| PCP working group | D. Wing, Ed. |
| Internet-Draft | Cisco |
| Intended status: Standards Track | S. Cheshire |
| Expires: October 28, 2011 | Apple |
| M. Boucadair | |
| France Telecom | |
| R. Penno | |
| Juniper Networks | |
| April 26, 2011 |
Port Control Protocol (PCP)
draft-ietf-pcp-base-09
Port Control Protocol allows a host to control how incoming IPv6 or IPv4 packets are translated and forwarded by a network address translator (NAT) or simple firewall to an IPv6 or IPv4 host, and also allows a host to optimize its NAT keepalive messages.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on October 28, 2011.
Copyright (c) 2011 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
The Port Control Protocol (PCP) provides a mechanism to control how incoming packets are forwarded by upstream devices such as NAT64, NAT44, and firewall devices, and a mechanism to reduce application keepalive traffic. PCP is primarily designed to be implemented in the context of both Carrier-Grade NATs (CGN) and small NATs (e.g., residential NATs). PCP allows hosts to operate server for a long time (e.g., a webcam) or a short time (e.g., while playing a game or on a phone call) when behind a NAT device, including when behind a CGN operated by their Internet service provider.
PCP allows applications to create mappings from an external IP address and port to an internal IP address and port. These mappings are required for successful inbound communications destined to machines located behind a NAT or a firewall.
After creating a mapping for incoming connections, it is necessary to inform remote computers about the IP address and port for the incoming connection. This is usually done in an application-specific manner. For example, a computer game would use a rendezvous server specific to that game (or specific to that game developer), and a SIP phone would use a SIP proxy. PCP does not provide this rendezvous function. The rendezvous function will support IPv4, IPv6, or both. Depending on that support and the application's support of IPv4 or IPv6, the PCP client will need an IPv4 mapping, an IPv6 mapping, or both.
Many NAT-friendly applications send frequent application-level messages to ensure their session will not be timed out by a NAT. These are commonly called "NAT keepalive" messages, even though they are not sent to the NAT itself (rather, they are sent 'through' the NAT). These applications can reduce the frequency of those NAT keepalive messages by using PCP to learn (and influence) the NAT mapping lifetime. This helps reduce bandwidth on the subscriber's access network, traffic to the server, and battery consumption on mobile devices.
Many NATs and firewalls have included application layer gateways (ALGs) to create mappings for applications that establish additional streams or accept incoming connections. ALGs incorporated into NATs additionally modify the application payload. Industry experience has shown that these ALGs are detrimental to protocol evolution. PCP allows an application to create its own mappings in NATs and firewalls, reducing the incentive to deploy ALGs in NATs and firewalls.
PCP can be used in various deployment scenarios, including:
The PCP OpCodes defined in this document are designed to support transport-layer protocols that use a 16-bit port number (e.g., TCP, UDP, SCTP, DCCP). Protocols that do not use a port number (e.g., IPsec ESP), and the ability to use PCP to forward all traffic to a single default host (often nicknamed a "DMZ"), are beyond the scope of this document.
The PCP machinery assumes a single-homed host model. That is, for a given IP version, only one default route exists to reach the Internet. This is important because after a PCP mapping is created and an inbound packet (e.g., TCP SYN) arrives at the host, the outbound response (e.g., TCP SYNACK) has to go through the same path so it is seen by the firewall or rewritten by the NAT. This restriction exists because otherwise there would need to be one PCP server for each egress, because the host could not reliably determine which egress path packets would take, so the client would need to be able to reliably make the same internal/external mapping in every NAT gateway, which in general is not possible (because the other NATs would likely have the necessary port mapped to another host).
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in "Key words for use in RFCs to Indicate Requirement Levels" [RFC2119].
The PCP server receives PCP requests. The PCP server might be integrated within the NAT or firewall device (as shown in Figure 1) which is expected to be a common deployment.
+-----------------+
+------------+ | NAT or firewall |
| PCP client |-<network>-+ with +---<Internet>
+------------+ | PCP server |
+-----------------+
It is also possible to operate the PCP server in a separate device from the NAT, so long as such operation is indistinguishable from the PCP client's perspective.
All PCP messages contain a request (or response) header containing an opcode, any relevant opcode-specific information, and zero or more options. The packet layout for the common header, and operation of the PCP client and PCP server are described in the following sections. The information in this section applies to all OpCodes. Behavior of the OpCodes defined in this document is described in Section 8 and Section 9.
All requests have the following format:
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Version = 1 |R| OpCode | Reserved (16 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Requested Lifetime | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | PCP Client's IP address (always 128 bits) | | | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : (optional) opcode-specific information : : : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : (optional) PCP Options : : : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
These fields are described below:
All responses have the following format:
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Version = 1 |R| OpCode | Reserved | Result Code | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Lifetime | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Epoch | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | PCP Client's IP address (always 128 bits) | | | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : (optional) OpCode-specific response data : : : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : (optional) Options : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
These fields are described below:
A PCP OpCode can be extended with an Option. Options can be used in requests and responses. The decision about whether to include a given piece of information in the base opcode format or in an option is an engineering trade-off between packet size and code complexity. For information that is usually (or always) required, placing it in the fixed opcode data results in simpler code to generate and parse the packet, because the information is a fixed location in the opcode data, but wastes space in the packet in the event that that field is all-zeroes because the information is not needed or not relevant. For information that is required less often, placing it in an option results in slightly more complicated code to generate and parse packets containing that option, but saves space in the packet when that information is not needed. Placing information in an option also means that an implementation that never uses that information doesn't even need to implement code to generate and parse it. For example, a client that never requests mappings on behalf of some other device doesn't need to implement code to generate the THIRD_PARTY option, and a PCP server that doesn't implement the necessary security measures to create third-party mappings safely doesn't need to implement code to parse the THIRD_PARTY option.
Options use the following Type-Length-Value format:
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Option Code | Reserved | Option-Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : (optional) data : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
The description of the fields is as follows:
A given Option MAY be included in a request containing a specific OpCode. The handling of an Option by the PCP client and PCP server MUST be specified in an appropriate document and MUST include whether the PCP Option can appear (one or more times) in a request and/or response, and indicate the contents of the Option in the request and in the response. If several Options are included in a PCP request or response, they MAY be encoded in any order by the PCP client and are processed in the order received.
If, while processing an option, an error is encountered that causes a PCP error response to be generated, the PCP request MUST cause no state change in the PCP server or the PCP-controlled device (i.e., it rolls back any changes it might have made while processing the request). The response MUST encode the Options in the same order, but may omit some PCP Options in the response, as is necessary to indicate the PCP server does not understand that Option or that Option is not permitted to be included in responses by the definition of the Option itself. Additional Options included in the response (if any) MUST be included at the end. A certain Option MAY appear more than once in a request or in a response, if permitted by the definition of the Option itself. If the Option's definition allows the Option to appear only once but it appears more than once in a request, the PCP server MUST respond with the MALFORMED_OPTION result code; if this occurs in a response, the PCP client processes the first occurrence and ignores the other occurrences as if they were not present.
If the "O" bit (high bit) in the OpCode is clear,
If the "O" bit is set, the PCP server MAY process or ignore this Option, entirely at its discretion.
Option definitions MUST include the information below:
The following result codes may be returned as a result of any OpCode received by the PCP server. The only success result code is 0, other values indicate an error. If a PCP server has encountered multiple errors during processing of a request, it SHOULD use the most specific error message.
