Internet-Draft | CONNECT-IP | July 2021 |
Chernyakhovsky, et al. | Expires 13 January 2022 | [Page] |
This document describes the CONNECT-IP HTTP method. CONNECT-IP is similar to CONNECT-UDP, but allows transmitting IP packets, without being limited to just TCP like CONNECT or UDP like CONNECT-UDP.¶
This note is to be removed before publishing as an RFC.¶
Discussion of this document takes place on the Multiplexed Application Substrate over QUIC Encryption Working Group mailing list (masque@ietf.org), which is archived at https://mailarchive.ietf.org/arch/browse/masque/.¶
Source for this draft and an issue tracker can be found at https://github.com/DavidSchinazi/draft-cms-masque-connect-ip.¶
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 13 January 2022.¶
Copyright (c) 2021 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 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.¶
This document describes the CONNECT-IP HTTP method. CONNECT-IP is similar to CONNECT-UDP, but allows transmitting IP packets, without being limited to just TCP like CONNECT or UDP like CONNECT-UDP.¶
CONNECT-IP allows endpoints to set up an IP tunnel between one another. This can be used to implement a consumer VPN, point-to-point, point-to-network, and network-to-network capabilities as described in [REQS].¶
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.¶
In this document, we use the term "proxy" to refer to the HTTP server that responds to the CONNECT-IP request. If there are HTTP intermediaries (as defined in Section 2.3 of [RFC7230]) between the client and the proxy, those are referred to as "intermediaries" in this document.¶
The CONNECT-IP method establishes a stream to an endpoint server that then permits the exchange of control data, such as IP address information, reachable IP ranges, and other relevant information for successfully transmitting IP datagrams between hosts.¶
The request-target of a CONNECT-IP request is a URI [URI] which uses the "https" scheme and a client-specified path. When using HTTP/2 [H2] or later, CONNECT-IP requests use HTTP pseudo-headers with the following requirements:¶
A CONNECT-IP request that does not conform to these restrictions is malformed (see [H2], Section 8.1.2.6).¶
Any 2xx (Successful) response indicates that the proxy is willing to open an IP tunnel between it and the client. Any response other than a successful response indicates that the tunnel has not yet been formed.¶
A proxy MUST NOT send any Transfer-Encoding or Content-Length header fields in a 2xx (Successful) response to CONNECT-IP. A client MUST treat a successful response to CONNECT-IP containing any Content-Length or Transfer-Encoding header fields as malformed.¶
A payload within a CONNECT-IP request message has no defined semantics; a CONNECT-IP request with a non-empty payload is malformed. Note that the CONNECT-IP stream is used to convey control messages, but they are not semantically part of the request or response themselves.¶
Responses to the CONNECT-IP method are not cacheable.¶
The lifetime of the tunnel is tied to the CONNECT-IP stream. Closing the stream (via the FIN bit on a QUIC STREAM frame, or a QUIC RESET_STREAM frame) closes the associated tunnel.¶
IP packets are sent using HTTP Datagrams [HTTP-DGRAM]. The HTTP Datagram Payload contains a full IP packet, from the IP Version field until the last byte of the IP Payload. In order to use HTTP Datagrams, the CONNECT-IP client will first decide whether or not to use HTTP Datagram Contexts and then register its context ID (or lack thereof) using the corresponding registration capsule, see [HTTP-DGRAM].¶
Since HTTP Datagrams require prior negotiation (for example, in HTTP/3 it is necessary to both send and receive the H3_DATAGRAM SETTINGS Parameter), clients MUST NOT send any HTTP Datagrams until they have established support on a given connection. If negotiation of HTTP Datagrams fails (for example if an HTTP/3 SETTINGS frame was received without the H3_DATAGRAM SETTINGS Parameter), the client MUST consider its CONNECT-IP request as failed.¶
