Skip to content

Wider QUIC adoption #90

Description

@yaroslavros

More QUIC, How?

Context

At IETF 124, Lars Eggert organised a side meeting titled "More QUIC, How?" to discuss the current state of QUIC adoption on the Internet. The discussion surfaced a number of useful data points and highlighted a broad set of deployment challenges across different parts of the ecosystem (clients, servers, networks, and intermediaries).

While the session was productive, there has been no structured follow-up or consolidation of findings that could inform broader architectural or operational guidance.

Why

QUIC is now a foundational transport for HTTP/3 and other emerging protocols, with clear technical advantages over TCP+TLS, including:

  • Improved latency and user experience
  • Resilience via connection migration
  • Native support for multiplexing without head-of-line blocking
  • Support for both reliable streams and unreliable datagrams within a single connection

Despite these benefits, adoption remains uneven across the Internet. Anecdotally and in operator discussions, barriers include:

  • Operational visibility and troubleshooting challenges
  • Middlebox interference and network ossification concerns
  • Deployment complexity (e.g., load balancing, observability, fallback strategies)
  • Unclear incentives or risk tradeoffs for certain operators and enterprises

This raises a broader architectural question: what is preventing QUIC from becoming the default transport where it is technically superior?

Next Step

A technical IAB discussion to understand and decide possible next steps

  • Whether this topic deserves a dedicated workshop
  • Scope and framing
  • Potential contributors and stakeholders
  • Expected outputs

Reference

Relevant discussion on QUIC mailing list

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions