Re: [tram] Oauth vs Origin [Was: Path Forward for STUN ORIGIN]

[Brandon Williams <[e-mail]>]

I agree that a kid must be unique within its scope. I disagree that the 
only possible scope is global. For example, as Oleg has called out, 
REALM _could_ be used to scope the kid. ORIGIN, if adopted, could also 
be used to scope the kid.

My point is that the RFC does not in fact speak to this question, which 
presents possible interoperability issues, since one relay service 
provider could choose to use REALM for this purpose and another could 
choose to require global scoping. Maybe that flexibility is OK, or maybe 
it's something that needs to be resolved if we decide to discard/limit 
ORIGIN.

--Brandon

On 09/01/2015 10:10 PM, Tirumaleswar Reddy (tireddy) wrote:
> Yes, kid must be unique. If relay server receives kid from an AS which conflicts with the kid already provisioned by another AS then relay server must request another kid using the mechanism specified in section 4.1.
>
> -Tiru
>
>> -----Original Message-----
>> From: tram [mailto:[e-mail]] On Behalf Of Oleg Moskalenko
>> Sent: Wednesday, September 02, 2015 7:30 AM
>> To: Brandon Williams
>> Cc: Justin Uberti; Alan Johnston; [e-mail]; draft-ietf-tram-stun-
>> [e-mail]
>> Subject: Re: [tram] Oauth vs Origin [Was: Path Forward for STUN ORIGIN]
>>
>> The kid value must be unique across the server (the relay ?).
>> Otherwise, it all does not make sense. Probably the RFC7635 writers even did
>> not consider possibility that it may be not unique... and I understand why.
>> Non-unique kids raise all sorts of weird questions.
>>
>> We can assign single unique REALM to each KID, and then we achieve "multi-
>> homeness".
>>
>> Thanks
>> Oleg
>>
>>
>> On Tue, Sep 1, 2015 at 3:00 PM, Brandon Williams
>> <[e-mail]> wrote:
>>> I agree that using the RFC7635 mechanism can work for #2, and I prefer
>>> that to ORIGIN for this purpose. That said, I think the mechanism is
>>> currently underspecified, in that there is nothing explicitly defined
>>> in RFC7635 that could be used to play the role of ORIGIN for auth
>>> purposes, so interoperability could be an issue.
>>>
>>>
>>> If the kid value delivered in the USERNAME attribute is unique to the
>>> relay, then the kid value alone is enough to allow ORIGIN
>>> identification. This rules out the method for key distribution
>>> describe in section 4.1.1 of the RFC though, since that method has the
>>> auth server selecting the keys/kid, not the relay server, and I don't
>>> know how the necessary uniqueness would be guaranteed in that case. As
>>> long as the kid is unique to the relay, though, the necessary keys can be
>> retrieved.
>>>
>>>
>>> If the kid is not guaranteed to be unique for the relay, then some
>>> method to qualify the kid and provide uniqueness would be necessary.
>>> Either ORIGIN or REALM could be used for this purpose, but whichever
>>> one it is would itself have to be unique for the relay in order to
>>> ensure that qualifier+kid is unique.
>>>
>>>
>>> RFC7635 doesn't impose requirements on kid uniqueness, nor does it
>>> require REALM, ORIGIN, or any other source of uniqueness to be present
>>> in the message that contains the token. Provided that either kid or
>>> realm+kid is unique to the relay, I agree that ORIGIN itself is not
>>> required for auth and the information could be delivered via some other
>> mechanism.
>>>
>>>
>>> All of this assumes that the STUN server name sent to the client from
>>> the relay as part of the initial unauthenticated exchange is a global
>>> value for the service and does not need to be allowed to be ORIGIN
>>> specific. In other words, any OAuth server that generates an access
>>> token for the 3rd party relay service would receive the same server
>>> name from a specific STUN server.
>>>
>>>
>>> --Brandon
>>>
>>>
>>> ----------------------------------------------------------------------
>>> --
>>> *From:* Justin Uberti <[e-mail]>
>>> *Sent:* Wednesday, August 26, 2015 7:08 PM
>>> *To:* Alan Johnston
>>> *Cc:* [e-mail]; [e-mail]
>>> *Subject:* [tram] Oauth vs Origin [Was: Path Forward for STUN ORIGIN]
>>> splitting thread
>>>
>>> On Wed, Aug 26, 2015 at 2:07 PM Alan Johnston
>>> <[e-mail] <mailto:[e-mail]>>
>> wrote:
>>>
>>>     Justin,
>>>
