SOFTWIRE Working Group D. Wing Internet-Draft Cisco Intended status: Standards Track R. Penno Expires: September 2, 2010 Juniper Networks March 1, 2010 Port Control Protocol (PCP) draft-wing-softwire-port-control-protocol-00 Abstract It is desirable to operate a server behind a device that does not normally permit operating a server, such as a carrier's NAT64, a carrier's NAT44, or an IPv6 customer premise router that filters incoming packets. This document proposes an application-level protocol, derived from NAT-PMP, to allow a host to request opening a port in such an upstream device. Requirements Language 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 RFC 2119 [RFC2119]. Status of this Memo This Internet-Draft is submitted to IETF in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. 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." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt. The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. This Internet-Draft will expire on September 2, 2010. Wing & Penno Expires September 2, 2010 [Page 1] Internet-Draft Port Control Protocol (PCP) March 2010 Copyright Notice Copyright (c) 2010 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 BSD License. Wing & Penno Expires September 2, 2010 [Page 2] Internet-Draft Port Control Protocol (PCP) March 2010 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Scope and Requirements . . . . . . . . . . . . . . . . . . . . 4 3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.1. Port Forwarding . . . . . . . . . . . . . . . . . . . . . 5 3.1.1. Basic Port Forwarding . . . . . . . . . . . . . . . . 5 3.1.2. Dynamic Port Forwarding . . . . . . . . . . . . . . . 5 3.1.3. Port Forwarding and Redirection . . . . . . . . . . . 6 4. Other Commonly-Deployed Port Reservation Mechanisms . . . . . 6 4.1. Comparison . . . . . . . . . . . . . . . . . . . . . . . . 6 4.2. Interoperation with commonly-deployed port-reservation protocols . . . . . . . . . . . . . . . . . . . . . . . . 6 5. Discovery of PCP Server . . . . . . . . . . . . . . . . . . . 7 6. Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 6.1. Protocol and Packet Format . . . . . . . . . . . . . . . . 8 6.1.1. Requests and Responses . . . . . . . . . . . . . . . . 9 6.1.2. Determining the External Address . . . . . . . . . . . 9 6.1.3. Returning a Mapping . . . . . . . . . . . . . . . . . 10 6.1.4. Destroying a Mapping . . . . . . . . . . . . . . . . . 13 6.1.5. Result Codes . . . . . . . . . . . . . . . . . . . . . 15 6.1.6. Seconds Since Start of Epoch . . . . . . . . . . . . . 16 6.2. Interactions with Outgoing Sessions . . . . . . . . . . . 16 7. PCP and Dual Stack-Lite . . . . . . . . . . . . . . . . . . . 16 8. PCP and NAT64 . . . . . . . . . . . . . . . . . . . . . . . . 16 9. PCP and IPv6 Simple Security . . . . . . . . . . . . . . . . . 17 10. PCP IANA-Assigned Address . . . . . . . . . . . . . . . . . . 17 11. Security Considerations . . . . . . . . . . . . . . . . . . . 17 12. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 17 13. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 18 14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 18 14.1. Normative References . . . . . . . . . . . . . . . . . . . 18 14.2. Informative References . . . . . . . . . . . . . . . . . . 18 Appendix A. PCP implemented on host . . . . . . . . . . . . . . . 19 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 19 Wing & Penno Expires September 2, 2010 [Page 3] Internet-Draft Port Control Protocol (PCP) March 2010 1. Introduction Port Control Protocol (PCP) provides a simple mechanism to request a TCP or UDP port from an upstream NAT (NAT64 or NAT44) or firewall (IPv6 or IPv4). PCP can be implemented within customer premise (CP) routers, or by an application running on a host. Such a protocol is necessary during the deployment of large-scale NATs (both NAT64 and NAT44) and with IPv6 CPE which filter incoming connections [I-D.ietf-v6ops-cpe-simple-security]. Both NAT-PMP [I-D.cheshire-nat-pmp] and SP-NAT-PMP [I-D.woodyatt-spnatpmp-appl] were used as the basis for the protocol described in this document. 2. Scope and Requirements PCP is not meant to be a general purpose NAT or Firewall port reservation protocol. PCP has a very focused goal of allowing port reservation in the context of Dual-Stack Lite [I-D.ietf-softwire-dual-stack-lite], NAT64 [I-D.ietf-behave-v6v4-framework], and IPv6 Simple CPE Security [I-D.ietf-v6ops-cpe-simple-security]. Other scenarios are out of scope. The requirements of the Port Control Protocol described in this document are: 1. trivial to implement in a large-scale NAT. This includes minimal state and timers on the NAT. 2. support IPv4 and IPv6. 