Lorenzo:
The "this" part is currently only in our minds for the moment, but if
you solve the issue of connecting our "minds" to your ethernet port,
I'll solve the 1ms latency part :)
A simple ping:
$ ping granat
PING granat.lcl.aspodata.se (192.168.93.37) 56(84) bytes of data.
64 bytes from granat.lcl.aspodata.se (192.168.93.37): icmp_seq=1 ttl=15 time=0.758 ms
gives me a round trip of < 1ms, so it should be possible.
So the time from analog, pushing out samples from the sender port,
till it is received the receiver port would be lower bounded by:
sample conversion time
lcl transport time to cpu
cpu processing
cpu to ethernet tranceiver
ethernet preamble (assume quiet net)
cable delay
packet bit speed * packet length
Add to this any processing on the receiving side.
To help achieve low (network) latency one can:
. set up fixed arp entries (see arp(8))
. use fixed ip-addresses, no dhcp, no mDNS, no autoconfiguration or
other talky protocols
. don't mix other hosts on the lan, make it as quiet as possible
. no DNS, use /ets/hosts files
. if possible use a crossover cable [1]
What kind of latency do you think would be a goal to strive to ?
Regards,
/Karl
[1] http://en.wikipedia.org/wiki/Ethernet_crossover_cable
-----------------------------------------------------------------------
Karl Hammar Asp
LINUX® is a registered trademark of Linus Torvalds in the USA and other countries.
Linuxaudio.org logo copyright Thorsten Wilms © 2006.
Hosting provided by the Virginia Tech Department of Music and DISIS.