Hi,
I wanted to build the Architecture below using LiCode but I don’t know what
each part would consist of and which API I would have to use.
In the System below the “Broadcasting Device” would use the webcam and
audio to stream live recording over to a place where “Clients” which don’t
use any devices or have an output stream, but only an input to watch the
video like you would on YouTube.
What extra information am I required to know?
Also what would be the maximum number of people able to connect to the
room/server and watch the video? (Using Erizo as an MCU) as I would want >
500 people to connect and watch.
https://lh6.googleusercontent.com/-L_6DPgIGckQ/UzSmTNm5m3I/AAAAAAAAAKg/KmnJdS6PJtg/s1600/arch.png
Thank You!
Thank you. I understand how it should work better now.
But what I wanted to know was, if one person creates a stream and like 20
people subscribe to it and watch and while that stream is live, then if
another stream is created will it be too heavy on the servers? or is this
not a problem?
Regards,
MichaelOn Monday, March 31, 2014 1:29:00 PM UTC+3, Pedro Rodriguez wrote:
Hi Michael,
In that scenario, in Licode, you would only have to create a local stream
for the broadcaster. That stream would then be published using room.publish
in the API. Each of the viewers would then only have to subscribe to that
stream once the stream-added event is received. The basicExample
(extras/basic_example) in Licode is a good place to start.
As far as performance goes, the amount of clients supported at the same
time will depend on the CPU available as in WebRTC we have to re-encrypt
the source stream for each participant. The quality (bandwidth) of the
video is also an important factor. There is no “hard” limit in Licode, as
long as you have the CPU/bandwidth/sockets needed, you can keep adding
viewers.
You may also have to update the maximum open files limit for the
erizoController process, as it would have to keep many open sockets (check
“ulimit -a” in Ubuntu)
On a quite old (2008) core2Duo@3.0GHz (E8400), we have tested broadcasts
for up to 30 people at 300Kbps, that takes up to 25% CPU of 1 of the cores
(12,5 in total). The CPU usage grows linearly with the amount of users.
Cheers!
–
Pedro Rodriguez
On 27 Mar 2014 at 23:30:09, Michael Awetahegn (michaelaw...@gmail.com<javascript:>) wrote:
Hi,
I wanted to build the Architecture below using LiCode but I don’t know
what each part would consist of and which API I would have to use.
In the System below the “Broadcasting Device” would use the webcam and
audio to stream live recording over to a place where “Clients” which don’t
use any devices or have an output stream, but only an input to watch the
video like you would on YouTube.
What extra information am I required to know?
Also what would be the maximum number of people able to connect to the
room/server and watch the video? (Using Erizo as an MCU) as I would want >
500 people to connect and watch.
https://lh6.googleusercontent.com/-L_6DPgIGckQ/UzSmTNm5m3I/AAAAAAAAAKg/KmnJdS6PJtg/s1600/arch.png
Thank You!
You received this message because you are subscribed to the Google Groups
“lynckia” group.
To unsubscribe from this group and stop receiving emails from it, send an
email to lynckia+u...@googlegroups.com <javascript:>.
For more options, visit https://groups.google.com/d/optout.
Hi Michael,
In that scenario, in Licode, you would only have to create a local stream for the broadcaster. That stream would then be published using room.publish in the API. Each of the viewers would then only have to subscribe to that stream once the stream-added event is received. The basicExample (extras/basic_example) in Licode is a good place to start.
As far as performance goes, the amount of clients supported at the same time will depend on the CPU available as in WebRTC we have to re-encrypt the source stream for each participant. The quality (bandwidth) of the video is also an important factor. There is no “hard” limit in Licode, as long as you have the CPU/bandwidth/sockets needed, you can keep adding viewers.
You may also have to update the maximum open files limit for the erizoController process, as it would have to keep many open sockets (check “ulimit -a” in Ubuntu)
On a quite old (2008) core2Duo@3.0GHz (E8400), we have tested broadcasts for up to 30 people at 300Kbps, that takes up to 25% CPU of 1 of the cores (12,5 in total). The CPU usage grows linearly with the amount of users.
Cheers!–
Pedro Rodriguez
On 27 Mar 2014 at 23:30:09, Michael Awetahegn (michaelawetahegn43@gmail.com) wrote:
Hi,
I wanted to build the Architecture below using LiCode but I don’t know what each part would consist of and which API I would have to use.
In the System below the “Broadcasting Device” would use the webcam and audio to stream live recording over to a place where “Clients” which don’t use any devices or have an output stream, but only an input to watch the video like you would on YouTube.
What extra information am I required to know?
Also what would be the maximum number of people able to connect to the room/server and watch the video? (Using Erizo as an MCU) as I would want > 500 people to connect and watch.
Thank You!
You received this message because you are subscribed to the Google Groups “lynckia” group.
To unsubscribe from this group and stop receiving emails from it, send an email to lynckia+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
