Status of This Memo

This is an Internet Standards Track document.¶

This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Further information on Internet Standards is available in Section 2 of RFC 7841.¶

Information about the current status of this document, any errata, and how to provide feedback on it may be obtained athttps://www.rfc-editor.org/info/rfc9298.¶

Copyright Notice

Copyright (c) 2022 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.¶

Table of Contents

1.Introduction

While HTTP provides the CONNECT method (seeSection 9.3.6 of [HTTP])for creating a TCP[TCP] tunnel to a proxy, it lacked a method fordoing so for UDP[UDP] traffic prior to this specification.¶

This document describes a protocol for tunneling UDP to a server acting as aUDP-specific proxy over HTTP. UDP tunnels are commonly used to create anend-to-end virtual connection, which can then be secured using QUIC[QUIC] or another protocol running over UDP. Unlike the HTTP CONNECTmethod, the UDP proxy itself is identified with an absolute URL containing thetraffic's destination. Clients generate those URLs using a URI Template[TEMPLATE], as described inSection 2.¶

This protocol supports all existing versions of HTTP by using HTTP Datagrams[HTTP-DGRAM]. When using HTTP/2[HTTP/2] or HTTP/3[HTTP/3], it uses HTTP Extended CONNECT as described in[EXT-CONNECT2]and[EXT-CONNECT3]. When using HTTP/1.x[HTTP/1.1], it uses HTTP Upgradeas defined inSection 7.8 of [HTTP].¶

2.Client Configuration

HTTP clients are configured to use a UDP proxy with a URI Template[TEMPLATE] that has the variables "target_host" and "target_port".Examples are shown below:¶

https://example.org/.well-known/masque/udp/{target_host}/{target_port}/https://proxy.example.org:4443/masque?h={target_host}&p={target_port}https://proxy.example.org:4443/masque{?target_host,target_port}

The following requirements apply to the URI Template:¶

• The URI TemplateMUST be a level 3 template or lower.¶
• The URI TemplateMUST be in absolute form andMUST include non-empty scheme,authority, and path components.¶
• The path component of the URI TemplateMUST start with a slash ("/").¶
• All template variablesMUST be within the path or query components of the URI.¶
• The URI TemplateMUST contain the two variables "target_host" and"target_port" andMAY contain other variables.¶
• The URI TemplateMUST NOT contain any non-ASCII Unicode characters andMUSTonly contain ASCII characters in the range 0x21-0x7E inclusive (note thatpercent-encoding is allowed; seeSection 2.1 of [URI]).¶
• The URI TemplateMUST NOT use Reserved Expansion ("+" operator), FragmentExpansion ("#" operator), Label Expansion with Dot-Prefix, Path SegmentExpansion with Slash-Prefix, nor Path-Style Parameter Expansion withSemicolon-Prefix.¶

ClientsSHOULD validate the requirements above; however, clientsMAY use ageneral-purpose URI Template implementation that lacks this specific validation.If a client detects that any of the requirements above are not met by a URITemplate, the clientMUST reject its configuration and abort the request withoutsending it to the UDP proxy.¶

The original HTTP CONNECT method allowed for the conveyance of the target hostand port, but not the scheme, proxy authority, path, or query. Thus, clientswith proxy configuration interfaces that only allow the user to configure theproxy host and the proxy port exist. Client implementations of thisspecification that are constrained by such limitationsMAY attempt to access UDPproxying capabilities using the default template, which is defined as"https://$PROXY_HOST:$PROXY_PORT/.well-known/masque/udp/{target_host}/{target_port}/",where $PROXY_HOST and $PROXY_PORT are the configured host and port of the UDPproxy, respectively. UDP proxy deploymentsSHOULD offer service at this locationif they need to interoperate with such clients.¶

3.Tunneling UDP over HTTP

To allow negotiation of a tunnel for UDP over HTTP, this document defines the"connect-udp" HTTP upgrade token. The resulting UDP tunnels use the CapsuleProtocol (seeSection 3.2 of [HTTP-DGRAM]) with HTTP Datagrams in the formatdefined inSection 5.¶

