High-Energy Physics Applications on the Grid - Grid Computing: Infrastructure, Service, and Applications - page 424

infrastructure at the scale of the grid should not only demonstrate its
value for a large number of users like a super computing center, but also
bring additional added value to its users.
On an infrastructure like the grid, the applications sit side by side and
benei t from each other's experience. The fact that every activity had some
specii c (possibly nongeneral) use case is largely counterbalanced by the
fact to i nd (in a sister application) colleagues sharing solutions, advising,
and so on. A team of scientists (or a company) should join the grid because
the balance between the advantages of the new technology is largely
exceeding the aggravation to change part of their working system (which
is at the base of their activity or their business). These examples and new
opportunities of collaboration are the real method for attracting new
applications and involving them in using grid computing.
The convergence among applications is not easy and cannot be estab-
lished “by decree.” There are very positive examples, even between differ-
ent communities as I mentioned, but should not be the only parameter for
success. The convergence on common solutions, even in the HEP commu-
nity, is not automatic and has not been achieved completely. There are
good reasons for this: computing is not a generic tool (at least not yet) and
it is on the critical path to get faster and better results. It is therefore under-
standable that (as we will describe in what follows) in a few cases we can
already observe full convergence. In other areas, more convergence will
be eventually achieved in the near future. Ultimately, some diversity
will stay.
I think it is difi cult to overestimate the importance of the visionary
power to the irrevocable move to the grid as the solution to all the comput-
ing of all leading-edge activities in LHC. This is something we have
observed only in HEP so far, namely to commit the success of the most
important scientii c program to the usage of this new technology. Due to
its history and computing requirements, HEP was the only science being
ready, both sociologically and technically, to move to the grid to build its
computing infrastructure. Often a comparison between the Web (invented
at CERN during the LEP period) and grid technologies is made. The next
few years will tell whether this comparison is appropriate.

17.2
In the case of HEP, a specii c effort was set up in the years 2004-2008 to
investigate the usage of the grid for the so-called end-user analysis: the
ARDA project. In the following we will use some of the activities of this
project to guide us in the HEP usage of grid technology and of the LCG
infrastructure in particular.

The ARDA Project in LCG/EGEE