RFC 1009 - Requirements for Internet gateways

[フレーム]

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
 community.
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
 network.
 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
 routers.
 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.
 Symbolically:
 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
 network.
 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
 pitfalls:
 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
 network).
 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
 addresses.
 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"
 network.
 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
 gateways.
 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> }
 or
 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
 send:
 * 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
 gateways.
 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
 message.
 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
 clocks.
 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
 systems.
 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
 restrictions.
 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
 delay).
 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
 destination.
 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
 gateways.
 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
 proposal.
 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
 resequencing.
 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].
 Headers:
 ...--------+--------+--------+
 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
 boundary.
 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
 (zero).
 The use of the IP LSAP (K1 = 6) is reserved for future
 development.
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
 used.
 * 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
 detection.
 * 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
 assumption).
 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
 time-out.
 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
 work.
 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
 stopped.
 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
 destination.
 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
 grows.
 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
 planning.
 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
 contribution.
 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
 problems.
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
 acceptable.
 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
 commands
 * 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
 redirects.
 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
 field.
 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
 destination.
 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
 data-base.
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
 independent.
 3. Ethernet-compatible systems are common and becoming more
 so.
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
 ("priority").
Braden & Postel [Page 49]

RFC 1009 - Requirements for Internet Gateways June 1987
Acknowledgments
 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]

AltStyle によって変換されたページ (->オリジナル) /