RV: Beowulf as a database server

Robert G. Brown rgb at phy.duke.edu
Sun Dec 3 13:03:39 PST 2000


On Sun, 3 Dec 2000, [iso-8859-1] Juan Manuel García García wrote:

>
> It is worth to use a Beowulf cluster as a database server?
> Somebody have any experience on this ?
> I will appreciate any suggestion.

It is worth it to use a cluster as a database server, but the cluster
would not be a "beowulf".  A beowulf is a high performance computing
cluster, not a data or web server cluster.  Beowulfs are typically used
to perform what are usually thought of a "supercomputing" applications
that are numerically intensive -- weather prediction, monte carlo
calculations, computational fluid dynamics, quantum field theoretic
calculations, and so forth.

This isn't an attempt to pick nits -- it's just that although all
beowulfs are clusters, not all clusters are beowulfs.  Server clusters
of any kind (except possibly computation server clusters) are not
beowulfs.

Now, with that said I will NOT say that one might not want to integrate
a parallelized database server with a parallelized numerical application
-- something like this might well facilitate work and resource division
in something easily parallelizable like making movies (rendering lots of
images that have to be kept sorted and crossreferenced) or processing
lots of data slices from e.g. CERN or Fermilab.  I know of at least one
High Energy physicist who is interested in trying to interface a
database of just plain old code with beowulf-style compute resources --
at this point HEP has such a huge base of contributed code, most of it
terribly documented and written by random graduate students over
decades, that connecting published results with the code used to
calculate them (so they can be checked!) is long since become a major
problem.

So databases (parallelized or otherwise) might be integrable with a
beowulf (parallelized at least if the good parallelized database
products weren't astronomically expensive) but database operations are,
for the most part, not viewed as part of the beowulf purpose.

    rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu







More information about the Beowulf mailing list