As a first step, you may try to use prebuilded Grid packages for SL
(Globus Toolkit, Sun Grid Engine) from our site:
ftp://ftp.linux-ink.ru/pub/SL/4.x/projects/GRID+HEP/
Best regards,
Oleg Sadov
В Пнд, 10/04/2006 в 13:16 -0600, Santiago Nuñez пишет:
> Hello everyone,
>
> A couple of months ago some people here began to think about enabling
> a national
> e-science infrastructure to aid researchers in their tasks. The main
> objetive is
> to make a definite step for basic and applied science using high-
> performance
> computing and setting up the basement for Grid integration.
>
> Well, things have gone very fast here with respect to clusters and we
> have come
> to a point where the grid becomes necessary. Now, in this field our
> experience
> is limited and we'd like some advice from people that have worked
> closely with
> grid models at many levels. I thought that I could ask some help based
> on your
> experience and work in the grid realms.
>
> We discovered that many, many people are interested in computational
> resource
> sharing and there are lots of applications suitable for the
> initiative. Also, we
> have at our full disposal a high-speed fiber optic interconnection
> linking the
> four national main universities, which enables a huge bandwidth that's
> been
> underused till now. In the present time, some funding has become
> available and
> the initial project has sufficient critical mass due to success
> achieved in the
> past months.
>
> Having this technological landscape, we had a meeting and decided to
> start the
> national grid project. In CR, there's one person which has direct
> experience in
> grids, he is Ph.D. Cesar Garita and his doctoral thesis was involved
> with VO's
> establishment in The Netherlands. Currently he is directing the School
> of
> Computer Engineering at the Costa Rica Institute of Technology (ITCR)
> and his
> time is quite short. In addition to our cluster (National Center for
> High
> Technology), we found some others that were hidden under the hood and
> had few
> projection in the local scientific community; now they will become a
> small grid
> with a third cluster in construction. They are part of the grid
> picture too. A
> fourth recent cluster at ITCR is also making the first steps.
>
> Let's go to the point. There are some elements which we need some
> advice (I know
> your time is also limited) or some contacts that may be interested in
> giving
> some help. Currently we can't finance anyone to be here but key ideas
> are
> important, particularly to learn from errors that happened in similar
> circumstances.
>
> These are the areas:
>
> - Technical aspects of grid infrastructure. Here the concern is to
> implement a
> low-cost solution, without trimming too much functionality and
> ensuring scalability.
> - Definition of the workflow model of computations.
> - Definition of administrative policies and processes.
> - Implementation of logistics and coordination for e-science
> activities (this is
> the priority)
> - Implementation of logistics and coordination for business activities
> - Cost management of services (the economic model for universities and
> research
> institutes)
> - Definition of roles, resposibilities and norms.
>
> I know that all of them are wide topics, all of them with their own
> subproblems
> and entire research activities. The initiative has a lot of strength
> and we
> don't want lose this inertia, so the hard part begins right now.
>
> The other person involved with this strategy is Ph.D. José Castro
> ([log in to unmask]).
>
> Thanks and if you have any questions or ideas, just write back.
>
>
> Santiago Núñez Corrales
>
> Co-coordinator, National Grid and Cluster Computing Strategy
> Computing and Advanced Informatics Reserach Center
> National Center for High Technology
|