I'm having some trouble getting the performance I want
out of bridging (and that I know is possible from
previous poster's results).

Here's my setup (machines connected only to each
other, "B" acting as bridge):

A <-> B <-> C

Here are the averages from tests I've run, using plain
binary mode FTP transfers as an informal benchmark (5
iterations).

 Kernel Bridge     User-space Bridge
(A <-> B <-> C)     (A <-> B <-> C)
-------------------------------------
  17.45 Mbps          16.6 Mbps


I would expect that the kernel bridge would have more
dramatic improvement, but the real kicker is when I
compare this with direct ethernet connections between
machines:

A <-> C (direct)        : 89 Mbps
A <-> B (eth0)          : 89 Mbps
C <-> B (eth1)          : 57 Mbps


The bridge has an epic100(eth1) and a 3C509B(eth0). 
Could the mismatch in hardware be causing the
slowdown?  Do I need to do some sort of tuning or
shaping to get better throughput with the bridge?  I
know I shouldn't expect to match the direct connection
speeds, but I've heard claims of > 70Mbps using the
kernel bridge.  I need to reach at least the high
40s... Any ideas?

Thanks!

  Al

__________________________________________________
Do You Yahoo!?
Great stuff seeking new owners in Yahoo! Auctions! 
http://auctions.yahoo.com
_______________________________________________
Bridge mailing list
[EMAIL PROTECTED]
http://www.math.leidenuniv.nl/mailman/listinfo/bridge

Reply via email to