443 is Enough: Guidance on Port Allocation for HTTP-based Services
draft-trammell-tsvwg-443-is-enough-00
This document is an Internet-Draft (I-D).
Anyone may submit an I-D to the IETF.
This I-D is not endorsed by the IETF and has no formal standing in the
IETF standards process.
| Document | Type | Active Internet-Draft (individual) | |
|---|---|---|---|
| Author | Brian Trammell | ||
| Last updated | 2026-07-23 | ||
| Replaces | draft-trammell-443-is-enough | ||
| RFC stream | (None) | ||
| Intended RFC status | (None) | ||
| Formats | |||
| Stream | Stream state | (No stream defined) | |
| Consensus boilerplate | Unknown | ||
| RFC Editor Note | (None) | ||
| IESG | IESG state | I-D Exists | |
| Telechat date | (None) | ||
| Responsible AD | (None) | ||
| Send notices to | (None) |
draft-trammell-tsvwg-443-is-enough-00
Transport and Services Working Group B. Trammell
Internet-Draft Google Switzerland GmbH
Intended status: Informational 23 July 2026
Expires: 24 January 2027
443 is Enough: Guidance on Port Allocation for HTTP-based Services
draft-trammell-tsvwg-443-is-enough-00
Abstract
[RFC7605] provides guidance on the use of port numbers and the
criteria for new port assignments, including a test for whether a
proposed service is distinct from an existing service. It gives the
example that "an automated system that happens to use HTTP framing --
but is not primarily accessed by a browser -- might be a new
service." It also might not. This document clarifies the
application of the distinct-protocol test in [RFC7605] Section 7.1 to
services built on HTTP as a substrate, in light of HTTP's evolution
since its publication, and provides guidance to applicants and
reviewers on when an HTTP-based service qualifies for a new port
assignment and when it does not.
About This Document
This note is to be removed before publishing as an RFC.
The latest revision of this draft can be found at
https://britram.github.io/443-is-enough/draft-trammell-tsvwg-443-is-
enough.html. Status information for this document may be found at
https://datatracker.ietf.org/doc/draft-trammell-tsvwg-443-is-enough/.
Source for this draft and an issue tracker can be found at
https://github.com/britram/443-is-enough.
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/.
Trammell Expires 24 January 2027 [Page 1]
Internet-Draft 443 is Enough July 2026
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 24 January 2027.
Copyright Notice
Copyright (c) 2026 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 . . . . . . . . . . . . . . . . . . . . . . . . 2
2. HTTP as an Application Transport Substrate . . . . . . . . . 3
3. Evaluating HTTP-Based Protocols for Distinctness . . . . . . 4
4. Application Naming Beyond Port Numbers . . . . . . . . . . . 5
5. Security Considerations . . . . . . . . . . . . . . . . . . . 6
6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 6
7. References . . . . . . . . . . . . . . . . . . . . . . . . . 6
7.1. Normative References . . . . . . . . . . . . . . . . . . 6
7.2. Informative References . . . . . . . . . . . . . . . . . 6
Disclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 8
Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 8
1. Introduction
[RFC7605] provides guidance on when a new port assignment is
warranted, including a distinctness test in Section 7.1: a new
service merits an assignment only if an unmodified client of an
existing service cannot interact with it.
In the decade since that document was published in 2015, HTTP has
become an overwhelmingly popular de facto substrate for application
protocol design -- a development that [RFC9205] both documents and
embraces. Section 7.1's observation that "an automated system that
happens to use HTTP framing -- but is not primarily accessed by a
Trammell Expires 24 January 2027 [Page 2]
Internet-Draft 443 is Enough July 2026
browser -- might be a new service" was intended to leave room for
novel cases. The evolution of and investment in the HTTP ecosystem
since then has only made the use of HTTP and the ecosystem
surrounding it as a substrate more attractive. One practical
consequence of this development has been some confusion about whether
new protocols over HTTP are new protocols in the sense of "requiring
a port assignment".
