Re: [http-auth] RFC 7804 One Round-Trip Reauthentication

Isen Ng <isen@treeboxsolutions.com> Thu, 01 December 2016 02:22 UTC

MIME-Version: 1.0
In-Reply-To: <81f2ff57-64a1-47cb-49bd-dfa17aaacdc8@isode.com>
References: <CABR9bnzG9Zv+YYN5psnLZzyHUBzg14WVH8VuJkG_w4YZb4PRhg@mail.gmail.com> <81f2ff57-64a1-47cb-49bd-dfa17aaacdc8@isode.com>
From: Isen Ng <isen@treeboxsolutions.com>
Date: Thu, 01 Dec 2016 10:22:07 +0800
Message-ID: <CABR9bnykwC_jcoHoG1g=3vHHrRx8VDDkV_LgBOHYY74E=EpLVw@mail.gmail.com>
To: Alexey Melnikov <alexey.melnikov@isode.com>
Content-Type: multipart/alternative; boundary="001a114fbfb8fcda8f05428f7cb1"
Archived-At: <https://mailarchive.ietf.org/arch/msg/http-auth/wOic9NWLjDBpcNdUT5B9TXb8FpY>
Cc: http-auth@ietf.org
Subject: Re: [http-auth] RFC 7804 One Round-Trip Reauthentication
Precedence: list

Hi Alexey,

Yeah, I have implemented the HTTP SCRAM-SHA256 protocol according to this
RFC7804 and is implementing the one round-trip re-authentication at the
moment.

My colleagues and I have agreed upon a workaround and that is to include
the original "sid" in the "client-final-message".
The reason we do so is because the server has to keep a cache of the
original c-nonce and s-nonce for nonce verification anyway. Instead of
trying to maintain a different state for this using a different identifier
and then butchering the "client-final-message" with extra variables, we
decided to simply reuse the session that is already maintained.

This way, the code written to process the "client-final-message" into the
"server-final-message" does need to change at all.
The same code that was used in an regular 2-round-trip authentication can
be used in this 1-trip re-authentication.

Regards,
Isen

*Isen Ng* |  Ninja

*TreeBox Solutions Pte Ltd*
1 Commonwealth Lane #03-01/02
Singapore 149544
*E:* isen@treeboxsolutions.com

website <http://treeboxsolutions.com/> |  linkedin
<https://www.linkedin.com/company/treebox-solutions-pte-ltd>  |  twitter
<https://twitter.com/TreeBoxNews>  |  facebook
<https://www.facebook.com/treeboxsolutions>

On Wed, Nov 30, 2016 at 8:42 PM, Alexey Melnikov <alexey.melnikov@isode.com>
wrote:

> Hi Isen,
>
> On 28/11/2016 06:11, Isen Ng wrote:
>
> Hi,
>
> I'm looking at RFC 7804, specifically at "One Round-Trip Reauthentication"
> (https://tools.ietf.org/html/rfc7804#page-10)
>
> It is said that
>
> "If the client has authenticated to the same realm before (i.e., it
> remembers "i" and "s" attributes for the user from earlier
> authentication exchanges with the server), it can respond to that
> with "client-final-message""
>
> And the example give:
>
> C: GET /resource HTTP/1.1
> C: Host: server.example.com
> C: Authorization: SCRAM-SHA-256 realm="testrealm@example.com",
>           data=Yz1iaXdzLHI9ck9wck5HZndFYmVSV2diTkVrcU80MDk2JWh2WURwV1VhM
>            lJhVENBZnV4RklsailoTmxGLHA9ZEh6YlphcFdJazRqVWhOK1V0ZTl5dGFnOX
>            pqZk1IZ3NxbW1pejdBbmRWUT0K
>
> C: [...]
>
> In this case, there no longer is an attribute named "sid" in this
> "client-final-message".
> How does the server know which user's secret to use to calculate the proof
> for matching?
>
> You are right, I messed it up! The client should be sending "n=<username>"
> somewhere in this response.
>
> Are you trying to implement this?
>
> Best Regards,
> Alexey
>
>
>

[http-auth] RFC 7804 One Round-Trip Reauthenticat… Isen Ng
Re: [http-auth] RFC 7804 One Round-Trip Reauthent… Alexey Melnikov
Re: [http-auth] RFC 7804 One Round-Trip Reauthent… Isen Ng