[omniORB] Re: omniorb-list-digest V1 #254

Chris Newbold cnewbold@laurelnetworks.com
Thu, 26 Oct 2000 11:58:37 -0400


From: "omniorb-list-digest" <owner-omniorb-list-digest@uk.research.att.com>
Sent: Wednesday, October 25, 2000 2:00 AM

> We are reworking all sorts of things for what will be omniORB 4. This
> is one of the things we will look into. Invocations though a servant
> manager of some kind will always be slower than ones to a servant
> which is already activated, though, since a lot more code has to run.

>From looking at the code path, it doesn't seem like there's really
that much extra overhead. A couple of extra function calls to get
to the servant manager, and of course what ever overhead I add myself
in my implementation.

I'm not concerned about this overhead so much as the hit of doing a
context switch and the trip through the loopback interface. It turns
out that there are a number of optimizations I can implement if local
calls are dispatched on the same thread (i.e. by keeping some state in
thread-local storage). Obviously I cannot depend on this for correct
operation, but it's a nice boost.

-Chris Newbold
Laurel Networks, Inc.