31].  It is described in RFC-791 [1] and also in MIL-STD-1777 [5]
      as clarified by RFC-963 [36] ([1] and [5] are intended to describe
      the same standard, but in quite different words).  The subnet
      extension is described in RFC-950 [21].

      With respect to current gateway requirements the following IP
      features can be ignored, although they may be required in the
      future:  Type of Service field, Security option, and Stream ID
      option.  However, if recognized, the interpretation of these
      quantities must conform to the standard specification.

      It is important for gateways to implement both the Loose and
      Strict Source Route options.  The Record Route and Timestamp
      options are useful diagnostic tools and must be supported in all

      The Internet model requires that a gateway be able to fragment
      datagrams as necessary to match the MTU of the network to which
      they are being forwarded, but reassembly of fragmented datagrams
      is generally left to the destination hosts.  Therefore, a gateway
      will not perform reassembly on datagrams it forwards.

      However, a gateway will generally receive some IP datagrams
      addressed to itself; for example, these may be ICMP Request/Reply
      messages, routing update messages (see Sections 2.3 and 2.6), or
      for monitoring and control (see Section 5).  For these datagrams,
      the gateway will be functioning as a destination host, so it must
      implement IP reassembly in case the datagrams have been fragmented
      by some transit gateway.  The destination gateway must have a
      reassembly buffer which is at least as large as the maximum of the
      MTU values for its network interfaces and 576.  Note also that it
      is possible for a particular protocol implemented by a host or
      gateway to require a lower bound on reassembly buffer size which
      is larger than 576.  Finally, a datagram which is addressed to a
      gateway may use any of that gateway's IP addresses as destination
      address, regardless of which interface the datagram enters.

      There are five classes of IP addresses:  Class A through
      Class E [23].  Of these, Class D and Class E addresses are

Braden & Postel                                                [Page 13]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      reserved for experimental use.  A gateway which is not
      participating in these experiments must ignore all datagrams with
      a Class D or Class E destination IP address.  ICMP Destination
      Unreachable or ICMP Redirect messages must not result from
      receiving such datagrams.

      There are certain special cases for IP addresses, defined in the
      latest Assigned Numbers document [23].  These special cases can be
      concisely summarized using the earlier notation for an IP address:

         IP-address ::=  { <Network-number>, <Host-number> }


         IP-address ::=  { <Network-number>, <Subnet-number>,
                                                         <Host-number> }

      if we also use the notation "-1" to mean the field contains all 1
      bits.  Some common special cases are as follows:

         (a)   {0, 0}

            This host on this network.  Can only be used as a source
            address (see note later).

         (b)   {0, <Host-number>}

            Specified host on this network.  Can only be used as a
            source address.

         (c)   { -1, -1}

            Limited broadcast.  Can only be used as a destination
            address, and a datagram with this address must never be
            forwarded outside the (sub-)net of the source.

         (d)   {<Network-number>, -1}

            Directed broadcast to specified network.  Can only be used
            as a destination address.

         (e)   {<Network-number>, <Subnet-number>, -1}

            Directed broadcast to specified subnet.  Can only be used as
            a destination address.

         (f)   {<Network-number>, -1, -1}

Braden & Postel                                                [Page 14]

RFC 1009 - Requirements for Internet Gateways                  June 1987

            Directed broadcast to all subnets of specified subnetted
            network.  Can only be used as a destination address.

         (g)   {127, <any>}

            Internal host loopback address.  Should never appear outside
            a host.

      The following two are conventional notation for network numbers,
      and do not really represent IP addresses.  They can never be used
      in an IP datagram header as an IP source or destination address.

         (h)   {<Network-number>, 0}

            Specified network (no host).

         (i)   {<Network-number>, <Subnet-number>, 0}

            Specified subnet (no host).

      Note also that the IP broadcast address, which has primary
      application to Ethernets and similar technologies that support an
      inherent broadcast function, has an all-ones value in the host
      field of the IP address.  Some early implementations chose the
      all-zeros value for this purpose, which is not in conformance with
      the specification [23, 49, 50].

   2.2.  Internet Control Message Protocol (ICMP)

      ICMP is an auxiliary protocol used to convey advice and error
      messages and is described in RFC-792 [2].

      We will discuss issues arising from gateway handling of particular
      ICMP messages.  The ICMP messages are grouped into two classes:
      error messages and information messages.  ICMP error messages are
      never sent about ICMP error messages, nor about broadcast or
      multicast datagrams.

         The ICMP error messages are: Destination Unreachable, Redirect,
         Source Quench, Time Exceeded, and Parameter Problem.

         The ICMP information messages are: Echo, Information,
         Timestamp, and Address Mask.

Braden & Postel                                                [Page 15]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      2.2.1.  Destination Unreachable

         The distinction between subnets of a subnetted network, which
         depends on the address mask described in RFC-950 [21], must not
         be visible outside that network.  This distinction is important
         in the case of the ICMP Destination Unreachable message.

         The ICMP Destination Unreachable message is sent by a gateway
         in response to a datagram which it cannot forward because the
         destination is unreachable or down.  The gateway chooses one of
         the following two types of Destination Unreachable messages to

            *  Net Unreachable

            *  Host Unreachable

         Net unreachable implies that an intermediate gateway was unable
         to forward a datagram, as its routing data-base gave no next
         hop for the datagram, or all paths were down.  Host Unreachable
         implies that the destination network was reachable, but that a
         gateway on that network was unable to reach the destination
         host.  This might occur if the particular destination network
         was able to determine that the desired host was unreachable or
         down.  It might also occur when the destination host was on a
         subnetted network and no path was available through the subnets
         of this network to the destination.  Gateways should send Host
         Unreachable messages whenever other hosts on the same
         destination network might be reachable; otherwise, the source
         host may erroneously conclude that ALL hosts on the network are
         unreachable, and that may not be the case.

      2.2.2.  Redirect

         The ICMP Redirect message is sent by a gateway to a host on the
         same network, in order to change the gateway used by the host
         for routing certain datagrams.  A choice of four types of
         Redirect messages is available to specify datagrams destined
         for a particular host or network, and possibly with a
         particular type-of-service.

         If the directly-connected network is not subnetted, a gateway
         can normally send a network Redirect which applies to all hosts
         on a specified remote network.  Using a network rather than a
         host Redirect may economize slightly on network traffic and on
         host routing table storage.  However, the saving is not
         significant, and subnets create an ambiguity about the subnet

Braden & Postel                                                [Page 16]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         mask to be used to interpret a network Redirect.  In a general
         subnet environment, it is difficult to specify precisely the
         cases in which network Redirects can be used.

         Therefore, it is recommended that a gateway send only host (or
         host and type-of-service) Redirects.

      2.2.3.  Source Quench

         All gateways must contain code for sending ICMP Source Quench
         messages when they are forced to drop IP datagrams due to
         congestion.  Although the Source Quench mechanism is known to
         be an imperfect means for Internet congestion control, and
         research towards more effective means is in progress, Source
         Quench is considered to be too valuable to omit from production

         There is some argument that the Source Quench should be sent
         before the gateway is forced to drop datagrams [62].  For
         example, a parameter X could be established and set to have
         Source Quench sent when only X buffers remain.  Or, a parameter
         Y could be established and set to have Source Quench sent when
         only Y per cent of the buffers remain.

         Two problems for a gateway sending Source Quench are: (1) the
         consumption of bandwidth on the reverse path, and (2) the use
         of gateway CPU time.  To ameliorate these problems, a gateway
         must be prepared to limit the frequency with which it sends
         Source Quench messages.  This may be on the basis of a count
         (e.g., only send a Source Quench for every N dropped datagrams
         overall or per given source host), or on the basis of a time
         (e.g., send a Source Quench to a given source host or overall
         at most once per T millseconds).  The parameters (e.g., N or T)
         must be settable as part of the configuration of the gateway;
         furthermore, there should be some configuration setting which
         disables sending Source Quenches.  These configuration
         parameters, including disabling, should ideally be specifiable
         separately for each network interface.

         Note that a gateway itself may receive a Source Quench as the
         result of sending a datagram targeted to another gateway.  Such
         datagrams might be an EGP update, for example.

      2.2.4.  Time Exceeded

         The ICMP Time Exceeded message may be sent when a gateway
         discards a datagram due to the TTL being reduced to zero.  It

Braden & Postel                                                [Page 17]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         may also be sent by a gateway if the fragments of a datagram
         addressed to the gateway itself cannot be reassembled before
         the time limit.

      2.2.5.  Parameter Problem

         The ICMP Parameter Problem message may be sent to the source
         host for any problem not specifically covered by another ICMP

      2.2.6.  Address Mask

         Host and gateway implementations are expected to support the
         ICMP Address Mask messages described in RFC-950 [21].

      2.2.7.  Timestamp

         The ICMP Timestamp message has proven to be useful for
         diagnosing Internet problems.  The preferred form for a
         timestamp value, the "standard value", is in milliseconds since
         midnight GMT.  However, it may be difficult to provide this
         value with millisecond resolution.  For example, many systems
         use clocks which update only at line frequency, 50 or 60 times
         per second.  Therefore, some latitude is allowed in a
         "standard" value:

            *  The value must be updated at a frequency of at least 30
               times per second (i.e., at most five low-order bits of
               the value may be undefined).

