[omniORB] ORB performance issues

[Duncan Grisby]

On Wednesday 13 February, Grealy Colin wrote:

> I have been doing some performance testing with omniorb4.0 on redhat. 

[...]
> We found that the time per call spiked at certain parameter sizes, which are
> shown on the graph. In some cases a particular size message took 2-3 times
> longer then a message a few bytes larger. Has anyone else experienced a
> similar problem or does anyone have any light to shine on this issue?

Are you running the client and server on the same machine?  There is
something very dodgy about the Linux TCP loopback interface. I have
seen differences in invocation times of up to ten times, with longer
transfers taking less time than shorter ones. If you are using
loopback, try enabling the Unix socket transport with command line
arguments -ORBendPoint giop:unix: -ORBendPoint giop:tcp:: . That
should make the problem go away.

If you aren't using the loopback, then there is obviously something
else going on. In this case, let me know and I'll look into it.

Cheers,

Duncan.

-- 
 -- Duncan Grisby  \  Research Engineer  --
  -- AT&T Laboratories Cambridge          --
   -- http://www.uk.research.att.com/~dpg1 --