3. cannot rely on a shared broadcast domain between the PCP client and its PCP server, as these are point-to-point links (in the case of Dual-Stack Lite) or a network of routers (in the case of NAT64). 4. for NATted environments (NAT44 and NAT64), allows the NAT to choose the public UDP port or public TCP port. This is necessary in large-scale NAT environments where a client might otherwise request ports that are already being used. 5. does not require modifying code in the application needing the incoming port; rather, another application can open the port. This permits a PC to use PCP to open ports for an IP-enabled webcam, and permits a CP router to open ports on behalf of Wing & Penno Expires September 2, 2010 [Page 4] Internet-Draft Port Control Protocol (PCP) March 2010 devices behind the CP. 6. for NATted environments (NAT44 and NAT64), the NAT must not forward traffic to other internal hosts (subscribers) until the mapping lifetime expires or the host expliticly releases it. This differs from UPnP IGD and NAT-PMP, which allow the NAT to destroy its binding if the NAT is reset (e.g., rebooted), and forward incoming traffic to those ports to other internal hosts. The following are also considered out of scope: 1. Multihoming is not supported. 3. Terminology 3.1. Port Forwarding Port forwading is normally used when several hosts behind a NAT device share a single external IP address to communicate. If an internal host is listening to connections on a specific port (that is, operating as a server), the external IP and port need to be port forwarded (also called "mapped") to the internal IP address and port. The key point is that the internal and external transport destination ports could be different; for example, a webcam might be listening on 192.168.1.1 and TCP/80, but the public address is 192.0.2.1 and TCP/ 12345; the NAT does 'port forwarding' of one to the other. There are some variations of port forwarding in respect to the fields in the IP and Transport headers that get modified as packets traverse a device and the nature of the mapping. These are discussed below. 3.1.1. Basic Port Forwarding In the most basic definition port forwarding means the static, pre- determined mapping of a transport destination port, protocol and IP address 3-tuple into another by a device. Following the nomenclature first introduced by RFC2633, this would be called basic-port- forwarding. Today, users typically configure basic port forwarding on their CP router using the CP router's HTTP configuration. 3.1.2. Dynamic Port Forwarding The mapping on the device could be installed dynamic based on application requests through protocols such as UPnP IGD [UPnP-IGD] or NAT-PMP. In this case these protocols in conjunction with the CPE Wing & Penno Expires September 2, 2010 [Page 5] Internet-Draft Port Control Protocol (PCP) March 2010 NAT device automatically take care of mapping management. 3.1.3. Port Forwarding and Redirection Certain protocols such as HTTP can redirect a client to use a different port without translation. This is widely used, specially in Content Delivery Networks, but not for the same purposes as port forwarding. On the other hand, port forwarding and HTTP redirection meet as a possible solution to manage inbound connections through Large Scale NATs (e.g., [I-D.wing-behave-http-46-relay]). 4. Other Commonly-Deployed Port Reservation Mechanisms 4.1. Comparison This section compares commonly-deployed port reservation mechanisms. UPnP IGD: IPv4 only. Relatively complicated to implement in the UPnP IGD server, due to reliable HTTPU (HTTP over reliable UDP), XML, and UPnP's multicast discovery. IPv6 support is not yet available. UPnP IGD 1.0 has the host decide and dictate the public port, which is not viable in a large scale NAT environment, as the port may already be allocated to another host. NAT-PMP: IPv4 only. NAT-PMP does not include discovery; it assumes the next-hop default router supports NAT-PMP. STUN: Requires a STUN server, and requires discoverying/learning the IP address of the STUN server. Requires using same source UDP/TCP port as being opened, which requires modifying the application itself (that is, the port opening can't be a separate application on the same system). 