This document clarifies how the [RFC7605] Section 7.1 distinctness
test applies to HTTP-based services, and provides guidance to
applicants and reviewers on when it is and is not satisfied. It does
not replace [RFC7605], but rather updates its application to reflect
the maturity of the HTTP ecosystem. Specifically, it addresses how
modern deployment patterns—such as ubiquitous TLS, SNI, ALPN, and
user-space demultiplexing—have resolved the practical cohabitation
and access control issues that have previously motivated requests for
dedicated port assignments.
2. HTTP as an Application Transport Substrate
HTTP has evolved since its origins as the basis of the World Wide
Web. HTTP/2 [RFC9113] redesigned HTTP's wire format around
multiplexed binary framing with non-browser use as an explicit design
goal; HTTP/3 [RFC9114] continues this evolution over QUIC [RFC9000].
[RFC9205] provides detailed guidance on building new protocols beyond
the web atop HTTP. The benefits of this approach are substantial:
HTTP-based services can leverage existing infrastructure including
reverse proxies, load balancers, content delivery networks, and
firewalls; they interoperate naturally with web clients; and they
inherit well-established security properties including TLS
certificate management and authentication frameworks.
Operating on standard web ports (80 and 443) also improves
compatibility with existing network tooling—such as packet analyzers
and diagnostic tools that are pre-configured for HTTP—and maximizes
the likelihood of traversing firewalls that restrict outbound traffic
to standard web ports. Such traversal is not guaranteed: firewalls
increasingly apply deep packet inspection and application-behavior
analysis to traffic on ports 80 and 443, so using these ports offers
the best chance of traversal rather than a certainty of it.
Furthermore, reusing these ports directly supports transport port
conservation, a key goal of [RFC7605].
The HTTP ecosystem also provides a rich set of mechanisms for service
differentiation, discovery, and coexistence that do not require
dedicated port assignments. Multiple independent services can share
ports 80 and 443 concurrently on the same host using path-based
routing (via reverse proxies or API gateways), host-based routing
Trammell Expires 24 January 2027 [Page 3]
Internet-Draft 443 is Enough July 2026
(via TLS Server Name Indication (SNI) [RFC6066]), or protocol-based
multiplexing (via Application-Layer Protocol Negotiation (ALPN)
[RFC7301]). These user-space demultiplexing techniques are standard
in modern deployments, resolving the "first binder wins" problem
inherent in OS-level transport-layer demultiplexing. Additionally,
sharing these ports allows network operators to leverage Layer 7
security policies (such as SNI- or ALPN-based filtering) rather than
relying on port-based firewall rules.
A service that requires a new ALPN identifier should register it in
the IANA TLS ALPN Protocol IDs registry, not seek a new port
assignment.
3. Evaluating HTTP-Based Protocols for Distinctness
Section 7.1 of [RFC7605] establishes one useful test for whether a
proposed service warrants a new port assignment: can an unmodified
client of an existing service interact with the proposed service?
Interoperability implies non-distinctness, and a non-distinct
protocol does not merit a new assignment.
For HTTP-based services, this test is easy to implement: can an
unmodified generic HTTP client tool such as curl issue requests to
and receive valid responses from the proposed service? Service
differentiation achieved through URL path structure, HTTP header
values, Content-Type negotiation, payload schema, or authentication
scheme does not constitute wire-level distinctness; these are
application-layer conventions carried within HTTP, not independent
protocols.
This does not mean that all HTTP-based protocols are indistinct.
Examples that might warrant an assignment include:
* a REST API running over the same TLS connection as a protocol with
a different wire format, using a protocol-specific multiplexing
scheme; or
* a protocol running on UDP or SCTP that uses a REST API over TCP as
a control or management plane.
The former would not interoperate with an unmodified client, and the
latter has incomplete semantics when used as such.