Additional result codes, specific to the OpCodes and Options defined in this document, are listed in Section 8.2 and Section 10.1.
PCP messages MUST be sent over UDP [RFC0768]. Every PCP request generates a response, so PCP does not need to run over a reliable transport protocol.
PCP is idempotent, so if the PCP client sends the same request multiple times and the PCP server processes those requests, the same result occurs. The order of operation is that a PCP client generates and sends a request to the PCP server, which processes the request and generates a response back to the PCP client.
This section details operation specific to a PCP client, for any OpCode. Procedures specific to the MAP OpCodes are described in Section 8, and procedures specific to the PEER OpCodes are described in Section 9.
Prior to sending its first PCP message, the PCP client determines which servers to use. The PCP client performs the following steps to determine its PCP server(s):
With that list of PCP servers, the PCP client formulates its PCP request. The PCP request contains a PCP common header, PCP OpCode and payload, and (possibly) Options. As with all UDP or TCP clients on any operating system, when several PCP clients are embedded in the same host, each uses a distinct source port number to disambiguate their requests and replies. The PCP client's source port SHOULD be randomly generated [RFC6056].
When attempting to contact a PCP server, the PCP client initializes a timer to 2 seconds. The PCP client sends a PCP message the first server in its list of PCP servers. If no response is received before the timer expires, the timer is doubled (to 4 seconds) and the request is re-transmitted. If no response is received before the timer expires, the timer is doubled again (to 8 seconds) and the request is re-transmitted. This procedure is repeated in parallel or in series to each PCP server in the list, on each interface, until a response is received from a PCP server. If the requests are sent in parallel and responses from multiple PCP servers are received, only the PCP server closest to the top of the list, on that interface, is used for subsequent requests; PCP requests which received a positive response and create state (e.g., MAP) SHOULD have their state cleared (e.g., lifetime set to 0).
Once a PCP client has successfully received a response from a PCP server on that interface, it sends subsequent PCP requests to that same server, with a retransmission timer of 2 seconds. If, after 2 seconds, a response is not received from that PCP server, the same back-off algorithm described above is performed.
Upon receiving a response (success or error), the PCP client does not change to a different PCP server. That is, it does not "shop around" trying to find a PCP server to service its (same) request.
This section details operation specific to a PCP server. Processing SHOULD be performed in the order of the following paragraphs.
A PCP server processes incoming requests on the PCP port from clients or an operator-configured interface (e.g., from the ISP's network operations center). The PCP server MUST drop (ignore) requests that arrive from elsewhere (e.g., the Internet).
Upon receiving a message, the PCP server parses and validates it. A valid request contains a valid PCP common header, one valid PCP Opcode, and zero or more Options (which the server might or might not comprehend). If an error is encountered during processing, the server generates an error response which is sent back to the PCP client. Processing an OpCode and the Options are specific to each OpCode.
If the received message is shorter than 4 octets or has the R bit set the message is simply dropped. If the length of the request exceeds 1024 octets or is not a multiple of 4 octets, it is invalid. Invalid requests are handled by copying up to 1024 octets of the request into the response, setting the result code to MALFORMED_REQUEST, and zero-padding the response to a multiple of 4 octets if necessary. If the version number is not supported, a response is generated with the UNSUPP_VERSION result code and the other steps detailed in Section 6.6. If the OpCode is not supported, a response is generated with the UNSUPP_OPCODE result code.
If the source IP address of the received packet does not match the contents of the PCP Client IP Address field, a response is generated with the ADDRESS_MISMATCH result code. This is done to detect and prevent accidental use of PCP where a non-PCP-aware NAT or NAPT exists between the PCP client and PCP server.
Error responses have the same packet layout as success responses, with fields from the request copied into the response, and fields assigned by the PCP server are set as indicated in Figure 3
The PCP client receives the response and verifies the source IP address and port belong to the PCP server of an outstanding PCP request. It validates the OpCode matches an outstanding PCP request. Responses shorter than 12 octets, longer than 1024 octets, or not a multiple of 4 octets are invalid and ignored, likely causing the request to be re-transmitted. The response is further matched by comparing fields in the response OpCode-specific data to fields in the request OpCode-specific data, as described by the processing for that OpCode. After these matches are successful, the PCP client checks the Epoch field to determine if it needs to restore its state to the PCP server (see Section 6.5).
If the result code is 0, the PCP client knows the request was successful.
If the result code is not 0, the request failed. If the result code is UNSUPP_VERSION, processing continues as described in Section 6.6. If the result code is SERVER_OVERLOADED, clients SHOULD NOT send *any* further requests to that PCP server for the indicated error lifetime. For other error result codes, The PCP client SHOULD NOT resend the same request for the indicated error lifetime. If a PCP server indicates an error lifetime in excess of 30 minutes, A PCP client MAY choose to set its retry timer to 30 minutes.
If the PCP client has discovered a new PCP server (e.g., connected to a new network), the PCP client MAY immediately begin communicating with this PCP server, without regard to hold times from communicating with a previous PCP server.
Hosts which desire a PCP mapping might be multi-interfaced (i.e., own several logical/physical interfaces). Indeed, a host can be configured with several IPv4 addresses (e.g., WiFi and Ethernet) or dual-stacked. These IP addresses may have distinct reachability scopes (e.g., if IPv6 they might have global reachability scope as for Global Unicast Address (GUA, [RFC3587]) or limited scope as for Unique Local Address (ULA) [RFC4193]).
IPv6 addresses with global reachability (e.g., GUA) SHOULD be used as the source address when generating a PCP request. IPv6 addresses without global reachability (e.g., ULA [RFC4193]), SHOULD NOT be used as the source interface when generating a PCP request. If IPv6 privacy addresses [RFC4941] are used for PCP mappings, a new PCP request will need to be issued whenever the IPv6 privacy address is changed. This PCP request SHOULD be sent from the IPv6 privacy address itself. It is RECOMMENDED that mappings to the previous privacy address be deleted.
Due to the ubiquity of IPv4 NAT, IPv4 addresses with limited scope (e.g., private addresses [RFC1918]) MAY be used as the source interface when generating a PCP request.
As mentioned in Section 2.3, only single-homed CP routers are in scope. Therefore, there is no viable scenario where a host located behind a CP router is assigned with two Global Unicast Addresses belonging to different global IPv6 prefixes.
Every PCP response sent by the PCP server includes an Epoch field. This field increments by 1 every second, and is used by the PCP client to determine if PCP state needs to be restored. If the PCP server resets or loses the state of its explicit dynamic Mappings (that is, those mappings created by PCP MAP requests), due to reboot, power failure, or any other reason, it MUST reset its Epoch time to 0. Similarly, if the public IP address(es) of the NAT (controlled by the PCP server) changes, the Epoch MUST be reset to 0. A PCP server MAY maintain one Epoch value for all PCP clients, or MAY maintain distinct Epoch values for each PCP client; this choice is implementation-dependent.
Whenever a client receives a PCP response, the client computes its own conservative estimate of the expected Epoch value by taking the Epoch value in the last packet it received from the gateway and adding 7/8 (87.5%) of the time elapsed since that packet was received. If the Epoch value in the newly received packet is less than the client's conservative estimate by more than one second, then the client concludes that the PCP server lost state, and the client MUST immediately renew all its active port mapping leases as described in Section 11.2.1.
