Communicating Proxy Configurations in Provisioning Domains

Status of This Memo

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 3 September 2024.¶

Copyright Notice

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.¶

Table of Contents

1. Introduction

HTTP proxies that use the CONNECT method Section 9.3.6 of [HTTP] (often referred to as "forward" proxies) allow clients to open connections to hosts via a proxy. These typically allow for TCP stream proxying, but can also support UDP proxying [CONNECT-UDP] and IP packet proxying [CONNECT-IP]. Such proxies are not just defined as hostnames and ports, but can use URI templates [URITEMPLATE].¶

In order to make use of multiple related proxies, clients need a way to understand which proxies are associated with one another.¶

Client can also benefit from learning about additional information associated with the proxy to optimize their proxy usage, such knowing that a proxy is configured to only allow access to a limited set of next hops.¶

These improvements to client behavior can be achieved through the use of Provisioning Domains. Provisioning Domains (PvDs) are defined in [PVD] as consistent sets of network configuration information, which can include proxy configuration details Section 2 of [PVD]. [PVDDATA] defines a JSON [JSON] format for describing Provisioning Domain Additional Information, which is an extensible dictionary of properties of the Provisioning Domain.¶

This document defines several mechanisms to use PvDs to help clients understand how to use proxies:¶

1. A way to fetch PvD Additional Information associated with a known proxy URI (Section 2)¶

2. A way to list one or more proxy URIs in a PvD, allowing clients to learn about other proxy options given a known proxy (Section 3).¶

3. A way to define a limited set of DNS zones that are accessible through the proxy (Section 4).¶

Additionally, this document partly describes how these mechanisms might be used to discover proxies associated with a network (Section 5).¶

2. Fetching PvD Additional Information for proxies

This document defines a way to fetch PvD Additional Information associated with a proxy. This PvD describes the properties of the network accessible through the proxy.¶

In order to fetch PvD Additional Information associated with a proxy, a client issues an HTTP GET request for the well-known PvD URI (".well-known/pvd") [PVDDATA] and the host authority of the proxy. This is applicable for both proxies that are identified by a host and port only (such as SOCKS proxies and HTTP CONNECT proxies) and proxies that are identified by a URI or URI template.¶

For example, a client would issue the following request for the PvD associated with "https://proxy.example.org/masque{?target_host,target_port}":¶

:method = GET
:scheme = https
:authority = proxy.example.org
:path = /.well-known/pvd
accept = application/pvd+json

For a HTTP CONNECT proxy on "proxy.example.org:8080", the client would send the following request:¶

:method = GET
:scheme = https
:authority = proxy.example.org:8080
:path = /.well-known/pvd
accept = application/pvd+json

Note that all proxies that are colocated on the same host and port share the same PvD Additional Information. Proxy deployments that need separate PvD configuration properties SHOULD use different hosts.¶

PvD Additional Information is required to contain the "identifier", "expires", and "prefixes" keys. For proxy PvDs as defined in this document, the "identifier" MUST match the hostname of the HTTP proxy. The "prefixes" array SHOULD be empty by default.¶

3. Enumerating proxies within a PvD

This document defines a new PvD Additional Information key, proxies, that is an array of dictionaries, where each dictionary in the array defines a single proxy that is available as part of the PvD (see Section 7.1). Each proxy is defined by a proxy protocol, a proxy location (i.e., a hostname and port or a URI template [URITEMPLATE]), along with potentially other keys.¶

This document defines two mandatory keys for the sub-dictionaries in the proxies array, protocol and proxy. These keys, defined below, are the initial contents of the proxy information key registry (Section 7.2). Other optional keys can be added to the dictionary to further define or restrict the use of a proxy. Clients that do not recognize or understand a key in a proxy sub-dictionary MUST ignore the entire proxy definition, since the proxy might be only applicable for particular uses.¶