4.2. Interoperation with commonly-deployed port-reservation protocols It is envisoned that customer premise equipment (CPE) routers will continue to offer popular LAN-based mechanisms to provide port mapping to computers, such as UPnP IGD, NAT-PMP, and HTTP-based configuration. PCP is expected to operate between the CPE router and the service provider's NAT44 or NAT64. This means the CPE router acts as client for HTML-based configuration and as a proxy between UPnP IGD and PCP, or between NAT-PMP and PCP, or both, as shown in the following diagrams. Links with PCP are shown with "=", links that aren't using PCP are shown with "-". Wing & Penno Expires September 2, 2010 [Page 6] Internet-Draft Port Control Protocol (PCP) March 2010 Scenario 1 +-------------+ +----------+ | PCP-proxying| |Dual-Stack| | customer +====+ Lite AFTR+-- IPv4 host using HTML-----+ premise | +----------+ | router | +-------------+ Figure 1: Scenario 1, CPE router as PCP client Scenario 2 +-------------+ +----------+ IPv4 host using NAT-PMP---+ PCP-proxying| |Dual-Stack| | customer +====+ Lite AFTR+-- IPv4 host using UPnP IGD--+ premise | +----------+ | router | +-------------+ Figure 2: Scenario 2, CPE router as PCP client, proxying UPnP IGD or NAT-PMP with PCP Note regarding UPnP IGD: for UPnP IGD to provide the necessary functionality, the AddAnyPortmapping from UPnP IGD:2 [IGD-2] is needed in (a) IPv4 host's UPnP IGD stack, (b) the IPv4 host's application calling the UPnP IGD stack, and (c) the PCP-proxying CP router. Without this, the host running UPnP IGD (which UPnP calls the 'control point') dictates the public port number; on a large NAT, that port number will likely already be used by another client. 5. Discovery of PCP Server NOTE: Analysis of these discovery mechanisms requires further discussion. The authors have considered two mechanisms for discovery of the PCP server: 1. a special-purpose IPv4 address, assigned by IANA, which is routed normally until it hits a PCP device, which responds. This works well with NAT64, as there are normally routers between the PCP client (the CP router) and the NAT64 (operated by the service provider). When a host wants to allow an incoming port, it sends a UDP packet to the IANA-assigned PCP discovery address and the PCP port. The upstream NAT44, NAT64, or IPv6-CPE will be listening for that address and port, and respond accordingly. Wing & Penno Expires September 2, 2010 [Page 7] Internet-Draft Port Control Protocol (PCP) March 2010 2. assume the default router is a PCP server, and send PCP packets to the IP address of the default router. This is how NAT-PMP operates today. An advantage is this is simple and does not rely on routing of the special-use address. Although multi-homing is out of scope of PCP, it is worth pointing out that both techniques work poorly with multi-homing. 6. Protocol PCP borrows heavily from NAT-PMP due to simplicity, available implementations and compatiblity. This also provides better support in the scenario where the CPE implements a first layer of NAT and supports NAT-PMP, with a mostly one to one message compatibility. The text from NAT-PMP is copied here for continuity purposes. The protocol specified below is written from a PCP client perspective ("Scenario 1" (Figure 1)), and could be implemented in a CPE triggered by HTML-based configuration. The applicability of PCP in a dual-stack lite or NAT64 network is discussed in further sections. A detailed description of proxying UPnP IGD or NAT-PMP to PCP ("Scenario 2" (Figure 2)) will be discussed in companion documents. 6.1. Protocol and Packet Format PCP runs over UDP. Every packet starts with an 8 bit version followed by an 8 bit operation code. This document specifies version 0 of the protocol. Any PCP gateway implementing this version of the protocol, receiving a packet with a version number other than 0, MUST return result code 1 (Unsupported Version), indicating the highest version number it does support (i.e. 0) in the version field of the reply. Opcodes between 0 and 127 are client requests. Opcodes from 128 to 255 are server responses. Responses always contain a 16 bit result code in network byte order. A result code of zero indicates success. Responses also contain a 32 bit unsigned integer corresponding to the number of seconds since the NAT gateway was rebooted or since its port mapping state was reset. Clients always send their Port Mapping Protocol requests to the IANA- assigned address or their default gateway, depending on the consensus for PCP discovery (Section 5). Wing & Penno Expires September 2, 2010 [Page 8] Internet-Draft Port Control Protocol (PCP) March 2010 6.1.1. Requests and Responses To determine the external IP address or request a port mapping, a PCP client sends its request packet to port TBD of its gateway address, and waits 250ms for a