When a client notices that the PCP server reduced its Epoch value, the PCP clients will send PCP requests to refresh their mappings. The PCP server needs to be scaled appropriately to accommodate this traffic. Because PCP lacks a mechanism to simultaneously inform all PCP clients of the Epoch value, the PCP clients will only flood the PCP server simultaneously when a power outage and restoration event causes state loss in both the PCP clients and PCP server.
A PCP client sends its requests using PCP version number 1. Should later updates to this document specify different message formats with a version number greater than 1 it is expected that PCP servers will still support version 1 in addition to the newer version(s). However, in the event that a server returns a response with error code UNSUPP_VERSION, the client MAY log an error message to inform the user that it is too old to work with this server.
When sending a response containing the UNSUPP_VERSION result code, the PCP message MUST be 12 octets long.
If future PCP versions greater than 1 are specified, version negotiation is expected to proceed as follows:
The following option can appear in certain PCP responses, without regard to the OpCode.
If the PCP server cannot process a mandatory-to-process option, for whatever reason, it includes the UNPROCESSED Option in the response, shown in Figure 5. This helps with debugging interactions between the PCP client and PCP server. This option MUST NOT appear more than once in a PCP response. The unprocessed options are listed once, and the option data is zero-filled to the necessary 32 bit boundary. If a certain Option appeared more than once in the PCP request, that Option value can appear once or as many times as it occurred in the request. The order of the Options in the PCP request has no relationship with the order of the Option values in this UNPROCESSED Option. This Option MUST NOT appear in a response unless the associated request contained at least one mandatory-to-process Option.
The UNPROCESSED option is formatted as follows, showing an example of two option codes that were unprocessed:
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | option-code-1 | option-code-2 | padding | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Padding: 0, 1, 2, or 3 octets. If the number of option-codes is not a multiple of 4, padding is used to make it 32-bit aligned. The padding MUST be on on sending, and MUST be ignored by the receiver.
There are four uses for the MAP and PEER OpCodes defined in this document: a host operating a server and wanting an incoming connection; a host operating a client and wanting to optimize the application keepalive traffic or restore lost state in its NAT; and a host operating a client and server on the same port. These are discussed in the following sections.
When operating a server (Section 7.1 and Section 7.4) the PCP client knows if it wants an IPv4 listener, IPv6 listener, or both on the Internet. The PCP client also knows if it has an IPv4 address on itself or an IPv6 interface on itself. It takes the union of this knowledge to decide to send a one or two MAP requests for each of its interfaces. Applications that embed IP addresses in payloads (e.g., FTP, SIP) will find it beneficial to avoid address family translation, if possible.
It is REQUIRED that the PCP-controlled device assign the same external IP address to PCP-created explicit dynamic mappings and to implicit dynamic mappings. It is RECOMMENDED that static mappings (e.g., those created by a command-line interface on the PCP server or PCP-controlled device) also be assigned to the same IP address. Once all internal addresses belonging to a given subscriber have no implicit dynamic mappings and have no explicit dynamic mappings in the PCP-controlled device, a subsequent PCP request for that internal address MAY be assigned to a different external IP address. Generally, this re-assignment would occur when a CGN device is load balancing newly-seen hosts to its public IPv4 address pool.
A host operating a server (e.g., a web server) listens for traffic on a port, but the server never initiates traffic from that port. For this to work across a NAT or a firewall, the host needs to (a) create a mapping from a public IP address and port to itself as described in Section 8 and (b) publish that public IP address and port via some sort of rendezvous server (e.g., DNS, a SIP message, a proprietary protocol). Publishing the public IP address and port is out of scope of this specification. To accomplish (a), the host follows the procedures described in this section.
As normal, the application needs to begin listening on a port. Then, the application constructs a PCP message with the appropriate MAP OpCode depending on if it is listening on an IPv4 or IPv6 address and if it wants a public IPv4 or IPv6 address.
The following pseudo-code shows how PCP can be reliably used to operate a server:
/* start listening on the local server port */
int s = socket(...);
bind(s, ...);
listen(s, ...);
getsockname(s, &internal_sockaddr, ...);
external_sockaddr = 0;
while (1)
{
/* Note: the "time_to_send_pcp_request" check below includes:
* 1. Sending the first request
* 2. Retransmitting requests due to packet loss
* 3. Resending a request due to impending lease expiration
* The PCP packet sent is identical in all cases, apart from the
* Suggested External Address and Port which may change over time
*/
if (time_to_send_pcp_request)
pcp_send_map_request(internal_sockaddr.sin_port,
internal_sockaddr.sin_addr,
&external_sockaddr, /* will be zero the first time */
requested_lifetime, &assigned_lifetime);
if (pcp_response_received)
update_rendezvous_server("Client Ident", external_sockaddr);
if (received_incoming_connection_or_packet)
process_it(s);
if (other_work_to_do)
do_it();
/* ... */
block_until_we_need_to_do_something_else();
}
A host operating a client (e.g., XMPP client, SIP client) sends from a port but never accepts incoming connections on this port. It wants to ensure the flow to its server is not terminated (due to inactivity) by an on-path NAT or firewall. To accomplish this, the application uses the procedure described in this section.
Middleboxes such as NATs or firewalls need to see occasional traffic or will terminate their session state, causing application failures. To avoid this, many applications routinely generate keepalive traffic for the primary (or sole) purpose of maintaining state with such middleboxes. Applications can reduce such application keepalive traffic by using PCP.
To use PCP for this function, the application first connects to its server, as normal. Afterwards, it issues a PCP request with the PEER4 or PEER6 OpCode as described in Section 9. The PEER4 OpCode is used if the host is using IPv4 for its communication to its peer; PEER6 if using IPv6. The same 5-tuple as used for the connection to the server is placed into the PEER4 or PEER6 payload.
The following pseudo-code shows how PCP can be reliably used with a dynamic socket, for the purposes of reducing application keepalive messages:
int s = socket(...);
connect(s, &remote_peer, ...);
getsockname(s, &internal_sockaddr, ...);
external_sockaddr = 0;
while (1)
{
/* Note: the "time_to_send_pcp_request" check below includes:
* 1. Sending the first request
* 2. Retransmitting requests due to packet loss
* 3. Resending a request due to impending lease expiration
* The PCP packet sent is identical in all cases, apart from the
* Suggested External Address and Port which may change over time
*/
if (time_to_send_pcp_request)
pcp_send_peer_request(internal_sockaddr.sin_port,
internal_sockaddr.sin_addr,
&external_sockaddr, /* will be zero the first time */
remote_peer, requested_lifetime, &assigned_lifetime);
if (data_to_send)
send_it(s);
if (other_work_to_do)
do_it();
/* ... */
block_until_we_need_to_do_something_else();
}
After a NAPT loses state (e.g., because of a crash or power failure), it is useful for clients to re-establish TCP mappings on the NAPT. This allows servers on the Internet to see traffic from the same IP address and port, so that sessions can be resumed exactly where they were left off. This can be useful for long-lived connections (e.g., instant messaging) or for connections transferring a lot of data (e.g., FTP). This can be accomplished by establishing a TCP connection normally and then sending a PEER request/response and remember the External Address and External Port. Later, when the NAPT has lost state, the client can send a PEER request with the Suggested External Port and Suggested External Address remembered from the previous session, which will create a mapping in the NAPT that functions exactly as an implicit dynamic mapping. The client then resumes sending TCP data to the server.
A host operating a client and server on the same port (e.g., Symmetric RTP [RFC4961] or SIP Symmetric Response Routing (rport) [RFC3581]) first establishes a local listener, (usually) sends the local and public IP addresses and ports to a rendezvous service (which is out of scope of this document), and initiates an outbound connection from that same source address and same port. To accomplish this, the application uses the procedure described in this section.