To initiate a UDP tunnel associated with a single HTTP stream, a client issues arequest containing the "connect-udp" upgrade token. The target of the tunnel isindicated by the client to the UDP proxy via the "target_host" and "target_port"variables of the URI Template; seeSection 2.¶

"target_host" supports using DNS names, IPv6 literals and IPv4 literals. Notethat IPv6 scoped addressing zone identifiers are not supported. Using the termsIPv6address, IPv4address, reg-name, and port from[URI], the "target_host" and"target_port" variablesMUST adhere to the format inFigure 2, usingnotation from[ABNF]. Additionally:¶

• both the "target_host" and "target_port" variablesMUST NOT be empty.¶
• if "target_host" contains an IPv6 literal, the colons (":")MUST bepercent-encoded. For example, if the target host is "2001:db8::42", it will beencoded in the URI as "2001%3Adb8%3A%3A42".¶
• "target_port"MUST represent an integer between 1 and 65535 inclusive.¶

target_host = IPv6address / IPv4address / reg-nametarget_port = port

When sending its UDP proxying request, the clientSHALL perform URI Templateexpansion to determine the path and query of its request.¶

If the request is successful, the UDP proxy commits to converting received HTTPDatagrams into UDP packets, and vice versa, until the tunnel is closed.¶

By virtue of the definition of the Capsule Protocol (seeSection 3.2 of [HTTP-DGRAM]), UDP proxying requests do not carry any message content.Similarly, successful UDP proxying responses also do not carry any messagecontent.¶

GET https://example.org/.well-known/masque/udp/192.0.2.6/443/ HTTP/1.1Host: example.orgConnection: UpgradeUpgrade: connect-udpCapsule-Protocol: ?1

HTTP/1.1 101 Switching ProtocolsConnection: UpgradeUpgrade: connect-udpCapsule-Protocol: ?1

HEADERS:method = CONNECT:protocol = connect-udp:scheme = https:path = /.well-known/masque/udp/192.0.2.6/443/:authority = example.orgcapsule-protocol = ?1

HEADERS:status = 200capsule-protocol = ?1

4.Context Identifiers

The mechanism for proxying UDP in HTTP defined in this document allows futureextensions to exchange HTTP Datagrams that carry different semantics from UDPpayloads. Some of these extensions can augment UDP payloads with additionaldata, while others can exchange data that is completely separate from UDPpayloads. In order to accomplish this, all HTTP Datagrams associated with UDPProxying request streams start with a Context ID field; seeSection 5.¶

Context IDs are 62-bit integers (0 to 2⁶²-1). Context IDs are encodedas variable-length integers; seeSection 16 of [QUIC]. The Context ID value of0 is reserved for UDP payloads, while non-zero values are dynamically allocated.Non-zero even-numbered Context IDs are client-allocated, and odd-numberedContext IDs are proxy-allocated. The Context ID namespace is tied to a givenHTTP request; it is possible for a Context ID with the same numeric value to besimultaneously allocated in distinct requests, potentially with differentsemantics. Context IDsMUST NOT be re-allocated within a given HTTP namespacebutMAY be allocated in any order. The Context ID allocation restrictions to theuse of even-numbered and odd-numbered Context IDs exist in order to avoid theneed for synchronization between endpoints. However, once a Context ID has beenallocated, those restrictions do not apply to the use of the Context ID; it canbe used by any client or UDP proxy, independent of which endpoint initiallyallocated it.¶

Registration is the action by which an endpoint informs its peer of thesemantics and format of a given Context ID. This document does not define howregistration occurs. Future extensionsMAY use HTTP header fields or capsules toregister Context IDs. Depending on the method being used, it is possible fordatagrams to be received with Context IDs that have not yet been registered. Forinstance, this can be due to reordering of the packet containing the datagramand the packet containing the registration message during transmission.¶