Since CONNECT-IP allows the transmission of IP packets over HTTP, CONNECT-IP endpoints will most often forward these packets to and from traditional IP interfaces. As such, CONNECT-IP endpoints act as IP routers. When a CONNECT-IP endpoint receives an HTTP Datagram containing an IP packet, it will parse the packet's IP header, perform any local policy checks (e.g., source address validation), check their routing table to pick an outbound interface, and then use an implementation-specific mechanism (such as raw sockets) to send the IP packet on that interface.¶
Conversely, when a CONNECT-IP endpoint receives an IP packet whose destination address does not match any local addresses, it consults its routing table to pick a forwarding destination, and if the table points to a CONNECT-IP tunnel, the endpoint performs the same forwarding checks before transmitting the packet inside the tunnel.¶
Note that CONNECT-IP endpoints will decrement the IP Hop Count (or TTL) upon encapsulation but not decapsulation. In other words, the Hop Count is decremented right before an IP packet is transmitted in an HTTP Datagram. This prevents infinite loops in the presence of routing loops, and matches the choices in IPsec [IPSEC].¶
Endpoints MAY implement additional filtering policies on the IP packets they forward.¶
Endpoints have the ability to advertise and reject routes using the ROUTE_ADVERTISEMENT (Section 6.3) and ROUTE_REJECTION (Section 6.3) capsule. Note that these capsules are purely informational: receipt of a ROUTE_ADVERTISEMENT capsule does not require the recipient to start routing traffic to its peer. Additionally, if an endpoint receives a ROUTE_REJECTION for a given prefix that it had previously received a ROUTE_ADVERTISEMENT capsule for, then the two cancel out and the endpoint MUST remove its state from the ROUTE_ADVERTISEMENT capsule instead of installing new state for the ROUTE_REJECTION capsule. Conversely, the same is true of a ROUTE_ADVERTISEMENT that matches a previous ROUTE_REJECTION. Routes are handled via longest-prefix-first preference, meaning that if a given IP prefix is covered by multiple route advertisement and route rejections, the one with the longest prefix is used.¶
When processing ROUTE_ADVERTISEMENT capsules, endpoints MUST check their local policy before deciding whether to forward packets to their peer. Since ignoring these capsules is allowed by the protocol, such policy decisions will not prevent interoperability.¶
The ADDRESS_ASSIGN capsule allows an endpoint to inform its peer that it has assigned an IP address to it. It allows assigning a prefix which can contain multiple addresses. This capsule uses a Capsule Type of 0xfff100. Its value uses the following format:¶
IP Version of this address assignment. MUST be either 4 or 6.¶
Assigned IP address. If the IP Version field has value 4, the IP Address field SHALL have a length of 32 bits. If the IP Version field has value 6, the IP Address field SHALL have a length of 128 bits.¶
Length of the IP Prefix assigned, in bits. MUST be lesser or equal to the length of the IP Address field, in bits.¶
The ADDRESS_REQUEST capsule allows an endpoint to request assignment of an IP address from its peer. It allows the endpoint to optionally indicate a preference for which address it would get assigned. This capsule uses a Capsule Type of 0xfff101. Its value uses the following format:¶
IP Version of this address request. MUST be either 4 or 6.¶
Requested IP address. If the IP Version field has value 4, the IP Address field SHALL have a length of 32 bits. If the IP Version field has value 6, the IP Address field SHALL have a length of 128 bits.¶
Length of the IP Prefix requested, in bits. MUST be lesser or equal to the length of the IP Address field, in bits.¶
Upon receiving the ADDRESS_REQUEST capsule, an endpoint SHOULD assign an IP address to its peer, and then respond with an ADDRESS_ASSIGN capsule to inform the peer of the assignment.¶
The ROUTE_ADVERTISEMENT capsule allows an endpoint to communicate to its peer that it is willing to route traffic to a given prefix. This indicates that the sender has an existing route to the prefix, and notifies its peer that if the receiver of the ROUTE_ADVERTISEMENT capsule sends IP packets for this prefix in HTTP Datagrams, the sender of the capsule will forward them along its preexisting route. This capsule uses a Capsule Type of 0xfff102. Its value uses the following format:¶
