RFC 3077

A Link-Layer Tunneling Mechanism for Unidirectional Links

📅 published March 2001, from ietf

Okay bestie, welcome to RFC 3077! 💖 It's a Standards Track doc, so it's a real spec that people are supposed to implement. It came out in March 2001 🌸 The authors are E. Duros (UDcast), W. Dabbous (INRIA Sophia-Antipolis), H. Izumiyama (no affiliation listed), N. Fujii (WIDE), and Y. Zhang (HRL). The header doesn't say it obsoletes or updates any other RFC, so it's standing on its own 🦋

Status of this Memo

This document specifies an Internet standards track protocol for the Internet community, and it's asking for discussion and suggestions for improvements 🫶 If you want to know where this protocol stands in the official list, check the current edition of the "Internet Official Protocol Standards" (STD 1). Distribution of this memo is unlimited ✨

Copyright (C) The Internet Society (2001). All Rights Reserved. 📜

Abstract

Okay so the big idea 💅 This doc describes a way to fake full two-way connectivity between every node that's directly hooked up to a one-way link. The "receiver" uses a link-layer tunneling mechanism to forward datagrams to the "feeds" over a separate two-way IP (Internet Protocol) network. Since it lives at the link layer, protocols other than IP can ride along too 🚀

1. Introduction ✨

Internet routing and upper-layer protocols assume that links are bidirectional. In other words, hosts that are directly connected can talk to each other over the same link 🌈

This doc describes a link-layer tunneling mechanism that lets a set of nodes (feeds and receivers, see Section 2 for terminology) that are directly connected by a unidirectional link send datagrams as if they were all connected by a bidirectional link 🦋 Section 3 walks through a generic topology with a tunneling mechanism that supports multiple feeds and receivers. Heads up, though: this mechanism is not designed for topologies where a pair of nodes is connected by 2 unidirectional links pointing in opposite directions ⚠️

The tunneling mechanism needs all nodes to have an extra interface to an IP interconnected infrastructure 🧚

The mechanism is implemented at the link layer of the interface of every node connected to the unidirectional link. The goal is to hide the unidirectional nature of the link from higher layers, meaning the network layer and above 🙈 It also includes an automatic tunnel configuration protocol, so nodes can come up and go down at any time 💜

Generic Routing Encapsulation [RFC2784] is suggested as the tunneling mechanism, because it can carry IP, ARP datagrams, and any other layer-3 protocol between nodes 🎀

The tunneling mechanism described here was discussed and agreed upon by the UDLR working group 👑

The keywords MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL, when they appear in this document, are to be interpreted as described in [RFC2119]. 📖

2. Terminology 🍓

  • Unidirectional link (UDL): A one-way transmission link, for example a broadcast satellite link 🛰️
  • Receiver: A router or a host that has receive-only connectivity to a UDL 👂
  • Send-only feed: A router that has send-only connectivity to a UDL 📤
  • Receive capable feed: A router that has send-and-receive connectivity to a UDL 🔁
  • Feed: A send-only or a receive capable feed 💌
  • Node: A receiver or a feed 🫧
  • Bidirectional interface: A typical communication interface that can send or receive packets, like an Ethernet card, a modem, and so on 🔌

3. Topology 🗺️

Feeds and receivers are connected via a unidirectional link. Send-only feeds can only send data over this link, and receivers can only receive data from it. Receive capable feeds have both send and receive capabilities 💫

This mechanism was designed to work with any topology with any number of receivers and one or more feeds. That said, the number of feeds is expected to stay small. The special case of a single send-only feed with multiple receivers is one of the topologies supported 🧁

A receiver has several interfaces: a receive-only interface and one or more additional bidirectional communication interfaces 📻

A feed also has several interfaces: a send-only or send-and-receive capable interface connected to the unidirectional link, and one or more additional bidirectional communication interfaces. A feed MUST be a router. 👑

Tunnels are built between the bidirectional interfaces of nodes, so those interfaces must be interconnected by an IP infrastructure. In this document we assume that infrastructure is the Internet 🌐

Figure 1 shows a generic topology with several feeds and several receivers 👀

                 Unidirectional Link

     ---->---------->------------------->------
      |          |               |           |
      |f1u       |f2u            |r2u        |r1u
  --------   --------        --------    --------   ----------
  |Feed 1|   |Feed 2|        |Recv 2|    |Recv 1|---|subnet A|
  --------   --------        --------    --------   ----------
      |f1b       |f2b            |r2b        |r1b      |
      |          |               |           |         |
     ----------------------------------------------------
     |                     Internet                     |
     ----------------------------------------------------
                 Figure 1: Generic topology

f1u (resp. f2u) is the IP address of the 'Feed 1' (resp. Feed 2) send-only interface. ✈️

f1b (resp. f2b) is the IP address of the 'Feed 1' (resp. Feed 2) bidirectional interface connected to the Internet. 🌍

r1u (resp. r2u) is the IP address of the 'Receiver 1' (resp. Receiver 2) receive-only interface. 📡

r1b (resp. r2b) is the IP address of the 'Receiver 1' (resp. Receiver 2) bidirectional interface connected to the Internet. 🌏