response. If no PCP response is received from the gateway after 250ms, the client retransmits its request and waits 500ms. The client SHOULD repeat this process with the interval between attempts doubling each time. If, after sending its 9th attempt (and then waiting for 64 seconds), the client has still received no response, then it SHOULD conclude that this gateway does not support PCP and MAY log an error message indicating this fact. In addition, if the PCP client receives an "ICMP Port Unreachable" message from the gateway for port TBD then it can skip any remaining retransmissions and conclude immediately that the gateway does not support PCP. As a performance optimization the client MAY record this information and use it to suppress further attempts to use PCP, but the client should not retain this information for too long. In particular, any event that may indicate a potential change of gateway or a change in gateway configuration (hardware link change indication, change of gateway MAC address, acquisition of new DHCP lease, receipt of PCP announcement packet from gateway, etc.) should cause the client to discard its previous information regarding the gateway's lack of PCP support, and send its next PCP request packet normally. When deleting a port mapping, the client uses the same initial 250ms timeout, doubling on each successive interval, except that clients may choose not to try the full nine times before giving up. This is because mapping deletion requests are in some sense advisory. They are useful for efficiency, but not required for correctness; it is always possible for client software to crash, or for power to fail, or for a client device to be physically unplugged from the network before it gets a chance to send its mapping deletion request(s), so NAT gateways already need to cope with this case. Because of this, it may be acceptable for a client to retry only once or twice before giving up on deleting its port mapping(s), but a client SHOULD always send at least one deletion request whenever possible, to reduce the amount of stale state that accumulates on NAT gateways. A client need not continue trying to delete a port mapping after the time when that mapping would naturally have expired anyway. 6.1.2. Determining the External Address To determine the external address, the client behind the NAT sends the following UDP payload to port TBD of the gateway address: Wing & Penno Expires September 2, 2010 [Page 9] Internet-Draft Port Control Protocol (PCP) March 2010 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Vers = 0 | OP = 0 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ A compatible NAT gateway MUST generate a response with 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 = 0 | OP = 128 + 0 | Result Code | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Seconds Since Start of Epoch | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | External IP Address (IPv4, 32 bits; IPv6 isn't NATted use "0")| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ This response indicates that the NAT gateway implements this version of the protocol and returns the external IP address of the NAT gateway. If the result code is non-zero, the value of External IP Address is undefined (MUST be set to zero on transmission, and MUST be ignored on reception). When PCP is used to punch holes in a simple firewall, there is no NAT being performed. Rather than indicating an IPv6 address (which is the same, as this isn't for NAT66), an IPv4 address of 0.0.0.0 is sent. The NAT gateway MUST fill in the "Seconds Since Start of Epoch" field with the time elapsed since its port mapping table was initialized on startup or reset for any other reason (see "Seconds Since Start of Epoch" (Section 6.1.6)). Upon receiving a response packet, the client MUST check the source IP address, and silently discard the packet if the address is not the address of the gateway to which the request was sent. 6.1.3. Returning a Mapping To create a mapping, the client sends a UDP packet to port TBD of the gateway's internal IP address with the following format: Wing & Penno Expires September 2, 2010 [Page 10] Internet-Draft Port Control Protocol (PCP) March 2010 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 = 0 | OP = x | Reserved (MUST be zero) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Internal Port | Requested External Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Requested Port Mapping Lifetime in Seconds | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Opcodes supported: 1 - Map UDP, 2 - Map TCP The Reserved