However, I realize that the problems with mobile Internet performance and reliability are real. So when HTTP/3 is stable, I'll do what I can to help it spread.
However, I realize that the problems with mobile Internet performance and reliability are real. So when HTTP/3 is stable, I'll do what I can to help it spread.
As for HTTP/3... so will we. See: https://blog.cloudflare.com/http-3-from-root-to-tip/, https://blog.cloudflare.com/the-road-to-quic/ and https://blog.cloudflare.com/head-start-with-quic/.
I like and use 1.1.1.1 though.
I obviously don't know how many Cloudflared sites I visit that don't pop up the nag. And Cloudflare's nag is certainly nicer than Google's more pervasive help-us-build-a-T-800 or Akamai's "just get lost". But that mode seemingly activates on light browsing just because it's coming from a slightly-less-trackable VPS address (non-shared), and that is a problem.
This might have changed but in the past it made using Tor for anything beyond onion sites extremely annoying.
Yes there's the argument that TOR provides protection for those in apressive states, but given the pros/cons of blocking TOR altogether I can at least understand the reasoning.
However, Cloudflare has also adopted and promoted at least one standard that adds complexity for dubious benefit, specifically DNSSEC, which tptacek has repeatedly criticized (e.g. [1]).
Moreover, Cloudflare is encouraging both providers and consumers to bypass the public Internet as much as possible in favor of Cloudflare's network and proprietary protocol(s). For providers, this is done through Argo and especially Argo Tunnel. And now for consumers, Warp is replacing the standard TCP with a proprietary protocol built on UDP.
Now that Cloudflare has proprietary replacements for the standard Internet on both sides, it can start taking advantage of network effects to make its proprietary network attractive to still more providers and consumers. As Cloudflare's power grows, it becomes harder to escape any future abuses of that power, as well as honest mistakes on Cloudflare's part.
I realize the standard Internet sucks in some ways, and Cloudflare is doing something about that. But I think the right answer is to improve the standards-based Internet, not offer a proprietary replacement. I suppose that's not compatible with running a VC-backed business, though.
I think that applies almost anywhere. One could say "don't trust Google or Facebook with personal data" merely based on the fact that almost all of their money comes from advertising.
And thanks for the interesting exchange!
FWIW, tptacek's argument in that thread seems to be premised on certificate pinning being widely deployed[1], which it's not, and it seems at this point like it never will be[2].
[1]: https://news.ycombinator.com/item?id=10553608
[2]: https://groups.google.com/a/chromium.org/forum/#!msg/blink-d...
It hardly matters at this point, though. DNSSEC is a dead letter. It's over. Stick a fork in it. It'll be around indefinitely for performative nerds to performatively noodle with --- lots of dead IETF protocols are! --- but Cloud Flare is likely to be the largest company ever to use it (and they're the exception that proves the rule, since they sell DNSSEC services).
If upstream is doing something you don't like and refusing to work with you, sure.
When upstream actively petitions you to not fork, asks you politely to work together, and you refuse to work with them, that is far, far from a "tried and true open source software process". That creates a fissure in the community and it generally ends up poorly for everyone involved.
My comment is far from inflammatory, it's a statement of fact, and something cloudflare has refused to acknowledge or respond to. Which just further drives the point home that they aren't acting in good faith.
EDIT: eastdakota filled me in, thanks John.
https://blog.cloudflare.com/boringtun-userspace-wireguard-ru...
We communicated with Jason throughout the process and have a ton of respect for him and the entire WireGuard community. In the short term, we need the flexibility to quickly update BoringTun's code base to support the project we built it for. That's harder when you need to coordinate with people outside Cloudflare and when we need to move as fast as we plan to. However, we really believe in Open Source and want the WireGuard community to thrive. We licensed the code very openly (3-paragraph BSD) and WireGuard may choose to fork it. If they do, we'll support it and plan to contribute any improvements in our own fork back. Over the long term, I think we're very open to merging this back into the upstream project.
>I thought the invitation to put their engineers as the head of a WireGuard subproject was a cool invitation, but alas.
https://lists.zx2c4.com/pipermail/wireguard/2019-March/00404...
I mean no offense, but the response comes off as corporate approved PR. "We need to move fast" when you haven't actually even tried engaging with the parent project and have no idea whether or not it would prohibit "moving fast" is disingenuous IMO.
More importantly, without having already tried it, it’s hard to predict how much overhead there will be.
Since CloudFlare had a (self-imposed) deadline, working fast had to take priority over optics. After all, the project can always be folded into the WireGuard organization later.
Implementing a standard without regard for the beliefs of other implementors is an action that supports a standard. Refusing to work with others does not implicitly harm a standard.
You assert that refusing to cooperate with another implementor is guaranteed to harm a standard. It is not guaranteed at all.
DJB has not destroyed DNS. BoringSSL has not destroyed TLS. A thousand reimplementations of standards in Rust have not destroyed a thousand standards.
You clearly believe that Cloudflare is acting in bad faith, and are constructing a worldview out of assumptions that you declare instead are facts. While I respect your right to hold those views, I do not respect your declaration of future outcomes as fact.
DJB didn't fork Bind and then refuse to work with them.
>BoringSSL has not destroyed TLS
BoringSSL didn't fork OpenSSL and then refuse to work with them.
About the closest modern comparison would be OpenOffice vs. LibreOffice - which created a complete mess like I mentioned before.
Except even THAT is a bad comparison because LibreOffice only forked when they were FORCED to fork.
WireGuard is written as Kernel Module in C, with a GPL licence; BoringTun is a user space program written in Rust with an MIT licence.
So it’s not really even a fork.
So one could argue you are both pushing the latest standards and the latest nonsense. ;)
Hence the ;) face, it's meant as a friendly jab, not a critical accusation. jgrahamc is awesome.
If they cannot then it is not the internet. It's more akin to a 'web' only service.
CGNAT means that the same is true of "mobile" connections in general, so it's not like Warp is changing anything for the worse here. Though the Tor network does allow you to host a .onion-linked service over such a connection, but that - while quite handy - seems more like a special case to me.
[1]: https://en.wikipedia.org/wiki/Embrace,_extend_and_extinguish
[2]: But then, sorry how this sounds, but pessimists tend to think, EEE perpetrators wouldn't publicly admit to it either...
Pretty much most if not all of Cloudflare's services and work suggest the complete opposite to me.
Like other commenters, Cloudflare for me is probably one of the only companies I truly trust. I'm not saying that because I'm a big user of there services in fact 1.1.1.1 is the only service I actively use.
Mind you, I'm still skeptical. I probably won't use Warp on my phone, or Cloudflare on my personal site. But I should have been more careful about how I expressed that skepticism in public. None of us want a world where we all assume the worst in each other without strong evidence. So again, I'm sorry.
Having worked in an ISP, only one thing mattered to costumers, and only one thing: YouTube.
Yes. Agreed. But if not Cloudflare as a pushback alternative to those trying to own the internet, then who?
It seems to me the "standard internet" is getting smaller and smaller. What other options do we have?
If anything, I've kinda been hoping Cloudflare would realize self-hosting and decentralization is what they should be supporting and pushing, as it's when using their CDN makes the most sense. And obviously, Amazon and Google and Microsoft all have their own CDN capabilities, so the less people using their cloud services, the better for Cloudflare.