both CORBA and ROOT have some "marshalling" (composing/decomposing) overhead.
CORBA however has to send many a number of messages around to find the
destination before it can contact the real destination (via the ORB).
CORBA and ROOT net packages are basically the same in case they both send
the same data. ROOT however can (optionally and transparently) compress large
messages before sending them (just like when writing to the database). In
that case ROOT messages will be much (2-5 times) smaller. However, when sending
small trivial objects around there is not much need for compression. As said
before, the main difference is in WHAT you can send around. CORBA does not
allow for the sending of objects. That is, you can not send, for example, a
complete histogram object from one process to another. You only can send a
histogram reference to another process. This other process can then execute
some member functions on the referenced object. With ROOT (and Java RMI)
you can send complete objects, which is often what you really want
(e.g. DAQ or monitoring systems).
Cheers, Fons.
Selim Issever wrote:
>
> On Fri, 30 Jan 1998, Fons Rademakers wrote:
>
> ]Hi Selim,
> ]
> ] Corba is between a factor 5 to 20 slower than ROOT's IPC system.
> ]This is mainly due to the large overhead the ORB has to find out where
> ]what object resides.
> [...]
> I still have a question left,..
>
> I guess this factor counts for composing/decomposing the Corba-request,..
> this means overhead _in_ the Server or Client,..
> But what about the bandwidth usage? are Corba net-packages much
> bigger than ROOT net-packages? Or is the overhead comparible? (Or is there
> just no overhead in both cases - sorry I am just very new to Corba and
> Root IPC)
>
> Best Regards
> Selim
>
> S E L I M I S S E V E R
> DESY-F15, Notkestr. 85, 22603 Hamburg, Germany; Tel/Fax: 040 8998-2843/4033
> http://www.physik.uni-dortmund.de/~issevers; selim.issever@desy.de
> Ete kemige burundum, Yunus diye gorundum. Yunus Emre
-- Org: CERN, European Laboratory for Particle Physics. Mail: 1211 Geneve 23, Switzerland Phone: +41 22 7679248 E-Mail: Fons.Rademakers@cern.ch Fax: +41 22 7677910