On Wed, 26 Jun 2024 at 01:34, Lucas Pardue <[email protected]> wrote:
> > This entire thread started from quite an odd position of question. With > one reading of it being there might be an expectation that all traffic > should run over QUIC on all networks. And by extension, any measurements > that don't show ~100% usage are some indicator of a problem. I believe that > to be a fallacy. > To be clear, I didn't expect 100% QUIC either. I was expecting consumer ISPs be over 50%, now that we're three years after RFC9000. I thought this would be driven by the actions of a few major clients and CDNs. The fact that it hasn't, gave me reason to ask what the obstacles might be. The QUIC WG standardised a new transport protocol and application protocol > that fits seamlessly into the mix of live Internet use cases and existing > technologies. The diversity of the Internet means there are endless > opportunities, constraints and rationale for selecting technologies at any > given moment. > Totally agree. > Looking at top level metrics and drawing conclusions risks mistaken > conclusions. Stuff is complicated and multi-faceted. There's no magic > bullet here. I encourage folks to dig deeper in order to find specific > challenges or opportunities that can be actioned through engineering, > standardisation or advocation. > I do realise this. Starting the thread was my attempt at beginning the dig. I think we've unearthed some objects for study: - Perception of risk (e.g. QUIC might degrade some corner cases) can lead to a model of requiring manual opt-in. I'm guessing that over time accumulated experience will reduce that perception. - There's room for clients to be more proactive in attempting H3. It's not clear why Alt-Svc plus HTTPS isn't leading to the selection of H3. - H3 has a "tough deployment journey" through the ecosystem. That's the way it is. I don't know how to make this easier, although others may do. Finally it's clear that Google and Meta are heavy users of QUIC (presumably successfully), and it's possible to use QUIC across ~97% of paths that Meta deliver over. Which is a little better than I was expecting. Chris
