hi,
fred, thank you very much for your full answer
I will verify these points in future conferences; top gives usually an
amount of 25% of usage CPU
Is a good plan to separate the server bbb and the server freeswitch ?
(to have a better control of freeswitch)
Because for us, the audio is the most important functionnality in
conferences, and a poor audio quality turns against bbb itself
In this case, is there a wiki to make the server bbb using a distinct
freeswitch server ?
Thanks a lot for all your work
Regards
On 28 août, 22:35, Fred Dixon <
ffdi...@gmail.com> wrote:
> Hi,
>
> There are a number of areas to check, and it may be a combination of
> problems that are causing the poor audio performance. The list below is
> not exhaustive, but will give us some ideas of where to check further.
>
> 1. Bandwidth available to your hosted zone server. We recommend, at a
> minimum, to have 100 Mbits/sec symmetrical bandwidth (both download and
> upload speeds are, at least, 100 Mbits/sec). You want to validate what the
> actual bandwidth to/from your hosted zone server is.
>
> To test: do you have access to a second, external server with good
> bandwidth? If so, can you try transferring a large file to/from your
> hosted zone server to the second external server using scp to get a rough
> idea what the maximum transfer rate is.
>
> 2. CPU available on your hosted zone server. In general, we recommend
> running BigBlueButton on a dedicated (non-virtual) server. See
>
>
http://code.google.com/p/bigbluebutton/wiki/FAQ#What_are_the_minimum_...
>
> To test: try logging into the hosted zone server and watching the CPU using
> top. Login with one, two, three, four, five, etc. external clients
> (external to your corporate network). Does the CPU go above 70%? In our
> experience, when the CPU consistently stays above 70%, you'll likely get
> audio troubles.
>
> This test may not reveal that your hosted zone VM is on a host server that
> is in heavy use by other virtual machines. In a virtualized environment,
> there is no guarantee on how much CPU time your server will get. If it
> can't keep up with incoming packets, audio will suffer.
>
> Can your ISP provide you a dedicated server for testing BigBlueButton? If
> so, that would help rule out running in a VM that is getting little CPU
> cycles.
>
> 3. Network connection to/from the hosted server from external clients is
> poor.
>
> To test: From the external clients (again external to your network), login
> to BigBlueButton on your DMZ server behind your firewall and test that the
> audio is good (this will calibrate that your external users
> have sufficient and stable network to your server).
>
> Then login to the hosted server and test the audio again. Is there a
> difference? If the same clients are hearing a poorer experience on the
> hosted zone VM, then the problem may not be the CPU of each server; rather,
> it's the network connection to each server.
>
> 4. Clients network connection is poor.
>
> To test: Have the clients usehttp://
speedtest.net/. They may say their
> ISP provides them a good connection, buthttp://
speedtest.net/will test
> their upstream/downstream bandwidth and give you a reading on their actual
> connection speeds.
>
> We recommend clients have 0.5 Mbits/second upload speed and 1.0
> Mbits/second download speed. Of course, these are not hard numbers, and
> BigBlueButton will certainly work with less bandwidth, but if your clients
> have bandwidth in this range, they should experience good audio (assuming
> that the problem is not 1, 2, or 3).
>
> If you can, try some of the above tests and let us know the results.
>
> Regards,.... Fred
> --
> BigBlueButton Developerhttp://
bigbluebutton.org/http://code.google.com/p/bigbluebutton
> >
http://groups.google.com/group/bigbluebutton-users/browse_thread/thre...