Subnet A is a local area network connected to recv1. 🏠

Here's the tea ☕ Nodes have IP addresses on both their unidirectional and their bidirectional interfaces. The addresses on the unidirectional interfaces (f1u, f2u, r1u, r2u) are drawn from the same IP network. The addresses on the bidirectional interfaces (f1b, f2b, r1b, r2b) are generally drawn from different IP networks, and the Internet routes between them 💖

Receive-only interfaces are "dumb" and send-only interfaces are "deaf". So a datagram handed to the link-layer driver of a receive-only interface is simply thrown away. The link layer of a send-only interface never receives anything 💀

The network layer doesn't know anything about the underlying transmission technology, except that it thinks its access is bidirectional. Basically, for outgoing datagrams, the network layer picks the correct first hop on the connected network using a routing table, then passes the packet(s) to the right link-layer driver 🧠

Looking at Figure 1, Recv 1 and Feed 1 belong to the same network. But if Recv 1 starts a ping f1u, it can't get a response from Feed 1. The network layer of Recv 1 hands the packet to the driver of the receive-only interface, which obviously can't send it to the feed 🙃

Lots of Internet protocols assume that links are bidirectional. Routing protocols used by directly connected routers, in particular, no longer behave properly when there's a unidirectional link in the mix 😤

5. Emulating a broadcast bidirectional network 📡💫

The simplest fix is to emulate a broadcast-capable link-layer network. That lets existing higher-level protocols be deployed right away without any changes 🙌 Other network structures, such as NBMA, could also be emulated, but a broadcast network is more generally useful. A layer 3 network could be emulated too, but a link-layer network lets you use any other network-layer protocols immediately, and most notably it lets you use ARP right away 🎀

The link-layer tunneling mechanism that emulates bidirectional connectivity in the presence of a unidirectional link is described in the next section. First, though, we look at the different communication scenarios that characterize a broadcast network. That tells us what the tunneling mechanism has to do in order to emulate a bidirectional broadcast link 🔍

Here are the scenarios that would be feasible on a broadcast network, meaning if feeds and receivers were connected by a bidirectional broadcast link 📋

  • Scenario 1: A receiver can send a packet to a feed (point-to-point communication between a receiver and a feed). 🎯
  • Scenario 2: A receiver can send a broadcast/multicast packet on the link to all nodes (point-to-multipoint). 📢
  • Scenario 3: A receiver can send a packet to another receiver (point-to-point communication between two receivers). 🤝
  • Scenario 4: A feed can send a packet to a send-only feed (point-to-point communication between two feeds). 🪄
  • Scenario 5: A feed can send a broadcast/multicast packet on the link to all nodes (point-to-multipoint). 🌟
  • Scenario 6: A feed can send a packet to a receiver or a receive capable feed (point-to-point). ✅

All of these scenarios are possible on a broadcast network. Scenario 6 is already feasible on the unidirectional link, so the link-layer tunneling mechanism should provide the functionality to support scenarios 1 to 5 🧚‍♀️

One more thing: regular IP forwarding over this emulated network (that is, using the emulated network as a transit network) works correctly. The next-hop address at the receiver will be the unidirectional link address of another router (a feed or a receiver), which will then relay the packet 🦋✨

okay so this one is the tunneling mechanism, and it lives underneath the network layer 🧚. its whole job is to fake bidirectional link-layer connectivity 💖. the best part is that the network layer has zero clue. to it, the link looks and acts bidirectional, periodt 💅

Figure 2 shows the layered picture for Scenario 1 🦋

           Send-only Feed                       Receiver

            decapsulation                     encapsulation
     /-----***************----\       /-->---***************--\
     |                        |       |                       |
     |                        |       |                       |
   --|----------------------  |       |  ---------------------|---
   | |    f1b  |  f1u      |  |       |  |    x  r1u | r1b    |  |
   | |         |       ^   |  |   IP  |  |    |      |        v  |
   | ^         |       |   |  v       |  |    |      |        |  |
   | |         |       |   |  |       |  |    v      |        |  |
   |-|---------|-------|---|  |       |  |----|------|--------|--|
   | |         |       |   |  |       ^  |    |      |        |  |
   | |         |       |   |  |   LL  |  |    |      |        |  |
   | |         |       |   |  |       |  |    |      |        |  |
   | |         |       O------/       \<------O      |        |  |
   |-|---------|-----------|             |-----------|--------|--|
   | |         |           |             |           |        |  |
   | |         |           |     PHY     |           |        |  |
   | |         |           |             |           |        v  |
   | |         | |         |             |         | |        |  |
   --|-----------|----------             ----------|----------|---
     | Bidir     | Send-Only             Recv-Only |   Bidir  |
     ^ Interf    | Interf        UDL      Interf   |   Interf |
     |           \------------>------->------------/          |
     \----------------------<------------------------<--------/
                          Bidirectional network

  x : IP layer at the receiver generates a datagram to be forwarded
      on the receive-only interface.
  O : Entry point where the link-layer tunneling mechanism is
      triggered.

