pull down to refresh
200 sats \ 2 replies \ @shafemtol 27 Feb \ parent \ on: Obscura VPN is live privacy
I haven't taken the time to ask the Obscura people about it or investigate their software at this point. Feel free to bring these concerns to their attention.
I do have some ideas on this kind of service and in particular how I would handle the Wireguard public keys. I'd be happy to work with Obscura or otherwise to implement these ideas once they are more complete.
As an aside, I am currently trying to create a Linux kernel patch or alternate Wireguard module to solve the timestamp issue by adding a randomly varying offset to the timestamp. Might publish it once I've got it working.
How it's going??
reply
Regarding the Wireguard patch, I've been mostly busy with other stuff lately, but it's progressing slowly. So far I've got the randomness algorithm working in kernel space. It's more complicated than one might imagine, as it's supposed to work as a noise source that is realistic (indistinguishable from natural clock noise) yet strong enough to make correlation difficult, and that is stateless yet consistent across reboots and reasonably fast to calculate.
The second part, which I'm currently working on, is an internal clock that is not directly affected whenever there's a jump in wall-clock time (e.g. by the user, NTP client or similar), but might slowly, or randomly, adjust over time. Finally this must all be integrated into the Wireguard module.
That said, Obscura and/or the kernel devs should rather just change to using the Wireguard timestamp field as a simple counter. They don't need to blend in to become part of any existing anonymity set.
reply