Comment by notpushkin
7 days ago
> So I set up my own self hosted provider
...get it certified, then transfer the keys to it. Or more likely you get an HSM and use its certification.
> I don't think consumer oriented protocols should support it.
Ideally yeah, that’s my point of view as well. In the corporate land however...
> binding credentials to some subset of physical devices
I don’t think that’s the goal. My guess is, they want to be able to transfer some passkeys but not allow any copies, and this only works if one device trusts another. It could work still with your scheme (store a private key on each device; when transferring a secret, make one device encrypt it to the other and immediately destroy its own copy, and make IT department sign any transfer requests so they can check whether they own both devices).
> My guess is, they want to be able to transfer some passkeys but not allow any copies,
Who are we taking about here?
For regular users it's no one's business if and when I copy my keys, for backups or any other purpose. Any scheme which attempts to prevent that must inherently infringe on my freedom.
For corporate I believe the goal is to know which registered (ie explicitly approved) device a login is coming from. Co-signing would accomplish that. The user key can be reused (not that you'd want to) and any session will still be tied to a known device. Plus you could bind a given user key to a given device on the server side so mission accomplished (assuming the hardware doesn't allow export of its key but then these are specifically approved devices so of course it won't if IT cares about that).
> Who are we taking about here?
The corporate.
> I believe the goal is to know which registered (ie explicitly approved) device a login is coming from.
They can do that with non-transferrable passkeys already (assuming they hand out the devices, and trust the manufacturer to actually prevent transferring the keys out). No cosigning required here.
Right but when it comes to that plus BYOD isn't that how we end up at attestation in the first place? I figure the alternative is to manually enroll each device as its own known entity, then require two "regular" signatures to authenticate. Basically it's just the observation that if you care about whitelisting the hardware then you must either have attestation or else accomplish the same thing manually. The former is less labor intensive but comes at the cost of enabling abuse.