            *  The origin of the value must be within a few minutes of
               midnight, i.e., the accuracy with which operators
               customarily set CPU clocks.

         To meet the second condition for a stand-alone gateway, it will
         be necessary to query some time server host when the gateway is
         booted or restarted.  It is recommended that the UDP Time
         Server Protocol [44] be used for this purpose.  A more advanced
         implementation would use NTP (Network Time Protocol) [45] to
         achieve nearly millisecond clock synchronization; however, this
         is not required.

         Even if a gateway is unable to establish its time origin, it
         ought to provide a "non-standard" timestamp value (i.e., with
         the non-standard bit set), as a time in milliseconds from
         system startup.

Braden & Postel                                                [Page 18]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         New gateways, especially those expecting to operate at T1 or
         higher speeds, are expected to have at least millisecond

      2.2.8.  Information Request/Reply

         The Information Request/Reply pair was intended to support
         self-configuring systems such as diskless workstations, to
         allow them to discover their IP network numbers at boot time.
         However, the Reverse ARP (RARP) protocol [15] provides a better
         mechanism for a host to use to discover its own IP address, and
         RARP is recommended for this purpose.  Information
         Request/Reply need not be implemented in a gateway.

      2.2.9.  Echo Request/Reply

         A gateway must implement ICMP Echo, since it has proven to be
         an extremely useful diagnostic tool.  A gateway must be
         prepared to receive, reassemble, and echo an ICMP Echo Request
         datagram at least as large as the maximum of 576 and the MTU's
         of all of the connected networks.  See the discussion of IP
         reassembly in gateways, Section 2.1.

         The following rules resolve the question of the use of IP
         source routes in Echo Request and Reply datagrams.  Suppose a
         gateway D receives an ICMP Echo Request addressed to itself
         from host S.

            1.  If the Echo Request contained no source route, D should
                send an Echo Reply back to S using its normal routing
                rules.  As a result, the Echo Reply may take a different
                path than the Request; however, in any case, the pair
                will sample the complete round-trip path which any other
                higher-level protocol (e.g., TCP) would use for its data
                and ACK segments between S and D.

            2.  If the Echo Request did contain a source route, D should
                send an Echo Reply back to S using as a source route the
                return route built up in the source-routing option of
                the Echo Request.

Braden & Postel                                                [Page 19]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   2.3.  Exterior Gateway Protocol (EGP)

      EGP is the protocol used to exchange reachability information
      between Autonomous Systems of gateways, and is defined in
      RFC-904 [11].  See also RFC-827 [51], RFC-888 [46], and
      RFC-975 [27] for background information.  The most widely used EGP
      implementation is described in RFC-911 [13].

      When a dynamic routing algorithm is operated in the gateways of an
      Autonomous System (AS), the routing data-base must be coupled to
      the EGP implementation.  This coupling should ensure that, when a
      net is determined to be unreachable by the routing algorithm, the
      net will not be declared reachable to other ASs via EGP.  This
      requirement is designed to minimize spurious traffic to "black
      holes" and to ensure fair utilization of the resources on other

      The present EGP specification defines a model with serious
      limitations, most importantly a restriction against propagating
      "third party" EGP information in order to prevent long-lived
      routing loops [27].  This effectively limits EGP to a two-level
      hierarchy; the top level is formed by the "core" AS, while the
      lower level is composed of those ASs which are direct neighbor
      gateways to the core AS.  In practice, in the current Internet,
      nearly all of the "core gateways" are connected to the ARPANET,
      while the lower level is composed of those ASs which are directly
      gatewayed to the ARPANET or MILNET.

      RFC-975 [27] suggested one way to generalize EGP to lessen these
      topology restrictions;  it has not been adopted as an official
      specification, although its ideas are finding their way into the
      new EGP developments.  There are efforts underway in the research
      community to develop an EGP generalization which will remove these

      In EGP, there is no standard interpretation (i.e., metric) for the
      distance fields in the update messages, so distances are
      comparable only among gateways of the same AS.  In using EGP data,
      a gateway should compare the distances among gateways of the same
      AS and prefer a route to that gateway which has the smallest
      distance value.

      The values to be announced in the distance fields for particular
      networks within the local AS should be a gateway configuration
      parameter; by suitable choice of these values, it will be possible
      to arrange primary and backup paths from other AS's.  There are
      other EGP parameters, such as polling intervals, which also need
      to be set in the gateway configuration.

Braden & Postel                                                [Page 20]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      When routing updates become large they must be transmitted in
      parts.  One strategy is to use IP fragmentation, another is to
      explicitly send the routing information in sections.  The Internet
      Engineering Task Force is currently preparing a recommendation on
      this and other EGP engineering issues.

   2.4.  Address Resolution Protocol (ARP)

      ARP is an auxiliary protocol used to perform dynamic address
      translation between LAN hardware addresses and Internet addresses,
      and is described in RFC-826 [4].

      ARP depends upon local network broadcast.  In normal ARP usage,
      the initiating host broadcasts an ARP Request carrying a target IP
      address; the corresponding target host, recognizing its own IP
      address, sends back an ARP Reply containing its own hardware
      interface address.

      A variation on this procedure, called "proxy ARP", has been used
      by gateways attached to broadcast LANs [14].  The gateway sends an
      ARP Reply specifying its interface address in response to an ARP
      Request for a target IP address which is not on the
      directly-connected network but for which the gateway offers an
      appropriate route.  By observing ARP and proxy ARP traffic, a
      gateway may accumulate a routing data-base [14].

      Proxy ARP (also known in some quarters as "promiscuous ARP" or
      "the ARP hack") is useful for routing datagrams from hosts which
      do not implement the standard Internet routing rules fully -- for
      example, host implementations which predate the introduction of
      subnetting.  Proxy ARP for subnetting is discussed in detail in
      RFC-925 [14].

      Reverse ARP (RARP) allows a host to map an Ethernet interface
      address into an IP address [15].  RARP is intended to allow a
      self-configuring host to learn its own IP address from a server at
      boot time.

   2.5.  Constituent Network Access Protocols

      See Section 3.

Braden & Postel                                                [Page 21]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   2.6.  Interior Gateway Protocols

      Distributed routing algorithms continue to be the subject of
      research and engineering, and it is likely that advances will be
      made over the next several years.  A good algorithm needs to
      respond rapidly to real changes in Internet connectivity, yet be
      stable and insensitive to transients.  It needs to synchronize the
      distributed data-base across gateways of its Autonomous System
      rapidly (to avoid routing loops), while consuming only a small
      fraction of the available bandwidth.

      Distributed routing algorithms are commonly broken down into the
      following three components:

         A.  An algorithm to assign a "length" to each Internet path.

            The "length" may be a simple count of hops (1, or infinity
            if the path is broken), or an administratively-assigned
            cost, or some dynamically-measured cost (usually an average

            In order to determine a path length, each gateway must at
            least test whether each of its neighbors is reachable; for
            this purpose, there must be a "reachability" or "neighbor
            up/down" protocol.

         B.  An algorithm to compute the shortest path(s) to a given

         C.  A gateway-gateway protocol used to exchange path length and
             routing information among gateways.

      The most commonly-used IGPs in Internet gateways are as follows.

      2.6.1.  Gateway-to-Gateway Protocol (GGP)

         GGP was designed and implemented by BBN for the first
         experimental Internet gateways [41].  It is still in use in the
         BBN LSI/11 gateways, but is regarded as having serious
         drawbacks [58].  GGP is based upon an algorithm used in the
         early ARPANET IMPs and later replaced by SPF (see below).

         GGP is a "min-hop" algorithm, i.e., its length measure is
         simply the number of network hops between gateway pairs.  It
         implements a distributed shortest-path algorithm, which
         requires global convergence of the routing tables after a
         change in topology or connectivity.  Each gateway sends a GGP

Braden & Postel                                                [Page 22]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         routing update only to its neighbors, but each update includes
         an entry for every known network, where each entry contains the
         hop count from the gateway sending the update.

      2.6.2.  Shortest-Path-First (SPF) Protocols

         SPF [40] is the name for a class of routing algorithms based on
         a shortest-path algorithm of Dijkstra.  The current ARPANET
         routing algorithm is SPF, and the BBN Butterfly gateways also
         use SPF.  Its characteristics are considered superior to
         GGP [58].

         Under SPF, the routing data-base is replicated rather than
         distributed.  Each gateway will have its own copy of the same
         data-base, containing the entire Internet topology and the
         lengths of every path.  Since each gateway has all the routing
         data and runs a shortest-path algorithm locally, there is no
         problem of global convergence of a distributed algorithm, as in
         GGP.  To build this replicated data-base, a gateway sends SPF
         routing updates to ALL other gateways; these updates only list
         the distances to each of the gateway's neighbors, making them
         much smaller than GGP updates.  The algorithm used to
         distribute SPF routing updates involves reliable flooding.