  Figure 2: Scenario 1 using the link-layer Tunneling Mechanism

6.1. Tunneling mechanism on the receiver

okay so on the receiver, a datagram gets handed down to the link layer of the unidirectional interface for transmission (see Figure 2) 🎀. then it gets wrapped in a MAC header for that unidirectional link. but this packet can't go straight over the link, so it goes through the tunneling mechanism next 🫶

the packet gets wrapped in an IP header whose destination is the IP address of a feed bidirectional interface (f1b or f2b) 💜. that destination address is also called the tunnel end-point. how a receiver learns these addresses and picks the feed is in Section 7 🔥. the encapsulation type is in Section 8.

in all cases the packet gets encapsulated, but the tunnel end-point (an IP address) depends on the encapsulated packet's destination MAC address. if the destination MAC address is:

  1. the MAC address of a feed interface connected to the unidirectional link (Scenario 1). the datagram is encapsulated, and the destination of the encapsulating datagram is the feed tunnel end-point (f1b referring to Figure 2).
  2. a MAC broadcast/multicast address (Scenario 2). the datagram is encapsulated, and the destination of the encapsulating datagram is the default feed tunnel end-point. see Section 7.4 for the deets on the default feed 🌈.
  3. a MAC address that belongs to the unidirectional network but is not a feed address (Scenario 3). the datagram is encapsulated, and the destination of the encapsulating datagram is the default feed tunnel end-point.

the encapsulated datagram goes to the network layer, which forwards it based on its destination address. that destination is a feed bidirectional interface that's reachable via the Internet. so in this case, the encapsulated datagram is forwarded via the receiver bidirectional interface (r1b) ✨

6.2. Tunneling mechanism on the feed

a feed handles unidirectional-link packets in two different ways 🧚:

  • packets generated by a local application or packets routed as usual by the IP layer may have to be forwarded over the unidirectional link (Section 6.2.1)
  • encapsulated packets received from another receiver or feed need tunnel processing (Section 6.2.2).

a feed cannot directly send a packet to a send-only feed over the unidirectional link (Scenario 4) 🙈. to fake this kind of communication, feeds have to tunnel packets to send-only feeds. A feed MUST maintain a list of all other feed tunnel end-points. This list MUST indicate which are send-only feed tunnel end-points. the local administrator configures this manually at the feed, as described in Section 7 💅.

when a datagram gets delivered to the link layer of a feed's unidirectional interface for transmission, what happens depends on the packet's destination MAC address. if it's:

  1. the MAC address of a receiver or a receive capable feed (Scenario 6). the packet is sent over the unidirectional link. this is classic "forwarding" 🍓.
  2. the MAC address of a send-only feed (Scenario 4). the packet is encapsulated and sent to the send-only feed tunnel end-point. the encapsulation type is described in Section 8.
  3. a broadcast/multicast destination (Scenario 5). the packet is sent over the unidirectional link. concurrently, a copy of this packet is encapsulated and sent to every feed on the list of send-only feed tunnel end-points. so the broadcast/multicast reaches all receivers and all send-only feeds 👑

6.2.2. Receiving encapsulated packets

feeds listen for incoming encapsulated datagrams on their tunnel end-points 👀. encapsulated packets will have been received on a bidirectional interface, and they traversed their way up the IP stack. then they enter a decapsulation process (see Figure 2).

decapsulation reveals the original link-layer packet. note that this has not been modified in any way by intermediate routers. in particular, the original MAC header will be intact 🫶

what happens next depends on the destination MAC address of the link-layer packet, which can be:

