Increasing Functional Specificity over Increasingly Commoditized Hardware - Guide to Cloud Computing for Business and Technology Managers - page 15

computational cosmology and astronomy, microprocessor research and
design, and large-scale data mining.
Experts continue to predict the merging or disappearance of
several computer categories. They think, for example, that many
midrange and mainframe systems have been made obsolete by
the power and versatility of networks composed of microcom-
puters and servers. Other industry experts have predicted that the
emergence of network computers and information appliances for applica-
tions on the Internet and corporate intranets will replace many per-
sonal computers, especially in large organizations and in the home
computer market.
Interconnecting microprocessors to create minisupercomputers is a reality.
The next wave was looking at harnessing the virtually infinite amount of
unused computing power that exists in the myriad of desktops and laptops
within the boundaries or outside of a modern organization. Distributed or
grid or cloud computing in general is a special type of parallel computing that
relies on complete or virtual computers (with onboard CPU, storage, power
supply, network interface, and so forth) connected to a network (private, pub-
lic, or the Internet) by a conventional or virtual network interface. This is
in contrast to the traditional notion of a supercomputer, which has many
processors connected together in a single machine. While the grid could be
formed by harnessing the unused CPU power in all of the desktops and lap-
tops in a single division of a company (or in the entire company, for that
matter), the cloud could be formed by harnessing CPU, storage, and other
resources in a company or external service providers. Chapter 12 discusses
Grid Computing, while Section III, Chapters 13 to 20, discusses the nature
and characteristics of Cloud Computing.
The primary advantage of distributed computing is that each node can be
purchased as commodity hardware; when combined, it can produce com-
puting resources similar to a multiprocessor supercomputer, but at a signifi-
cantly lower cost. This is due to the economies of scale of producing desktops
and laptops, compared with the lower efficiency of designing and construct-
ing a small number of custom supercomputers.
1.4 Parallel Processing
Parallel processing is performed by the simultaneous execution of pro-
gram instructions that have been allocated across multiple processors with