      2.6.3.  Routing Information (RIP)

         RIP is the name often used for a class of routing protocols
         based upon the Xerox PUP and XNS routing protocols.  These are
         relatively simple, and are widely available because they are
         incorporated in the embedded gateway code of Berkeley BSD
         systems.  Because of this simplicity, RIP protocols have come
         the closest of any to being an "Open IGP", i.e., a protocol
         which can be used between different vendors' gateways.
         Unfortunately, there is no standard, and in fact not even a
         good document, for RIP.

         As in GGP, gateways using RIP periodically broadcast their
         routing data-base to their neighbor gateways, and use a
         hop-count as the metric.

         A fixed value of the hop-count (normally 16) is defined to be
         "infinity", i.e., network unreachable.  A RIP implementation
         must include measures to avoid both the slow-convergence
         phenomen called "counting to infinity" and the formation of
         routing loops.  One such measure is a "hold-down" rule.  This
         rule establishes a period of time (typically 60 seconds) during
         which a gateway will ignore new routing information about a
         given network, once the gateway has learned that network is

Braden & Postel                                                [Page 23]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         unreachable (has hop-count "infinity").  The hold-down period
         must be settable in the gateway configuration; if gateways with
         different hold-down periods are using RIP in the same
         Autonomous System, routing loops are a distinct possibility.
         In general, the hold-down period is chosen large enough to
         allow time for unreachable status to propagate to all gateways
         in the AS.

      2.6.4.  Hello

         The "Fuzzball" software for an LSI/11 developed by Dave Mills
         incorporated an IGP called the "Hello" protocol [39].  This IGP
         is mentioned here because the Fuzzballs have been widely used
         in Internet experimentation, and because they have served as a
         testbed for many new routing ideas.

   2.7.  Monitoring Protocols

      See Section 5 of this document.

   2.8.  Internet Group Management Protocol (IGMP)

      An extension to the IP protocol has been defined to provide
      Internet-wide multicasting, i.e., delivery of copies of the same
      IP datagram to a set of Internet hosts [47, 48].  This delivery is
      to be performed by processes known as "multicasting agents", which
      reside either in a host on each net or (preferably) in the

      The set of hosts to which a datagram is delivered is called a
      "host group", and there is a host-agent protocol called IGMP,
      which a host uses to join, leave, or create a group.  Each host
      group is distinguished by a Class D IP address.

      This multicasting mechanism and its IGMP protocol are currently
      experimental; implementation in vendor gateways would be premature
      at this time.  A datagram containing a Class D IP address must be
      dropped, with no ICMP error message.

Braden & Postel                                                [Page 24]

RFC 1009 - Requirements for Internet Gateways                  June 1987

3.  Constituent Network Interface

   This section discusses the rules used for transmission of IP
   datagrams on the most common types of constituent networks.  A
   gateway must be able to send and receive IP datagrams of any size up
   to the MTU of any constituent network to which it is connected.

   3.1.  Public data networks via X.25

      The formats specified for public data networks accessed via X.25
      are described in RFC-877 [8].  Datagrams are transmitted over
      standard level-3 virtual circuits as complete packet sequences.
      Virtual circuits are usually established dynamically as required
      and time-out after a period of no traffic.  Link-level
      retransmission, resequencing and flow control are performed by the
      network for each virtual circuit and by the LAPB link-level
      protocol.  Note that a single X.25 virtual circuit may be used to
      multiplex all IP traffic between a pair of hosts.  However,
      multiple parallel virtual circuits may be used in order to improve
      the utilization of the subscriber access line, in spite of small
      X.25 window sizes; this can result in random resequencing.

      The correspondence between Internet and X.121 addresses is usually
      established by table-lookup.  It is expected that this will be
      replaced by some sort of directory procedure in the future.  The
      table of the hosts on the Public Data Network is in the Assigned
      Numbers [23].

      The normal MTU is 576; however, the two DTE's (hosts or gateways)
      can use X.25 packet size negotiation to increase this value [8].

   3.2.  ARPANET via 1822 LH, DH, or HDH

      The formats specified for ARPANET networks using 1822 access are
      described in BBN Report 1822 [3], which includes the procedures
      for several subscriber access methods.  The Distant Host (DH)
      method is used when the host and IMP (the Defense Communication
      Agency calls it a Packet Switch Node or PSN) are separated by not
      more than about 2000 feet of cable, while the HDLC Distant Host
      (HDH) is used for greater distances where a modem is required.
      Under HDH, retransmission, resequencing and flow control are
      performed by the network and by the HDLC link-level protocol.

      The IP encapsulation format is simply to include the IP datagram
      as the data portion of an 1822 message.  In addition, the
      high-order 8 bits of the Message Id field (also known as the
      "link" field") should be set to 155 [23].  The MTU is 1007 octets.

Braden & Postel                                                [Page 25]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      While the ARPANET 1822 protocols are widely used at present, they
      are expected to be eventually overtaken by the DDN Standard X.25
      protocol (see Section 3.3).  The original IP address mapping
      (RFC-796 [38]) is in the process of being replaced by a new
      interface specification called AHIP-E; see RFC-1005 [61] for the

      Gateways connected to ARPANET or MILNET IMPs using 1822 access
      must incorporate features to avoid host-port blocking (i.e., RFNM
      counting) and to detect and report as ICMP Unreachable messages
      the failure of destination hosts or gateways (i.e., convert the
      1822 error messages to the appropriate ICMP messages).

      In the development of a network interface it will be useful to
      review the IMP end-to-end protocol described in RFC-979 [29].

   3.3.  ARPANET via DDN Standard X.25

      The formats specified for ARPANET networks via X.25 are described
      in the Defense Data Network X.25 Host Interface Specification [6],
      which describes two sets of procedures: the DDN Basic X.25, and
      the DDN Standard X.25.  Only DDN Standard X.25 provides the
      functionality required for interoperability assumptions of the
      Internet protocol.

      The DDN Standard X.25 procedures are similar to the public data
      network X.25 procedures, except in the address mappings.
      Retransmission, resequencing and flow control are performed by the
      network and by the LAPB link-level protocol.  Multiple parallel
      virtual circuits may be used in order to improve the utilization
      of the subscriber access line; this can result in random

      Gateways connected to ARPANET or MILNET using Standard X.25 access
      must detect and report as ICMP Unreachable messages the failure of
      destination hosts or gateways (i.e., convert the X.25 diagnostic
      codes to the appropriate ICMP messages).

      To achieve compatibility with 1822 interfaces, the effective MTU
      for a Standard X.25 interface is 1007 octets.

Braden & Postel                                                [Page 26]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   3.4.  Ethernet and IEEE 802

      The formats specified for Ethernet networks are described in
      RFC-894 [10].  Datagrams are encapsulated as Ethernet packets with
      48-bit source and destination address fields and a 16-bit type
      field (the type field values are listed in the Assigned
      Numbers [23]).  Address translation between Ethernet addresses and
      Internet addresses is managed by the Address Resolution Protocol,
      which is required in all Ethernet implementations.  There is no
      explicit link-level retransmission, resequencing or flow control,
      although most hardware interfaces will retransmit automatically in
      case of collisions on the cable.

      The IEEE 802 networks use a Link Service Access Point (LSAP) field
      in much the same way the ARPANET uses the "link" field.  Further,
      there is an extension of the LSAP header called the Sub-Network
      Access Protocol (SNAP).

      The 802.2 encapsulation is used on 802.3, 802.4, and 802.5 network
      by using the SNAP with an organization code indicating that the
      following 16 bits specify the Ether-Type code [23].


          MAC Header|      Length     |                  802.{3/4/5} MAC

         | DSAP=K1| SSAP=K1| control|                          802.2 SAP

         |protocol id or org code=K2|    Ether-Type   |       802.2 SNAP

      The total length of the SAP Header and the SNAP header is
      8-octets, making the 802.2 protocol overhead come out on a 64-bit

      K1 is 170.  The IEEE likes to talk about things in bit
      transmission order and specifies this value as 01010101.  In
      big-endian order, as used in the Internet specifications, this
      becomes 10101010 binary, or AA hex, or 170 decimal.  K2 is 0

      The use of the IP LSAP (K1 = 6) is reserved for future

Braden & Postel                                                [Page 27]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      The assigned values for the Ether-Type field are the same for
      either this IEEE 802 encapsulation or the basic Ethernet
      encapsulation [10].

      In either Ethernets or IEEE 802 nets, the IP datagram is the data
      portion of the packet immediately following the Ether-Type.

      The MTU for an Ethernet or its IEEE-standard equivalent (802.3) is
      1500 octets.

   3.5.  Serial-Line Protocols

      In some configurations, gateways may be interconnected with each
      other by means of serial asynchronous or synchronous lines, with
      or without modems.  When justified by the expected error rate and
      other factors, a link-level protocol may be required on the serial
      line.  While there is no single Internet standard for this
      protocol, it is suggested that one of the following protocols be

         *  X.25 LAPB  (Synchronous Lines)

            This is the link-level protocol used for X.25 network
            access.  It includes HDLC "bit-stuffing" as well as
            rotating-window flow control and reliable delivery.

