Re: [rtcweb] realiable data service

[Magnus Westerlund <[e-mail]>]

On 2011-07-18 20:21, Cullen Jennings wrote:
> 
> Tim, I agree with your concern here. My largest concern of not
> specifying something is that it just means we will see a million
> different things done in JavaScript. That's pretty much what we have
> in the IETF today. Lots of protocols implement a different
> reliability layer on top of UDP. 

I would note that the the JS library at least is going to run on top of
a congestion controlled and secured datagram service. Thus, any
implementor can primarily hurt themselves rather than the network.

> However, it has not been all that
> bad and it has been very hard to get the transport guys to go and
> define one that everyone can use (I've presented in transport asking
> for this in the past). The end result has not really been all that
> bad. The protocols work even thought each one did there own
> reliability systems so I have not been all that worried about it. I'm
> not sure DTMF is a fair example to compare to...

I can't resist this bate as a transport guy. Well, things don't happen
for free, and no one was willing to stick in there for the long haul to
actually define it.

There also has been a lot of political resistance, and the fact that
transport development in OS etc happens at glacier speed, unless someone
well connected enough truly wants something.

I think a reasonable, re-use as much as possible, sane proposal has some
chance. And we WG chairs should try to get early feedback from TSV-Dir
etc so we avoid late surprises.

Please note that this is not an a comment in any direction around the
support for reliable data service. Just trying to provide some
background and feel for the land.

cheers

Magnus Westerlund

----------------------------------------------------------------------
Multimedia Technologies, Ericsson Research EAB/TVM
----------------------------------------------------------------------
Ericsson AB                | Phone  [phone]
Färögatan 6                | Mobile [phone]
SE-164 80 Stockholm, Sweden| mailto: [e-mail]
----------------------------------------------------------------------