[omniORB] omniorb and omniorbpy speed

Duncan Grisby dgrisby@uk.research.att.com
Mon, 04 Dec 2000 10:01:10 +0000


On Tuesday 28 November, Stefan Seefeld wrote:

> sorry for the plug, but optimization interests me as well...  Didn't
> omniorb try a shm or unix domain socket transport as well ? I
> remember some discussion about that. More specifically, it appeared
> you had reasons not to use it (the tcp/ip transport was faster, or
> something).

There were some experiments with using other transports in omniORB a
while ago, but they were very experimental, and never found their way
into the release. They did make things go faster, but not enormously
so. The two technical reports linked the following page explain more:

  http://www.uk.research.att.com/omniORB/doc/

As part of the work on omniORB 4, the transport code is being
restructured to make it easier to add new transports. A shared memory
transport could well crop up then.

Cheers,

Duncan.

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