               A gateway must be configurable to play the role of either
               the DCE or the DTE.

         *  HDLC Framing  (Synchronous Lines)

            This is just the bit-stuffing and framing rules of LAPB.  It
            is the simplest choice, although it provides no flow control
            or reliable delivery; however, it does provide error

         *  Xerox Synchronous Point-to-Point  (Synchronous Lines)

            This Xerox protocol is an elaboration upon HDLC framing that
            includes negotiation of maximum packet sizes, dial-up or
            dedicated circuits, and half- or full-duplex operation [12].

         *  Serial Line Framing Protocol  (Asynchronous Lines)

            This protocol is included in the MIT PC/IP package for an
            IBM PC and is defined in Appendix I to the manual for that
            system [20].

Braden & Postel                                                [Page 28]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      It will be important to make efficient use of the bandwidth
      available on a serial line between gateways.  For example, it is
      desirable to provide some form of data compression.  One possible
      standard compression algorithm, "Thinwire II", is described in
      RFC-914 [42].  This and similar algorithms are tuned to the
      particular types of redundancy which occur in IP and TCP headers;
      however, more work is necessary to define a standard serial-line
      compression protocol for Internet gateways.  Until a standard has
      been adopted, each vendor is free to choose a compression
      algorithm; of course, the result will only be useful on a serial
      line between two gateways using the same compression algorithm.

      Another way to ensure maximum use of the bandwidth is to avoid
      unnecessary retransmissions at the link level.  For some kinds of
      IP traffic, low delay is more important than reliable delivery.
      The serial line driver could distinguish such datagrams by their
      IP TOS field, and place them on a special high-priority,
      no-retransmission queue.

      A serial point-to-point line between two gateways may be
      considered to be a (particularly simple) network, a "null net".
      Considered in this way, a serial line requires no special
      considerations in the routing algorithms of the connected
      gateways, but does need an IP network number.  To avoid the
      wholesale consumption of Internet routing data-base space by null
      nets, we strongly recommend that subnetting be used for null net
      numbering, whenever possible.

         For example, assume that network 128.203 is to be constructed
         of gateways joined by null nets; these nets are given (sub-)net
         numbers 128.203.1, 128.203.2, etc., and the two interfaces on
         each end of null net 128.203.s might have IP addresses
         128.203.s.1 and 128.203.s.2.

      An alternative model of a serial line is that it is not a network,
      but rather an internal communication path joining two "half
      gateways".  It is possible to design an IGP and routing algorithm
      that treats a serial line in this manner [39, 52].

Braden & Postel                                                [Page 29]

RFC 1009 - Requirements for Internet Gateways                  June 1987

4.  Gateway Algorithms

   Gateways are general packet-switches that forward packets according
   to the IP address, i.e., they are IP routers.   While it is beyond
   the scope of this document to specify the details of the mechanisms
   used in any particular, perhaps proprietary, gateway architecture,
   there are a number of basic algorithms which must be provided by any
   acceptable design.

   4.1.  Routing Algorithm

      The routing mechanism is fundamental to Internet operation.  In
      all but trivial network topologies, robust Internet service
      requires some degree of routing dynamics, whether it be effected
      by manual or automatic means or by some combination of both.  In
      particular, if routing changes are made manually, it must be
      possible to make these routing changes from a remote Network
      Operation Center (NOC) without taking down the gateway for
      reconfiguration.  If static routes are used, there must be
      automatic fallback or rerouting features.

      Handling unpredictable changes in Internet connectivity must be
      considered the normal case, so that systems of gateways will
      normally be expected to have a routing algorithm with the
      capability of reacting to link and other gateway failures and
      changing the routing automatically.

      This document places no restriction on the type of routing
      algorithm, e.g., node-based, link-based or any other algorithm, or
      on the routing distance metric, e.g., delay or hop-count.
      However, the following features are considered necessary for a
      successful gateway routing algorithm:

         1.  The algorithm must sense the failure or restoration of a
             link or other gateway and switch to appropriate paths.  A
             design objective is to switch paths within an interval less
             than the typical TCP user time-out (one minute is a safe

         2.  The algorithm must suppress routing loops between neighbor
             gateways and must contain provisions to avoid or suppress
             routing loops that may form between non-neighbor gateways.
             A design objective is for no loop to persist for longer
             than an interval greater than the typical TCP user

         3.  The control traffic necessary to operate the routing
             algorithm must not significantly degrade or disrupt normal

Braden & Postel                                                [Page 30]

RFC 1009 - Requirements for Internet Gateways                  June 1987

             network operation.  Changes in state which might
             momentarily disrupt normal operation in a local-area must
             not cause disruption in remote areas of the network.

         4.  As the size of the network increases, the demand on
             resources must be controlled in an efficient way.  Table
             lookups should be hashed, for example, and data-base
             updates handled piecemeal, with only incremental changes
             broadcast over a wide-area.

         5.  The size of the routing data-base must not be allowed to
             exceed a constant, independent of network topology, times
             the number of nodes times the mean connectivity (average
             number of incident links).  An advanced design might not
             require that the entire routing data-base be kept in any
             particular gateway, so that discovery and caching
             techniques would be necessary.

         6.  Reachability and delay metrics, if used, must not depend on
             direct connectivity to all other gateways or on the use of
             network-specific broadcast mechanisms.  Polling procedures
             (e.g., for consistency checking) must be used only
             sparingly and in no case introduce an overhead exceeding a
             constant, independent of network topology, times the
             longest non-looping path.

         7.  Default routes (generally intended as a means to reduce the
             size of the routing data-base) must be used with care,
             because of the many problems with multiple paths, loops,
             and mis-configurations which routing defaults have caused.

             The most common application of defaults is for routing
             within an Internet region which is connected in a strictly
             hierarchical fashion and is a stub from the rest of the
             Internet system.  In this case, the default is used for
             routing "up" the tree.  Unfortunately, such restricted
             topology seldom lasts very long, and defaults cease to

             More generally, defaults could be used for initial routing
             guesses, with final routes to be discovered and cached from
             external or internal data-bases via the routing algorithm
             or EGP.

Braden & Postel                                                [Page 31]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   4.2.  Subnets and Routing

      We will call a gateway "subnetted" if at least one of its
      interfaces is connected to a subnet; the set of gateways directly
      connected to subnets of the same network will be referred to as a
      "subnet cluster".  For example, in the following diagram, network
      2 is subnetted, with subnets 2.1 and 2.2, but network 1 is not;
      gateways 1, 2, and 3 are subnetted and are members of the same
      subnet cluster.

         (Net 1) === [Gwy 1] === (Net 2.1) === [Gwy 2] === (Net 2.2)
            |                                                   |
            |                                                   |
             =================== [Gwy 3] =======================

      Subnets have the following effects on gateway routing:

         A.  Non-subnetted gateways are not affected at all.

         B.  The routing data-base in a subnetted gateway must consider
             the address mask for subnet entries.

         C.  Routing updates among the gateways in the same subnet
             cluster must include entries for the various subnets.  The
             corresponding address mask(s) may be implicit, but for full
             generality the mask needs to be given explicitly for each
             entry.  Note that if the routing data-base included a full
             32-bit mask for every IP network, the gateway could deal
             with networks and subnets in a natural way.  This would
             also handle the case of multiple subnet masks for the same
             subnetted network.

         D.  Routing updates from a subnetted gateway to a gateway
             outside the cluster can contain nets, never subnets.

         E.  If a subnetted gateway (e.g., gateway 2 above) is unable to
             forward a datagram from one subnet to another subnet of the
             same network, then it must return a Host Unreachable, not a
             Net Unreachable, as discussed in Section 2.2.1.

      When considering the choice of routing protocol, a gateway builder
      must consider how that protocol generalizes for subnets.  For some
      routing protocols it will be possible to use the same procedures
      in a regular gateway and a subnetted gateway, with only a change
      of parameters (e.g., address masks).

      A different subnet address mask must be configurable for each

Braden & Postel                                                [Page 32]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      interface of a given gateway.  This will allow a subnetted gateway
      to connect to two different subnetted networks, or to connect two
      subnets of the same network with different masks.

   4.3   Resource Allocation

      In order to perform its basic datagram-forwarding functions, a
      gateway must allocate resources; its packet buffers and CPU time
      must be allocated to packets it receives from connected networks,
      while the bandwidth to each of the networks must also be allocated
      for sending packets.  The choice of allocation strategies will be
      critical when a particular resource is scarce.  The most obvious
      allocation strategy, first-come-first-served (FCFS), may not be
      appropriate under overload conditions, for reasons which we will
      now explore.

      A first example is buffer allocation.  It is important for a
      gateway to allocate buffers fairly among all of its connected
      networks, even if these networks have widely varying bandwidths.
      A high-speed interface must not be allowed to starve slower
      interfaces of buffers.  For example, consider a gateway with a
      10 Mbps Ethernet connection and two 56 Kbps serial lines.  A buggy
      host on the Ethernet may spray that gateway interface with packets
      at high speed.  Without careful algorithm design in the gateway,
      this could tie up all the gateway buffers in such a way that
      transit traffic between the serial lines would be completely

      Allocation of output bandwidth may also require non-FCFS
      strategies.  In an advanced gateway design, allocation of output
      bandwidth may depend upon Type-of-Service bits in the IP headers.
      A gateway may also want to give priority to datagrams for its own
      up/down and routing protocols.