The values for the protocol key are defined in the proxy protocol registry (Section 7.3). For consistency, any new proxy types that use HTTP Upgrade Tokens (and use the :protocol pseudo-header) SHOULD define the protocol value to match the Upgrade Token / :protocol value.¶

The value of proxy depends on the Proxy Location Format defined by proxy protocol. The types defined here either use a hostname and port, or a full URI template.¶

When a PvD that contains the proxies key is fetched from a known proxy using the method described in Section 2 the proxies list describes equivalent proxies (potentially supporting other protocols) that can be used in addition to the known proxy.¶

Such cases are useful for informing clients of related proxies as a discovery method, with the assumption that the client already is aware of one proxy. Many historical methods of configuring a proxy only allow configuring a single FQDN hostname for the proxy. A client can attempt to fetch the PvD information from the well-known URI to learn the list of complete URIs that support non-default protocols, such as [CONNECT-UDP] and [CONNECT-IP].¶

:method = GET
:scheme = https
:authority = proxy.example.org
:path = /.well-known/pvd
accept = application/pvd+json

:status = 200
content-type = application/pvd+json
content-length = 222

{
  "identifier": "proxy.example.org.",
  "expires": "2023-06-23T06:00:00Z",
  "prefixes": [],
  "proxies": [
    {
      "protocol": "http-connect",
      "proxy": "proxy.example.org:80"
    },
    {
      "protocol": "connect-udp",
      "proxy": "https://proxy.example.org/masque{?target_host,target_port}"
    }
  ]
}

4. Split DNS information for proxies

Split DNS configurations are cases where only a subset of domains is routed through a VPN tunnel or a proxy. For example, IKEv2 defines split DNS configuration in [IKEV2SPLIT].¶

PvD Additional Information can be used to indicate that a proxy PvD has a split DNS configuration.¶

Section 4.3 of [PVDDATA] defines the optional dnsZones key, which contains searchable and accessible DNS zones as an array of strings.¶

When present in a PvD Additional Information dictionary that is retrieved for a proxy as described in Section 2, domains in the dnsZones array indicate specific zones that are accessible using the proxy. If a hostname is not included in the enumerated zones, then a client SHOULD assume that the hostname will not be accessible through the proxy.¶

Entries listed in dnsZones MUST NOT expand the set of domains that a client is willing to send to a particular proxy. The list can only narrow the list of domains that the client is willing to send through the proxy. For example, if the client has a local policy to only send requests for "example.com" to a proxy "proxy.example.com", and the dnsZones array contains "internal.example.com" and "other.company.com", the client would end up only proxying "internal.example.com" through the proxy.¶

:method = GET
:scheme = https
:authority = proxy.example.org
:path = /.well-known/pvd
accept = application/pvd+json

:status = 200
content-type = application/pvd+json
content-length = 135

{
  "identifier": "proxy.example.org.",
  "expires": "2023-06-23T06:00:00Z",
  "prefixes": [],
  "dnsZones": ["internal.example.org"]
}

5. Discovering proxies from network PvDs

[PVDDATA] defines how PvD Additional Information is discovered based on network advertisements using Router Advertisements [RFC4861]. A network defining its configuration via PvD information can include the proxies key (Section 3) to inform clients of a list of proxies available on the network.¶

This association of proxies with the network's PvD can be used as a mechanism to discover proxies, as an alternative to PAC files. However, client systems MUST NOT automatically send traffic over proxies advertised in this way without explicit configuration, policy, or user permission. For example, a client can use this mechanism to choose between known proxies, such as if the client was already proxying traffic and has multiple options to choose between.¶

Further security and experience considerations are needed for these cases.¶

6. Security Considerations

Configuration advertised via PvD Additional Information, such DNS zones or associated proxies, can only be safely used when fetched over a secure TLS-protected connection, and the client has validated that that the hostname of the proxy, the identifier of the PvD, and the validated hostname identity on the certificate all match.¶

7. IANA Considerations

Authors' Addresses