Re: [Rats] Call for adoption (after draft rename) for Yang module draft

[Dave Thaler <[e-mail]>]

Challenge response can be done in a protocol initiated by the Attester, whether the Attester talks directly to the Verifier (passport model),
or indirectly via a Relying Party (normal background check model), without having the device needing to run a server of some sort like netconf or restconf.

At the hackathon, Nancy provided the first actual answer to my question I've seen, so I'll try to summarize it here.
I'm also going to use "device initiated" instead of "push", and "verifier initiated" instead of "pull" in case others have different definitions of push/pull than I intend.

Here's the two cases I've heard for device attestations that are verifier initiated (or "network initiated" if you prefer since these are network use cases):
Case 1) The network notices anomalous traffic coming from a device already on the network, which triggers a verifier to ask the device to attest to its health (which may have changed since it was last attested).  Here there might even be no Relying Party involved per se. 
Case 2) The network has not noticed anything odd, but wants to proactively query a device anyway, e.g., because the network's appraisal policy of what is considered trustworthy has just changed.  Again there might even be no Relying Party involved.

It was also pointed out at the hackathon that using a L2 protocol like EAP to trigger the device to attest, can work in cases where a host firewall is present.
However, not all devices/links use EAP.

Dave

-----Original Message-----
From: Schönwälder, Jürgen <[e-mail]> 
Sent: Sunday, November 17, 2019 10:41 PM
To: Dave Thaler <[e-mail]>
Cc: Laurence Lundblade <[e-mail]>; Nancy Cam-Winget (ncamwing) <[e-mail]>; Oliver, Ian (Nokia - FI/Espoo) <[e-mail]>; Smith, Ned <[e-mail]>; Henk Birkholz <[e-mail]>; [e-mail]
Subject: Re: [Rats] Call for adoption (after draft rename) for Yang module draft

On Fri, Nov 15, 2019 at 07:38:32AM +0000, Dave Thaler wrote:
> 
> If there is a compelling reason to support a pull-based mechanism, and we get consensus that we need it, then great.
> But so far I haven’t heard one.
>

I hope I get the terminology right...

Research papers I have seen often use a challenge response model where the 'verifier' sends a specific challenge that the 'attester' has to answer (often with time constraints, minimizing the chance to relay the challenge and such things). I think I pointed this out before.

Perhaps RATS does not need this form of challenge response attestation and perhaps RATS is fine with the assumption that the attester somehow knows which claims a verifier needs. But then it would be nice to spell this out clearly so that people looking for challenge response attestation attestation flows know that RATS is not for them.

/js

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
Fax:   [phone]         <https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.jacobs-university.de%2F&amp;data=02%7C01%7Cdthaler%40microsoft.com%7C6dddd58781f6453012ea08d76b6c40b3%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637095985013049843&amp;sdata=0LoEUlR96Nh%2BAQfVufe7EWD7yY7qQkLyklZphmltb8c%3D&amp;reserved=0>