      Finally, Nagle [24] has suggested that gateways implement "fair
      queueing", i.e., sharing output bandwidth equitably among the
      current traffic sources.  In his scheme, for each network
      interface there would be a dynamically-built set of output queues,
      one per IP source address; these queues would be serviced in a
      round-robin fashion to share the bandwidth.  If subsequent
      research shows fair queueing to be desirable, it will be added to
      a future version of this document as a universal requirement.

Braden & Postel                                                [Page 33]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   4.4.  Special Addresses and Filters

      Section 2.1 contained a list of the 32-bit IP addresses which have
      special meanings.  They do not in general represent unique IP
      addresses of Internet hosts, and there are restrictions on their
      use in IP headers.

      We can distinguish two classes of these special cases.  The first
      class (specifically, cases (a), (b), (c), (g), (h), and (i) in
      section 2.1) contains addresses which should never appear in the
      destination address field of any IP datagram, so a gateway should
      never be asked to route to one of these addresses.  However, in
      the real world of imperfect implementations and configuration
      errors, such bad destination addresses do occur.  It is the
      responsibility of a gateway to avoid propagating such erroneous
      addresses; this is especially important for gateways included in
      the global interconnect system.  In particular, a gateway which
      receives a datagram with one of these forbidden addresses should:

         1.  Avoid inserting that address into its routing database, and
             avoid including it in routing updates to any other gateway.

         2.  Avoid forwarding a datagram containing that address as a

      To enforce these restrictions, it is suggested that a gateway
      include a configurable filter for datagrams and routing updates.
      A typical filter entry might consist of a 32-bit mask and value
      pair.  If the logical AND of the given address with the mask
      equals the value, a match has been found.  Since filtering will
      consume gateway resources, it is vital that the gateway
      configuration be able to control the degree of filtering in use.

      There is a second class of special case addresses (cases (d), (e),
      and (f) in section 2.1), the so-called "directed broadcasts".  A
      directed broadcast is a datagram to be forwarded normally to the
      specified destination (sub-)net and then broadcast on the final
      hop.  An Internet gateway is permitted, but not required, to
      filter out directed broadcasts destined for any of its
      locally-connected networks.  Hence, it should be possible to
      configure the filter to block the delivery of directed broadcasts.

      Finally, it will also be useful for Internet O&M to have a
      configurable filter on the IP source address.  This will allow a
      network manager to temporarily block traffic from a particular
      misbehaving host, for example.

Braden & Postel                                                [Page 34]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   4.5.  Redirects

      The ICMP Redirect message is specified only for use by a gateway
      to update the routing table of a host on the same connected net.
      However, the Redirect message is sometimes used between gateways,
      due to the following considerations:

         The routing function in a host is very much like that in a
         "dumb gateway" (i.e., a gateway having only static routes).  It
         is desirable to allow the routing tables of a dumb gateway to
         be changed under the control of a dynamic gateway (i.e., a
         gateway with full dynamic routing) on the same network.  By
         analogy, it is natural to let the dynamic gateway send ICMP
         Redirect messages to dumb gateway.

      The use of ICMP Redirect between gateways in this fashion may be
      considered to be part of the IGP (in fact, the totality of the
      IGP, as far as the dumb gateway is concerned!) in the particular
      Autonomous System.   Specification of an IGP is outside the scope
      of this document, so we only note the possibility of using
      Redirect in this fashion.  Gateways are not required to receive
      and act upon redirects, and in fact dynamic gateways must ignore
      them.  We also note that considerable experience shows that dumb
      gateways often create problems resulting in "black holes"; a full
      routing gateway is always preferable.

      Routing table entries established by redirect messages must be
      removed automatically, either by a time-out or when a use count
      goes to zero.

   4.6.  Broadcast and Multicast

      A host which is connected to a network (generally a LAN) with an
      intrinsic broadcast capability may want to use this capability to
      effect multidestination delivery of IP datagrams.  The basic
      Internet model assumes point-to-point messages, and we must take
      some care when we incorporate broadcasting.  It is important to
      note that broadcast addresses may occur at two protocol levels:
      the local network header and the IP header.

      Incorrect handling of broadcasting has often been the cause of
      packet avalanches (sometimes dubbed "meltdown") in LANs.  These
      avalanches are generally caused by gratuitous datagram-forwarding
      by hosts, or by hosts sending ICMP error messages when they
      discard broadcast datagrams.

      Gateways have a responsibility to prevent avalanches, or datagrams
      which can trigger avalanches, from escaping into another network.

Braden & Postel                                                [Page 35]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      In general, a gateway must not forward a datagram which arrives
      via local network broadcast, and must not send an ICMP error
      message when dropping the datagram.  A discussion of the rules
      will be found in Appendix A; see also [50].

      As noted in Section 4.4, a gateway is permitted to filter out
      directed broadcasts.  Hence, directed broadcasts will only be
      useful in limited Internet regions (e.g., the within the subnets
      of a particular campus) in which delivery is supported by the
      gateway administrators.  Host group multicasting (see Sections 2.8
      and 4.6) will soon provide a much more efficient mechanism than
      directed broadcasting.  Gateway algorithms for host group
      multicasting will be specified in future RFC's.

   4.7.  Reachability Procedures

      The architecture must provide a robust mechanism to establish the
      operational status of each link and node in the network, including
      the gateways, the links connecting them and, where appropriate,
      the hosts as well.  Ordinarily, this requires at least a
      link-level reachability protocol involving a periodic exchange of
      messages across each link.  This function might be intrinsic to
      the link-level protocols used (e.g., LAPB).  However, it is in
      general ill-advised to assume a host or gateway is operating
      correctly even if its link-level reachability protocol is
      operating correctly.  Additional confirmation is required in the
      form of an operating routing algorithm or peer-level reachability
      protocol (such as used in EGP).

      Failure and restoration of a link and/or gateway are considered
      network events and must be reported to the control center.  It is
      desirable, although not required, that reporting paths not require
      correct functioning of the routing algorithm itself.

   4.8.  Time-To-Live

      The Time-to-Live (TTL) field of the IP header is defined to be a
      timer limiting the lifetime of a datagram in the Internet.  It is
      an 8-bit field and the units are seconds.  This would imply that
      for a maximum TTL of 255 a datagram would time-out after about 4
      and a quarter minutes.  Another aspect of the definition requires
      each gateway (or other module) that handles a datagram to
      decrement the TTL by at least one, even if the elapsed time was
      much less than a second.  Since this is very often the case, the
      TTL effectively becomes a hop count limit on how far a datagram
      can propagate through the Internet.

Braden & Postel                                                [Page 36]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      As the Internet grows, the number of hops needed to get from one
      edge to the opposite edge increases, i.e., the Internet diameter

      If a gateway holds a datagram for more than one second, it must
      decrement the TTL by one for each second.

      If the TTL is reduced to zero, the datagram must be discarded, and
      the gateway may send an ICMP Time Exceeded message to the source.
      A datagram should never be received with a TTL of zero.

      When it originates a datagram, a gateway is acting in the role of
      a host and must supply a realistic initial value for the TTL.

Braden & Postel                                                [Page 37]

RFC 1009 - Requirements for Internet Gateways                  June 1987

5.  Operation and Maintenance

   5.1.  Introduction

      Facilities to support operation and maintenance (O&M) activities
      form an essential part of any gateway implementation.  The
      following kinds of activity are included under gateway O&M:

         *  Diagnosing hardware problems in the gateway processor, in
            its network interfaces, or in the connected networks,
            modems, or communication lines.

         *  Installing a new version of the gateway software.

         *  Restarting or rebooting a gateway after a crash.

         *  Configuring (or reconfiguring) the gateway.

         *  Detecting and diagnosing Internet problems such as
            congestion, routing loops, bad IP addresses, black holes,
            packet avalanches, and misbehaved hosts.

         *  Changing network topology, either temporarily (e.g., to
            diagnose a communication line problem) or permanently.

         *  Monitoring the status and performance of the gateways and
            the connected networks.

         *  Collecting traffic statistics for use in (Inter-)network

      Gateways, packet-switches, and their connected communication lines
      are often operated as a system by a centralized O&M organization.
      This organization will maintain a (Inter-)network operation
      center, or NOC, to carry out its O&M functions.  It is essential
      that gateways support remote control and monitoring from such a
      NOC, through an Internet path (since gateways might not be
      connected to the same network as their NOC).  Furthermore, an IP
      datagram traversing the Internet will often use gateways under the
      control of more than one NOC; therefore, Internet problem
      diagnosis will often involve cooperation of personnel of more than
      one NOC.  In some cases, the same gateway may need to be monitored
      by more than one NOC.