  1. the MAC address of the feed interface connected to the unidirectional link, i.e., own MAC address (Scenarios 1 and 4). the packet is passed to the link layer of the interface connected to the unidirectional link, which can then deliver it up to higher layers. so the datagram is processed as if it was coming from the unidirectional link and being delivered locally. Scenarios 1 and 4 are now feasible: a receiver or a feed can send a packet to a feed 🎀.
  2. a receiver address (Scenario 3). the packet is passed to the link layer of the interface connected to the unidirectional link. it is directly sent over the unidirectional link, to the indicated receiver. note, the packet must not be delivered locally. Scenario 3 is now feasible: a receiver can send a packet to another receiver 🦋.
  3. a broadcast/multicast address. this corresponds to Scenarios 2 and 5. we have to tell two cases apart: either (i) the encapsulated packet was sent from a receiver or (ii) from a feed (encapsulated broadcast/multicast packet sent to a send-only feed). these cases are told apart by examining the source address of the encapsulating packet and comparing it with the configured list of feed IP addresses. the action then taken is:
    • i) the feed was designated as a default feed by a receiver to forward the broadcast/multicast packet. the feed is then in charge of sending the multicast packet to all nodes. delivery to all nodes is accomplished by executing all 3 of the following actions:
      • the packet is encapsulated and sent to the list of send-only feed tunnel end-points.
      • also, the packet is passed to the link layer of the interface which forwards it directly over the unidirectional link (all receivers and receive capable feeds receive it).
      • also, the link layer delivers it locally to higher layers.
Caution: a receiver which sends an encapsulated broadcast/multicast packet to a default feed will receive its own packet via the unidirectional link. Correct filtering as described in [RFC1112] must be applied. 🙀
  • ii) the feed receives the packet and keeps it for local delivery. the packet is passed to the link layer of the interface connected to the unidirectional link, which delivers it to higher layers.

Scenario 2 is now feasible: a receiver can send a broadcast/multicast packet over the unidirectional link and it will be heard by all nodes 📣

7. Dynamic Tunnel Configuration Protocol (DTCP) 🎀

Okay bestie, so receivers and feeds need to know the feed tunnel end-points before they can forward encapsulated datagrams (see Scenarios 1 and 4) 💅

The number of feeds is expected to be pretty small (see Section 3), so every feed gets its full list of feeds configured by hand 📝 That list should note which feeds are send-only and which are receive capable. The administrator sets up tunnels to all the send-only feeds. A tunnel end-point is an IP address of a bidirectional link on a send-only feed 🦋

Receivers can't be configured by hand, though, bc scalability 😭 So tunnels have to be configured and maintained dynamically by receivers, both for scalability and to handle these events:

  1. New feed detection. When a new feed comes up, every receiver must create a tunnel to enable bidirectional communication with it. ✨
  2. Loss of unidirectional link detection. When the unidirectional link is down, receivers must disable their tunnels. The tunneling mechanism emulates bidirectional connectivity between nodes. Therefore, if the unidirectional link is down, a feed should not receive datagrams from the receivers. Protocols that consider a link as operational if they receive datagrams from it (e.g., the RIP protocol [RFC2453]) require this behavior for correct operation. 🚫
  3. Loss of feed detection. When a feed is down, receivers must disable their corresponding tunnel. This prevents unnecessary datagrams from being tunneled, which might overload the Internet. For instance, there is no need for receivers to forward a broadcast message through a tunnel whose end-point is down. 💀

The DTCP protocol gives receivers a way to dynamically discover feeds and keep a list of operational tunnel end-points 🌈 Feeds periodically announce their tunnel end-point addresses over the unidirectional link, and receivers listen to those announcements and maintain their list of tunnel end-points. It's giving organised 💜

7.1. The HELLO message

DTCP is a 'unidirectional protocol' 🫶 Messages are only sent from feeds to receivers, never the other way around.

The packet format is shown in Figure 3. Fields contain binary integers, in normal Internet order with the most significant bit first. Each tick mark represents one bit. 🎀

   0                   1                   2                   3
   0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   | Vers  |  Com  |    Interval   |           Sequence            |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   | res |F|IP Vers|  Tunnel Type  |   Nb of FBIP  |    reserved   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                   Feed  BDL IP addr (FBIP1)    (32/128 bits)  |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                             .....                             |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                   Feed  BDL IP addr (FBIPn)    (32/128 bits)  |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Figure 3: Packet Format

Every datagram contains the following fields. Constants are written in uppercase and are defined in Section 7.5:

  • Vers (4 bit unsigned integer): DTCP version number. It MUST be DTCP_VERSION.
  • Com (4 bit unsigned integer): Command field, possible values are:
    • 1 - JOIN: A message announcing that the feed sending this message is up and running.
    • 2 - LEAVE: A message announcing that the feed sending this message is being shut down.
  • Interval (8 bit unsigned integer): Interval in seconds between HELLO messages for the IP protocol in "IP Vers". It has to be > 0. The recommended value is HELLO_INTERVAL. If this value is increased, the feed MUST continue to send HELLO messages at the old rate for at least the old HELLO_LEAVE period. ⏱️
  • Sequence (16 bit unsigned integer): Random value initialized at boot time and incremented by 1 every time a value of the HELLO message is modified. 🔢
  • res (3 bits): Reserved/unused field, MUST be zero.
  • F (1 bit): bit indicating the type of feed:
    • 0 = Send-only feed
    • 1 = Receive-capable feed
  • IP Vers (4 bit unsigned integer): IP protocol version of the feed bidirectional IP addresses (FBIP):
    • 4 = IP version 4
    • 6 = IP version 6
  • Tunnel Type (8 bit unsigned integer): tunneling protocol supported by the feed. This value is the IP protocol number defined in [RFC1700] [iana/protocol-numbers] and their legitimate descendents. Receivers MUST use this form of tunnel encapsulation when tunneling to the feed.

