Re: [OAUTH-WG] A few comments on draft-ietf-oauth-rar-01

[Neil Madden <[e-mail]>]

> On 9 Jul 2020, at 07:33, Torsten Lodderstedt <[e-mail]> wrote:
> 
> 
> 
>>> On 8. Jul 2020, at 23:52, Neil Madden <[e-mail]> wrote:
>>> 
>>> 
>>>> On 8 Jul 2020, at 20:56, Torsten Lodderstedt <[e-mail]> wrote:
>>> 
>>>> Am 08.07.2020 um 20:46 schrieb Neil Madden <[e-mail]>:
>>>> 
>>>> On 8 Jul 2020, at 19:03, Torsten Lodderstedt <[e-mail]> wrote:
>>>>>>> 
>>>>>>> What in particular should the use consent with in this step?
>>>>>> 
>>>>>> “FooPay would like to:
>>>>>> - initiate payments from your account (you will be asked to approve each one)”
>>>>>> 
>>>>>> The point is that a client that I don’t have any kind of relationship with can’t just send me a request to transfer $500 to some account. 
>>>>> 
>>>>> Are we talking about legal consent or a security measures here?
>>>> 
>>>> Normal OAuth consent. My phone is my resource, and I am its resource owner. If a client wants to send payment requests to my phone (e.g. via CIBA backchannel) then it should have to get my permission first. Even without backchannel requests, I’d much rather that only the three clients I’ve explicitly consented to can ask me to initiate payments rather than the hundreds/thousands clients my bank happens to have a relationship with.
>>> 
>>> To me it sounds like you would like to require a client to get user authorization to send an authorization request. Would you require the same if I would use scope values to encode a payment initiation request?
>> 
>> Yes. If something is sufficiently high value to require per-transaction authorization then initiating transactions itself becomes a privileged operation. 
> 
> The per transaction authorization alone is a significant increase in security. What is the added value of requiring an authorization to send a per-transaction authorisation request in an additional step?

Because Open Banking allows any client at any time to send an asynchronous back channel request to my phone to approve a payment. This is pretty risky. 

I can’t think of another transactional auth system that allows this without some kind of initial indication of user consent. For example, in Apple Pay all payment requests must be initiated from an explicit user gesture, providing some indication that the user wants to use this. The Dropbox Chooser and Saver APIs also have to be triggered from a user gesture. Again, this provides some confirmation that the user actually initiated the interaction. 

In OAuth, the AS doesn’t have this level of integration into the client’s UI so it needs some other way to establish user consent. By the time the user has a payment confirmation request on their screen it’s too late. 


>>>>> In case of open banking the user legally consents to this process at the client (TPP) even before the OAuth/Payment Initiation dance starts. 
>>>> 
>>>> How does the bank (ASPSP) confirm that this actually happened?
>>> 
>>> It does not because it is not the responsibility of the ASPSP. The TPP is obliged by law to obtain consent.
>> 
>> If the TPP can be trusted to obey the law about this, why not also trust them to be honest about transactions? Why enforce one thing with access tokens but take the other on trust? Especially as the actual transactions are more likely to have a rigorous audit trail. 
>> 
>> If we could trust clients to obtain consent we wouldn’t need OAuth at all. 
> 
> I thought the same initially, but we must distinguish between legal consent and strong authentication/transaction authorization in such a case. Legal consent can be obtained in various ways including the traditional OAuth user consent but also in other places. Authenticating the user (probably with 2FA) and getting authorization for a certain transaction (the meaning of PSD2 SCA) must be conducted by the AS. 
> 

Do you mean legal protection for the bank or their users? As a user, if an OB client acts in a way that I don’t like, but doesn’t break any actual laws or policies, what’s my recourse? In normal OAuth I can revoke the grant to that client. This is not possible in transactional uses of RAR, and that seems like a big difference that significantly changes the relationship between users and clients. 

— Neil