Re: [hybi] Framing take IV

[Yves Lafon <[e-mail]>]

On Wed, 4 Aug 2010, Mike Belshe wrote:

> I'm going to write a bunch of stuff which makes it seem like I'm in favor of
> variable length encoding.  My true position:  it's not materially different
> to use a fixed length + frags vs a variable-length-encoding which supports
> large sizes.

variable-lenght-encoding + fragmentation allows to send data before having 
everything ready, it can reduce the perceived latency (like for HTTP chunk 
encoding), after that it is one of the trade-offs to think about.

>> - Performance cost of implementing multiplexing at the application level
>> compared to doing it at the protocol level (either in terms of more
>> connections resulting and the actual cost of that, or in terms of additional
>> overhead on the client side from using an iframe or shared worker or
>> whatever to share the connection.)
>>
>
> Multiplexing is relatively easy (not trivial, though).  It's the flow
> control that is hard.   And unfortunately, I don't think you can seriously
> claim to have multiplexing unless you address flow control (SPDY version 1
> did it, it falls back to TCP flow control, which can stall out streams
> inappropriately and is generally not-good).  But when you add flow control
> to the multiplexed data, it's very very easy to kill performance
> (underthrottle).

During the past experiments on WebMux, that was the main conclusion, 
messing with TCP flow control by adding another flow control was not easy 
and could lead to some disasters, it would be far better to have mux at 
the TCP level, but I see that you already mentionned SCTP :)

-- 
Baroula que barouleras, au tiéu toujou t'entourneras.

         ~~Yves