Scientific Computing Environment: CORBA Components for the GRID
Today's scientific applications require access to clusters of computers,
supercomputers, storage systems, data sources and different devices so
researchers can work on a particular topic. The requirements to develop
such applications are paramount. Not only are they distributed,
heterogeneous and use different data sources, but they should be of high
performance, maintainable, adaptable and user friendly. The programmer of
such an application must be knowledgeable of many programming techniques
(e.g. multi-threading, transaction, persistency) and has to know about
distribution issues (e.g. global time, location, message ordering) as
well as how to write algorithms. This can only be achieved by an
environment which offesr extensibility and reuse through the synergetic
integration of different tools (e.g. compilers, debuggers), libraries,
executables.
Recently many scientific domains (e.g. High Energy Physics and Earth
Observation) have decided to move towards a form of computing envisioned
by the Cluster Computing (CC) community, called the GRID, as a possible
solution. The GRID is analogous to the electric power grid and aims to
couple geographically distributed data and computing resources and offer
easy access to them.
On the other hand the Object-Oriented (OO) Computing community have come
up with the idea of the Object Request Broker as an abstraction layer for
distributed client-server applications. This led to the development of
CORBA and Components (widely used in the industry) as the solution for
extensibility and reuse.
Until now the CC and OO communities have developed their solutions
independently but the organisers believe that both communities would
benefit from a common discussion.
The Call for Papers is here.
- Zsolt Kovács
- zsolt.kovacs@cern.ch
- CERN/UWE
- Lászlo Varga
- laszlo.varga@sztaki.hu
- MTA SZTAKI