2018-01-11 20:36 GMT+01:00 Ilari Liusvaara <[email protected]>:

> On Thu, Jan 11, 2018 at 08:23:26PM +0100, Sophie Herold wrote:
> > Hi,
> >
> > challenge tokens "MUST have at least 128 bits of entropy", at the same
> > time it seems trivial to guess order and authorization URLs like the
> > ones used in the examples. It seems natural, that URLs MUST be generated
> > with the same amount of entropy. But I couldn't find that in the draft.
> >
> >
> > For account objects, GET request are not allowed:
> >
> >    Servers SHOULD NOT respond to GET requests for account resources as
> >    these requests are not authenticated.
> >
> > This suggests that all non-expiring URLs should be protected in this
> > way. At least for orders lists, this protection is missing.
> >
>
> The token entropy requirement is to render those tokens unguessable
> before the validation request is received. This is to protect against
> careless servers. The token is not actually secret after it has been
> generated.
>

I don't think it is for careless servers, the token is provided in the
request, so any server can respond with it. Rather it's required to fulfill
the CA/Browser Forum rules.

Regards, Niklas


> Now, in some approved CA validation methods, the tokens actually are
> secret, but none of those is used in ACME.
>
>
> -Ilari
>
> _______________________________________________
> Acme mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/acme
>
_______________________________________________
Acme mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/acme

Reply via email to