An application that is using the same port for outgoing connections as well as incoming connections MUST first signal its operation of a server using the PCP MAP OpCode, as described in Section 8, and receive a positive PCP response before it sends any packets from that port.
The following pseudo-code shows how PCP can be used to operate a symmetric client and server:
/* start listening on the local server port */
int s = socket(...);
bind(s, ...);
listen(s, ...);
getsockname(s, &internal_sockaddr, ...);
external_sockaddr = 0;
while (1)
{
/* Note: the "time_to_send_pcp_request" check below includes:
* 1. Sending the first request
* 2. Retransmitting requests due to packet loss
* 3. Resending a request due to impending lease expiration
* The PCP packet sent is identical in all cases, apart from the
* Suggested External Address and Port which may change over time
*/
if (time_to_send_pcp_request)
pcp_send_map_request(internal_sockaddr.sin_port,
internal_sockaddr.sin_addr,
&external_sockaddr, /* will be zero the first time */
requested_lifetime, &assigned_lifetime);
if (pcp_response_received)
update_rendezvous_server("Client Ident", external_sockaddr);
if (received_incoming_connection_or_packet)
process_it(s);
if (need_to_make_outgoing_connection)
make_outgoing_connection(s, ...);
if (data_to_send)
send_it(s);
if (other_work_to_do)
do_it();
/* ... */
block_until_we_need_to_do_something_else();
}
This section defines two OpCodes which control forwarding from a NAT (or firewall) to an internal host. They are:
The internal address is the source IP address of the PCP request message itself, unless the THIRD_PARTY option is used.
Note that all mappings created by PCP MAP requests are, by definition, Endpoint Independent Mappings (even on a NAT that usually creates Endpoint Dependent Mappings for outgoing connections) since the purpose of a MAP mapping is to receive inbound traffic from any remote endpoint, not from only one specific remote endpoint.
Note also that all NAT mappings (created by PCP or otherwise) are by necessity bidirectional and symmetrical. For any packet going in one direction (in or out) that is translated by the NAT, a reply going in the opposite direction needs to have the corresponding opposite translation done so that the reply arrives at the right endpoint. This means that if a client creates a MAP mapping, and then later sends an outgoing packet using the mapping's internal source port, the NAT should translate that packet's Internal Address and Port to the mapping's External Address and Port, so that replies addressed to the External Address and Port are correctly translated to the mapping's Internal Address and Port.
The operation of the MAP OpCodes is described in this section.
The two MAP OpCodes (MAP4, MAP6) share a similar packet layout for both requests and responses. Because of this similarity, they are shown together. For both of the MAP OpCodes, if the assigned external IP address and assigned external port match the request's Internal IP address and port, the functionality is purely a firewall; otherwise it pertains to a network address translator which might also perform firewall-like functions.
The following diagram shows the OpCode-specific information format in a request for the MAP4 and MAP6 OpCodes.
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Protocol | Reserved (24 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Internal Port | Suggested External Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : Suggested External IP Address (32 or 128, depending on OpCode): : : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
These fields are described below:
The following diagram shows the OpCode-specific information format in a response packet for the MAP4 and MAP6 OpCodes:
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Protocol | Reserved (24 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Internal Port | Assigned External Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : Assigned External IP Address (32 or 128, depending on OpCode) : : : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
These fields are described below:
In addition to the general PCP result codes (Section 5.4), the following additional result codes may be returned as a result of the four MAP OpCodes received by the PCP server. These errors are considered 'long lifetime' or 'short lifetime', which provides guidance to PCP server developers for the value of the Lifetime field for these errors. It is RECOMMENDED that short lifetime errors use 30 second lifetime and long lifetime errors use 30 minute lifetime.
Additional result codes may be returned if the THIRD_PARTY option is used, see Section 10.1.
This section describes the operation of a PCP client when sending requests with OpCodes MAP4 and MAP6.
The request MAY contain values in the suggested-external-ip-address and suggested-external-port fields. This allows the PCP client to attempt to rebuild the PCP server's state, so that the PCP client could avoid having to change information maintained at the rendezvous server. Of course, due to other activity on the network (e.g., by other users or network renumbering), the PCP server may not be able to fulfill the request.
An existing mapping can have its lifetime extended by the PCP client. To do this, the PCP client sends a new MAP request indicating the internal port. The PCP MAP request SHOULD also include the currently allocated external IP address and port as the suggested external IP address and port, so that if the NAT gateway has lost state it can recreate the lost mapping with the same parameters.
The PCP client SHOULD renew the mapping before its expiry time, otherwise it will be removed by the PCP server (see Section 8.6). In order to prevent excessive PCP chatter, it is RECOMMENDED to send a single renewal request packet when a mapping is halfway to expiration time, then, if no SUCCESS result is received, another single renewal request 3/4 of the way to expiration time, and then another at 7/8 of the way to expiration time, and so on, subject to the constraint that renewal requests MUST NOT be sent less than four seconds apart (a PCP client MUST NOT send ever-closer-together requests in the last few seconds before a mapping expires).
This section describes the operation of a PCP server when processing a request with the OpCodes MAP4 or MAP6. Processing SHOULD be performed in the order of the following paragraphs.
If the server is overloaded by requests (from a particular client or from all clients), it MAY simply discard requests, as the requests will be retried by PCP clients, or MAY generate the SERVER_OVERLOADED error response, or both.
If the request contains internal-port=0 and the lifetime is non-zero, the server MUST generate a MALFORMED_REQUEST error.
If the requested lifetime is not zero, it indicates a request to create a mapping or extend the lifetime of an existing mapping.
Processing of the lifetime is described in Section 8.6.
If the PCP-controlled device is stateless (that is, it does not establish any per-flow state, and simply rewrites the address and/or port in a purely algorithmic fashion), the PCP server simply returns an answer indicating the external IP address and port yielded by this stateless algorithmic translation. This allows the PCP client to learn its external IP address and port as seen by remote peers. Examples of stateless translators include stateless NAT64 and 1:1 NAT44, both of which modify addresses but not port numbers.
If an Option with value less than 128 exists (i.e., mandatory to process) but that option does not make sense (e.g., the PREFER_FAILURE option is included in a request with lifetime=0), the request is invalid and generates a MALFORMED_OPTION error.
If the PCP server can allocate the suggested external port, and the request did not contain the PREFER_FAILURE Option, it SHOULD do so. This is beneficial for re-establishing state lost when the PCP server loses its state (e.g., due to a reboot). If the PCP server cannot allocate the suggested external port but can allocate some other port and the request did not contain the PREFER_FAILURE Option, the PCP server MUST do so and return the allocated port in the response. Cases where a NAT gateway cannot allocate the suggested external port include:
By default, a PCP-controlled device MUST NOT create mappings for a protocol not indicated in the request. For example, if the request was for a TCP mapping, a UDP mapping MUST NOT be created.
If the THIRD_PARTY option is not present in the request, the source IP address of the PCP packet is used when creating the mapping. If the THIRD_PARTY option is present, the PCP server validates that the client is authorized to make mappings on behalf of the indicated internal IP address. This validation depends on the PCP deployment scenario; see Section 13.3 for the validation procedure. If the internal IP address in the PCP request is not authorized to make mappings on behalf of the indicated internal IP address, an error response MUST be generated with result code NOT_AUTHORIZED.
