UMS Auto generate Live Broadcast

Various experience with applications based on Live Server

UMS Auto generate Live Broadcast

Postby justin1997 » Sun Mar 24, 2019 7:43 pm

Hello there,

I am new to Unreal Media server and am very pleased with the test result so far.
I am now about to purchase the unlimited license for my service but I just have couple of questions about the capabilities.

1. I need a way to connect my web app with UMS. The desired situation is: when the user creates a new job on the web app, it will create a unique alias name for the job which will then be read by UMS. UMS will need to automatically create a new RTMP live broadcast with this alias. Is this possible with UMS APK?
2. I also need to save the recording to the local server UMS is running on.
3. How does UMS perform in terms of scaling? Can a single server handle 300 camera sources with maybe 1000 viewers at a time? If so what are the required hardware spec needed for the server?

Looking forward to adapt the server into our business.

kind regards,
Justin
justin1997
 
Posts: 0
Joined: Thu Mar 21, 2019 11:39 pm

Re: UMS Auto generate Live Broadcast

Postby admin » Sun Mar 24, 2019 9:51 pm

Hello,

1. You need to use ServerConfig COM dll from our SDK: http://umediaserver.net/umediaserver/source.html

2. The SDK also has COM objects for Archival Server automation, for recording purposes. You will need to run Unreal Archival Server as well.

3. Hardware specs and performance graphs can be found at http://umediaserver.net/umediaserver/Un ... 0Specs.pdf
Your main limiting factor here is bandwidth; depending on your stream bitrate, the server will need to have 1Gb, 10Gb or higher network connection speed, in order to handle such numbers. Streaming to 1000 concurrent players is not an issue and doesn't take much CPU with the exception of WebRTC players - these use 5 times more server resources than RTMP players, so realistically i7 quad core CPU will handle only about 500 WebRTC concurrent players. But again, it is very likely that 500 concurrent streams will use all your bandwidth, so most probably the bandwidth will be your bottleneck.
Same story about 300 cameras publishing to the same server - it's not a CPU or memory intensive process, but bandwidth will set your limits.
If you also want to use Archival Server to record all these streams, you will need a high performance external NAS drive able to record in parallel.

Realistically, from our customers experience, taking 720p 1000 kbps stream as a typical stream and 1Gb network card, you should plan to have a maximum of 100 concurrently publishing cameras on the same server, and streaming to 500 - 800 concurrent players on that server.
You can scale by adding another network card, more CPU cores (for WebRTC players), or just running another server machine.
admin
Site Admin
 
Posts: 1042
Joined: Fri Aug 21, 2009 10:13 am


Return to Live applications

Who is online

Users browsing this forum: No registered users and 0 guests

cron