Definition: Metacomputing involves using multiple computers linked
by high-speed networks to solve application problems; they appear to the user as
a single computer but provide a level of performance far in excess of individual
computers in the network.
Metacomputers or computational grids function like a networked virtual
supercomputer. Metacomputing was born out of a need to utilize greater
processing power than was available in a single site, and to combine the power
of computers with different architecture.
Metacomputing is also useful for collecting, manipulating and analyzing data
databases and instruments like microscopes, telescopes and satellite
downlinks. Large-scale data intensive applications require high-performance
computing, high-speed networking,
data storage facilities and interactive software.
In addition to science and engineering, high-performance computing is now
required for creating new films, weather forecasting, designing new drugs, etc.,
and system simulation is used by engineering companies for rapid prototyping and
to reduce the time to market for new products.
Metacomputing provides users with the power of a supercomputer, without the high
cost. Small companies need high-performance computing for developing new
products, but their limited requirement may not justify the high installation
and maintenance costs of such systems.
Metacomputing offers an economical solution and allows small companies to access
high-performance computing capabilities, as and when they are required. Users
can pay for their actual usage and avoid the burden of high ownership costs.
Metacomputing allows users to gain access to desktop supercomputing with
powerful graphics capabilities and to use distributed supercomputing to solve
complex problems. People can interact and collaborate with other users in
different geographical locations.
When high performance computing facilities are centralized there may be a lack
of flexibility and redundancy. A failure could cause major disruption to the
Distributed heterogeneous computing uses computing installations in different
locations, linked by a high-speed network. There is a big reduction in the
requirement of bandwidth and a failure in one component is a local problem and
does not cause the collapse of entire grid. Different service providers can be a
part of the network and compete for customers, leading to a reduction in prices
and enhanced service quality.