IP Version of this route advertisement. MUST be either 4 or 6.¶
IP address of the advertised route. If the IP Version field has value 4, the IP Address field SHALL have a length of 32 bits. If the IP Version field has value 6, the IP Address field SHALL have a length of 128 bits.¶
Length of the IP Prefix of the advertised route, in bits. MUST be lesser or equal to the length of the IP Address field, in bits.¶
Upon receiving the ROUTE_ADVERTISEMENT capsule, an endpoint MAY start routing IP packets in that prefix to its peer.¶
The ROUTE_REJECTION capsule allows an endpoint to communicate to its peer that it is not willing to route traffic to a given prefix. This capsule uses a Capsule Type of 0xfff103. Its value uses the following format:¶
IP Version of this route rejection. MUST be either 4 or 6.¶
IP address of the rejected route. If the IP Version field has value 4, the IP Address field SHALL have a length of 32 bits. If the IP Version field has value 6, the IP Address field SHALL have a length of 128 bits.¶
Length of the IP Prefix of the advertised route, in bits. MUST be lesser or equal to the length of the IP Address field, in bits.¶
Upon receiving the ROUTE_REJECTION capsule, an endpoint MUST stop routing IP packets in that prefix to its peer. Note that this capsule can be reordered with DATAGRAM frames, and therefore an endpoint that receives packets for routes it has rejected MUST NOT treat that as an error.¶
The ROUTE_RESET capsule allows an endpoint to cancel any routes it had previously advertised or denied. This capsule uses a Capsule Type of 0xfff104. Its value uses the following format:¶
Upon receiving the ROUTE_RESET capsule, an endpoint MUST stop routing IP packets to its peer. Note that this capsule can be reordered with DATAGRAM frames, and therefore an endpoint that receives packets for routes it has rejected MUST NOT treat that as an error.¶
The main purpose of the ROUTE_RESET capsule is to allow endpoints to not have to remember the full list of routes they have shared with their peer. In practice, it is expected that ROUTE_RESET capsules will be closely followed by ROUTE_ADVERTISEMENT capsules that will refill the routing table that was just cleared.¶
The SHUTDOWN capsule allows an endpoint to communicate to its peer that it is about to close the CONNECT-IP stream, with a string explaining the reason for the shutdown. This capsule uses a Capsule Type of 0xfff105. Its value uses the following format:¶
Additional diagnostic information for the shutdown. This SHOULD be a UTF-8 encoded string [UTF8], though the frame does not carry information, such as language tags, that would aid comprehension by any entity other than the one that created the text.¶
Note that the SHUTDOWN capsule is informational, the tunnel is only closed when its corresponding CONNECT-IP stream is closed. Endpoints MAY close the tunnel with a reason phrase by sending the SHUTDOWN capsule with the FIN bit set on the underlying QUIC STREAM frame that carried it.¶
The ATOMIC_START capsule allows an endpoint to create an atomic set of capsules. This capsule uses a Capsule Type of 0xfff106. Its value uses the following format:¶
Upon receiving an ATOMIC_START capsule, an endpoint MUST buffer all incoming known CONNECT-IP-specific capsules (i.e., capsules defined in this document) until it receives an ATOMIC_END capsule. Endpoints MUST NOT send two ATOMIC_START capsules without an ATOMIC_END capsule between them.¶
Endpoints MUST NOT buffer unknown capsules. Endpoints MAY choose to immediately process IP_PACKET and SHUTDOWN capsules instead of buffering them. Capsules defined in other documents are by default not buffered by ATOMIC_START. Extensions that register new capsule types MAY specify that these capsules should be buffered by ATOMIC_START, and whether it is allowed to skip buffering for them.¶