Mappings typically consume state on the PCP-controlled device, and it is RECOMMENDED that a per-subscriber or per-host limit be enforced by the PCP server to prevent exhausting the mapping state. If this limit is exceeded, the result code USER_EX_QUOTA is returned.
If all of the proceeding operations were successful (did not generate an error response), then the requested mappings are created or refreshed as described in the request and a SUCCESS response is built. This SUCCESS response contains the same OpCode as the request, but with the "R" bit set.
This section describes the operation of the PCP client when it receives a PCP response for the OpCodes MAP4 or MAP6.
After performing common PCP response processing, the response is further matched with an outstanding request by comparing the protocol, internal IP address, internal port. On error responses, the assigned external address and assigned external port can also be used to match the responses (which is useful if several requests with the PREFER_FAILURE option are outstanding). Other fields are not compared, because the PCP server sets those fields.
If a successful response, the PCP client can use the external IP address and port(s) as desired. Typically the PCP client will communicate the external IP address and port(s) to another host on the Internet using an application-specific rendezvous mechanism such as DNS SRV records.
On an error response, clients SHOULD NOT repeat the same request to the same PCP server within the lifetime returned in the response.
The PCP client requests a certain lifetime, and the PCP server responds with the assigned lifetime. The PCP server MAY grant a lifetime smaller or larger than the requested lifetime. The PCP server SHOULD be configurable for permitted minimum and maximum lifetime, and the RECOMMENDED values are 120 seconds for the minimum value and 24 hours for the maximum. It is RECOMMENDED that the server be configured to restrict lifetimes to less than 24 hours, because they will consume ports even if the internal host is no longer interested in receiving the traffic or no longer connected to the network. These recommendations are not strict, and deployments should evaluate the tradeoffs to determine their own minimum and maximum lifetime values.
Once a PCP server has responded positively to a mapping request for a certain lifetime, the port forwarding is active for the duration of the lifetime unless the lifetime is reduced by the PCP client (to a shorter lifetime or to zero) or until the PCP server loses its state (e.g., crashes). Mappings created by PCP MAP requests are not special or different to mappings created other ways. In particular, it is implementation-dependent if outgoing traffic extends the lifetime of such mappings. PCP clients MUST NOT depend on this behavior to keep mappings active, and MUST explicitly renew their mappings as required by the Lifetime field in PCP response messages.
If the requested lifetime is zero (lifetime==0) then:
The suggested external address and port fields are ignored in requests where the requested lifetime is 0.
If the PCP client attempts to delete a single static mapping (i.e., a mapping created outside of PCP itself), the error NOT_AUTHORIZED is returned. If the PCP client attempts to delete an implicit dynamic mapping (e.g., created by a TCP SYN), the PCP server deletes the mapping and responds with the SUCCESS result code. If the PCP client attempts to delete a mapping that does not exist, the SUCCESS result code is returned (this is necessary for PCP to be idempotent). If the PCP MAP request was for port=0 (indicating 'all ports'), the PCP server deletes all of the explicit dynamic mappings it can (but not any implicit mappings), and returns a SUCCESS response. If the deletion request was properly formatted and successfully processed, a SUCCESS response is generated with lifetime of 0 and the server copies the protocol and internal port number from the request into the response. An explicit dynamic mapping MUST NOT have its lifetime reduced by transport protocol messages (e.g., TCP RST, TCP FIN).
An application that forgets its PCP-assigned mappings (e.g., the application or OS crashes) will request new PCP mappings. This may consume port mappings, if the application binds to a different Internal Port every time it runs. The application will also likely initiate new implicit dynamic mappings (e.g., TCP connections) without using PCP, which will also consume port mappings. If there is a port mapping quota for the internal host, frequent restarts such as this may exhaust the quota. PCP provides some protections against such port consumption: When a PCP client first acquires a new IP address (e.g., reboots or joins a new network), it SHOULD remove mappings that may already be instantiated for that new Internal Address. To do this, the PCP client sends a MAP request with protocol, internal port, and lifetime set to 0. Some port mapping APIs (e.g., the "DNSServiceNATPortMappingCreate" API provided by Apple's Bonjour on Mac OS X, iOS, Windows, Linux) automatically monitor for process exit (including application crashes) and automatically send port mapping deletion requests if the process that requested them goes away without explicitly relinquishing them.
To reduce unwanted traffic and data corruption, UDP and TCP ports should not be immediately re-used for an interval (TIME_WAIT interval as discussed in [RFC0793]). However, the PCP server MUST allow the same subscriber and same internal address to re-acquire the same port during that interval.
As a side-effect of creating a mapping, ICMP messages associated with the mapping MUST be forwarded (and also translated, if appropriate) for the duration of the mapping's lifetime. This is done to ensure that ICMP messages can still be used by hosts, without application programmers or PCP client implementations needing to signal PCP separately to create ICMP mappings for those flows.
The following list summarizes the sentinel values when deleting a mapping using lifetime=0:
The customer premises router might obtain a new IP address. This can occur because of a variety of reasons including a reboot, power outage, DHCP lease expiry, or other action by the ISP. If this occurs, traffic forwarded to the subscriber might be delivered to another customer who now has that address. This affects both implicit dynamic mappings and explicit dynamic mappings. However, this same problem occurs today when a subscriber's IP address is re-assigned, without PCP and without an ISP-operated CGN. The solution is the same as today: the problems associated with subscriber renumbering are caused by subscriber renumbering and are eliminated if subscriber renumbering is avoided. PCP defined in this document does not provide machinery to reduce the subscriber renumbering problem.
When a new Internal Address is assigned to a host embedding a PCP client, the NAT (or firewall) controlled by the PCP server will continue to send traffic to the old IP address. Typically, the PCP client will no longer receive traffic sent to that old IP address. Assuming the PCP client wants to continue receiving traffic, it needs to install new mappings for its new IP address. The suggested external port field will not be fulfilled by the PCP server, in all likelihood, because it is still being forwarded to the old IP address. Thus, a mapping is likely to be assigned a new external port number and/or public IP address. Note that this scenario is not expected to happen routinely on a regular basis for most hosts, since most hosts renew their DHCP leases before they expire (or re-request the same address after reboot) and most DHCP servers honor such requests and grant the host the same address it was previously using before the reboot.
A host might gain or lose interfaces while existing mappings are active (e.g., Ethernet cable plugged in or removed, joining/leaving a WiFi network). Because of this, if the PCP client is sending a PCP request to maintain state in the PCP server, it SHOULD ensure those PCP requests continue to use the same interface (e.g., when refreshing mappings). If the PCP client is sending a PCP request to create new state in the PCP server, it MAY use a different source interface or different source address.
This section defines two OpCodes for controlling dynamic connections. They are:
The operation of these OpCodes is described in this section.
The PEER OpCodes provide a single function: the ability for the PCP client to query and (possibly) extend the lifetime of an existing mapping.
The two PEER OpCodes (PEER4 and PEER6) share a similar packet layout for both requests and responses. Because of this similarity, they are shown together.
The following diagram shows the request packet format for PEER4 and PEER6. This packet format is aligned with the response packet format:
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Protocol | Reserved (24 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Internal Port | Suggested External Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Remote Peer Port | Reserved (16 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : Remote Peer IP Address (32 bits if PEER4, 128 bits if PEER6) : : : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Suggested External IP address (always 128 bits) | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
These fields are described below:
The following diagram shows the response packet format for PEER4 and PEER6:
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Protocol | External_AF | Reserved (16 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Internal Port | External Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Remote Peer Port | Reserved (16 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : Remote Peer IP Address (32 bits if PEER4, 128 bits if PEER6) : : : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : External IP Address (32 bits if External_AF indicates IPv4 : : 128 bits if External_AF indicates IPv6) : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
This section describes the operation of a client when generating the OpCodes PEER4 or PEER6.
