Expo/Computation/The Metacomputer
| Back |
Map |
Glossary |
Information |
The Parts of a Metacomputer
The beauty of the metacomputing concept is that the user shouldn't have
to know about its components or how they work together, anymore than you need to know
about internal combustion engines to drive a car. But if you're the curious
type who likes to know what makes the wheels of your car turn whenever you
press the accelerator, then this section is for you!
NCSA's System Diagram
JPEG Image (61.1 KB); Credits and Copyrights
The most obvious component of the metacomputer is a computer or, more likely,
an array of computers, also known as "processors." Think of processors as
the "number crunchers."
The researcher uses one type of computer--typically a desktop workstation--to
write, debug and compile codes. A high-performance computer or supercomputer
uses those codes to crunch the data fed to it. A third computer may be
used to
render the results in a graphical form or to power a virtual reality display
device. A
metacomputer, then, assembles the characteristic strengths of individual
computers to complete a certain task.
Getting such a disparate collection of computers to work together would
be like
holding a session of the United Nations without live translation--little
prospect of coordinated action. Making the whole collection user-friendly
requires yet another layer of software. Communications Software
bridges all of the gaps--between different computers, between
computers
and people, even between different people.
A collection of individual computers is just that, without fast, high
band-width networks that provide rapid and reliable connections between the
machines, whether they're in the same building or dispersed across a
continent.
Multiple supercomputers churning gigaflops? Cross-country networks
bristling with gigabytes of data streaming in and out storage? Sounds like a recipe for data
overload!
Handling all the data pouring from countless sites nationwide poses a major
challenge to the emerging technologies of metacomputing. The challenge can
only intensify as scientists demand realtime interaction and control of their
simulations and instruments.
When the number crunchers finish crunching, the user is facedwith the
mammoth
task of making sense of the data. Enter scientific visualization, the
representation of data in meaningful images. But the goals of scientific
visualization are changing. No longer is it enough to generate marvellous
images, even movies, as end-products of research. Now scientists and
engineers
want to see, even "hear" or "feel" their data and interact with it in
realtime.
As visualization and computation become ever more closely coupled, new
environments for scientific discovery emerge: virtual environments.
Return to What is a Metacomputer
Exhibit Map
Glossary
Information Center
Copyright, (c) 1995: Board of Trustees, University of
Illinois
NCSA. Last modified 10/19/95.
Processors