      The tools available for monitoring at a NOC may cover a wide range
      of sophistication.  Proposals have included multi-window, dynamic
      displays of the entire gateway system, and the use of AI
      techniques for automatic problem diagnosis.

Braden & Postel                                                [Page 38]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      Gateway O&M facilities discussed here are only a part of the large
      and difficult problem of Internet management.  These problems
      encompass not only multiple management organizations, but also
      multiple protocol layers.  For example, at the current stage of
      evolution of the Internet architecture, there is a strong coupling
      between host TCP implementations and eventual IP-level congestion
      in the gateway system [9].  Therefore, diagnosis of congestion
      problems will sometimes require the monitoring of TCP statistics
      in hosts.  Gateway algorithms also interact with local network
      performance, especially through handling of broadcast packets and
      ARP, and again diagnosis will require access to hosts (e.g.,
      examining ARP caches).  However, consideration of host monitoring
      is beyond the scope of this RFC.

      There are currently a number of R&D efforts in progress in the
      area of Internet management and more specifically gateway O&M.  It
      is hoped that these will lead quickly to Internet standards for
      the gateway protocols and facilities required in this area.  This
      is also an area in which vendor creativity can make a significant

   5.2.   Gateway O&M Models

      There is a range of possible models for performing O&M functions
      on a gateway.  At one extreme is the local-only model, under which
      the O&M functions can only be executed locally, e.g., from a
      terminal plugged into the gateway machine.  At the other extreme,
      the fully-remote model allows only an absolute minimum of
      functions to be performed locally (e.g., forcing a boot), with
      most O&M being done remotely from the NOC.  There intermediate
      models, e.g., one in which NOC personnel can log into the gateway
      as a host, using the Telnet protocol, to perform functions which
      can also be invoked locally.  The local-only model may be adequate
      in a few gateway installations, but in general remote operation
      from a NOC will be required, and therefore remote O&M provisions
      are required for most gateways.

      Remote O&M functions may be exercised through a control agent
      (program).  In the direct approach, the gateway would support
      remote O&M functions directly from the NOC using standard Internet
      protocols (e.g., UDP or TCP); in the indirect approach, the
      control agent would support these protocols and control the
      gateway itself using proprietary protocols.  The direct approach
      is preferred, although either approach is acceptable.  The use of
      specialized host hardware and/or software requiring significant
      additional investment is discouraged; nevertheless, some vendors
      may elect to provide the control agent as an integrated part of
      the network in which the gateways are a part.  If this is the

Braden & Postel                                                [Page 39]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      case, it is required that a means be available to operate the
      control agent from a remote site using Internet protocols and
      paths and with equivalent functionality with respect to a local
      agent terminal.

      It is desirable that a control agent and any other NOC software
      tools which a vendor provides operate as user programs in a
      standard operating system.  The use of the standard Internet
      protocols UDP and TCP for communicating with the gateways should
      facilitate this.

      Remote gateway monitoring and (especially) remote gateway control
      present important access control problems which must be addressed.
      Care must also be taken to ensure control of the use of gateway
      resources for these functions.  It is not desirable to let gateway
      monitoring take more than some limited fraction of the gateway CPU
      time, for example.  On the other hand, O&M functions must receive
      priority so they can be exercised when the gateway is congested,
      i.e., when O&M is most needed.

      There are no current Internet standards for the control and
      monitoring protocols, although work is in progress in this area.
      The Host Monitoring Protocol (HMP) [7] could be used as a model
      until a standard is developed; however, it is strongly recommended
      that gateway O&M protocol be built on top of one of the standard
      Internet end-to-end protocols UDP or TCP. An example of a very
      simple but effective approach to gateway monitoring is contained
      in RFC-996 [43].

   5.3.   Gateway O&M Functions

      The following O&M functions need to be performed in a gateway:

         A.  Maintenance -- Hardware Diagnosis

            Each gateway must operate as a stand-alone device for the
            purposes of local hardware maintenance.  Means must be
            available to run diagnostic programs at the gateway site
            using only on-site tools, which might be only a diskette or
            tape and local terminal.  It is desirable, although not
            required, to be able to run diagnostics or dump the gateway
            via the network in case of fault.  Means should be provided
            to allow remote control from the NOC of of modems attached
            to the gateway.  The most important modem control capability
            is entering and leaving loopback mode, to diagnose line

Braden & Postel                                                [Page 40]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         B.  Control -- Dumping and Rebooting

            It must be possible to dump and reboot a stand-alone gateway
            upon command from the NOC.  In addition, a stand-alone
            gateway must include a watchdog timer that either initiates
            a reboot automatically or signals a remote control site if
            not reset periodically by the software.  It is desirable
            that the boot data involved reside at an Internet host
            (e.g., the NOC host) and be transmitted via the net;
            however, the use of local devices at the gateway site is

         C.  Control -- Configuring the Gateway

            Every gateway will have a number of configuration parameters
            which must be set (see the next section for examples).  It
            must be possible to update the parameters without rebooting
            the gateway; at worst, a restart may be required.

         D.  Monitoring -- Status and Performance

            A mechanism must be provided for retrieving status and
            statistical information from a gateway.  A gateway must
            supply such information in response to a polling message
            from the NOC.  In addition, it may be desirable to configure
            a gateway to transmit status spontaneously and periodically
            to a NOC (or set of NOCs), for recording and display.

            Examples of interesting status information include: link
            status, queue lengths, buffer availability, CPU and memory
            utilization, the routing data-base, error counts, and packet
            counts.  Counts should be kept for dropped datagrams,
            separated by reason.  Counts of ICMP datagrams should be
            kept by type and categorized into those originating at the
            gateway, and those destined for the gateway.  It would be
            useful to maintain many of these statistics by network
            interface, by source/destination network pair, and/or by
            source/destination host pair.

            Note that a great deal of useful monitoring data is often to
            be found in the routing data-base.  It is therefore useful
            to be able to tap into this data-base from the NOC.

         E.  Monitoring -- Error Logging

            A gateway should be capable of asynchronously sending
            exception ("trap") reports to one or more specified Internet
            addresses, one of which will presumably be the NOC host.

Braden & Postel                                                [Page 41]

RFC 1009 - Requirements for Internet Gateways                  June 1987

            There must also be a mechanism to limit the frequency of
            such trap reports, and the parameters controlling this
            frequency must be settable in the gateway configuration.

            Examples of conditions which should result in traps include:
            datagrams discarded because of TTL expiration (an indicator
            of possible routing loops); resource shortages; or an
            interface changing its up/down status.

   5.4.   Gateway Configuration Parameters

      Every gateway will have a set of configuration parameters
      controlling its operation.  It must be possible to set these
      parameters remotely from the NOC or locally at any time, without
      taking the gateway down.

      The following is a partial but representative list of possible
      configuration parameters for a full-function gateway.  The items
      marked with "(i)" should be settable independently for each
      network interface.

         * (i)  IP (sub-) network address

         * (i)  Subnet address mask

         * (i)  MTU of local network

         * (i)  Hardware interface address

         * (i)  Broadcast compatibility option (0s or 1s)

         *      EGP parameters -- neighbors, Autonomous System number,
                and polling parameters

         *      Static and/or default routes, if any

         *      Enable/Disable Proxy ARP

         *      Source Quench parameters

         *      Address filter configuration

         *      Boot-host address

         *      IP address of time server host

         *      IP address(es) of logging host(s)

Braden & Postel                                                [Page 42]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         *      IP address(es) of hosts to receive traps

         *      IP address(es) of hosts authorized to issue control

         *      Error level for logging

         *      Maximum trap frequency

         *      Hold-down period (if any)

Braden & Postel                                                [Page 43]

RFC 1009 - Requirements for Internet Gateways                  June 1987

Appendix A.  Technical Details

   This Appendix collects a number of technical details and rules
   concerning datagram forwarding by gateways and datagram handling by
   hosts, especially in the presence of broadcasting and subnets.

   A.1.  Rules for Broadcasting

      The following rules define how to handle broadcasts of packets and
      datagrams [50]:

         a.  Hosts (which do not contain embedded gateways) must NEVER
             forward any datagrams received from a connected network,
             broadcast or not.

             When a host receives an IP datagram, if the destination
             address identifies the host or is an IP broadcast address,
             the host passes the datagram to its appropriate
             higher-level protocol module (possibly sending ICMP
             protocol unreachable, but not if the IP address was a
             broadcast address).  Any other IP datagram must simply be
             discarded, without an ICMP error message.  Hosts never send

         b.  All packets containing IP datagrams which are sent to the
             local-network packet broadcast address must contain an IP
             broadcast address as the destination address in their IP
             header.  Expressed in another way, a gateway (or host) must
             not send in a local-network broadcast packet an IP datagram
             that has a specific IP host address as its destination

         c.  A gateway must never forward an IP datagram that arrives
             addressed to the IP limited broadcast address {-1,-1}.
             Furthermore, it must must not send an ICMP error message
             about discarding such a datagram.

         d.  A gateway must not forward an IP datagram addressed to
             network zero, i.e., {0, *}.

         e.  A gateway may forward a directed broadcast datagram, i.e.,
             a datagram with the IP destination address:

            { <Network-number>, -1}.

