This document might eventually specify a way to mark messages (or parts of them) for unreliable delivery in HTTP. For now, it just enumerates some design assumptions in Section 1.1 to seed discussion.
RFC EDITOR: please remove this section before publication
The issues list for this draft can be found at https://github.com/mnot/I-D/labels/unreliable-http.
The most recent (often, unpublished) draft is at https://mnot.github.io/I-D/unreliable-http/.
Recent changes are listed at https://github.com/mnot/I-D/commits/gh-pages/unreliable-http.
See also the draft’s current status in the IETF datatracker, at https://datatracker.ietf.org/doc/draft-nottingham-unreliable-http/.
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 http://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 October 1, 2020.
Copyright © 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 (http://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.
Some HTTP applications might benefit from being able to tell the protocol that delivery of data need not be reliable; that is, rather than offering guaranteed end-to-end transmission of the entire message payload, it is preferable to forgo this, because the retransmission of lost data is harmful (or just not useful).
For example, in some cases retransmission of audio or video data might be counterproductive, because use of the data is time-sensitive, and retransmission of loss only competes with more current data “on the wire.”
This document might eventually specify a way to do this in HTTP. For now, it just enumerates some design assumptions in Section 1.1 to seed discussion.
This document asserts that the following assumptions are necessary for a successful unreliable HTTP mechanism. If you disagree, please discuss on the HTTP WG mailing list, or the issues list above.
An unreliable HTTP delivery mechanism should:
- Allow unreliable delivery both on requests and responses. While the response side is the most obvious target, requests such as POST can support interesting use cases too.
- Additionally, provide a way to establish a channel within the connection for routing unreliable data within.
- Focus on unreliable delivery of the message body. Header and trailers need to be reliable.
- Be triggered by an in-protocol mechanism, like a header field or request method. Requiring implementations to have out-of-band knowledge hurts deployment.
- Be able to fall back to “normal” reliable HTTP on hops that don’t support unreliable delivery. This implies that unreliable delivery is an optimisation, not an application semantic.
- Provide a way for endpoints to advise their peer whether they will be sending unreliable data on this connection; likewise, whether they are capable of receiving it.
- Give some level of feedback to both ends about whether unreliable delivery is in use, end-to-end. Probably also optional loss stats, hop-by-hop. This might be through a header like [I-D.ietf-httpbis-proxy-status].
- Provide some way to guarantee that application data is delimited at certain boundaries, to add application loss handling. This might be just a convention that DATAGRAM frames are never combined or split, or it might be something in-protocol.
- Provide out-of-order delivery by default. Possibly, provide a mechanism (e.g., sequence numbers) for applications that desire in-order delivery.
3.1. Normative References
- Bradner, S., “Key words for use in RFCs to Indicate Requirement Levels”, BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997, <https://www.rfc-editor.org/info/rfc2119>.
- Leiba, B., “Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words”, BCP 14, RFC 8174, DOI 10.17487/RFC8174, May 2017, <https://www.rfc-editor.org/info/rfc8174>.