HTTP Resources with User Names
draft-vanrein-http-unauth-user-04

Abstract

Many protocols support users under domain names, but HTTP does not. Usage patterns in the wild do suggest a desire to have this facility. This specification defines a header for user names, orthogonal to authenticated identities.

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 August 3, 2020.

Copyright Notice

Copyright (c) 2020 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.

Table of Contents

• 1. Introduction
• 2. The HTTP User Header
• 3. URI with User Information
• 4. Protocol Handling of URI and User
• 5. Environment Variable with Local User
• 6. Orthogonality of Authentication (Example)
• 7. IANA Considerations
• 8. Security Considerations
• 9. References
• 9.1. Normative References
• 9.2. Informative References
• Appendix A. Compatibility with Basic Authentication
• Appendix B. Compatibility with RESTful Patterns
• Appendix C. Compatibility with Caching
• Appendix D. Acknowledgements
• Author's Address

1. Introduction

Many protocols support global identities like john@example.com to identify users like john under domains such as example.com. The URI format for HTTP can express [Section 2.7.1 of [RFC7230]] such authority components, with an intended interpretation of locating a user-specific resources. Many online applications publish about individual users, but there is no standard in HTTP to express user names to address them. This specification therefore introduces a header "User", in close parallel to the "Host" header.

Some current URIs for HTTP have used the userinfo field in the URI to express an authentication user, in spite of the intended use to refine the authority information. This conflates the user part of the resource with the (authenticated) client-side identity. This specification makes authentication and resource user two orthogonal concepts, and specifies a clear relation to the URI format.

Orthogonality yields a generalisation, but voluntary conflation of the user concepts remains possible, and may even be useful as default behaviour. Servers can be configured with resource users that demand authentication with the same client name; they may even accept authentication as a bypass to a resource user. Clients may accept authentication requests and use their submitted resource user as a hint to the authentication user.

Orthogonal concepts can however be untied to support advanced use cases, of which these practices are a specific form. Servers may welcome members to authenticate against a shared/group resource. Clients treat the resource user as part of the realm identity and can use it to locate the customary authentication user for that realm. Credentials need not be tied to the server, so it is possible to "bring your own identity" by facilitating realm crossover for credentials.

The purpose of this specification is to define clear meaning for HTTP URIs with a userinfo component and their mapping to HTTP.

2. The HTTP User Header

The "User" header field provides an aspect of the desired resource name scope. The value can be inserted by software or be taken from the visible portion of the userinfo component [Section 3.2 of [RFC3986]] of the target URI but it MUST NOT include a ":" colon (U+003a) character.

User = *( unreserved / pct-encoded / sub-delims )

The User header value holds precisely one value with the following ABNF grammar: [RFC3986] and can be directly included in the quoted-string header form; a plain token cannot express "(", ")", "=", ";" and "," without escaping [Section 3.2.6 of [RFC7230]].

3. URI with User Information

This section is informative.

The use of a user name in the authority component of a URI is a general concept, already used for addressing users with SMTP, SIP, XMPP and many other protocols. By allowing users in HTTP resource names, better crossover of identities with these protocols can be achieved. Unlike server-specific user name mapping conventions, the specified generic meaning of URI userinfo as part of the authority information enables such crossovers to be automated.

There is a current practice of writing an authentication name in the userinfo portion of a URI. This can be useful if it adds Basic authentication to the first request; Basic can do this because it does not incorporate a server-sent challenge.

Having never been standardised, the mechanistic side of this practice is highly diverse, and URIs are far from portable between browsers or even the various places where they occur within one browser. As a result, these URIs cannot be spread freely and their usage pattern is dedicated to the client software in use. On grounds of this immobility, it would have been sufficient to not break this practice. What this specification does is completely support [Appendix A] this pattern as a special case.

This specification follows the URI's intention of the userinfo field, and prescribes copying its value into the User header. It will however remove anything from a colon onward, to suppress the portion of userinfo that is prescribed to not be rendered.