>>>     It is an important question for the WG on whether ORIGIN is a useful
>>>     solution to the use cases. Hopefully we'll get more feedback on this.
>>>
>>>     Your comments seem mainly directed towards a solution for use case
>>>     #2 below (REALM selection for TURN authentication), right?  I'm not
>>>     sure if use case #3 applies or not.  I don't think OAuth helps with
>>>     #1 or #4 or #5.
>>>
>>>
>>> I think OAuth can solve #1, #2, and #3. The app's interaction with the
>>> AS can give the AS context, which can be encoded into the returned
>>> token. This then allows the TURN server to have this context when
>>> receiving the token, such that it can make informed decisions for #1/#2/#3.
>>>
>>>
>>>     I know we have discussed it before, but I also don't think that
>>>     everyone believes that OAuth means that ORIGIN is no longer needed.
>>>     There have been suggestions that a multi-tenanted TURN server using
>>>     OAuth might still use ORIGIN.
>>>
>>>
>>> Since the origin can be determined by the AS, it's not clear why this
>>> would also be needed at the TURN server. One might ask, why it is
>>> better to use origin to the AS versus the TURN server? The upsides of
>>> origin over HTTP to AS are that 1) it is encrypted and 2) it's easier
>>> to deal with the privacy implications of origin in a single protocol rather
>> than 2 protocols.
>>>
>>>
>>>     And using domain-specific USERNAME attributes sounds a bit like the
>>>     problem that Martin mentioned of simply working around the approach.
>>>
>>> Yes - that's what I was trying to say as well.
>>>
>>>     - Alan -
>>>
>>>     On Tue, Aug 25, 2015 at 3:50 PM, Justin Uberti <[e-mail]
>>>     <mailto:[e-mail]>> wrote:
>>>
>>>         Overall, I think most/all of the goals that ORIGIN is trying to
>>>         accomplish would be better handled by STUN OAuth
>>>
>>> (https://datatracker.ietf.org/doc/draft-ietf-tram-turn-third-party-aut
>>> hz/
>>>
>>> <https://urldefense.proofpoint.com/v2/url?u=https-3A__datatracker.ietf
>>> .org_doc_draft-2Dietf-2Dtram-2Dturn-2Dthird-2Dparty-
>> 2Dauthz_&d=BQMFaQ&
>>> c=96ZbZZcaMF4w0F4jpN6LZg&r=bwZ-nnRGWmcGKRIuadq6-
>> NSnsgwbBVUJa4mZfmEIBXg
>>>
>> &m=YAVYqLAvlnjV5WFeb0DMhGTziYu4jLvpB3UTCbrR8T4&s=uYVEJtj6D0D8A
>> ALXY1GfS
>>> zY-RO-q9CHJJjr2mCjvWlk&e=>)
>>>
>>>
>>>         Generally, it's not clear that ORIGIN is a great solution for
>>>         any of the use cases listed.
>>>         1) is easily accomplished via oauth, without any of the
>>>         restrictions present here
>>>         2) can be dealt with using other means (e.g. oauth, or supplying
>>>         domain info in USERNAME)
>>>         3) is fairly weak control, and can be done better through oauth,
>>>         or domained USERNAMEs
>>>
>>>         Of course, as noted by others, use of domained USERNAMEs creates
>>>         a quasi-ORIGIN value within the USERNAME attribute, which makes
>>>         OAuth more attractive.
>>>
>>>         On Tue, Aug 25, 2015 at 12:57 PM Alan Johnston
>>>         <[e-mail] <mailto:[e-mail]>>
>>>         wrote:
>>>
>>>             All,
>>>
>>>             Based on feedback during IETF-93 in Prague, the authors of
>>>             draft-ietf-stun-origin are looking for feedback on a
>>>             possible path forward to solve the privacy/meta data issues
>>>             and clear the DISCUSSes (see
>>>
>>> https://datatracker.ietf.org/doc/draft-ietf-tram-stun-origin/ballot/
>>>
>>> <https://urldefense.proofpoint.com/v2/url?u=https-
>> 3A__datatracker.ietf.org_doc_draft-2Dietf-2Dtram-2Dstun-
>> 2Dorigin_ballot_&d=BQMFaQ&c=96ZbZZcaMF4w0F4jpN6LZg&r=bwZ-
>> nnRGWmcGKRIuadq6-
>> NSnsgwbBVUJa4mZfmEIBXg&m=YAVYqLAvlnjV5WFeb0DMhGTziYu4jLvpB3UT
>> CbrR8T4&s=lx5ApLOv-KiXbtLZppLxrkE5vHcqXhUqPSXlDcF-O28&e=>).
>>>
>>>             There are two areas of feedback: Use Cases and Proposed
>>>             Solution.
>>>
>>>             Use Cases
>>>             ========
>>>
>>>             First, we want to be sure that we are considering all the
>>>             use cases for STUN ORIGIN.  In the draft, three are described:
>>>
>>>             1.  Logging of usage of STUN/TURN server. An operator of a
>>>             STUN or TURN server would learn the Origin of the web, SIP,
>>>             and XMPP users.
>>>
