Re: [TLS] Please discuss: draft-housley-evidence-extns-00

[Martin Rex <martin.rex@xxxxxxx>]

Mark Brown wrote:
> 
> The point of TLS Evidence's use within this illustration is that anyone,
> even a third party, can easily show that a particular web server gave some
> app data to a particular client (to the extent the PKCs can show
> authenticity), and that app data is not a forgery (unless the digital
> signature crypto was cracked).  The goal here is not perfection, but to
> still be helpful.

The point is, that this is entirely pointless.

TLS is typically used on a hop-by-hop basis, and the only connection
where it would make any sense is the one between the last hop on the
client side and the initial entry point to the server side.

In pretty much every serious deployment the front-door of a business
system runs on a host in the DMZ, and the only "valuable" thing
that this possesses is a keypair suitable for online authentication
of the server.  A business/organization would be quite stupid to
deploy keys with digital signature capabilities in any legally
binding sense on such a host.

Furthermore:
The last two Online-shops where I ordered computer parts explicitly
indicated that neither the successful completion of the "checkout",
and not even the "autometic confirmation email" sent as a result of
the checkout would be legally binding in any way for them to ship
the requested product for the displayed amount and conditions.

Only an explicit confirmation Email generated in their backend
system (usually sent within a business day) will confirm the
contract. (I don't know whether those involve human action(s).

TLS-based Evidence would be entirely useless in these scenarios,
because they already have had issues when the intermediate/frontend
systems failed (or were hacked), and the businesses don't want to
be held responsible for that.  A very common error is a typo in
the price of the web shop, and it did happen to companies like Dell.

It would be a pretty stupid idea to include the easiest and well
known point of failure (and point of attack) into the creation
of legally binding digital signatures.


TLS/SSL has been about online authentication, and it should firmly
stick to that archtecture.


-Martin 

_______________________________________________
TLS mailing list
TLS@xxxxxxxxxxxxxx
https://www1.ietf.org/mailman/listinfo/tls