Other protocol types allowing link-layer encapsulation are permitted. Obtaining new values is documented in [RFC2780].

  • Nb of FBIP (8 bit unsigned integer): Number of bidirectional IP feed addresses which are enumerated in the HELLO message. 🔢
  • reserved (8 bits): Reserved/unused field, MUST be zero.
  • Feed BDL IP addr (32 or 128 bits): The bidirectional IP address feed is the IP address of a feed bidirectional interface (tunnel end-point) reachable via the Internet. A feed has Nb of FBIP IP addresses which are operational tunnel end-points. They are enumerated in preferred order, with FBIP1 being the most suitable tunnel end-point. 👑

7.2. DTCP on the feed: sending HELLO packets

DTCP runs on top of UDP 🦋 Packets are sent to the "DTCP announcement" multicast address over the unidirectional link on port HELLO_PORT with a TTL of 1. Due to existing deployments, a feed SHOULD also support the use of the old DTCP announcement address, as described in Appendix B.

The source address of the HELLO packet is set to the IP address of the feed interface connected to the unidirectional link. In the rest of the document, this value is called FUIP (Feed Unidirectional IP address). 📡

The process in charge of sending HELLO packets fills every field of the datagram according to the description given in Section 7.1. 💅

As long as a feed is up and running, it periodically announces its presence to receivers. It MUST send HELLO packets containing a JOIN command every HELLO_INTERVAL over the unidirectional link. 📣

Referring to Figure 1 in Section 3, Feed 1 (resp. Feed 2) sends HELLO messages with the FBIP1 field set to f1b (resp. f2b). 🌸

When a feed is about to be shut down, or when routing over the unidirectional link is about to be intentionally interrupted, it is recommended that feeds:

  1. stop sending HELLO messages containing a JOIN command.
  2. send a HELLO message containing a LEAVE command to inform receivers that the feed is no longer performing routing over the unidirectional link. 👋

7.3. DTCP on the receiver: receiving HELLO packets

Based on the reception of HELLO messages, receivers discover the presence of feeds, maintain a list of active feeds, and keep track of the tunnel end-points for those feeds. 🕵️‍♀️

For each active feed, and each IP protocol supported, at least the following information will be kept:

  • FUIP: feed unidirectional link IP address
  • FUMAC: MAC address corresponding to the above IP address
  • (FBIP1,...,FBIPn): list of tunnel end-points
  • tunnel type: tunnel type supported by this feed
  • Sequence: "Sequence" value from the last HELLO received from this feed
  • timer: used to timeout this entry

The FUMAC value for an active feed is needed for the operation of this protocol. However, the method of discovery of this value is not specified here. 🤷‍♀️

Initially, the list of active feeds is empty. 🫙

When a receiver is started, it MUST run a process which joins the "DTCP announcement" multicast group and listens to incoming packets on the HELLO_PORT port from the unidirectional link. 👂

Upon the reception of a HELLO message, the process checks the version number of the protocol. If it is different from HELLO_VERSION, the packet is discarded and the process waits for the next incoming packet. 🗑️

After successfully checking the version number, further action depends on the type of command:

  • JOIN:

The process verifies if the address FUIP already belongs to the list of active feeds.

If it does not, a new entry, for feed FUIP, is created and added to the list of active feeds. The number of feed bidirectional IP addresses to read is deduced from the Nb of FBID field. These tunnel end-points (FBIP1,...,FBIPn) can then be added to the new entry. The tunnel Type and Sequence values are also taken from the HELLO packet and recorded in the new entry. A timer set to HELLO_LEAVE is associated with this entry. ⏲️

If it does, the sequence number is compared to the sequence number contained in the previous HELLO packet sent by this feed. If they are equal, the timer associated with this entry is reset to HELLO_LEAVE. Otherwise, all the information corresponding to FUIP is set to the values from the HELLO packet.

Referring to Figure 1 in Section 3, both receivers (recv 1 and recv 2) have a list of active feeds containing two entries: Feed 1 with a FUIP of f1u and a list of tunnel end-points (f1b); and Feed 2 with a FUIP of f2u and a list of tunnel end-points (f2b). 🧚

  • LEAVE:

The process checks if there is an entry for FUIP in the list of active feeds. If there is, the timer is disabled and the entry is deleted from the list. The LEAVE message provides a means of quickly updating the list of active feeds. 🏃‍♀️💨

A timeout occurs for either of two reasons:

  1. A feed went down without sending a LEAVE message. As JOIN messages are no longer sent from this feed, a timeout occurs at HELLO_LEAVE after the last JOIN message.
  2. The unidirectional link is down. Thus, no more JOIN messages are received from any of the feeds, and they will each time out independently. The timeout of each entry depends on its individual HELLO_LEAVE value, and on when the last JOIN message was sent by that feed, before the unidirectional link went down. 📉
In either case, bidirectional connectivity can no longer be ensured between the receiver and the feed (FUIP): either the feed is no longer routing datagrams over the unidirectional link, or the link is down. Thus the associated entry is removed from the list of active feeds, whatever the cause. As a result, the list only contains operational tunnel end-points. 🚨

The HELLO protocol provides receivers with a list of feeds, and a list of usable tunnel end-points (FBIP1,..., FBIPn) for each feed. In the following Section, we describe how to integrate the HELLO protocol into the tunneling mechanism described in Sections 6.1 and 6.2. Tunnels, but make them fully automatic 🌈✨

7.4. Tunneling mechanism using the list of active feeds 💅✨

okay bestie, this section is about how the tunneling mechanism uses the list of active feeds to decide where datagrams that need tunneling go 🦋 It also points back to Section 6.1 to show how feed tunnel end-points get picked 🎀

the default feed gets chosen separately at each receiver 🌸 that choice is totally local policy, and this document doesn't cover that policy 🙈 as an example, the default feed may be the feed with the lowest round trip time to the receiver, just saying 💜

when a receiver sends a packet to a feed, it has to pick a tunnel end-point from inside the FBIP list 💖 the "preferred FBIP" is usually FBIP1 (Section 7.1). but for various reasons a receiver might decide to use a different FBIP, like FBIPi instead of FBIP1, as the tunnel end-point 👑 for example, the receiver might have better connectivity to FBIPi. and that call is made by the receiver administrator, not the protocol 🫶

now let's look at how the list of active feeds shows up when a receiver tunnels a link-layer packet 🧚 Section 6.1 listed these cases, depending on whether the MAC destination address of the packet is:

  1. the MAC address of a feed interface connected to the unidirectional link 🎯 this is TRUE if the address matches a FUMAC address in the list of active feeds. the packet gets tunneled to the preferred FBIP of the matching feed.
  2. the broadcast address of the unidirectional link, or a multicast address 📣 this is decided by the MAC address format rules, and the list of active feeds is not involved at all. the packet gets tunneled to the preferred FBIP of the default feed.
  3. an address that belongs to the unidirectional network but is not a feed address 🌈 this is TRUE if the address is neither broadcast nor multicast, and also not found in the list of active feeds. the packet gets tunneled to the preferred FBIP of the default feed.

no matter which case it is, the encapsulation type depends on the tunnel type required by whichever feed gets selected 🍓

7.5. Constant definitions 🔢

these are the constants, bestie, copy them exactly 💜

DTCP_VERSION is 1.

HELLO_INTERVAL is 5 seconds.

"DTCP announcement" multicast group is 224.0.0.36, assigned by IANA.

HELLO_PORT is 652.  It is a reserved system port assigned by IANA, no
   other traffic must be allowed.

HELLO_LEAVE is 3*Interval, as advertised in a HELLO packet, i.e., 15
   seconds if the default HELLO_INTERVAL was advertised.

8. Tunnel encapsulation format 🎀

okay so the tunneling mechanism works at the link layer and makes receivers and feeds look like they're bidirectionally connected 🦋 we assume the hardware on the unidirectional link supports broadcast and unicast MAC addressing 💖 that means a feed can send a packet to one particular receiver using a unicast MAC destination address, or to a whole set of receivers using a broadcast or multicast destination address ✨ then the hardware (or the driver) on the receiver can filter the incoming packets coming over the unidirectional link, without assuming anything about what kind of data is inside 🙈

a receiver should be able to send unicast and broadcast MAC packets through its tunnels too 🌸 the link-layer packets get encapsulated. after the feed decapsulates an incoming packet, it can do link-layer filtering just like the data came straight off the unidirectional link (see Figure 2) 🫶

Generic Routing Encapsulation (GRE) [RFC2784] fits what we need, because it defines a way to encapsulate arbitrary packets and lets you use IP as the delivery protocol 👑

the feed's local administrator decides which encapsulation the receivers have to use, and sets the tunnel type field in the HELLO message to match 💅 the value 47 (decimal) means GRE 🌈 other values can be used, but feeds and receivers have to agree on what they mean. that kind of usage isn't defined here 🧚

8.1. Generic Routing Encapsulation on the receiver 🍓

a GRE packet has a header whose type field says which protocol is inside (ARP, IP, IPX, etc.) 💜 check [RFC2784] for the full details of the encapsulation 🙈 for our case, only support for the MAC addressing scheme of the unidirectional link MUST be implemented 💖

so a tunneled packet looks like this: the delivery header is an IP header whose destination is the tunnel end-point (FBIP), followed by a GRE header that says what link-layer type the unidirectional link uses 🦋 Figure 4 shows the whole encapsulated packet ✨

