Re: [pcp] Comparison of PCP authentication

[Alper Yegin <[e-mail]>]

On Aug 7, 2012, at 1:07 PM, Margaret Wasserman wrote:

> 
> On Aug 7, 2012, at 3:48 AM, Alper Yegin wrote:
>> That's right.
>> PANA session will generate a Session-Id, Key-Id, and a PCP key. We should use the two identifiers along with the MIC generated using the PCP key in Authentication Tag Option. That's how PANA session and PCP messages get crypto-bound to each other.
> 
> We should also do something (a hash, or whatever) to bind those keys to the contents of a specific request,

Authentication Data field of the Authentication Tag Option in your I-D does that.


> so that we can be sure that the contents of the request haven't been modified/replaced.  If we include a time and/or the transaction ID in the hash, I think we can guard against replay attacks that way.  I think that would do it.
> 

Yes.

Alper

> Much of this is the same, whether we use a demultiplexing approach or an encapsulation approach, and much of it would apply to a PCP-specific mechanism as well.
> 
> Margaret
> 
> 
>