5.HTTP Datagram Payload Format

When HTTP Datagrams (seeSection 2 of [HTTP-DGRAM]) are associated with UDPProxying request streams, the HTTP Datagram Payload field has the format definedinFigure 7, using notation fromSection 1.3 of [QUIC]. Note that whenHTTP Datagrams are encoded using QUIC DATAGRAM frames[QUIC-DGRAM],the Context ID field defined below directly follows the Quarter Stream ID field,which is at the start of the QUIC DATAGRAM frame payload; seeSection 2.1 of [HTTP-DGRAM].¶

UDP Proxying HTTP Datagram Payload {  Context ID (i),  UDP Proxying Payload (..),}

Context ID:

A variable-length integer (seeSection 16 of [QUIC]) that contains the valueof the Context ID. If an HTTP/3 Datagram that carries an unknown Context ID isreceived, the receiverSHALL either drop that datagram silently or buffer ittemporarily (on the order of a round trip) while awaiting the registration ofthe corresponding Context ID.¶

UDP Proxying Payload:

The payload of the datagram, whose semantics depend on the value of theprevious field. Note that this field can be empty.¶

UDP packets are encoded using HTTP Datagrams with the Context ID field set tozero. When the Context ID field is set to zero, the UDP Proxying Payload fieldcontains the unmodified payload of a UDP packet (referred to as data octets in[UDP]).¶

By virtue of the definition of the UDP header[UDP], it is not possible toencode UDP payloads longer than 65527 bytes. Therefore, endpointsMUST NOT sendHTTP Datagrams with a UDP Proxying Payload field longer than 65527 using ContextID zero. An endpoint that receives an HTTP Datagram using Context ID zero whoseUDP Proxying Payload field is longer than 65527MUST abort the correspondingstream. If a UDP proxy knows it can only send out UDP packets of a certainlength due to its underlying link MTU, it has no choice but to discard incomingHTTP Datagrams using Context ID zero whose UDP Proxying Payload field is longerthan that limit. If the discarded HTTP Datagram was transported by a DATAGRAMcapsule, the receiverSHOULD discard that capsule without buffering the capsulecontents.¶

If a UDP proxy receives an HTTP Datagram before it has received thecorresponding request, itSHALL either drop that HTTP Datagram silently orbuffer it temporarily (on the order of a round trip) while awaiting thecorresponding request.¶

Note that buffering datagrams (either because the request was not yet receivedor because the Context ID is not yet known) consumes resources. Receivers thatbuffer datagramsSHOULD apply buffering limits in order to reduce the risk ofresource exhaustion occurring. For example, receivers can limit the total numberof buffered datagrams or the cumulative size of buffered datagrams on aper-stream, per-context, or per-connection basis.¶

A clientMAY optimistically start sending UDP packets in HTTP Datagrams beforereceiving the response to its UDP proxying request. However, implementers shouldnote that such proxied packets may not be processed by the UDP proxy if itresponds to the request with a failure or if the proxied packets are received bythe UDP proxy before the request and the UDP proxy chooses to not buffer them.¶

6.Performance Considerations

Bursty traffic can often lead to temporally correlated packet losses; in turn,this can lead to suboptimal responses from congestion controllers in protocolsrunning over UDP. To avoid this, UDP proxiesSHOULD strive to avoid increasingburstiness of UDP traffic; theySHOULD NOT queue packets in order to increasebatching.¶

When the protocol running over UDP that is being proxied uses congestion control(e.g.,[QUIC]), the proxied traffic will incur at least two nested congestioncontrollers. The underlying HTTP connectionMUST NOT disable congestion controlunless it has an out-of-band way of knowing with absolute certainty that theinner traffic is congestion-controlled.¶

