[foms] WebM Manifest

[Thomas Vander Stichele]

Hi everyone,

(I realize I am late to this email party, apologies!)



> We have been talking about how to implement client-side HTTP adaptive
> streaming for WebM for a few months. We have focused on VOD first
> (with live to come later).

Am I really the only one who considers the 'with live to come later' a
mistake ?

A good design should evaluate all existing use cases.  A lot of existing
streaming designs have made the mistake of 'doing live later' and then
never handling correctly some of the finer, important points.

A good design should adapt itself to handle those existing use cases.
What good does it do us if all the hard work we do for the VoD case does
not work for the live case ? As a reminder, even Google's own Chrome
browser was unable to play live WebM until more than a month after we
(Flumotion) implemented the first live WebM streaming.

A good design should be able to adapt itself to the changes in the
market out there, or have compelling counter-offers.  The biggest reason
why the three big streaming software implementors (Microsoft, Apple,
Adobe) switched to chunked streaming for live is because CDN's have told
them to do so - it is much easier for them to deploy reverse proxy
infrastructure (which they already have) than it is to deploy streaming
servers.  For example, Akamai typically only has about 10% of their
edges with streaming capabilities.

I realize all of you think live is 'not that important', and VoD is by
nature easier to do, but I don't think it would be that hard to at least
take it into account the use case.  

A single 'file' per track is surely possible for live - in fact, we've
always done that in Flumotion, and so does Icecast - but it guarantees
that the end result is simply *not* cacheable by the internet at large,
as it is an ever changing ill-defined 'resource'.

The reality today is that there isn't a single CDN out there (except
maybe ours, but we're not even in the US) that is able to stream live
non-adaptive WebM.  Not even Google seems to plan to do it.  

Not even considering live at this point just means that this whole cycle
of getting it specced/implemented in browsers/adopted/available to
actual customers will have to be repeated all over again.

Just my 2 cents,

Thomas

-- 

-- 
Words are the part of silence that can be spoken.
--
Flumotion - the only way to stream!
http://www.flumotion.net/