Sure!
Hello,
Earlier this year, the team wrapped up support for Arti Onion Services connectivity such that an Arti client can connect to Onion Services within the Tor network.
Last month, an exciting new milestone was reached by the team: we now can use Arti to connect to an Onion Service that was also made available via Arti’s API. Since then, the team has worked on stabilization and eliminating bugs that will make the Onion Services feature work as you would expect from well-tested code.
The team is working hard now to wrap up the service side, such that Onion Services can be hosted internally via Arti’s API or using a reverse proxy similar to that seen in the C implementation of Tor.
However, we will not be able to meet the deadline we originally had for Onion Services support (Deliverable 1.1). We will detail here where we are at with the project so you will better understand the causes of the delay and when we expect to be completed.
Beyond the client-side and service side Onion Service work that was done, the team also did work that was not originally in the grant but that we identified as being essential to make the availability of the Onion Services feature better for end-users and third-party integrators:
-
We added an internal key management subsystem in Arti for handling cryptographic material that Arti needs to store (Onion Services identity keys, Client Auth keys, Certificates, etc.). Since this is a better format and design than C Tor’s, the team will be working on a command-line interface tool to generate and work with these objects.
-
As we have mentioned before, the Arti team at Tor also wanted to work on the RPC/FFI milestone from an early point of view so that we could gather some feedback, both internally and externally, on designs. This should make the engineering work more approachable once the team continues this engineering deliverable after the Onion Services work is wrapped up. This work continued from where we previously reported last time.
We additionally had some other unexpected delays come up when we had to move some staff around internally earlier in the year, resulting in a ramp-up phase. Fortunately, that was significantly shorter than we thought it would be back in March.
We believe the team will be able to wrap up Onion Services work around February 2024 with inclusion in the Arti release that will happen at the end of February/early March 2024.
This, of course, impacts the timeline for the entire duration of the grant, but we will return in January 2024 with some better estimates for when we think the grant as a whole can be finalized. As some of you may be aware, we recently got external funding to deliver a Relay implementation of Arti, which means we can finally move the majority of the Network Team at Tor to Arti engineering in 2024, this will likely slow us down a little as the C engineers transition into the Arti project, but once they do we anticipate an increased pace.
Update: Arti 1.1.11 is released: More onion progress | The Tor Project