Hello,
Will Tor Browser ever support newer protocols like QUIC and HTTPS/3?
Also, will there be an implementation that uses GNUnet and GNS?
Will Arti make the Tor browser obsolete or will it eventually be integrated into the Tor Project at a later stage?
Will Tor Browser ever support newer protocols like QUIC and HTTPS/3?
Probably. Right now, the Tor network only allows relaying TCP streams, not UDP datagrams which are quired for QUIC and HTTP/3, but there are people working on making it possible.
Also, will there be an implementation that uses GNUnet and GNS?
As far as I know, no such thing is planned.
Will Arti make the Tor browser obsolete or will it eventually be integrated into the Tor Project at a later stage?
The plan is for Arti to replace tor, which is a component of Tor Browser as well as something you can use standalone. In fact, it’s already possible to use Tor Browser with Arti, though that’s mostly for testing, Arti isn’t ready for prime time just yet.
1 Like
Hello Trinity,
First of all, thank you for that information you just provided - and thanks to TorOps as well, without this thread I wouldn’t have discovered Arti.
I’m really curious about what the actual state of Arti is, what problems have you faced that is slowing down the project and what ideas are yet to be developed since late 2022 was the time to get Arti ready for production client use. Are there any plans to upload more Arti Development meetings on youtube? Last one was in August 2021.
What about us, relay operators?
If Arti is going to eventually replace tor, aren’t there any worries the change will affect tor’s image? Many people have already listened something about tor in their lives (for good or bad), a change of name may scare off new users.
I’m really thankful for you guys, the work that is being put on tor is just amazing. Hopefully, I will soon be able to contribute with more relays or bridges to the network.
And also, will the iconic onion logo be replaced with something more…Arti - like? It just scares me to know that it will replace that name and logo that I’ve known and used for so long. And what name would you give for the onion sites? Will you replace the .onion TLD with something else? It’s just so painful to know that one day the onion name for Tor will be gone. @trinity-1686a , what do you have to say about this?
what problems have you faced that is slowing down the project
As far as I know, there is no major problem slowing down development, however tor wasn’t built in a day. It’s some few hundreds thousands lines of code wrote over about two decades. It takes time to (re)build anything that complex.
what ideas are yet to be developed since late 2022 was the time to get Arti ready for production client use.
You can read release notes of Arti on TorProject blog. Depending on when you last looked, it probably gained support for pluggable transport, and a lot of work has been put into supporting onion services (though that part is still in progress, and should be finished later this year).
Are there any plans to upload more Arti Development meetings on youtube? Last one was in August 2021.
I wasn’t aware they were available on Youtube. I think they are still getting recorded, but I’m not quite sure where they are published. I’ll try to find that for you.
aren’t there any worries the change will affect tor’s image
This isn’t a rebranding, Tor (the network) will keep its name, TorProject too. The only thing which will change is that tor, little-t-tor or c-tor (same thing, different names, because tor on its own is often ambiguous) getting replaced. I think there are two case to differentiate: most users, which don’t know about tor, they have heard about the network, and the browser, but aren’t impacted by implementation details. And on the other hand, power-users, relays operators… who are probably tech-savy enough to understand what these changes mean and what they don’t.
If anything, I think it will make some forms of communication easier, in particular when trying to help people. When someone comes and says they need help with tor, you often have to guess if that’s with the network daemon, or the browser.
2 Likes
Thanks for answering back Trinity,
Now I better understand the work it is needed to be put to get Arti ready. I hope you guys keep getting funded to keep the project on top of your to-do list. I’m really excited about it.
If you can find whether or not meetings are still being published that’d be useful. If you were wondering where they were published, here’s the playlist on youtube.
Alright, since it isn’t a rebranding, I guess anything related to onion won’t change, such as .onion sites.
I don’t get it clear about relay operators. Arti is able to connect to the tor network (which implies relays that we run), however, relays work by configuring tor (which is written in C). When tor reaches it’s EoL, will we have to make any changes? As I understand relays will end up using Arti’s code, Rust. Or will the transition be completely smooth?
I doubt the transition will be completely smooth in the “there is nothing to change sense”. We (relay operators) will have to make some changes, though hopefully they should be easy enough. Arti doesn’t use the same configuration format, and it won’t implement the control port (though should have something else in place, which will hopefully be easier to work with) so tools like nyx won’t work unless they get updated.
But that transition won’t go overnight. It’ll be years before Arti can run as a relay, and then, there will be some time again before tor finally reaches EoL, during which both will coexist.
While Arti will definitely change things for us (no more need for X relays per ip because tor is single-threaded, change in configuration format…), that’s honestly not something to think about too much right now, it will come, but it’s far from being there already.
1 Like
meeting records are available at the same place the meetings take place, which is here:
3 Likes
@TorOps or anyone else here, I’m interested in knowing more about why you want QUIC support. Are there advantages over HTTP/2 that you expect to gain? Are there apps that require QUIC with no fallback?
I’m working on UDP, and specifically I’m working on more carefully defining the benefits and risks we expect. There’s often a tradeoff to make between application compatibility and retaining the best privacy properties. I want to make sure we can define specific applications and the properties we expect them to have when tunneled over Tor.