Thomas Eizinger
Thomas Eizinger
> DNS already runs over UDP and handles unreliability for us Only if there is no response. Currently, the client instantly responds to the DNS query with new IPs. We...
> > Would we not want to log an access for each subdomain of a wildcard resource? > > I suppose this is the tradeoff -- logging means writing to...
Another option for a reliable transport could be to use SCTP. That is part of the WebRTC stack and is used for data channels there. str0m uses https://docs.rs/sctp-proto internally which...
@ReactorScram I am merging based on @conectado 's review and to move my stacked PRs forward. I'll implement any feedback in follow-up PRs!
Outdated.
It might be better to still send them but delay _acting_ on them.
> It might be better to still send them but delay _acting_ on them. We !need to add them to the agent before sending so we know that they are...
> potentially also one by one, i.e. a new candidate every ~100ms. To start with, it is probably easier to do all relay candidates at once if we don't see...
Moving this to TODO because I think it will help with fixing https://github.com/firezone/firezone/issues/4290.
> Following up on that, perhaps every ~5 minutes we should run ICE again to see if we can switch to a Direct from our current Relayed. So just force-close...