The PEER4 or PEER6 OpCodes MAY be sent before or after establishing bi-directional communication with the remote peer. If sent before, PEER4 or PEER6 OpCodes will create a mapping in the PCP-controlled device that functions exactly as if an implicit dynamic connection were made (e.g., TCP SYN). If sent after, the PEER4 or PEER6 OpCodes query (and control) the implicit dynamic mapping.
The PEER4 and PEER6 OpCodes contain a description of the remote peer address, from the perspective of the PCP client. This is important when the PCP-controlled device is performing address family translation (NAT46 or NAT64), because the destination address from the perspective of the PCP client is different from the destination address on the other side of the address family translation device. For this reason, the PEER4 and PEER6 responses contain an External_AF field.
This section describes the operation of a server when receiving a request with the OpCode PEER4 or PEER6. Processing SHOULD be performed in the order of the following paragraphs.
On receiving the PEER4 or PEER6 OpCode, the PCP server examines the mapping table. If the described mapping does not yet exist yet, it is created, honoring and the Suggested External Port and Suggested External IP Address are honored (if possible; if not possible, a mapping on a different IP address or different port is created). By having PEER create such a mapping, we avoid a race condition between the PEER request or the initial outgoing packet arriving at the NAT gateway first, and allow PEER to be used to recreate an implicit dynamic mapping (see last paragraph of Section 11.2.1).
The PEER4 or PEER6 OpCode MAY reduce the lifetime of an existing mapping; this is implementation-dependent.
If the PCP-controlled device can extend the lifetime of a mapping, the PCP server uses the smaller of its configured maximum lifetime value and the requested lifetime from the PEER request, and sets the lifetime to that value.
If all of the proceeding operations were successful (did not generate an error response), then a SUCCESS response is generated, with the Lifetime field containing the lifetime of the mapping.
After a successful PEER response is sent, it is implementation-specific if the PCP-controlled device destroys the mapping when the lifetime expires, or if the PCP-controlled device's implementation allows traffic to keep the mapping alive. Thus, if the PCP client wants the mapping to persist beyond the lifetime, it MUST refresh the mapping (by sending another PEER message) prior to the expiration of the lifetime. If the mapping is terminated by the TCP client or server (e.g., TCP FIN or TCP RST), the mapping will be destroyed normally; the mapping will not persist for the time indicated by Lifetime. This means the Lifetime in a PEER response indicates how long the mapping will persist in the absence of a transport termination message (e.g., TCP RST).
This section describes the operation of a client when processing a response with the OpCode PEER4 or PEER6.
After performing common PCP response processing, the response is further matched with a request by comparing the protocol, external AF, internal IP address, internal port, remote peer address and remote peer port. Other fields are not compared, because the PCP server changes those fields to provide information about the mapping created by the OpCode.
If a successful response, the application can use the assigned lifetime value to reduce its frequency of application keepalives for that particular NAT mapping. Of course, there may be other reasons, specific to the application, to use more frequent application keepalives. For example, the PCP assigned lifetime could be one hour but the application may want to maintain state on its server (e.g., "busy" / "away") more frequently than once an hour.
If the PCP client wishes to keep this mapping alive beyond the indicated lifetime, it SHOULD issue a new PCP request prior to the expiration. That is, inside->outside traffic is not sufficient to ensure the mapping will continue to exist. It is RECOMMENDED to send a single renewal request packet when a mapping is halfway to expiration time, then, if no SUCCESS response is received, another single renewal request 3/4 of the way to expiration time, and then another at 7/8 of the way to expiration time, and so on, subject to the constraint that renewal requests MUST NOT be sent less than four seconds apart (a PCP client MUST NOT ever-closer-together requests in the last few seconds before a mapping expires).
This section describes Options for the MAP4, MAP6, PEER OpCodes. These Options MUST NOT appear with other OpCodes, unless permitted by those OpCodes.
This Option is used when a PCP client wants to control a mapping to an internal host other than itself. This is used with both MAP and PEER OpCodes.
A THIRD_PARTY Option MUST NOT contain the same address as the source address of the packet. A PCP server receiving a THIRD_PARTY Option specifying the same address as the source address of the packet MUST return a MALFORMED_REQUEST result code. This is because many PCP servers may not implement the THIRD_PARTY Option at all, and a client using the THIRD_PARTY Option to specify the same address as the source address of the packet will cause mapping requests to fail where they would otherwise have succeeded.
Where possible, it may beneficial if a client using the THIRD_PARTY option to create and maintain mappings on behalf of some other device can take steps to verify that the other device is still present and active on the network. Otherwise the client using the THIRD_PARTY option to maintain mappings on behalf of some other device risks maintaining those mappings forever, long after the device that required them has gone. This would defeat the purpose of PCP mappings having a finite lifetime so that they can be automatically deleted after they are no longer needed.
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : Internal IP Address (32 bits of 128 bits, depending : : on Option length) : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
The fields are described below:
This Option:
The following additional result codes may be returned as a result of using this Option.
A PCP server MAY be configured to permit or to restrict the use of the THIRD_PARTY option. If this option is permitted, any host can create, modify, or destroy mappings for another host on the subscriber's network. If third party mappings are restricted, only authorized clients can perform these operations. If a PCP server is configured to restrict third party mappings, and receives a PCP MAP request with a THIRD_PARTY option, it MUST generate a UNAUTH_THIRD_PARTY_INTERNAL_ADDRESS response. Determining which PCP clients are authorized to use the THIRD_PARTY option depends on the deployment scenario. For Dual-Stack Lite deployments, the PCP server only supports this option if the source IPv6 address is the B4's source IP address. For home deployments (where the PCP server is embedded in the NAT device), this option MUST NOT be processed. For scenarios where the subscriber has only one IP address (e.g., typical residential ISP service) this Option serves no purpose (and will only generate error messages from the server). If a subscriber has more than one IP address the ISP MUST determine its own policy for how to identify the trusted device within the subscriber's home. This might be, for example, the lowest- or highest-numbered host address for that user's IPv4 prefix. A cryptographic authentication and authorization model is outside the scope of this specification.
It is RECOMMENDED that PCP servers embedded into customer premise equipment be configured to refuse third party mappings by default. With this default, if a user wants to create a third party mapping, the user needs to interact out-of-band with their customer premise router (e.g., using its HTTP administrative interface).
It is RECOMMENDED that PCP servers embedded into service provider NAT and firewall devices be configured to permit the THIRD_PARTY option, when sent by the customer premise router. With this configuration, if a user wants to create an explicit dynamic mapping or query an implicit dynamic mapping for another host within their network, the user needs to interact out-of-band with their customer premise router (e.g., using its HTTP administrative interface). To accomplish this, the PCP server in the ISP's network processes requests with the THIRD_PARTY option if they arrived from the IP address of the customer premise router. In deployments with only one IP address (e.g., which is common in residential networks), the PCP messages will -- by necessity -- arrive from the IP address of the customer premise router router. In networks where users have multiple IPv4 or multiple IPv6 addresses, the PCP server MUST only allow the THIRD_PARTY option if the PCP message was sent by the IP address of the subscriber's customer premise router. In Dual-Stack Lite, this would be the B4 element's IPv6 address. If the packet arrived from a different address, the PCP server MUST generate an UNAUTH_THIRD_PARTY_INTERNAL_ADDRESS error.
