Hannes Tschofenig <[email protected]> 写于 2012-09-10 18:26:56:

> I guess I figured it out. H(R) is the hash of a random number. 
Yes. 
> 
> The issue with this approach is that the Client can only use the 
> Access Token once with your approach. 
access token generally have time limit from a few minutes to 1 hour. 
according to oauth2, access token must be transported in TLS for 
confidentiality,
attacker can only obtain H(R), and it is difficult  to reverse a hash 
result in one hour. 
In this case, R can be used more than once.


> Of course, one could extend the approach to a hash chain and then 
> disclose the reverse hash chain. 
That is a better method to extend the above access token to more than 
once. 

> Still, this approach requires that the hash chain is bound to the 
> access token in some way and this requires a key. In general, to 
> secure the Access Token it is necessary to protect the token anyway 
> (for other reasons as well). 
On the path from AS to client , the access token and H(R) can be bound by 
a key shared between AS and client, which may be inevitably needed for 
client authentication to AS.
On the path from Client to RS, access token is transported in 
confifential, the validity of access token can be checked as usual(how is 
a access token verified when there is no shared key 
between AS and RS?), plus a verification of H(R).


> While I like these hash chain proposals (and I have seen these 
> already back in the Mobile IPv6 days) I don't think they make 
> anything simpler in the end. 
It provides an option where AS and RS don't need shared key and where 
access  time is quite limited especially when access limitation is done by 
access number. 

> I prefer to go for a well-established scheme
> 
> On Sep 10, 2012, at 12:32 PM, [email protected] wrote:
> 
> > client send H(R) in token request to AS, AS includes the H(R) in 
> the token, and client sends (token,R) to RS, 
> > RS can verify the access capablity by recalculating H(R) and 
> checking access toekn, 
> > by feature of hash, RS can trust R provider,this method does not 
> use pre-shared key between AS  and RS.
> 
> 

_______________________________________________
OAuth mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to