> In the same spirit of keeping clarity, I'd say that Additional-Authorization 
> would be more correct choice, unless adding a new Service-Type would somehow 
> complicate adoption...

Agreed.

> 
>> - the first Access-Accept must contain a State attribute
>> for is already required for use of Authorize-Only
>> 
>> - additional authorization attributes are received via a series
>> of Access-Request / Access-Challenge
>> 
>> - each Access-Request MUST contain Service-Type = Authorize-Only
>>   and a State
> 
> Or, respectively, Additional-Authorization, isn't it?

No it's an extension/reworking of the concept described in RFC 3576 3.1, but 
the trigger is an Access-Accept packet with the 
Service-Type/Additional-Authorization attribute.

> 
>> - the State MUST change for each Access-Challenge response
>>   I can get into that later
> 
> I can imagine this is with the intention of guaranteeing order and avoiding 
> (or alleviating) MITM attacks, right?

Yes, correct. The packets don't contain any explicit ordering information, so 
the attributes are just concatenated together in the order they're received.

-Arran
_______________________________________________
abfab mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/abfab

Reply via email to