If authorized to do so, a PCP client can delete all the PCP-created explicit dynamic mappings (i.e., those created by PCP MAP requests) for all hosts belonging to the same subscriber. This is done by sending a PCP MAP request including the THIRD_PARTY option with its Internal Address field set to 0.
This option is only used with the MAP4 and MAP6 OpCodes.
This option indicates that if the PCP server is unable to MAP the suggested port, then instead of returning an available port that it *can* allocate, the PCP server should instead allocate no port and return result code CANNOT_PROVIDE_EXTERNAL_PORT.
This option is intended solely for use by UPnP IGD interworking [I-D.bpw-pcp-upnp-igd-interworking], where the semantics of UPnP IGD version 1 do not provide any way to indicate to an UPnP IGD client that any port is available other than the one it wanted. A PCP server MAY support this option, if its designers wish to support downstream devices that perform UPnP IGD interworking. PCP servers MAY choose to rate-limit their handling of PREFER_FAILURE requests, to protect themselves from a rapid flurry of 65535 consecutive PREFER_FAILURE requests from clients probing to discover which external ports are available. PCP servers that are not intended to support downstream devices that perform UPnP IGD interworking are not required to support this option. PCP clients other than UPnP IGD interworking clients SHOULD NOT use this option because it results in inefficient operation, and they cannot safely assume that all PCP servers will implement it. It is anticipated that this option will be deprecated in the future as more clients adopt PCP natively and the need for UPnP IGD interworking declines.
This Option indicates filtering incoming packets is desired. The remote peer port and remote peer IP Address indicate the permitted remote peer's source IP address and port for packets from the Internet. The remote peer prefix length indicates the length of the remote peer's IP address that is significant; this allows a single Option to permit an entire subnet. After processing this MAP request containing the FILTER option and generating a successful response, the PCP-controlled device will drop packets received on its public-facing interface that don't match the filter fields. After dropping the packet, if its security policy allows, the PCP-controlled device MAY also generate an ICMP error in response to the dropped packet.
The FILTER packet layout is described below:
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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Reserved | prefix-length | Remote Peer Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ : : : Remote Peer IP address (32 bits if MAP4, : : 128 bits if MAP6) : : : +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
These fields are described below:
This Option:
The prefix-length indicates how many bits of the IPv6 address or IPv4 address are used for the filter. For MAP4, a prefix-length of 32 indicates the entire IPv4 address is used. For MAP6, a prefix-length of 128 indicates the entire IPv6 address is used. For MAP4 the minimum prefix-length value is 0 and the maximum value is 32. For MAP6 the minimum prefix-length value is 0 and the maximum value is 128. Values outside those range cause an MALFORMED_OPTION result code.
If multiple occurrences of the FILTER option exist in the same MAP request, they are processed in the same order received, and they MUST all be successfully processed or return an error (e.g., MALFORMED_OPTION if one of the options was malformed), and they MAY overlap the filtering requested. As with other PCP errors, returning an error causes no state to be changed in the PCP server or in the PCP-controlled device. If an existing mapping exists (with or without a filter) and the server receives a MAP request with FILTER, the filters indicated in the new request are added to any existing filters. If a MAP request has a lifetime of 0 and contains the FILTER option, the error MALFORMED_OPTION is returned.
To remove all existing filters, the prefix-length 0 is used. There is no mechanism to remove a specific filter.
To change an existing filter, the PCP client sends a MAP request containing two FILTER options, the first option containing a prefix-length of 0 (to delete all existing filters) and the second containing the new remote peer's IP address and port. Other FILTER options in that PCP request, if any, add more allowed remote hosts.
The PCP server or the PCP-controlled device is expected to have a limit on the number of remote peers it can support. This limit might be as small as one. If a MAP request would exceed this limit, the entire MAP request is rejected with the result code EXCESSIVE_REMOTE_PEERS, and the state on the PCP server is unchanged.
This section provides non-normative guidance that may be useful to implementors.
For implicit dynamic mappings, some existing NAT devices have endpoint-independent mapping (EIM) behavior while other NAT devices have non-endpoint-independent mapping (non-EIM) behavior. NATs which have EIM behavior do not suffer from the problem described in this section. EIM behavior is strongly encouraged by both [RFC4787] and [RFC5382].
In such non-EIM NAT devices, the same external port may be used by an implicit dynamic connection (from the same internal host or from a different internal host) and an explicit dynamic connection. This complicates the interaction with the MAP4 and MAP6 OpCodes. With such NAT devices, there are two ways envisioned to implement the MAP4 and MAP6 OpCodes:
If an event occurs that causes the PCP server to lose explicit dynamic mapping state (such as a crash or power outage), the mappings created by PCP are lost. Such loss of state is rare in a service provider environment (due to redundant power, disk drives for storage, etc.), but more common in a residential NAT device which does not write information to its non-volatile memory. Of course, due to outright failure of service provider equipment (e.g., software malfunction), state may still be lost.
The Epoch allows a client to deduce when a PCP server may have lost its state. When the Epoch value is smaller than expected, the PCP client can attempt to recreate the mappings following the procedures described in this section.
When the PCP server loses state and begins processing new PCP messages, its Epoch is reset to zero (per the procedure of Section 6.5). A mapping renewal packet is formatted identically to an original mapping request; from the point of view of the client it is a renewal of an existing mapping, but from the point of view of the PCP server it appears as a new mapping request. In the normal process of routinely renewing its mappings before they expire, a PCP client will automatically recreate all its lost mappings.
In addition, as the result of receiving a packet where the Epoch field indicates that a reboot or similar loss of state has occurred, the client can renew its port mappings sooner, without waiting for the normal routine renewal time.
A PCP client can refresh a mapping by sending a new PCP request containing information from the earlier PCP response. The PCP server will respond indicating the new lifetime. It is possible, due to failure of the PCP server, that the public IP address and/or public port, or the PCP server itself, has changed (due to a new route to a different PCP server). To detect such events more quickly, the PCP client may find it beneficial to use shorter lifetimes (so that it communicates with the PCP server more often). If the PCP client has several mappings, the Epoch value only needs to be retrieved for one of them to verify the PCP server has not lost explicit dynamic mapping state.
If the client wishes to check the PCP server's Epoch, it sends a PCP request for any one of the client's mappings. This will return the current Epoch value. In that request the PCP client could extend the mapping lifetime (by asking for more time) or maintain the current lifetime (by asking for the same number of seconds that it knows are remaining of the lifetime).
If an internal IP address is no longer valid (e.g., because the internal host has moved to a new network), and the PCP client wishes to still receive incoming traffic, it needs create a new mapping on that new network. A new mapping will also require an update to the application-specific rendezvous server (see Section 7.1 and Section 8.7).
To prevent spoofing of PCP requests, ingress filtering [RFC2827] MUST be performed by devices between the PCP clients and PCP server. For example, with a PCP server integrated into a customer premise router, the Ethernet switch needs to perform ingress filtering. As another example, with a PCP server deployed by a service provider, the service provider's aggregation router (the first device connecting to subscribers) needs to do ingress filtering.
On PCP-controlled devices that create state when a mapping is created (e.g., NAPT), the PCP server SHOULD maintain a per-subscriber quota for explicit dynamic mappings. It is implementation-specific if the PCP server has a separate or combined quota for both implicit dynamic mappings (e.g., created by TCP SYNs) and explicit dynamic mappings (created using PCP).