             However, it must not send such a directed broadcast out the
             same interface it came in, if this interface has
             <Network-number> as its network number.  If the code in the

Braden & Postel                                                [Page 44]

RFC 1009 - Requirements for Internet Gateways                  June 1987

             gateway making this decision does not know what interface
             the directed-broadcast datagram arrived on, the gateway
             cannot support directed broadcast to this connected network
             at all.

         f.  A gateway is permitted to protect its connected networks by
             discarding directed broadcast datagrams.

      A gateway will broadcast an IP datagram on a connected network if
      it is a directed broadcast destined for that network.  Some
      gateway-gateway routing protocols (e.g., RIP) also require
      broadcasting routing updates on the connected networks.  In either
      case, the datagram must have an IP broadcast address as its

         Note:  as observed earlier, some host implementations (those
         based on Berkeley 4.2BSD) use zero rather than -1 in the host
         field.  To provide compatibility during the period until these
         systems are fixed or retired, it may be useful for a gateway to
         be configurable to send either choice of IP broadcast address
         and accept both if received.

   A.2.  ICMP Redirects

      A gateway will generate an ICMP Redirect if and only if the
      destination IP address is reachable from the gateway (as
      determined by the routing algorithm) and the next-hop gateway is
      on the same (sub-)network as the source host.  Redirects must not
      be sent in response to an IP network or subnet broadcast address
      or in response to a Class D or Class E IP address.

      A host must discard an ICMP Redirect if the destination IP address
      is not its own IP address, or the new target address is not on the
      same (sub-)network.  An accepted Redirect updates the routing
      data-base for the old target address.  If there is no route
      associated with the old target address, the Redirect is ignored.
      If the old route is associated with a default gateway, a new route
      associated with the new target address is inserted in the

Braden & Postel                                                [Page 45]

RFC 1009 - Requirements for Internet Gateways                  June 1987

Appendix B.  NSFNET Specific Requirements

   The following sections discuss certain issues of special concern to
   the NSF scientific networking community.  These issues have primary
   relevance in the policy area, but also have ramifications in the
   technical area.

   B.1.  Proprietary and Extensibility Issues

      Although hosts, gateways and networks supporting Internet
      technology have been in continuous operation for several years,
      vendors users and operators must understand that not all
      networking issues are fully resolved.  As a result, when new needs
      or better solutions are developed for use in the NSF networking
      community, it may be necessary to field new protocols or augment
      existing ones.  Normally, these new protocols will be designed to
      interoperate in all practical respects with existing protocols;
      however, occasionally it may happen that existing systems must be
      upgraded to support these new or augmented protocols.

      NSF systems procurements may favor those vendors who undertake a
      commitment to remain aware of current Internet technology and be
      prepared to upgrade their products from time to time as
      appropriate.  As a result, vendors are strongly urged to consider
      extensibility and periodic upgrades as fundamental characteristics
      of their products.  One of the most productive and rewarding ways
      to do this on a long-term basis is to participate in ongoing
      Internet research and development programs in partnership with the
      academic community.

   B.2.  Interconnection Technology

      In order to ensure network-level interoperability of different
      vendor's gateways within the NSFNET context, we specify that a
      gateway must at a minimum support Ethernet connections and serial
      line protocol connections.

      Currently the most important common interconnection technology
      between Internet systems of different vendors is Ethernet.  Among
      the reasons for this are the following:

         1.  Ethernet specifications are well-understood and mature.

         2.  Ethernet technology is in almost all aspects vendor

         3.  Ethernet-compatible systems are common and becoming more

Braden & Postel                                                [Page 46]

RFC 1009 - Requirements for Internet Gateways                  June 1987

      These advantages combined favor the use of Ethernet technology as
      the common point of demarcation between NSF network systems
      supplied by different vendors, regardless of technology.  It is a
      requirement of NSF gateways that, regardless of the possibly
      proprietary switching technology used to implement a given
      vendor-supplied network, its gateways must support an Ethernet
      attachment to gateways of other vendors.

      It is expected that future NSF gateway requirements will specify
      other interconnection technologies.  The most likely candidates
      are those based on X.25 or IEEE 802, but other technologies
      including broadband cable, optical fiber, or other media may also
      be considered.

   B.3.  Routing Interoperability

      The Internet does not currently have an "open IGP" standard, i.e.,
      a common IGP which would allow gateways from different vendors to
      form a single Autonomous System.  Several approaches to routing
      interoperability are currently in use among vendors and the NSF
      networking community.

      *  Proprietary IGP

         At least one gateway vendor has implemented a proprietary IGP
         and uses EGP to interface to the rest of the Internet.

      *  RIP

         Although RIP is undocumented and various implementations of it
         differ in subtle ways, it has been used successfully for
         interoperation among multiple vendors as an IGP.

      *  Gateway Daemon

         The NSF networking community has built a "gateway daemon"
         program which can mediate among multiple routing protocols to
         create a mixed-IGP Autonomous System.  In particular, the
         prototype gateway daemon executes on a 4.3BSD machine acting as
         a gateway and exchanges routing information with other
         gateways, speaking both RIP and Hello protocols; in addition,
         it supports EGP to other Autonomous Systems.

Braden & Postel                                                [Page 47]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   B.4.  Multi-Protocol Gateways

      The present NSF gateway requirements specify only the Internet
      protocol IP.  However, in a few years the Internet will begin a
      gradual transition to the functionally-equivalent subset of the
      ISO protocols [17].  In particular, an increasing percentage of
      the traffic will use the ISO Connectionless Mode Network Service
      (CLNS, but commonly called "ISO IP") [33] in place of IP.  It is
      expected that the ISO suite will eventually become the dominant
      one; however, it is also expected that requirements to support
      Internet IP will continue, perhaps indefinitely.

      To support the transition to ISO protocols and the coexistence
      stage, it is highly desirable that a gateway design provide for
      future extensions to support more than one protocol simultaneous,
      and in particular both IP and CLNS [18].

      Present NSF gateway requirements do not include protocols above
      the network layer, such as TCP, unless necessary for network
      monitoring or control.  Vendors should recognize that future
      requirements to interwork between Internet and ISO applications,
      for example, may result in an opportunity to market gateways
      supporting multiple protocols at all levels up through the
      application level [16].  It is expected that the network-level NSF
      gateway requirements summarized in this document will be
      incorporated in the requirements document for these
      application-level gateways.

      Internet gateways function as intermediate systems (IS) with
      respect to the ISO connectionless network model and incorporate
      defined packet formats, routing algorithms and related procedures
      [33, 34].  The ISO ES-IS [37] provides the functions of ARP and
      ICMP Redirect.

   B.5.  Access Control and Accounting

      There are no requirements for NSF gateways at this time to
      incorporate specific access-control and accounting mechanisms in
      the design;  however, these important issues are currently under
      study and will be incorporated into a subsequent edition of this
      document.  Vendors are encouraged to plan for the introduction of
      these mechanisms into their products.  While at this time no
      definitive common model for access control and accounting has
      emerged, it is possible to outline some general features such a
      model is likely to have, among them the following:

Braden & Postel                                                [Page 48]

RFC 1009 - Requirements for Internet Gateways                  June 1987

         1.  The primary access control and accounting mechanisms will
             be in the service hosts themselves, not the gateways,
             packet-switches or workstations.

         2.  Agents acting on behalf of access control and accounting
             mechanisms may be necessary in the gateways, to collect
             data, enforce password protection, or mitigate resource
             priority and fairness.  However, the architecture and
             protocols used by these agents may be a local matter and
             cannot be specified in advance.

         3.  NSF gateways may be required to incorporate access control
             and accounting mechanisms based on datagram
             source/destination address, as well as other fields in the
             IP header.

         4.  NSF gateways may be required to enforce policies on access
             to gateway and communication resources.  These policies may
             be based upon equity ("fairness") or upon inequity

Braden & Postel                                                [Page 49]

RFC 1009 - Requirements for Internet Gateways                  June 1987


   An earlier version of this document (RFC-985) [60] was prepared by
   Dave Mills in behalf of the Gateway Requirements Subcommittee of the
   NSF Network Technical Advisory Group, in cooperation with the
   Internet Activities Board, Internet Architecture Task Force, and
   Internet Engineering Task Force.  This effort was chaired by Dave
   Mills, and contributed to by many people.

   The authors of current document have also received assistance from
   many people in the NSF and ARPA networking community.  We thank you,
   one and all.

Braden & Postel                                                [Page 50]

RFC 1009 - Requirements for Internet Gateways                  June 1987

References and Bibliography

   Many of these references are  available from the DDN Network
   Information Center, SRI International, 333 Ravenswood Avenue, Menlo
   Park, California 94025 (telephone: 800-235-3155).

   [1]   Postel, J., "Internet Protocol", RFC-791, USC Information
         Sciences Institute, September 1981.

   [2]   Postel, J., "Internet Control Message Protocol", RFC-792, USC
         Information Sciences Institute, September 1981.