The purpose of this frame is to avoid timing issues where an endpoint installs a route before an important route rejection was received. Endpoints SHOULD group their initial configuration into an atomic block to allow their peer to mark the tunnel as operational once the whole block is parsed.¶
The ATOMIC_END capsule allows an endpoint to end an atomic set of capsules. This capsule uses a Capsule Type of 0xfff107. Its value uses the following format:¶
Upon receiving an ATOMIC_END capsule, an endpoint MUST parse all previously buffered capsules, in order of receipt. Endpoints MUST NOT send an ATOMIC_END capsule without a preceding ATOMIC_START capsule.¶
CONNECT-IP can be extended via multiple mechanisms to increase functionality. There are three main ways to extend CONNECT-IP: HTTP headers, Capsule Types, and HTTP Datagram Registration Extension Data. For example, an authentication extension could define an HTTP header that allows endpoints to send authentication credentials to their peer during the creation of the tunnel. Alternatively, one could specify an extension that defines a new Capsule Type which allows exchanging DNS configuration between endpoints. Additionally, an extension to CONNECT-IP can use multiple HTTP Datagram Contexts [HTTP-DGRAM] simultaneously to compress some IP packets by associating the compression context with an HTTP Datagram Context ID.¶
There are significant risks in allowing arbitrary clients to establish a tunnel to arbitrary servers, as that could allow bad actors to send traffic and have it attributed to the proxy. Proxies that support CONNECT-IP SHOULD restrict its use to authenticated users. The HTTP Authorization header [AUTH] MAY be used to authenticate clients. More complex authentication schemes are out of scope for this document but can be implemented using CONNECT-IP extensions.¶
Since CONNECT-IP endpoints can proxy IP packets send by their peer, they SHOULD follow the guidance in [BCP38] to help prevent denial of service attacks.¶
In theory, endpoints could use ROUTE_ADVERTISEMENT capsules to divert traffic from naive endpoints. To avoid this, receivers of ROUTE_ADVERTISEMENT capsules MUST check their local policy before acting on such capsules, see Section 5.¶
This document will request IANA to register "CONNECT-IP" in the HTTP Method Registry (IETF review) maintained at <https://www.iana.org/assignments/http-methods>.¶
+-------------+------+------------+---------------+ | Method Name | Safe | Idempotent | Reference | +-------------+------+------------+---------------+ | CONNECT-IP | no | no | This document | +-------------+------+------------+---------------+¶
This document will request IANA to add the following values to the "HTTP Capsule Types" registry created by [HTTP-DGRAM]:¶
+----------+---------------------+---------------------+---------------+ | Value | Type | Description | Reference | +----------+---------------------+---------------------+---------------+ | 0xfff100 | ADDRESS_ASSIGN | Address Assignment | This document | | 0xfff101 | ADDRESS_REQUEST | Address Request | This document | | 0xfff102 | ROUTE_ADVERTISEMENT | Route Advertisement | This document | | 0xfff103 | ROUTE_REJECTION | Route Rejection | This document | | 0xfff104 | ROUTE_RESET | Route Reset | This document | | 0xfff105 | SHUTDOWN | Shutdown Reason | This document | | 0xfff106 | ATOMIC_START | Atomic Start | This document | | 0xfff107 | ATOMIC_END | Atomic End | This document | +----------+---------------------+---------------------+---------------+¶
In this scenario, the client will typically receive a single IP address that the proxy has picked from a pool of addresses it maintains. The client will route all traffic through the tunnel. The exchange could look as follows:¶
Client Server ADDRESS_REQUEST --------> IP Version = 4 IP Address = 0.0.0.0 IP Prefix Length = 0 <-------- ADDRESS_ASSIGN IP Version = 4 IP Address = 192.0.2.42 IP Prefix Length = 32 <-------- ROUTE_ADVERTISEMENT IP Version = 4 IP Address = 0.0.0.0 IP Prefix Length = 0¶
The design of CONNECT-IP was inspired by discussions in the MASQUE working group around [REQS]. The authors would like to thank participants in those discussions for their feedback.¶