This document defines Port Control Protocol and two types of OpCodes, PEER and MAP. The PEER OpCode allows querying and extending (if permitted) the lifetime of an existing implicit dynamic mapping, so a host can reduce its keepalive messages. The MAP OpCode allows creating a mapping so a host can receive incoming unsolicited connections from the Internet in order to run a server.
The PEER OpCode can create a mapping (which behaves exactly as if an implicit dynamic mapping were created (e.g., by a TCP SYN)). In that case, the security implications for PEER are similar to MAP, described below. When PEER is used to set (or query) an existing mapping, it does not introduce any new security considerations, unless the THIRD_PARTY Option is included. Discussion of the THIRD_PARTY Option is below.
With the exception of wireless providers (who are interested in protecting their radio access network), Internet service providers do not typically filter traffic from the Internet towards their subscribers. However, when an ISP introduces stateful address sharing with a NAPT device, such filtering will occur as a side effect of the NAPT device. Filtering will also occur with an IPv6 CPE [RFC6092]. The MAP OpCode allows a PCP client to create a mapping so that a host can receive inbound traffic and operate a server. Security considerations for the MAP OpCode are described in the following sections.
Because of the state created in a NAPT or firewall, a per-subscriber quota will likely exist for both implicit dynamic mappings (e.g., outgoing TCP connections) and explicit dynamic mappings (PCP). A subscriber might make an excessive number of implicit or explicit dynamic mappings, consuming an inordinate number of ports, causing a denial of service to other subscribers. Thus, Section 12.2 recommends that subscribers be limited to a reasonable number of explicit dynamic mappings.
It is important to prevent a subscriber from creating a mapping for another subscriber (or for another host), because this allows incoming packets from the Internet and consumes the other user's mapping quota. Both implicit dynamic mappings (e.g., outgoing TCP connections) and explicit dynamic mappings (PCP) need ingress filtering. Thus, PCP relies on the same ingress filtering as implicit dynamic mappings and does not create a new requirement for ingress filtering.
The THIRD_PARTY Option contains a Internal Address field, which allows a PCP client to create, extend, or delete an implicit or explicit dynamic mapping for another host.
In scenarios where the subscriber has one IP address (e.g., as commonly occurs with IPv4 residential deployments) or the subscriber has multiple IP addresses and a CP router enforces a PCP policy (by operating its own PCP server or performing filtering [RFC6092]), the PCP server in both the CP router and the ISP's equipment will both reject any message containing THIRD_PARTY. Thus, PCP cannot be used by a host to create, modify, or delete mappings of other hosts, except by using the administrative interface of the customer premise router (e.g., HTTP interface), as described in Section 10.1.
In other scenarios, where the subscriber has multiple IP addresses and the subscriber CP router is not filtering, but the ISP is providing filtering, the ISP should only accept PCP messages containing the THIRD_PARTY Option from the IP address of the customer's router, as described in Section 10.1.
In the time between when a PCP server loses state and the PCP client notices the lower-than-expected Epoch value, it is possible that the PCP client's mapping will be acquired by another host (via an explicit dynamic mapping or implicit dynamic mapping). This means incoming traffic will be sent to a different host ("theft"). A mechanism to immediately inform the PCP client of state loss would reduce this interval, but would not eliminate this threat. The PCP client can reduce this interval by using a relatively short lifetime; however, this increases the amount of PCP chatter. This threat is eliminated by using persistent storage of explicit dynamic mappings in the PCP server (so it does not lose explicit dynamic mapping state), or by ensuring the previous external IP address and port cannot be used by another host (e.g., by using a different IP address pool). This threat can be mitigated by authenticating the data connection between the hosts (e.g., using TLS).
IANA is requested to perform the following actions:
PCP will use port 5351 (currently assigned by IANA to NAT-PMP). We request that IANA re-assign that same port number to PCP, and relinquish UDP port 44323.
IANA shall create a new protocol registry for PCP OpCodes, initially populated with the values in Section 8 and Section 9. The values 0 and 127 are reserved.
Additional OpCodes in the range 4-95 can be created via Standards Action [RFC5226], and the range 96-126 is for Private Use [RFC5226].
IANA shall create a new registry for PCP result codes, numbered 0-255, initially populated with the result codes from Section 5.4, Section 8.2, Section 10.3, and Section 10.1. The values 0 and 255 are reserved.
Additional Result Codes can be defined via Specification Required [RFC5226].
IANA shall create a new registry for PCP Options, numbered 0-255 with an associated mnemonic. The values 0-127 are mandatory-to-process, and 128-255 are optional-to-process. The initial registry contains the options described in Section 10 and Section 10.1. The option values 127 and 255 are reserved.
Additional PCP option codes in the ranges 5-63 and 128-191 can be created via Standards Action [RFC5226], and the ranges 64-126 and 192-254 are for Private Use [RFC5226].
Thanks to Alain Durand, Christian Jacquenet, Jacni Qin, Simon Perreault, Paul Selkirk, and James Yu for their comments and review. Thanks to Simon Perreault for highlighting the interaction of dynamic connections with PCP-created mappings.
Thanks to Francis Dupont for his several thorough reviews of the specification, which improved the protocol significantly.
| [RFC2119] | Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. |
| [RFC5226] | Narten, T. and H. Alvestrand, "Guidelines for Writing an IANA Considerations Section in RFCs", BCP 26, RFC 5226, May 2008. |
| [RFC0768] | Postel, J., "User Datagram Protocol", STD 6, RFC 768, August 1980. |
| [RFC2827] | Ferguson, P. and D. Senie, "Network Ingress Filtering: Defeating Denial of Service Attacks which employ IP Source Address Spoofing", BCP 38, RFC 2827, May 2000. |
| [I-D.ietf-behave-v6v4-xlate] | Li, X, Bao, C and F Baker, "IP/ICMP Translation Algorithm", Internet-Draft draft-ietf-behave-v6v4-xlate-23, September 2010. |
| [RFC4193] | Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast Addresses", RFC 4193, October 2005. |
| [RFC6056] | Larsen, M. and F. Gont, "Recommendations for Transport-Protocol Port Randomization", BCP 156, RFC 6056, January 2011. |
| [proto_numbers] | IANA, , "Protocol Numbers", 2010. |
The Port Control Protocol (PCP) is a successor to the NAT Port Mapping Protocol (NAT-PMP), and shares similar semantics, concepts, and packet formats. Because of this NAT-PMP and PCP both use the same port, and use the NAT-PMP and PCP's version negotiation capabilities to determine which version to use. This section describes how an orderly transition may be achieved.
A client supporting both NAT-PMP and PCP SHOULD send its request using the PCP packet format. This will be received by a NAT-PMP server or a PCP server. If received by a NAT-PMP server, the response will be as indicated by [I-D.cheshire-nat-pmp], which will cause the client to downgrade to NAT-PMP and re-send its request in NAT-PMP format. If received by a PCP server, the response will be as described by this document and processing continues as expected.
A PCP server supporting both NAT-PMP and PCP can respond to requests in either format. The first byte of the packet indicates if it is NAT-PMP (first byte zero) or PCP (first byte non-zero).
A PCP-only gateway receiving a NAT-PMP request (identified by the first byte being zero) will interpret the request as a version mismatch. Normal PCP processing will emit a PCP response that is compatible with NAT-PMP, without any special handling by the PCP server.
[Note to RFC Editor: Please remove this section prior to publication.]