Increasing Functional Specificity over Increasingly Commoditized Hardware - Guide to Cloud Computing for Business and Technology Managers - page 10

punched with holes, it was possible for the Jacquard loom to weave fab-
rics in a variety of patterns. Jacquard's loom was controlled by a program
encoded into the punched cards. The operator created the program once
and was able to duplicate it many times with consistency and accuracy.
Herman Hollerith eventually adapted Jacquard's concept of the punched
card to record census data in the late 1880s. Hollerith's machine was highly
successful; it cut the time it took to tabulate the result of the census by two-
thirds, and it made money for the company that manufactured it. In 1911,
this company merged with its competitor to form International Business
Machines (IBM).
First Generation: Vacuum Tube Technology, 1946-1956
The first generation of computers relied on vacuum tubes to store and pro-
cess information. These tubes consumed huge amount of power, were short-
lived, and generated a great deal of heat. First-generation computers were
colossal in size, had extremely limited memory and processing capability,
and their usage was restricted to limited areas in science and engineering.
The maximum main memory size was approximately 5000 bytes (5 kilo-
bytes), with a processing speed of 10 kilo instructions per second. This gen-
eration employed rotating magnetic drums for internal storage and punched
cards for external storage. Jobs such as running programs or printing reports
had to be coordinated manually.
Second Generation: Transistors, 1957-1963
In the second generation, the vacuum tubes were replaced by transistors
for storing and processing information. Transistors were much more stable
and reliable than vacuum tubes; they generated less heat and consumed less
power. However, each transistor had to be custom made and wired into a
printed circuit board—a slow and tedious process. Magnetic core memory
was the primary storage technology of this period. It was composed of small
magnetic doughnuts with 1 mm diameter, which could be polarized in either
of two possible directions to represent a bit ( binary digit ) of data. This whole
system had to be assembled by hand and, hence, was time consuming and
very expensive. Second-generation computers had a random access memory
(RAM) of up to 32 kilobytes and processing speeds of about 200 kilo instruc-
tions per sec to 300 kilo instructions per second. The enhanced processing
power and memory of the second-generation computers enabled them to be
used most widely not only for scientific and engineering work, but also for
business work (like payroll and billing).
Third Generation: Integrated Circuits, 1964-1979
Third-generation computers relied on integrated circuits, which were made
by printing hundreds and later thousands of tiny transistors on small silicon