field MUST be set to zero on transmission and MUST be ignored on reception. The Internal Port is set to the local port on which the client is listening. If the client would prefer to have a high-numbered "anonymous" external port assigned, then it should set the Requested External Port to zero, which indicates to the gateway that it should allocate a high-numbered port of its choosing. If the client would prefer instead to have the mapped external port be the same as its local Internal Port if possible (e.g. a web server listening on port 80 that would ideally like to have external port 80) then it should set the Requested External Port to the desired value. However, the gateway is not obliged to assign the port requested, and may choose not to, either for policy reasons (e.g. port 80 is reserved and clients may not request it) or because that port has already been assigned to some other client. Because of this, some product developers have questioned the value of having the Requested External Port field at all. The reason is for failure recovery. Most low- cost home NAT gateways do not record temporary port mappings in persistent storage, so if the gateway crashes or is rebooted, all the mappings are lost. A renewal packet is formatted identically to an initial mapping request packet, except that for renewals the client sets the Requested External Port field to the port the gateway actually assigned, rather than the port the client originally wanted. When a freshly-rebooted NAT gateway receives a renewal packet from a client, it appears to the gateway just like an ordinary initial request for a port mapping, except that in this case the Requested External Port is likely to be one that the NAT gateway *is* willing to allocate (it allocated it to this client right before the reboot, so it should presumably be willing to allocate it again). The RECOMMENDED Port Mapping Lifetime is 3600 seconds. If the client is allocated fixed ports, the PCP server simply returns a UDP or TCP port that allocated to the client. Wing & Penno Expires September 2, 2010 [Page 11] Internet-Draft Port Control Protocol (PCP) March 2010 After sending the port mapping request, the client then waits for the NAT gateway to respond. If after 250ms, the gateway doesn't respond, the client SHOULD re-issue its request as described above in "Requests and Responses" (Section 6.1.1). The NAT gateway responds with the following 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Version = 0 | OP = 128 + x | Result Code | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Seconds Since Start of Epoch | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Internal Port | Mapped External Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Port Mapping Lifetime in Seconds | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The 'x' in the OP field MUST match what the client requested. Some NAT gateways are incapable of creating a UDP port mapping without also creating a corresponding TCP port mapping, and vice versa, and these gateways MUST NOT implement NAT Port Mapping Protocol until this deficiency is fixed. A NAT gateway which implements this protocol MUST be able to create TCP-only and UDP-only port mappings. If a NAT gateway silently creates a pair of mappings for a client that only requested one mapping, then it may expose that client to receiving inbound UDP packets or inbound TCP connection requests that it did not ask for and does not want. While a NAT gateway MUST NOT automatically create mappings for TCP when the client requests UDP, and vice versa, the NAT gateway MUST reserve the companion port so the same client can choose to map it in the future. For example, if a client requests to map TCP port 80, as long as the client maintains the lease for that TCP port mapping, another client with a different IP address MUST NOT be able to successfully acquire the mapping for UDP port 80. The client normally requests the external port matching the internal port. If that external port is not available, the NAT gateway MUST return an available external port or return an error code if no ports are available. The source address of the packet MUST be used for the internal address in the mapping. This protocol is not intended to facilitate one device behind a NAT creating mappings for other devices. If there are legacy devices that require inbound mappings, permanent mappings can be created manually by the administrator, just as they are today. Wing & Penno Expires September 2, 2010 [Page 12] Internet-Draft Port Control Protocol (PCP) March 2010 If a mapping already