[icecast-dev] Any buffering on server?

[Ricardo Galli]

On Monday 10 February 2003 04:29, Michael Smith shaped the electrons to 
shout:
> Other than the headers, there is no deliberate burst at initial connect
> (but there will be an option for this soon).

You can try it (48 KB of startup buffers) at:

http://mcrg.uib.es:8000/live.ogg
and
http://mcrg.uib.es:8000/high.ogg

The only think I'm still not very convinced is that clients has to wait 
for at least 1 "source iteration" (and 2 in the worst case) in order to 
beging to receive any data at all. This is due to the algorith:

while (source)
    wait for source
    read from source
    send to clients
    check new clients
    add buffers
    etc..
end

I think the following will behave better for startups:

while (source)
    wait for source
    check new clients
    add buffers
    read from source
    send to clients
    etc..
end

It will decrease in a few tenth of seconds the time needed to send data, 
but will increase complexity in tree locks.

<p>
-- 
  ricardo galli       GPG id C8114D34

--- >8 ----
List archives:  http://www.xiph.org/archives/
icecast project homepage: http://www.icecast.org/
To unsubscribe from this list, send a message to 'icecast-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.