Evaluating the Performance of CORBA for Distributed and Grid Computing
Applications
T. Es-sqalli, E. Fleury, J. Guyard, S. Bhiri
RESEDAS Project, LORIA, UHP Campus Scientifique, BP 239
F-54506 VANDŒUVRE-LÈS-NANCY cedex France
Abstract
Distributed computing applications rely on transmis-
sions of messages between processes. However, as work-
stations are not intended to manage this kind of communi-
cations, it is necessary to use some communication tools
known as Message Passing libraries. Currently, MPI (Mes-
sage Passing Interface) and PVM (Parallel Virtual Ma-
chine) are the most used.
In addition, the need for interoperability among the
rapidly proliferating number of hardwares and softwares
led to the definition of CORBA (Common Object Request
Broker Architecture), which standardizes the execution sup-
port of distributed object applications. This architecture
offers a good interoperability and encapsulation support.
Moreover, its method invocation mechanism, offers sev-
eral features and can be used as a good communication
model. Consequently, the use of such architecture for dis-
tributed and grid computing applications seems interesting
and promising. However, CORBA is always criticized on
the performances level.
By taking account of all these elements, a performance
evaluation of CORBA seems necessary.
The topic of
this paper is to present our benchmark results of some
CORBA implementations (ORBacus and TAO) and to com-
pare them with those of the two message passing libraries
MPI (MPICH and LAM) and PVM, by using a Network of
Work Stations (NOW).
1. Introduction
CORBA [6] is a distributed object system that defines ser-
vices and supports which an application needs in order to
be carried out in a distributed environment. The commu-
nication platform, or ORB (Object Request Broker), is the
middle-ware that establishes the client-server relationships
between objects. Using an ORB, a client can transparently
invoke a method on a server object. The client does not
have to be aware of where the object is loc