        ----------------------------------------
        |           IP delivery header         |
        |        destination addr = FBIP       |
        |          IP proto = GRE (47)         |
        ----------------------------------------
        |             GRE Header               |
        |      type = MAC type of the UDL      |
        ----------------------------------------
        |            Payload packet            |
        |             MAC packet               |
        ----------------------------------------

              Figure 4: Encapsulated packet

9. Issues 🚨

9.1. Hardware address resolution 🔍

okay this one's a little spicy 🌶️ regardless of whether the link is unidirectional or bidirectional, if a feed sends a packet over a non-point-to-point network, it needs the data link address of the destination 💅 ARP [RFC826] is what Ethernet networks use for this 🎀

the link-layer mechanism pretends there's a bidirectional network even though the link is unidirectional 🦋 but there are asymmetric delays between every (feed, receiver) pair 😭 the backchannel between a receiver and a feed has delays that keep changing, since the packets go through the Internet 🌈 and a typical unidirectional link, like a GEO satellite link, has a delay of about 250 milliseconds 🛰️

because the round trip delays are so long, reactive address resolution methods like ARP [RFC826] might not work well 💀 for example, a feed might have to forward packets at high data rates to a receiver whose hardware address it doesn't know yet. the stream of packets gets passed to the link-layer driver of the feed's send-only interface 🫠 when the first packet shows up, the link layer realizes it doesn't have the hardware address of the next hop, so it sends an ARP request 📨 while the link layer waits for the response (at least 250 ms in the GEO satellite case), the feed buffers the IP packets ⏳ and if it runs out of space before the ARP response arrives, the IP packets get dropped 😭

this address resolution problem isn't solved in this document 💜 an ad-hoc fix is possible when the MAC address is configurable, which some satellite receiver cards allow ✨ then a simple transformation (maybe null) of the IP address can be used as the MAC address 🧚 in that case senders don't need to "resolve" an IP address to a MAC address. they just do the simple transformation 🫶

9.2. Routing protocols 🛣️

the link-layer tunneling mechanism hides the fact that feeds and receivers are connected by a unidirectional link from the network and the higher layers 👑 communication is bidirectional, just asymmetric in bandwidth and delay 🌸

to get unidirectional links into the Internet, feeds and receivers might have to run routing protocols in some topologies 💅 these protocols work fine, because the tunneling mechanism gives bidirectional connectivity between all feeds and receivers 🦋 so routing messages can be exchanged just like on any bidirectional network ✨

the tunneling mechanism allows any IP traffic, not only routing protocol messages, to be forwarded between receivers and feeds 🍓 receivers can route datagrams on the Internet using the most suitable feed or receiver as a next hop 🎯 administrators may want to set the metrics their routing protocols use, so the routing tables reflect the asymmetric characteristics of the link, and traffic goes over the right paths 🌈

feeds and receivers may implement multicast routing, so dynamic multicast routing can run over the unidirectional link 📣 but issues related to multicast routing (for example, protocol configuration) are not covered in this document 🙈

9.3. Scalability 📈

the DTCP protocol doesn't generate a lot of traffic, no matter how many nodes there are 💖 the problem with lots of nodes isn't really about this protocol. it's about bigger questions, like the maximum number of nodes that can be connected to any link 🫶 that's out of scope for this document 🌸

10. IANA Considerations 📋

IANA has reserved the address 224.0.0.36 for the "DTCP announcement" multicast address, as defined in Section 7 🎀

IANA has reserved the udp port 652 for the HELLO_PORT, as defined in Section 7 ✨

11. Security Considerations 🔐

okay this section is serious, bestie 😭 many unidirectional link technologies are known for how easily anyone can receive what's on the link 👀 so if sensitive or valuable information is being sent, link-layer security mechanisms are an appropriate measure 🛡️ for UDLR itself, the feed tunnel end-point addresses sent in HELLO messages may be considered sensitive 💅 in those cases, link-layer security mechanisms may be used 🫶

security in a network using the link-layer tunneling mechanism should be pretty similar to security in a normal IPv4 network 🌸 but since the tunneling mechanism requires tunnels, it opens the door for unauthorized access to the service 😱 in particular, ARP and IP spoofing are real threats, because nodes may not be authorized to tunnel packets 💀 this can be countered by authenticating all tunnels 🔒 the authentication mechanism isn't specified in this document. it can happen either in the delivery IP protocol (for example, AH[RFC2402]) or in an authentication protocol built into the tunneling mechanism 🧚

at a higher level, receivers may not be authorized to provide routing information, even though they're connected to the unidirectional link 🚫 to stop unauthorized receivers from giving fake routing information, routing protocols running on top of the link-layer tunneling mechanism MUST use authentication mechanisms when available 🔐👑

