On Oct 21, 2018, at 12:05, Jordan Erickson <jordan at coolmic.net> wrote:> 283000/5000=56.6 > > So with ~283Mb/s upstream, if you wanted 5,000 listener capacity your > streams would have to have a bitrate of 56.6kb/s or lower.Speaking conservatively, I would de-rate that number by 50%, or in other words: 0.5*283000/5000=28.3 kb/sec. This to account for various overheads (TCP/IP packet framing and negotiation, data bursting at stream startup, administrative interface traffic, etc. It all adds up. Cheers! |----------------------------------------------------------------------| | Frederick F. Gleason, Jr. | Chief Developer | | | Paravel Systems | |----------------------------------------------------------------------| | A room without books is like a body without a soul. | | -- Cicero | |----------------------------------------------------------------------| -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.xiph.org/pipermail/icecast/attachments/20181021/ce7c186a/attachment.html>
Jordan Erickson
2018-Oct-21 17:35 UTC
[Icecast] Configure Ubuntu Server 16.04 for icecast2
Hi Frederick, Wouldn't overhead (at least partly) already be factored in with the speed test data/results? Obviously not 100% comparable to audio streaming, but.. Cheers, Jordan On 10/21/18 9:10 AM, Fred Gleason wrote:> On Oct 21, 2018, at 12:05, Jordan Erickson <jordan at coolmic.net > <mailto:jordan at coolmic.net>> wrote: > >> 283000/5000=56.6 >> >> So with ~283Mb/s upstream, if you wanted 5,000 listener capacity your >> streams would have to have a bitrate of 56.6kb/s or lower. > > Speaking conservatively, I would de-rate that number by 50%, or in other > words: > > 0.5*283000/5000=28.3 kb/sec. > > This to account for various overheads (TCP/IP packet framing and > negotiation, data bursting at stream startup, administrative interface > traffic, etc. It all adds up. > > Cheers! > > > |----------------------------------------------------------------------| > | Frederick F. Gleason, Jr. | Chief Developer | > | | Paravel Systems | > |----------------------------------------------------------------------| > | A room without books is like a body without a soul. | > | -- Cicero | > |----------------------------------------------------------------------| > > _______________________________________________ > Icecast mailing list > Icecast at xiph.org > http://lists.xiph.org/mailman/listinfo/icecast >
Those speed tests are best case scenario - one big old file where the only overhead would be tcp syn/acks. 5000 x syn/acks and handshakes everytime a client connects would quite a lot of overhead I imagine. Note that this is an educated guess - I don't run /anything/ on the scale of 5000 users lol ---- Jordan Erickson wrote ---->Hi Frederick, > >Wouldn't overhead (at least partly) already be factored in with the >speed test data/results? Obviously not 100% comparable to audio >streaming, but.. > > >Cheers, >Jordan > >On 10/21/18 9:10 AM, Fred Gleason wrote: >> On Oct 21, 2018, at 12:05, Jordan Erickson <jordan at coolmic.net >> <mailto:jordan at coolmic.net>> wrote: >> >>> 283000/5000=56.6 >>> >>> So with ~283Mb/s upstream, if you wanted 5,000 listener capacity your >>> streams would have to have a bitrate of 56.6kb/s or lower. >> >> Speaking conservatively, I would de-rate that number by 50%, or in other >> words: >> >> 0.5*283000/5000=28.3 kb/sec. >> >> This to account for various overheads (TCP/IP packet framing and >> negotiation, data bursting at stream startup, administrative interface >> traffic, etc. It all adds up. >> >> Cheers! >> >> >> |----------------------------------------------------------------------| >> | Frederick F. Gleason, Jr. | Chief Developer | >> | | Paravel Systems | >> |----------------------------------------------------------------------| >> | A room without books is like a body without a soul. | >> | -- Cicero | >> |----------------------------------------------------------------------| >> >> _______________________________________________ >> Icecast mailing list >> Icecast at xiph.org >> http://lists.xiph.org/mailman/listinfo/icecast >> >_______________________________________________ >Icecast mailing list >Icecast at xiph.org >http://lists.xiph.org/mailman/listinfo/icecast-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.xiph.org/pipermail/icecast/attachments/20181022/07dd0f5e/attachment-0001.html>