Both the general userinfo format and the removal of anything starting from a colon may lead to an empty string as the User header value. Servers may interpret this empty User header as they see fit; the underlying resource may be a default page, for example one that asks for a user name to redirect to. After client authentication, such a page may list user names that would be accessible.

4. Protocol Handling of URI and User

User agents are supposed to render userinfo in the target URI (up to the first colon ":" (U+003a) if it contains any [Section 3.2.1 of [RFC3986]]) where it may be helpful if it stands out graphically [Section 7.6 of [RFC3986]]. User agents SHOULD NOT remove userinfo from the target URI.

Compliant user agents MUST pass userinfo in the target URI (up to the first colon ":" (U+003a) if it contains any) as a User header field if, and only if, the target URI contains a userinfo part.

The User header MAY appear in requests and MUST NOT occur in responses. When sending it, the user agent SHOULD generate User as the next header field after Host. Transparent intermediates such as proxies and caches MUST NOT add, remove or modify the User header. The CONNECT method and Host header both exclude this information, so the User header complements them.

Compliant servers MAY ignore the User header [Section 3.2.1 of [RFC7230]] and they MAY impose a more restrictive grammar (like a NAI's utf8-username [RFC7542]) than the URI syntax before further processing it. When they do use it, the Effective Request URI [Section 5.5 of [RFC7230]] MUST be constructed with the userinfo and the "@" at delimiter (U+0040) prefixed to the host name and optional port. Realms are specific to an authority section [Section 2.2 of [RFC7235]] and so a realm never spans across different userinfo values.

As a result of a consistent translation of any User header value into the Effective Request URI, the server would map consistently to resources. It is merely enabled to include a User header as an extra input variable to this mapping to resources.

HTTP caches [RFC7234] derive no privacy or security concerns from the User header, but they do need to differentiate requests based on it. To accommodate that, the Vary header [Section 7.1.4 of [RFC7231]] MUST be generated by the server in the matching response, and the header MUST either be a single "*" star (U+002a) or list the "User" name, for all responses whose processing was influenced by the User header. This requirement has no bearing on software and configurations that ignore the User header.

During redirects or other traversals to relative HTTP URIs, compliant user agents MUST overwrite the userinfo in the target URI when the new URI specifies an authority component, and MUST keep it otherwise.

5. Environment Variable with Local User

The following variable SHOULD be passed to applications that run on top of the HTTP stack in a server:

LOCAL_USER

gives the HTTP User header value after syntax checking and percent-decoding. Like the customary variables HTTP_HOST and PATH_INFO, this specifies the resource being requested. The HTTP_USER header does not describe the identity of the client-side user, which usually lands in a variable REMOTE_USER after authentication.

6. Orthogonality of Authentication (Example)

This section is informative.

The authority user in a URI refines the server's resource name space, and the server is free to distinguish this resource user from an authenticated client identity. The example below demonstrates the expressive power gained by treating these user concepts as separate, in an advanced pattern of use in which an individual user accesses a group resource. This advanced binding between the group resource and an authenticated client can build on access control lists.

GET /docs HTTP/1.1
Host: example.com
User: sales

HTTP/1.1 301 Moved Permanently
Location: /docs/
Vary: User

GET /docs/ HTTP/1.1
Host: example.com
User: sales

HTTP/1.1 401 Unauthorized
WWW-Authenticate: KnockKnock realm="Documents"
Vary: User

GET /docs/ HTTP/1.1
Host: example.com
User: sales
Authorization: KnockKnock realm="Documents", user="mary", ...

HTTP/1.1 200 OK
Vary: User
Content-Type: text/html

...
<a href="/docs/review.cgi?docid=123">Review 123 now</a>
...

GET /docs/review.cgi?docid=123 HTTP/1.1
Host: example.com
User: sales
Authorization: KnockKnock realm="Documents", user="mary", ...