exists for a given internal port on a given local client (whether that mapping was created explicitly using PCP, implicitly as a result of an outgoing TCP SYN packet, or manually by a human administrator) and that client requests another mapping for the same internal port (possibly requesting a different external port) then the mapping request should succeed, returning the already- assigned external port. This is necessary to handle the case where a client requests a mapping with requested external port X, and is granted a mapping with actual external port Y, but the acknowledgement packet gets lost. When the client retransmits its mapping request, it should get back the same positive acknowledgement as was sent (and lost) the first time. The NAT gateway MUST NOT accept mapping requests destined to the NAT gateway's external IP address or received on its external network interface. Only packets received on the internal interface(s) with a destination address matching the internal address(es) of the NAT gateway should be allowed. The NAT gateway MUST fill in the "Seconds Since Start of Epoch" field with the time elapsed since its port mapping table was initialized on startup or reset for any other reason (see Section "Seconds Since Start of Epoch" (Section 6.1.6)). The Port Mapping Lifetime is an unsigned integer in seconds. The NAT gateway MAY reduce the lifetime from what the client requested. The NAT gateway SHOULD NOT offer a lease lifetime greater than that requested by the client. Upon receiving the response packet, the client MUST check the source IP address, and silently discard the packet if the address is not the address of the gateway to which the request was sent. The client SHOULD begin trying to renew the mapping halfway to expiry time, like DHCP. The renewal packet should look exactly the same as a request packet, except that the client SHOULD set the requested external port to what the NAT gateway previously mapped, not what the client originally requested. As described above, this enables the gateway to automatically recover its mapping state after a crash or reboot. 6.1.4. Destroying a Mapping A mapping may be destroyed in a variety of ways. If a client fails to renew a mapping, then when its lifetime expires the mapping MUST be automatically deleted. In the common case where the gateway device is a combined DHCP server and NAT gateway, when a client's DHCP address lease expires, the gateway device MAY automatically Wing & Penno Expires September 2, 2010 [Page 13] Internet-Draft Port Control Protocol (PCP) March 2010 delete any mappings belonging to that client. Otherwise a new client being assigned the same IP address could receive unexpected inbound UDP packets or inbound TCP connection requests that it did not ask for and does not want. A client MAY also send an explicit packet to request deletion of a mapping that is no longer needed. A client requests explicit deletion of a mapping by sending a message to the NAT gateway requesting the mapping, with the Requested Lifetime in Seconds set to 0. The requested external port MUST be set to zero by the client on sending, and MUST be ignored by the gateway on reception. When a mapping is destroyed successfully as a result of the client explicitly requesting the deletion, the NAT gateway MUST send a response packet which is formatted as defined in Section 6.1.3. The response MUST contain a result code of 0, the internal port as indicated in the deletion request, an external port of 0, and a lifetime of 0. The NAT gateway MUST respond to a request to destroy a mapping that does not exist as if the request were successful. This is because of the case where the acknowledgement is lost, and the client retransmits its request to delete the mapping. In this case the second request to delete the mapping MUST return the same response packet as the first request. If the deletion request was unsuccessful, the response MUST contain a non-zero result code and the requested mapping; the lifetime is undefined (MUST be set to zero on transmission, and MUST be ignored on reception). If the client attempts to delete a port mapping which was manually assigned by some kind of configuration tool, the NAT gateway MUST respond with a 'Not Authorized' error, result code 2. When a mapping is destroyed as a result of its lifetime expiring or for any other reason, if the NAT gateway's internal state indicates that there are still active TCP connections traversing that now- defunct mapping, then the NAT gateway SHOULD send appropriately- constructed TCP RST (reset) packets both to the local client