[Red5] FMS, Windows Media Server and Red5 - LONG

[electroteque]

Hi, firstly Id like to say, well done for the open source addition of a 
flash media server. We are wanting to move away from using Windows 
Media, because of the lack of support for other browsers/platforms than 
IE/PC, and start streaming Flash VOD content and webcam/video 
conferencing applications.

I had a few questions to ask, basically is there any kind of 
administration/server monitoring console available for RED5 yet ?, just 
like with the windows media console snap-in which lets you edit its 
properties. Not even the Flash Media Server lets you do this, its just 
a monitoring admin application, there is an admin API which requires 
you to code your own admin interface and for that kind of money 
($4500USD) you would expect one supplied, like the windows media 
snap-in it comes supplied with the windows server they would be nuts to 
expect you to use their C# .NET API for instance to code your own 
admin, even though every single part of the server is exposed via a 
scriptable .NET API which is nice.

Im assuming ever part of this server is full scriptable and extendable 
via java which is using the spring framework ? This would mean then, 
that the differences between the application code for the flash media 
server is in actionscript2/3 and the code used for server side 
applications in RED5 is java based, but Im assuming they both hold the 
same principles, possibly using similar package names etc ?

Ideally we need to build a VOD server side playlist when a webcam is 
not showing and then when a webcam is on, it will switch to the 
stream-name of the webcam from the playlist and update all the clients, 
and then when it stops broadcasting and disconnects it will return back 
the VOD playlist.

The live stream also needs to be archived on the server during that 
period, is there any application examples in RED5 which shows how to 
create server side playlists and archive live streams ? In windows 
media broadcast publishing points (application end points) can be setup 
with a server side playlist which points to VOD content in a directory 
which will play the files then switch to the next element in the 
playlist which is the webcam broadcast publishing point (application 
point) when it is turned on by the encoder, so yes i guess can RED5 
switche to the next xml element which is connecting to the webcam 
application on some local protocol as with windows media
lpp://theapp.

Our other question was regarding logging, I noticed log4j is the logger 
of choice, which is a great tool, ive used that aswell as log4net. Is 
it able to use W3C style formatting, which is then easily parsable 
using the LogParser library for .NET ? This will mean there is not much 
migration required with our current setup which parses windows media 
logs in W3C format for collecting bandwidth used from each publishing 
point (stream name/ application).

The final question is, how can we replicate windows media publishing 
points which separate and individualize streams which is useful for 
logging people separately, and how is some kind of authentication 
plugin able to be setup on each "publishing point" to only enable for 
instance a website to access it via a webpage using some kind of key 
authentication, im assuming session keys can be built into the java 
application , so it prevents stuff like hot-linking.

Regarding hot-linking, i had built a session based system to try and 
keep malicious users out of getting access to the video-links sent to a 
windows media object, however as it was streamed based, the only way to 
get the url to output when loaded in the plugin was outputting the link 
to the server via an asx playlist or whatever, using packet sniffers 
this was easy to find, even the player exposed the url.

On testing flash video with connections to the media server I was able 
to obtain the media server url, application and stream-name exposed in 
the metadata, and load the stream in a test receiver projector 
application. So I cannot see how macromedia/adobe can claim flash video 
is secure maybe progressive downloadable media accessed via a proxy 
script and sessions yes, so my question here would be what kind of 
communication can be enabled to Red5 to not expose this information or 
use some kind of key authentication calling an application method to 
not expose some information.

We offer free and paid services, so obviously the free stuff isn't a 
problem, but being able to packet sniff exposed information in flash 
makes you wonder.

Im sorry about the long set of questions, it was needed, as we've had 
to ditch Adobe being they were rude and arrogant in returning technical 
information like my questions above and sales quotes to a possible 
customer which they have lost.

My preference is to support open source firstly at all times, I can 
even convince my boss to donate a fair sum of money if he is happy with 
it considering the stupid license for FMS. My boss is fascinating over 
Adobe's origin/edge load balanced FMS systems, would there be an easy 
way to kludge something similar in RED5 using a load balancer to a few 
machines accessing a shared network drive on a video content server for 
the content etc ?

Let me know, and sorry about the essay thanks.

Daniel