   [3]   BBN, "Interface Message Processor - Specifications for the
         Interconnection of a Host and an IMP", Report 1822, Bolt
         Beranek and Newman, December 1981.

   [4]   Plummer, D., "An Ethernet Address Resolution Protocol",
         RFC-826, Symbolics, September 1982.

   [5]   DOD, "Military Standard Internet Protocol", Military Standard
         MIL-STD-1777, United States Department of Defense, August 1983.

   [6]   BBN, "Defense Data Network X.25 Host Interface Specification",
         Report 5476, Bolt Beranek and Newman, December 1983.

   [7]   Hinden, R., "A Host Monitoring Protocol", RFC-869, BBN
         Communications, December 1983.

   [8]   Korb, J.T., "A Standard for the Transmission of IP Datagrams
         over Public Data Networks", RFC-877, Purdue University,
         September 1983.

   [9]   Nagle, J., "Congestion Control in IP/TCP Internetworks",
         RFC-896, Ford Aerospace, January 1984.

   [10]  Hornig, C., "A Standard for the Transmission of IP Datagrams
         over Ethernet Networks", RFC-894, Symbolics, April 1984.

   [11]  Mills, D.L., "Exterior Gateway Formal Specification", RFC-904,
         M/A-COM Linkabit, April 1984.

   [12]  Xerox, "Xerox Synchronous Point-to-Point Protocol", Xerox
         System Integration Standard 158412, December 1984.

   [13]  Kirton, P., "EGP Gateway under Berkeley UNIX 4.2", RFC-911, USC
         Information Sciences Institute, August 1984.

Braden & Postel                                                [Page 51]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   [14]  Postel, J., "Multi-LAN Address Resolution", RFC-925, USC
         Information Sciences Institute, October 1984.

   [15]  Finlayson, R., T. Mann, J. Mogul, and M. Theimer, "A Reverse
         Address Resolution Protocol", RFC-904, Stanford University,
         June 1984.

   [16]  NRC, "Transport Protocols for Department of Defense Data
         Networks", RFC-942, National Research Council, March 1985.

   [17]  Postel, J., "DOD Statement on NRC Report", RFC-945, USC
         Information Sciences Institute, April 1985.

   [18]  ISO, "Addendum to the Network Service Definition Covering
         Network Layer Addressing", RFC-941, International Standards
         Organization, April 1985.

   [19]  Leiner, B., J. Postel, R. Cole and D. Mills, "The DARPA
         Internet Protocol Suite", Proceedings INFOCOM 85, IEEE,
         Washington DC, March 1985.  Also in: IEEE Communications
         Magazine, March 1985.  Also available as ISI-RS-85-153.

   [20]  Romkey, J., "PC/IP Programmer's Manual", MIT Laboratory for
         Computer Science, pp. 57-59, April 1986.

   [21]  Mogul, J., and J. Postel, "Internet Standard Subnetting
         Procedure", RFC-950, Stanford University, August 1985.

   [22]  Reynolds, J., and J. Postel, "Official Internet Protocols",
         RFC-1011, USC Information Sciences Institute, May 1987.

   [23]  Reynolds, J., and J. Postel, "Assigned Numbers", RFC-1010, USC
         Information Sciences Institute, May 1987.

   [24]  Nagle, J., "On Packet Switches with Infinite Storage", RFC-970,
         Ford Aerospace, December 1985.

   [25]  SRI, "DDN Protocol Handbook", NIC-50004, NIC-50005, NIC-50006,
         (three volumes), SRI International, December 1985.

   [26]  SRI, "ARPANET Information Brochure", NIC-50003, SRI
         International, December 1985.

   [27]  Mills, D.L., "Autonomous Confederations", RFC-975, M/A-COM
         Linkabit, February 1986.

   [28]  Jacobsen, O., and J. Postel, "Protocol Document Order
         Information",  RFC-980, SRI International, March 1986.

Braden & Postel                                                [Page 52]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   [29]  Malis, A.G., "PSN End-to-End Functional Specification",
         RFC-979, BBN Communications, March 1986.

   [30]  Postel, J, "Internetwork Applications using the DARPA Protocol
         Suite", Proceedings INFOCOM 85, IEEE, Washington DC,
         March 1985.  Also available as ISI-RS-85-151.

   [31]  Postel, J, C. Sunshine, and D. Cohen, "The ARPA Internet
         Protocol", Computer Networks, Vol. 5, No. 4, July 1981.

   [32]  Cerf, V., and R. Kahn, "A Protocol for Packet Network
         Intercommunication", IEEE Transactions on Communication,
         May 1974.

   [33]  ISO, "Protocol for Providing the Connectionless-mode Network
         Service", RFC-994, DIS-8473, International Standards
         Organization, March 1986.

   [34]  ANSI, "Draft Network Layer Routing Architecture", ANSI X3S3.3,
         86-215R, April 1987.

   [35]  Rosen, E., "Exterior Gateway Protocol (EGP)", RFC-827, Bolt
         Beranek and Newman, October 1982.

   [36]  Sidhu, D., "Some Problems with the Specification of the
         Military Standard Internet Protocol", RFC-963, Iowa State
         University, November 1985.

   [37]  ISO, "End System to Intermediate System Routing Exchange
         Protocol for use in conjunction with ISO 8473", RFC-995,
         April 1986.

   [38]  Postel, J., "Address Mappings", RFC-796, USC/Information
         Sciences Institute, September 1981.

   [39]  Mills, D., "DCN Local Network Protocols", RFC-891, M/A-COM
         Linkabit, December 1983.

   [40]  McQuillan, J. M., I. Richer, and E. C. Rosen, "The New Routing
         Algorithm for the ARPANET",  IEEE Transactions on
         Communications, May 1980.

   [41]  Hinden, R., and A. Sheltzer, "The DARPA Internet Gateway",
         RFC-823, Bolt Beranek and Newman, September 1982.

   [42]  Farber, D., G. Delp, and T. Conte, "A Thinwire Protocol for
         Connecting Personal Computers to the Internet", RFC-914,
         University of Delaware, September 1984.

Braden & Postel                                                [Page 53]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   [43]  Mills, D., "Statistics Server", RFC-996, University Of
         Delaware, February 1987.

   [44]  Postel, J. and K. Harrenstien, "Time Protocol", RFC-868,
         May 1983.

   [45]  Mills, D., "Network Time Protocol (NTP)", RFC-958, M/A-Com
         Linkabit, September 1985.

   [46]  Seamonson, L., and E. Rosen, "Stub Exterior Gateway Protocol",
         RFC-888, Bolt Beranek And Newman, January 1984.

   [47]  Deering, S., and D. Cheriton, "Host Groups: A Multicast
         Extension to the Internet Protocol", RFC-966, Stanford
         University, December 1985.

   [48]  Deering, S., "Host Extensions for IP Multicasting", RFC-988,
         Stanford University, July 1986.

   [49]  Mogul, J., "Broadcasting Internet Datagrams", RFC-919, Stanford
         University, October 1984.

   [50]  Mogul, J., "Broadcasting Internet Datagrams in the Presence of
         Subnets", RFC-922, Stanford University, October 1984.

   [51]  Rosen, E., "Exterior Gateway Protocol", RFC-827, Bolt Beranek
         and Newman, October 1982.

   [52]  Rose, M., "Low Tech Connection into the ARPA Internet: The Raw
         Packet Split Gateway", Technical Report 216, Department of
         Information and Computer Science, University of California,
         Irvine, February 1984.

   [53]  Rosen, E., "Issues in Buffer Management", IEN-182, Bolt Beranek
         and Newman, May 1981.

   [54]  Rosen, E., "Logical Addressing", IEN-183, Bolt Beranek and
         Newman, May 1981.

   [55]  Rosen, E., "Issues in Internetting - Part 1: Modelling the
         Internet", IEN-184, Bolt Beranek and Newman, May 1981.

   [56]  Rosen, E., "Issues in Internetting - Part 2: Accessing the
         Internet", IEN-187, Bolt Beranek and Newman, June 1981.

   [57]  Rosen, E., "Issues in Internetting - Part 3: Addressing",
         IEN-188, Bolt Beranek and Newman, June 1981.

Braden & Postel                                                [Page 54]

RFC 1009 - Requirements for Internet Gateways                  June 1987

   [58]  Rosen, E., "Issues in Internetting - Part 4: Routing", IEN-189,
         Bolt Beranek and Newman, June 1981.

   [59]  Sunshine, C., "Comments on Rosen's Memos", IEN-191, USC
         Information Sciences Institute, July 1981.

   [60]  NTAG, "Requirements for Internet Gateways -- Draft", RFC-985,
         Network Technical Advisory Group, National Science Foundation,
         May 1986.

   [61]  Khanna, A., and Malis, A., "The ARPANET AHIP-E Host Access
         Protocol (Enhanced AHIP)", RFC-1005, BBN Communications,
         May 1987

   [62]  Nagle, J., "Congestion Control in IP/TCP Internetworks", ACM
         Computer Communications Review, Vol.14, no.4, October 1984.

Braden & Postel                                                [Page 55]