Similarly, protocols that run natively over QUIC [RFC9000] but do not
use HTTP semantics are distinct from HTTP. While HTTP/3 [RFC9114]
runs over QUIC on port 443, a protocol that uses QUIC as a transport
layer directly (without the HTTP mapping defined in [RFC9114]) is a
different service. For example, DNS-over-QUIC (DoQ) [RFC9250] runs
Trammell Expires 24 January 2027 [Page 4]
Internet-Draft 443 is Enough July 2026
directly over QUIC and is assigned a dedicated port (853), whereas
DNS-over-HTTPS (DoH) [RFC8484] layers DNS queries within HTTP
sessions and runs over standard web ports.
That a protocol is wire-distinct from HTTP does not by itself imply
that it requires a new port assignment, however. Because ALPN allows
multiple protocols to share a port, a non-HTTP protocol running
directly over QUIC may coexist with HTTP/3 on port 443, selected by
its own ALPN identifier rather than "h3"; media delivery protocols
that can operate either over HTTP/3 or directly over QUIC are one
example. Such a protocol should generally register an ALPN
identifier rather than request a port. DoQ's assignment of a
dedicated port (853) reflects its role as a transport-layer companion
to DNS-over-TLS, which already occupies that port, rather than a
general expectation that protocols running natively over QUIC receive
dedicated ports.
A related case involves hybrid protocols that use a UDP-based
transport for primary data transfer but rely on an HTTP-based REST
API for control, management, or bootstrap operations. In these
cases, if the UDP component clearly warrants a dedicated port
assignment on its own, the protocol designer should consider the
tradeoffs of using the corresponding TCP port for the control plane
versus the advantages of using standard web ports.
4. Application Naming Beyond Port Numbers
A service built on these substrates may be identified in more than
one IANA registry: a service name in the Service Name and Transport
Protocol Port Number Registry (used, for example, in DNS SRV records
[RFC2782]); an ALPN protocol identifier in the TLS Application-Layer
Protocol Negotiation (ALPN) Protocol IDs registry [RFC7301]; and an
underscored node name in the Underscored and Globally Scoped DNS Node
Names registry [RFC8552]. These registries evolved independently and
are not coordinated. A single service may require entries in several
of them, with no guarantee that a chosen name is available or
consistent across all three. Rationalizing these namespaces is out
of scope for this document; the point here is that a port assignment
is only one of several forms of registration a service may need, and
often not the one most relevant to how the service is actually
selected on the wire.
Trammell Expires 24 January 2027 [Page 5]
Internet-Draft 443 is Enough July 2026
5. Security Considerations
The intended effect of the guidance given by this document is
effectively to reduce port assignments for HTTP-based services,
directing these services to use port 443 rather than dedicated port
assignments. This has implications for overall network security:
traffic from non-Web HTTP applications running on port 443 is less
distinguishable from other traffic in the face of metadata
examination. TLS deployment is more likely to be properly configured
when services share the standard HTTPS port and its associated
certificate management infrastructure, and network operators can
apply consistent security policy across all services on that port.
[RFC9205] Section 4.4.3 notes that deploying an HTTP-based
application on a non-default port carries privacy implications
because the protocol becomes distinguishable from other traffic; the
guidance in this document is consistent with minimizing that
distinguishability.
6. IANA Considerations
This document has no IANA actions. It is intended as guidance for
IANA Transport Port Expert Reviewers.
7. References
7.1. Normative References
[RFC7605] Touch, J., "Recommendations on Using Assigned Transport
Port Numbers", BCP 165, RFC 7605, DOI 10.17487/RFC7605,
August 2015, <https://www.rfc-editor.org/rfc/rfc7605>.
[RFC9205] Nottingham, M., "Building Protocols with HTTP", BCP 56,
RFC 9205, DOI 10.17487/RFC9205, June 2022,
<https://www.rfc-editor.org/rfc/rfc9205>.
