Re: Minimizing the protocol (Re: Data API)

-----Original Message-----
From: Randell Jesup <randell-ietf@jesup.org>
Date: Thu, 1 Mar 2012 18:08:24 +0100
To: W3C WEBRTC <public-webrtc@w3.org>
Subject: Re: Minimizing the protocol (Re: Data API)

>On 3/1/2012 10:17 AM, Cullen Jennings wrote:
>> Hmm - All the Facetime I have seen does not use SCTP so uh, not sure
>>what to say about that  - but I do agree it looks like in many cases
>>Facetime does set up a communication channel between the endpoints then
>>use that channel to directly do the SDP negotiate media.
>
>Yes, I think that was Harald's point.
>
>> One of the apps I have been thinking about would use the web server to
>>instantly bring up a data channel - this does not need much user
>>authorization as no media is being sent so it can happen fairly early
>>and quickly. From that point on, all SDP to add more media channels and
>>such would be passed back and forth between the browsers over the direct
>>data channel. (This is the "low path" from the slides I had long ago).
>>It's not like this data channel has to be channel 0 or anything - it's
>>just a data channel like any other. And it's not like the browser has to
>>have any awareness of this - the JS app is just set up a data channel
>>and using it. The only thing the JS app has to be careful of is that it
>>does not tear down the data channel.
>
>Yes, exactly - an excellent use of the data channel for many possible
>apps.  We do need to keep the server channel available, since I think we
>need it to handle reconnection of the data channels on IP address change
>(to exchange new ICE candidates).  As I mentioned,
>pc.createDataChannel("signaling",...) :-)

+1

And since many apps often include other "signaling" as well, the data
channel comes handy for this traffic as well.

>
>-- 
>Randell Jesup
>randell-ietf@jesup.org
>
>

Received on Thursday, 1 March 2012 21:02:21 UTC