| Internet-Draft | QUIC Address Discovery | August 2024 |
| Seemann | Expires 21 February 2025 | [Page] |
Unless they have out-of-band knowledge, QUIC endpoints have no information about their network situation. They neither know their external IP address and port, nor do they know if they are directly connected to the internet or if they are behind a NAT. This QUIC extension allows nodes to determine their public IP address and port for any QUIC path.¶
This note is to be removed before publishing as an RFC.¶
The latest revision of this draft can be found at https://marten-seemann.github.io/draft-seemann-quic-address-discovery/draft-seemann-quic-address-discovery.html. Status information for this document may be found at https://datatracker.ietf.org/doc/draft-seemann-quic-address-discovery/.¶
Discussion of this document takes place on the QUIC Working Group mailing list (mailto:quic@ietf.org), which is archived at https://mailarchive.ietf.org/arch/browse/quic/. Subscribe at https://www.ietf.org/mailman/listinfo/quic/.¶
Source for this draft and an issue tracker can be found at https://github.com/marten-seemann/draft-seemann-quic-address-discovery.¶
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 https://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 21 February 2025.¶
Copyright (c) 2024 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 (https://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 Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
STUN ([RFC8489]) allows nodes to discover their reflexive transport address by asking a remote server to report the observed source address. While the QUIC ([RFC9000]) packet header was designed to allow demultiplexing from STUN packets, moving address discovery into the QUIC layer has a number of advantages:¶
STUN traffic is unencrypted, and can be observed and modified by on-path observers. By moving address discovery into QUIC's encrypted envelope it becomes invisible to observers.¶
When located behind a load balancer, QUIC packets may be routed based on the QUIC connection ID. Depending on the architecture, not using STUN might simplify the routing logic.¶
If QUIC traffic doesn't need to be demultiplexed from STUN traffic, implementations can enable QUIC bit greasing ([RFC9287]).¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
Endpoints advertise their support of the extension by sending the address_discovery (0x9f81a174) transport parameter (Section 7.4 of [RFC9000]) with a variable-length integer value. The value determines the behavior with respect to address discovery:¶
0: The node is willing to provide address observations to its peer, but is not interested in receiving address observations itself.¶
1: The node is interested in receiving address observations, but it is not willing to provide address observations.¶
2: The node is interested in receiving address observations, and it is willing to provide address observations.¶
Implementations that understand this transport parameter MUST treat the receipt of any other value than these as a connection error of type TRANSPORT_PARAMETER_ERROR.¶
When using 0-RTT, both endpoints MUST remember the value of this transport parameter. This allows sending the frame defined by this extension in 0-RTT packets. If 0-RTT data is accepted by the server, the server MUST NOT disable this extension or change the value on the resumed connection.¶
This extension defines the OBSERVED_ADDRESS frame.¶
OBSERVED_ADDRESS Frame {
Type (i) = 0x9f81a2..0x9f81a3,
[ IPv4 (32) ],
[ IPv6 (128) ],
Port (16),
}
¶
The OBSERVED_ADDRESS frame contains the following fields:¶
The IPv4 address. Only present if the least significant bit of the frame type is 0.¶
The IPv6 address. Only present if the least significant bit of the frame type is 1.¶
The port number, in network byte order.¶
This frame MUST only appear in the handshake and in the application data packet number space. It is a "probing frame" as defined in Section 9.1 of [RFC9000]. OBSERVED_ADDRESS frames are ack-eliciting, and SHOULD be retransmitted if lost. Retransmissions MUST happen on the same path as the original frame was sent on.¶
An endpoint MUST NOT send an OBSERVED_ADDRESS frame to a node that did not request the receipt of address observations as described in Section 3. A node that did not request the receipt of address observations MUST close the connection with a PROTOCOL_VIOLATION error if it receives an OBSERVED_ADDRESS frame.¶
An endpoint that negotiated (see Section 3) this extension and offered to provide address observations to the peer MUST send an OBSERVED_ADDRESS frame on every new path. This also applies to the path used for the QUIC handshake.¶
The OBSERVED_ADDRESS frame SHOULD be sent as early as possible. However, during the handshake an endpoint SHOULD prioritize frames that lead to handshake progress (CRYPTO and ACK frames in particular) over sending of the OBSERVED_ADDRESS frame.¶
For paths used after completion of the handshake, endpoints SHOULD bundle the OBSERVED_ADDRESS frame with probing packets. This is possible, since the frame is defined to be a probing frame (Section 8.2 of [RFC9000]).¶
Additionally, the sender SHOULD send an OBSERVED_ADDRESS frame when it detects a change in the remote address on an existing path. This could be indicative of a NAT rebindings.¶
In general, nodes cannot be trusted to report the correct address in OBSERVED_ADDRESS frames. If possible, endpoints might decide to only request address observations when connecting to trusted peers, or if that is not possible, define some validation logic (e.g. by asking multiple untrusted peers and observing if the responses are consistent). This logic is out of scope for this document.¶
Depending on the routing setup, a node might not be able to observe the peer's reflexive transport address, and attempts to do so might reveal details about the internal network. In these cases, the node SHOULD NOT offer to provide address observations.¶
TODO: fill out registration request for the transport parameter and frame types¶
TODO acknowledge.¶