12. Acknowledgments 🫶

we'd like to thank Tim Gleeson (Cisco Japan) for his valuable editing and technical input during the finalization phase of the document 💖

we'd also like to thank Patrick Cipiere (UDcast) for his valuable input on the design of the encapsulation mechanism 🦋

we'd like to thank everyone who took part too: Akihiro Tosaka (IMD), Akira Kato (Tokyo Univ.), Hitoshi Asaeda (IBM/ITS), Hiromi Komatsu (JSAT), Hiroyuki Kusumoto (Keio Univ.), Kazuhiro Hara (Sony), Kenji Fujisawa (Sony), Mikiyo Nishida (Keio Univ.), Noritoshi Demizu (Sony CSL), Jun Murai (Keio Univ.), Jun Takei (JSAT) and Harri Hakulinen (Nokia) 🌈✨

Appendix A: Conformance and interoperability 🎀

okay so this appendix is about making sure everybody actually plays nice 🫶 the whole mechanism lets two nodes that are hooked up by a one-way link act like they have two-way connectivity. You get to pick from a bunch of key system parameters, so it works across loads of different equipment and setups. main character energy ✨

BUT (and this part matters, bestie) just saying "I do the mechanism!" is not enough for interoperability 🙈 a usage profile for a specific environment needs you to define a few things first 📝

  • the MAC format used
  • the tunneling mechanism to be used (GRE is recommended)
  • the "tunnel type" indication if GRE is not used

For example, a system might claim to implement "the link-layer tunneling mechanism for unidirectional links, using IEEE 802 LLC, and GRE encapsulation for the tunnels." 💅

Appendix B: DTCP announcement address transition plan 🦋

some older receivers are still listening for DTCP announcements on the 224.0.1.124 multicast address, aka the "old DTCP announcement" address 🕰️ to keep those legacy receivers happy, feeds SHOULD be configurable so they send every announcement to both the "DTCP announcement" address and the "old DTCP announcement" address at the same time. the default is just the "DTCP announcement" address, so that's the baseline 🌸

to push the transition along, the "old" feeds MUST be updated to send DTCP announcements the way this section describes 🚀 there aren't a ton of "old" feeds out there, so the update should be pretty easy, no cap ✨ and "new" receivers only support "new" feeds, i.e., they listen to DTCP announcements on the "DTCP announcement" address. simple and iconic 👑

References 📚

  • [RFC826] Plummer, D., "An Ethernet Address Resolution Protocol", STD 37, RFC 826, November 1982.
  • [RFC1112] Deering, S., "Host Extensions for IP Multicasting", STD 5, RFC 1112, August 1989
  • [RFC1700] Reynolds, J. and J. Postel, "ASSIGNED NUMBERS", STD 2, RFC 1700, October 1994.
  • [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.
  • [RFC2402] Kent, S. and R. Atkinson, "IP Authentication Header", RFC 2402, November 1998.
  • [RFC2453] Malkin, G., "RIP Version 2", STD 56, RFC 2453, November 1998.
  • [RFC2780] Bradner, S. and V. Paxson, "IANA Allocation Guidelines For Values In the Internet Protocol and Related Headers", BCP 37, RFC 2780, March 2000.
  • [RFC2784] Farinacci, D., Hanks, S., Meyer, D. and P. Traina, "Generic Routing Encapsulation (GRE)", RFC 2784, March 2000.

Authors' Addresses 💌

okay the people who wrote this, and how to reach them 🌈

Emmanuel Duros
UDcast
1681, route des Dolines
Les Taissounieres - BP 355
06906 Sophia-Antipolis Cedex
France

Phone : +33 4 93 00 16 60
Fax   : +33 4 93 00 16 61
EMail : Emmanuel.Duros@UDcast.com
Walid Dabbous
INRIA Sophia Antipolis
2004, Route des Lucioles BP 93
06902 Sophia Antipolis
France

Phone : +33 4 92 38 77 18
Fax   : +33 4 92 38 79 78
EMail : Walid.Dabbous@inria.fr
Hidetaka Izumiyama
JSAT Corporation
Toranomon 17 Mori Bldg.5F
1-26-5 Toranomon, Minato-ku
Tokyo 105
Japan

Phone : +81-3-5511-7568
Fax   : +81-3-5512-7181
EMail : izu@jsat.net
Noboru Fujii
Sony Corporation
2-10-14 Osaki, Shinagawa-ku
Tokyo 141
Japan

Phone : +81-3-3495-3092
Fax   : +81-3-3495-3527
EMail : fujii@dct.sony.co.jp
Yongguang Zhang
HRL
RL-96, 3011 Malibu Canyon Road
Malibu, CA 90265,
USA

Phone : 310-317-5147
Fax   : 310-317-5695
EMail : ygz@hrl.com

the legal bit, read it bestie 💖 it's kept word for word on purpose so nothing gets lost 🔒

Copyright (C) The Internet Society (2001). All Rights Reserved.

This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English.

The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns.

This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Acknowledgement 🫶

funding for the RFC Editor function is currently provided by the Internet Society 💜

📦 where this page comes from