>>>             2.  REALM selection for multi-tenanted TURN server. This
>>>             avoids having to run multiple instances on different ports
>>>             or IP addresses when more than one REALM is in use.
>>>
>>>             3.  Service determination after authentication.  Bandwidth
>>>             or QoS policy can be applied, after authentication, based on
>>>             the Origin.
>>>
>>>             Also, the following use cases have also been discussed but
>>>             are not in the draft
>>>
>>>             4.  Authorization at border TURN server after
>>>             authentication.  An enterprise border TURN server can apply
>>>             policy and allow usage for certain Origins (whitelist) or
>>>             block usage of others (blacklist).
>>>
>>>             5. draft-jennings-behave-rtcweb-firewall-01 discusses how a
>>>             firewall with STUN server could use ORIGIN in authorization
>>>             decisions.
>>>
>>>             Are there any other use cases we should consider?
>>>
>>>             Possible Solution
>>>             =============
>>>
>>>             The solution that was discussed in Prague that seems to be
>>>             the best tradeoff between privacy and functionality is to
>>>             only include ORIGIN information in a STUN or TURN request if
>>>             the domain of the Origin matches the domain of the STUN or
>>>             TURN server in the STUN or TURN URI.  Since web and STUN and
>>>             TURN run on different ports, and often will resolve to
>>>             different IP addresses, a full host comparison is not
>>>             possible.  This proposal is not perfect: it does not provide
>>>             perfect privacy for subdomains (e.g. sub1.example.com
>>>
>>> <https://urldefense.proofpoint.com/v2/url?u=http-
>> 3A__sub1.example.com&d=BQMFaQ&c=96ZbZZcaMF4w0F4jpN6LZg&r=bwZ-
>> nnRGWmcGKRIuadq6-
>> NSnsgwbBVUJa4mZfmEIBXg&m=YAVYqLAvlnjV5WFeb0DMhGTziYu4jLvpB3UT
>> CbrR8T4&s=rmZxvMmZrZEWtX42fGrg7dcvFr_RUv9ngo2yod2ZHPo&e=>
>>>             and sub2.example.com
>>>
>>> <https://urldefense.proofpoint.com/v2/url?u=http-
>> 3A__sub2.example.com&
>>> d=BQMFaQ&c=96ZbZZcaMF4w0F4jpN6LZg&r=bwZ-nnRGWmcGKRIuadq6-
>> NSnsgwbBVUJa4
>>>
>> mZfmEIBXg&m=YAVYqLAvlnjV5WFeb0DMhGTziYu4jLvpB3UTCbrR8T4&s=AyGt
>> dO3Zale2
>>> SSP5BSnM64KJGpiFT9ICrzuSzeBVj7k&e=>
>>>
>>>             will match with this rule, even if they are separate
>>>             entities.  Of course, subdomain users could always register
>>>             a new domain and point DNS records to their sub hoster.)
>>>
>>>             In comparing this solution against the use cases above:
>>>
>>>             1. Allows logging of your own users on your own TURN server
>>>             (or one that you setup DNS SRV records to point towards).
>>> Other users of your STUN or TURN server will not be
>>>             identified by an Origin, but you will at least know if
>>>             others are using your server.
>>>
>>>             2. Works for REALM selection as long as you use your own
>>>             TURN server (or one that you setup DNS SRV records to point
>>>             towards).
>>>
>>>             3. Again, works for service determination as long as you use
>>>             your own TURN server (or one that you setup DNS SRV records
>>>             to point towards). Having service-specific TURN credentials
>>>             is another solution that doesn't rely on ORIGIN.
>>>
>>>             4. & 5. Does not seem provide a good solution for these use
>>>             cases.
>>>
>>>             Does this analysis make sense?
>>>
>>>             Feedback most welcome so we can move forward with this draft.
>>>
>>>             - Alan -
>>>
>>>
>>>
>>>
>>> --
>>> Brandon Williams; Chief Architect
>>> Cloud Networking; Akamai Technologies Inc.
>>>
>>> _______________________________________________
>>> tram mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/tram
>>
>> _______________________________________________
>> tram mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/tram
>
> _______________________________________________
> tram mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/tram
>

-- 
Brandon Williams; Chief Architect
Cloud Networking; Akamai Technologies Inc.