For mp3 streaming this is true (simultaneous streams), i added a log
message at the constructor of the MP3Reader and for each Stream (at
client side a NetStream Object) a new MP3Reader with complete mapping of
the required file is instantiated (even if the required files are the
same).



-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Dan Rossi
Sent: Freitag, 6. Juli 2007 10:02
To: [email protected]
Subject: Re: [Red5] first immpression

Mees, Stefan wrote:
> Hello,
> first thing i have to say i'm really amazed about this great open 
> source community and the professional work that have been done till 
> yet. We evalutate red5 for a streaming solution of plain mp3 audio 
> files via rtmp to a flash frontend. This functionality seems to be 
> stable (a little bug exists, discussed little bit earlier on this list

> with the topic "problems after NetStream.Play.Stop is issued").
>
> I had take a intensive look at relevant code, here are some peanuts 
> that catch my eyes ;-)
>
> org.red5.io.mp3.impl.MP3Reader
>
> - every streaming file are mapped completely at memory (that means if 
> i have 100 simultaneous streams where each file has a size of 4 - 5 MB

> -> 400 - 500 MB of memory are allocated, even if 50 Streams use the 
> same
> file)
>   
Is this true ? We have files over 200MB each, with thousands of
connections to them, hasnt maxed out 1.5GB yet ..



_______________________________________________
Red5 mailing list
[email protected]
http://osflash.org/mailman/listinfo/red5_osflash.org

_______________________________________________
Red5 mailing list
[email protected]
http://osflash.org/mailman/listinfo/red5_osflash.org

Reply via email to