7.2. Informative References
[RFC2782] Gulbrandsen, A., Vixie, P., and L. Esibov, "A DNS RR for
specifying the location of services (DNS SRV)", RFC 2782,
DOI 10.17487/RFC2782, February 2000,
<https://www.rfc-editor.org/rfc/rfc2782>.
[RFC6066] Eastlake 3rd, D., "Transport Layer Security (TLS)
Extensions: Extension Definitions", RFC 6066,
DOI 10.17487/RFC6066, January 2011,
<https://www.rfc-editor.org/rfc/rfc6066>.
Trammell Expires 24 January 2027 [Page 6]
Internet-Draft 443 is Enough July 2026
[RFC6455] Fette, I. and A. Melnikov, "The WebSocket Protocol",
RFC 6455, DOI 10.17487/RFC6455, December 2011,
<https://www.rfc-editor.org/rfc/rfc6455>.
[RFC7301] Friedl, S., Popov, A., Langley, A., and E. Stephan,
"Transport Layer Security (TLS) Application-Layer Protocol
Negotiation Extension", RFC 7301, DOI 10.17487/RFC7301,
July 2014, <https://www.rfc-editor.org/rfc/rfc7301>.
[RFC8484] Hoffman, P. and P. McManus, "DNS Queries over HTTPS
(DoH)", RFC 8484, DOI 10.17487/RFC8484, October 2018,
<https://www.rfc-editor.org/rfc/rfc8484>.
[RFC8552] Crocker, D., "Scoped Interpretation of DNS Resource
Records through "Underscored" Naming of Attribute Leaves",
BCP 222, RFC 8552, DOI 10.17487/RFC8552, March 2019,
<https://www.rfc-editor.org/rfc/rfc8552>.
[RFC8615] Nottingham, M., "Well-Known Uniform Resource Identifiers
(URIs)", RFC 8615, DOI 10.17487/RFC8615, May 2019,
<https://www.rfc-editor.org/rfc/rfc8615>.
[RFC9000] Iyengar, J., Ed. and M. Thomson, Ed., "QUIC: A UDP-Based
Multiplexed and Secure Transport", RFC 9000,
DOI 10.17487/RFC9000, May 2021,
<https://www.rfc-editor.org/rfc/rfc9000>.
[RFC9110] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke,
Ed., "HTTP Semantics", STD 97, RFC 9110,
DOI 10.17487/RFC9110, June 2022,
<https://www.rfc-editor.org/rfc/rfc9110>.
[RFC9113] Thomson, M., Ed. and C. Benfield, Ed., "HTTP/2", RFC 9113,
DOI 10.17487/RFC9113, June 2022,
<https://www.rfc-editor.org/rfc/rfc9113>.
[RFC9114] Bishop, M., Ed., "HTTP/3", RFC 9114, DOI 10.17487/RFC9114,
June 2022, <https://www.rfc-editor.org/rfc/rfc9114>.
[RFC9250] Huitema, C., Dickinson, S., and A. Mankin, "DNS over
Dedicated QUIC Connections", RFC 9250,
DOI 10.17487/RFC9250, May 2022,
<https://www.rfc-editor.org/rfc/rfc9250>.
Trammell Expires 24 January 2027 [Page 7]
Internet-Draft 443 is Enough July 2026
Disclosure
LLM-based tools (Claude Sonnet in Claude Code, Gemini Flash in
Antigravity) were used in the workflow management, reference and
archival research, initial draft generation, and editorial review of
this document, in part as an evaluation of the readiness of these
tools for such tasks.
Acknowledgments
The author would like to thank Wesley Eddy, Michael Scharf, Joe
Touch, and Christian Huitema for the feedback and input that improved
this document.
Author's Address
Brian Trammell
Google Switzerland GmbH
Gustav-Gull-Platz 1
CH-8004 Zürich
Switzerland
Email: ietf@trammell.ch
Trammell Expires 24 January 2027 [Page 8]