and to the remote peer on the Internet to terminate that TCP connection. A client can request the explicit deletion of all its UDP or TCP mappings by sending the same deletion request to the NAT gateway with external port, internal port, and lifetime set to 0. A client MAY choose to do this when it first acquires a new IP address in order to protect itself from port mappings that were performed by a previous owner of the IP address. After receiving such a deletion request, the gateway MUST delete all its UDP or TCP port mappings (depending on the opcode). The gateway responds to such a deletion request with a response as described above, with the internal port set to zero. Wing & Penno Expires September 2, 2010 [Page 14] Internet-Draft Port Control Protocol (PCP) March 2010 If the gateway is unable to delete a port mapping, for example, because the mapping was manually configured by the administrator, the gateway MUST still delete as many port mappings as possible, but respond with a non-zero result code. The exact result code to return depends on the cause of the failure. If the gateway is able to successfully delete all port mappings as requested, it MUST respond with a result code of 0. 6.1.5. Result Codes Currently defined result codes: 0 - Success 1 - Unsupported Version 2 - Not Authorized/Refused (e.g. box supports mapping, but user has turned feature off) 3 - Network Failure (e.g. NAT box itself has not obtained a DHCP lease) 4 - Out of resources (NAT box cannot create any more mappings at this time) 5 - Unsupported opcode 6 - Per-subscriber resource limit would be exceeded (from SPNATPMP) If the result code is non-zero, the format of the packet following the result code may be truncated. For example, if the client sends a request to the server with an opcode of 17 and the server does not recognize that opcode, the server SHOULD respond with a message where the opcode is 17 + 128 and the result code is 5 (opcode not supported). Since the server does not understand the format of opcode 17, it may not know what to place after the result code. In some cases, relevant data may follow the opcode to identify the operation that failed. For example, a client may request a mapping but that mapping may fail due to resource exhaustion. The server SHOULD respond with the result code to indicate resource exhaustion (4) followed by the requested port mapping so the client may identify which operation failed. Clients MUST be able to properly handle result codes not defined in Wing & Penno Expires September 2, 2010 [Page 15] Internet-Draft Port Control Protocol (PCP) March 2010 this document. Undefined results codes MUST be treated as fatal errors of the request. 6.1.6. Seconds Since Start of Epoch Every packet sent by the NAT gateway includes a "Seconds since start of epoch" field (SSSOE). If the NAT gateway resets or loses the state of its port mapping table, due to reboot, power failure, or any other reason, it MUST reset its epoch time and begin counting SSSOE from 0 again. Whenever a client receives any packet from the NAT gateway, either gratuitously or in response to a client request, the client computes its own conservative estimate of the expected SSSOE value by taking the SSSOE 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 SSSOE in the newly received packet is less than the client's conservative estimate by more than one second, then the client concludes that the NAT gateway has undergone a reboot or other loss of port mapping state, and the client MUST immediately renew all its active port mapping leases. 6.2. Interactions with Outgoing Sessions There are a few important considerations when port forwarding is combained with a NAT that uses address pools. First and foremost, if there is a port forwarding mapping between certain external and internal addresses, all sessions originated by the host associated with the internal address should use the same external address present in the port forwarding mapping. 7. PCP and Dual Stack-Lite In the case of Dual Stack-Lite deployments, PCP packets triggered by HTML-based configuration would be crafted as described in Section 6 and encapsulated in IPv4. The source IPv4 would be the internal host used in the port forwarding configuration and the destination IPv4 is based on the decision of Section 5. The UDP destination port MUST be set fo the IANA allocated desitnation port for PCP. Then, the PCP request is encapsulated in an IPv6 packet following RFC2473 and sent to the softwire concentrator (which is called "AFTR" in Dual-Stack Lite). The AFTR decapsulates the IPv4 packet and processes the PCP packet. 