John and Mary both work at the Sales department of Example, Inc and John has written a document that he wants Mary to review. Mary opens a link to the document name space under the shared sales account at https://sales@example.com/docs and her user agent sends:

7. IANA Considerations

Header Field Name   Template   Protocol   Status    Reference
------------------  ---------  ---------  -------   ----------
User                           http       TBD       TBD:THIS_SPEC

IANA adds the following entry to the Message Headers registry:

8. Security Considerations

The User header field as defined herein is orthogonal to issues of authentication and authorisation, and adds no security concerns.

9. References

9.1. Normative References

9.2. Informative References

Appendix A. Compatibility with Basic Authentication

This appendix is informative.

Basic authentication is regularly used as a quick and easy HTTP authentication technique. Several user agents continue to support it with the "user:password@" URI prefix to a hostname, despite its deprecation [Section 3.2.1 of [RFC3986]]. This specification imposes no new constraints on this practice; it merely prescribes to send the User header field, and leaves it to client software whether it also sends Basic authentication.

The mapping from authority names to resources is the prerogative of the server. A server supportive of resource selection through Basic authentication could ignore the User header field and still comply with this specification. A server that does recognise the User header field would use it to locate a resource, before deciding about access control to that resource; it may subsequently require authentication, and select schemes that could be supported. At this time, it may or may not welcome an added Basic authentication attempt. All this depends on server configuration.

This flexibility can support a transition from Basic authentication to User headers on the server, and allows client software to also migrate by first adding the User header, and later supporting the advanced uses by allowing differentiation between resource user and authentication user. Server administrators have a free choice whether to gradually phase out older clients or to continue to support them.

Sending both the User header and Basic authentication is only to be expected from user agents who conflate authority userinfo with authentication user. Such user agents will be less flexible, and will not be able to support more advanced usage patterns that separate these concepts, such as shared/group resources addressed with the User header field and, when desired, authentication through an independent set of headers. Among the tools to drive separate authentication could be the negotiation of more advanced mechanisms, even just Digest authentication.

Appendix B. Compatibility with RESTful Patterns

This appendix is informative.

Whether and how the User header is interpreted is the prerogative of the server. The server will translate resources in the same manner when provided with the same User header, and may do so without regard for the method used. The main concern is now if it will be addressed in the same manner in every case. This depends on the user agents.

Development environments make sending the User header field simple, so application support is as easy as the applications are flexible. Binary user agents and ones that may lag behind in updates do however call for backward compatibility support of consistent translation to resources.

Backward compatibility can be guaranteed for host names that always require a User header; such resources may be described with a URI holding a (possibly empty) userinfo field. Failure to send a User header to such resources when the URI holds one can be assumed to have originated from a user agent that fails to support it. When an offer for Basic authentication is presented, it may be interpreted as the conflated approach to userinfo, and treated as a substitute for the User header. If neither is offered, then an error may be reported or control redirected to another means of selecting a resource user, perhaps through an alternate local naming convention.

This indicates that the server is able to detect inconsistent tranlation risks, and avoid accidentally binding a request to an unintended resource as a result of a missing User header.

Appendix C. Compatibility with Caching

This appendix is informative.

Whether and how the User header is used to find resources is the prerogative of their server. A conservative cache design might insert the User header value in request URIs, but lose the capability of seeing the equivalence of a resource as perceived by the server. The inclusion of the name "User" in the Vary header of the response adds explicit non-equivalence information, and thereby provides a more accurate cache controlling instruction.

Whether a result is "private" is independent of the User header, as that only signifies a refinement of the resource name space on the server. The rules that signify authentication as default indicator of privacy is orthogonal to the User header. Independent inclusion of Basic authentication may still invalidate caching, but not as a result of this specification.

Appendix D. Acknowledgements

This specification could be improved thanks to input from Daniel Stenberg, Amos Jeffries, Paul Vixie, James Fuller and Henri Manson.

Author's Address