If a client or UDP proxy with a connection containing a UDP Proxying requeststream disables congestion control, itMUST NOT signal Explicit CongestionNotification (ECN)[ECN] support on that connection. That is, itMUSTmark all IP headers with the Not-ECT codepoint. ItMAY continue to report ECNfeedback via QUIC ACK_ECN frames or the TCP ECE bit, as the peer may not havedisabled congestion control.¶

When the protocol running over UDP that is being proxied uses loss recovery(e.g.,[QUIC]), and the underlying HTTP connection runs over TCP, the proxiedtraffic will incur at least two nested loss recovery mechanisms. This can reduceperformance as both can sometimes independently retransmit the same data. Toavoid this, UDP proxyingSHOULD be performed over HTTP/3 to allow leveraging theQUIC DATAGRAM frame.¶

7.Security Considerations

There are significant risks in allowing arbitrary clients to establish a tunnelto arbitrary targets, as that could allow bad actors to send traffic and have itattributed to the UDP proxy. HTTP servers that support UDP proxying ought torestrict its use to authenticated users.¶

There exist software and network deployments that perform access control checksbased on the source IP address of incoming requests. For example, some softwareallows unauthenticated configuration changes if they originated from 127.0.0.1.Such software could be running on the same host as the UDP proxy or in the samebroadcast domain. Proxied UDP traffic would then be received with a source IPaddress belonging to the UDP proxy. If this source address is used for accesscontrol, UDP proxying clients could use the UDP proxy to escalate their accessprivileges beyond those they might otherwise have. This could lead tounauthorized access by UDP proxying clients unless the UDP proxy disallows UDPproxying requests to vulnerable targets, such as the UDP proxy's own addressesand localhost, link-local, multicast, and broadcast addresses. UDP proxies canuse the destination_ip_prohibited Proxy Error Type fromSection 2.3.5 of [PROXY-STATUS] when rejecting such requests.¶

UDP proxies share many similarities with TCP CONNECT proxies when consideringthem as infrastructure for abuse to enable denial-of-service (DoS) attacks. Bothcan obfuscate the attacker's source address from the attack target. In the caseof a stateless volumetric attack (e.g., a TCP SYN flood or a UDP flood), bothtypes of proxies pass the traffic to the target host. With stateful volumetricattacks (e.g., HTTP flooding) being sent over a TCP CONNECT proxy, the proxywill only send data if the target has indicated its willingness to accept databy responding with a TCP SYN-ACK. Once the path to the target is flooded, theTCP CONNECT proxy will no longer receive replies from the target and will stopsending data. Since UDP does not establish shared state between the UDP proxyand the target, the UDP proxy could continue sending data to the target in sucha situation. While a UDP proxy could potentially limit the number of UDP packetsit is willing to forward until it has observed a response from the target, thatprovides limited protection against DoS attacks when attacks target open UDPports where the protocol running over UDP would respond and that would beinterpreted as willingness to accept UDP by the UDP proxy. Such a packet limitcould also cause issues for valid traffic.¶

The security considerations described inSection 4 of [HTTP-DGRAM] also applyhere. Since it is possible to tunnel IP packets over UDP, the guidance in[TUNNEL-SECURITY] can apply.¶

8.IANA Considerations

Acknowledgments

This document is a product of the MASQUE Working Group, and the author thanks all MASQUE enthusiasts for their contributions. This proposal was inspired directly or indirectly by prior work from many people, in particular[HELIUM] byBen Schwartz,[HiNT] byLucas Pardue, and the original MASQUE Protocol[MASQUE-ORIGINAL] by the author of this document.¶

The author would like to thankEric Rescorla for suggesting the use of an HTTP method to proxy UDP. The author is indebted toMark Nottingham andLucas Pardue for the many improvements they contributed to this document. The extensibility design in this document came out of the HTTP Datagrams Design Team, whose members wereAlan Frindell,Alex Chernyakhovsky,Ben Schwartz,Eric Rescorla,Lucas Pardue,Marcus Ihlar,Martin Thomson,Mike Bishop,Tommy Pauly,Victor Vasiliev, and the author of this document.¶

Author's Address