8. PCP and NAT64 Hosts behind a NAT64 device can make use of PCP in order to perform Wing & Penno Expires September 2, 2010 [Page 16] Internet-Draft Port Control Protocol (PCP) March 2010 port reservation (to get a publicly-routable IPv4 port). If in "Discovery" (Section 5) we choose to have an IANA-assigned IP address for discovery of the PCP server, that IPv4 address can be placed into the IPv6 destination address following that particular network's well-known prefix or network-specific prefix, per [I-D.ietf-behave-address-format]. 9. PCP and IPv6 Simple Security [[text to be added.]] 10. PCP IANA-Assigned Address Using an IANA-assigned PCP address allows the device performing PCP to be located anywhere (rather than solely on the local LAN or on the next-hop router). The special-use address MUST NOT be advertised in the global routing table. Packets with that destination address SHOULD be filtered so they are not transmitted on the Internet. 11. Security Considerations Any software on the host can open a UDP or TCP port on an upstream NAT or upstream firewall, permitting incoming connections. At first glance, this seems risky, as malicious software running on a host could allow that host's web server to be accessible from the Internet, for example. However, that same malicious software, if it were restricted to only open incoming connections for itself could do so, and could then relay incoming traffic to the host's own webserver. Thus, security is no worse by allowing an application to open other arbitrary ports. 12. Acknowledgements PCP is very closely related to NAT-PMP [I-D.cheshire-nat-pmp] (Stuart Cheshire, Marc Krochmal, Kiren Sekar) and SP-NAT-PMP [I-D.woodyatt-spnatpmp-appl] (James Woodyatt). Our thanks to authors of those documents. Wing & Penno Expires September 2, 2010 [Page 17] Internet-Draft Port Control Protocol (PCP) March 2010 13. IANA Considerations Assign an IPv4 and IPv6 address for PCP, should we decide to use the IP address discovery mechanism Section 5. Assign a UDP port for PCP communication, preferably from the well- known port range (0-1023). 14. References 14.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. 14.2. Informative References [I-D.cheshire-nat-pmp] Cheshire, S., "NAT Port Mapping Protocol (NAT-PMP)", draft-cheshire-nat-pmp-03 (work in progress), April 2008. [I-D.ietf-behave-address-format] Huitema, C., Bao, C., Bagnulo, M., Boucadair, M., and X. Li, "IPv6 Addressing of IPv4/IPv6 Translators", draft-ietf-behave-address-format-04 (work in progress), January 2010. [I-D.ietf-behave-v6v4-framework] Baker, F., Li, X., Bao, C., and K. Yin, "Framework for IPv4/IPv6 Translation", draft-ietf-behave-v6v4-framework-07 (work in progress), February 2010. [I-D.ietf-softwire-dual-stack-lite] Durand, A., Droms, R., Haberman, B., Woodyatt, J., Lee, Y., and R. Bush, "Dual-stack lite broadband deployments post IPv4 exhaustion", draft-ietf-softwire-dual-stack-lite-03 (work in progress), February 2010. [I-D.ietf-v6ops-cpe-simple-security] Woodyatt, J., "Recommended Simple Security Capabilities in Customer Premises Equipment for Providing Residential IPv6 Internet Service", draft-ietf-v6ops-cpe-simple-security-09 (work in progress), February 2010. [I-D.wing-behave-http-46-relay] Wing & Penno Expires September 2, 2010 [Page 18] Internet-Draft Port Control Protocol (PCP) March 2010 Wing, D., "Relaying HTTP from IPv4 to IPv6", draft-wing-behave-http-46-relay-02 (work in progress), October 2009. [I-D.woodyatt-spnatpmp-appl] Woodyatt, J., "Applicability of NAT-PMP with Service Provider Deployments of Network Address Translation", draft-woodyatt-spnatpmp-appl-01 (work in progress), November 2008. [IGD-2] UPnP Gateway Committee, "IGD:2 improvements over IGD:1", 2009, . [UPnP-IGD] UPnP Forum, "Universal Plug and Play Internet Gateway Device", 2000, . Appendix A. PCP implemented on host NOTE: This is for future study. It is also possible for PCP to be implemented directly on a host, as shown below. Scenario 3 +-------------+ +----------+ IPv4 host using PCP=======+ | |Dual-Stack| | customer +====+ Lite AFTR+-- | premise | +----------+ | router | +-------------+ Figure 3: CPE router unaware PCP Wing & Penno Expires September 2, 2010 [Page 19] Internet-Draft Port Control Protocol (PCP) March 2010 Authors' Addresses Dan Wing Cisco Systems, Inc. 170 West Tasman Drive San Jose, California 95134 USA Email: dwing@cisco.com Reinaldo Penno Juniper Networks 1194 N Mathilda Avenue Sunnyvale, California 94089 USA Phone: Fax: Email: rpenno@juniper.net URI: Wing & Penno Expires September 2, 2010 [Page 20]