Subject: Re: [linux-audio-dev] Distributed audio processing
From: Benno Senoner (sbenno_AT_gardena.net)
Date: Thu Oct 05 2000 - 23:17:21 EEST
On Thu, 05 Oct 2000, John Lazzaro wrote:
> > Theory apart, using current infrastructures (error correction modems,
> > etc.), and following your experience, UDP has shown useful to obtain
> > better latencies?
>
> The zero-order issue is this: lets say you have a nominally low-latency
> link (for example, Stanford-Berkeley round-trip times from Linux machine
> to Linux machine is < 4ms now), but 1 in 100 packets get delayed to
> 20-30ms or dropped altogether (which is how the Stanford-Berkeley link
> seems to operate during heavy-load time periods).
>
ok, but what about the following ?
a cluster of machines directly interconnected with crosscables,
or high quality switches, with no other disturbing "random" traffic on
the LAN.
Assume all links to be at least 100Mbit.
I have seen graphs on the myrinet.com site about latencies well
below the 50-100usec mark.
So I was wondering if one could achieve worst case latencies of 5-10msec
on a 100Mbit Lan.
(5msec "guaranteed" would be ideal, since it would allow us to run
a MIDI synthesis engine on a beowulf cluster)
:-)
Benno.
This archive was generated by hypermail 2b28 : Thu Oct 05 2000 - 21:54:21 EEST