Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Date: Tue, 11 Oct 2011 17:35:07 +0530
Message-ID: <1D062974A4845E4D8A343C6538049202068FED61@XMB-BGL-414.cisco.com>
In-Reply-To: <4E935A8B.8020700@alvestrand.no>
Thread-Topic: [rtcweb] Sebastien Cubaud's non-ICE mechanism (Re: Summary of
 ICE discussion)
Thread-Index: AcyHjkLROLg5C5rpSKevWSMpyvuZJwAYOygA
References: <4E8B192E.80809@ericsson.com><E6AA070839B987489960B202AD80E18D019D9119@ftrdmel0.rd.francetelecom.fr>
 <4E935A8B.8020700@alvestrand.no>
From: "Muthu Arul Mozhi Perumal (mperumal)" <mperumal@cisco.com>
To: "Harald Alvestrand" <harald@alvestrand.no>, <rtcweb@ietf.org>
Subject: Re: [rtcweb] Sebastien Cubaud's non-ICE mechanism (Re: Summary of ICE
 discussion)
Precedence: list

|> - Assuming B receives these packets, it then sends=20
|> via the signalling channel an information from the=20
|> media path unknown from S (i.e. not accessible via JS).

A requirement hidden here is that B shouldn't alert the user till the =
entire consent verification is complete (otherwise B would have to =
buffer the media coming from the user). If B is a PSTN gateway this =
isn't going to be trivial to implement.

If we go down this path, I think we would end up reinventing =
ICE/ICE-Lite in a different form (which doesn't look worth the effort), =
so we should just make use of ICE.

Muthu

|-----Original Message-----
|From: rtcweb-bounces@ietf.org [mailto:rtcweb-bounces@ietf.org] On =
Behalf Of Harald Alvestrand
|Sent: Tuesday, October 11, 2011 2:20 AM
|To: rtcweb@ietf.org
|Subject: [rtcweb] Sebastien Cubaud's non-ICE mechanism (Re: Summary of =
ICE discussion)
|
|Changing the subject to keep threads separate....
|
|On 10/09/2011 12:00 PM, sebastien.cubaud@orange.com wrote:
|> Hi there,
|>
|> Re Magnus's call for proposal (If anyone can find a solution that =
fulfill
|>   the security goals and have improved legacy interoperability people =
would
|> be interested in that solution. So far RTCP has been discarded as =
insufficient.)
|> on media consent verification, here is below a suggestion for the =
group:
|>
|> Basically, the idea, compared to Hadriel's RTCP proposal, relies on =
the use
|> of the signalling path instead of RTCP as a feedback channel for =
media
|> consent verification.
|>
|> Here are the steps I foresee before allowing the establishment of a =
media session:
|>
|> - Let's consider A (a RTC-Web compliant browser) connected to server =
S and
|> wishing to share real-time media with destination B (potentially a =
SIP
|> endpoint or a browser)
|> - A&  B learn via the signalling channel the triple @IP, transport =
proto and
|> associated listening port of the remote media
|> - A sends a few RTP packets to B (3 as in RFC 2833/4733 or more?).- =
This would
|>   allow the mechanism to resist against packet loss -. The format of =
such packets
|>   are to be discussed
|> - Assuming B receives these packets, it then sends via the signalling =
channel
|> an information from the media path unknown from S (i.e. not =
accessible via JS).
|> I propose to use the min of the sequence number of the RTP packets =
received
|> (which is random per RFC 3550)
|The sequence number is a 16-bit number, so there are 16 bits of
|randomness to play with here.
|An attack based on just returning a random number will succeed 1 out of
|65.536 times; if any of the 3 packets' sequence numbers are acceptable,
|it will succeed 1 out of 21.845 times.
|
|A browser can reasonably limit the number of attempts per second =
(either
|overall or per-script); multiple browsers will not be able to =
coordinate
|on any such limit.
|
|Is this defense strong enough to warrant considering this further?
|
|> - A receives via S this info granting B's consent to receiving media =
from A
|> - A now can start sending media to B
|>
|> When B uses SIP as its signalling channel, Step 4 would probably =
require a
|> SDP extension to convey this info, as well as possibly extended =
H.248/Diameter
|>   informations would be required in a decoupled sig./media =
architecture.
|> Of course, a solution avoiding extensions would be more than welcome.
|>
|> The benefit of this mechanism largely relies on the assumption that =
the
|> transmission of information from the media termination of the =
destination
|> to its associated signalling channel is less costly than implementing =
ICE connectivity
|> checks. Which is still to be checked.
|>
|> This mechanism provides basic media consent verification and if it =
turns out
|> that it provides less security properties than ICE or other drawbacks =
(e.g.
|> increasing media session establishment delay, issues with the few =
initial RTP
|> packets,..), it could be seen as a fallback media consent =
verification
|> when ICE is not implemented at each end of media terminations.
|>
|> Looking forward to hearing feedbacks from the group on this.
|>
|> Cheers
|> Sebastien
|>
|> -----Message d'origine-----
|> De : rtcweb-bounces@ietf.org [mailto:rtcweb-bounces@ietf.org] De la =
part de Magnus Westerlund
|> Envoy=E9 : mardi 4 octobre 2011 16:33
|> =C0 : rtcweb@ietf.org
|> Objet : [rtcweb] Summary of ICE discussion
|>
|> WG,
|>
|> I have bellow tired to summarize the result of the ICE discussion. =
This
|> is intended as furthering this discussion and form a basis for going
|> forward in the consensus process. I do expect people that disagree =
with
|> my summary of the discussion to speak up.
|>
|> Major requirements
|>
|> - Need for data transmission consent for protocols using UDP as the
|> traffic receiver needs to consent to receiving the data
|>
|> - Perform NAT and FW traversal when ever needed
|>
|> - Support IPv4 to IPv6 transition
|>
|> Desired behavior:
|>
|> - Be interoperable with deployed legacy systems as SIP Trunk, PSTN
|> gateways, VoIP phones.
|>
|> WG chairs conclusion of discussion so far:
|>
|> - ICE is so far the only solution that provides the security goals =
and
|> have any legacy deployment.
|>
|> - ICE usage will require that STUN connectivity MUST have succeeded
|> prior to any data transmission to fulfill security goals.
|>
|>    * The Browser will enforce this requirement to prevent being an =
attack
|> vector in all cases.
|>
|> - If anyone can find a solution that fulfill the security goals and =
have
|> improved legacy interoperability people would be interested in that
|> solution. So far RTCP has been discarded as insufficient.
|>
|> - Media Gateway can support a reduced functionality set from Full ICE
|>
|> Cheers
|>
|> Magnus Westerlund
|> as WG chair
|>
|> =
----------------------------------------------------------------------
|> Multimedia Technologies, Ericsson Research EAB/TVM
|> =
----------------------------------------------------------------------
|> Ericsson AB                | Phone  +46 10 7148287
|> F=E4r=F6gatan 6                | Mobile +46 73 0949079
|> SE-164 80 Stockholm, Sweden| mailto: magnus.westerlund@ericsson.com
|> =
----------------------------------------------------------------------
|>
|> _______________________________________________
|> rtcweb mailing list
|> rtcweb@ietf.org
|> https://www.ietf.org/mailman/listinfo/rtcweb
|> _______________________________________________
|> rtcweb mailing list
|> rtcweb@ietf.org
|> https://www.ietf.org/mailman/listinfo/rtcweb
|>
|
|_______________________________________________
|rtcweb mailing list
|rtcweb@ietf.org
|https://www.ietf.org/mailman/listinfo/rtcweb