architectural question regarding the use of wireguard

[Alex Giurgiu <alex@giurgiu.io>]

Dear wg list,

First of all, thank you for this wonderful piece of software!

I am working on a software project which manages Linux containers on one or multiple machines, and I would like to implement a networking model where a Wireguard interface is created for every application that runs on this platform. Each application will be identified by a wg key + an IPv6 address deterministically derived from that key (similarly to what yggdrasil does).

Do you think this approach makes sense and will it scale from wg's point of view for situations where there are thousands of applications and hundreds of users each having one or two devices, with a complex matrix of permissions between the users and the applications? The key mapping and synchronisation between users and apps seems straightforward to me but I can't figure out if using wg for this scenario makes sense.

Thanks!

Alex