From the Computer Utility to Cloud Computing - To the Cloud: Big Data in a Turbulent World - page 67

CIA planned to tender AWS a $600 million contract until IBM blew the
whistle, raising questions about how the federal government handles cloud
contracts, and a review of the agreement with AWS (Woodall 2013). While
waiting to learn whether its bid for the CIA's cloud business would suc-
ceed, IBM won the largest government cloud-computing contract, worth
$1 billion, from the Interior Department (Miller and Strohm 2013). That
helped cushion the blow for IBM when Amazon was oficially awarded
the CIA contract (Babcock 2013a).
These moves are not very surprising, particularly in light of the history
of the U.S. government's relationship to large communication companies
(Mazzucato 2013). For years, government agencies, including the Depart-
ment of Defense, had a very close relationship with IBM for computing
and an even closer one with AT&T for telecommunications services.
Even as business consumers lined up to support breaking up AT&T and
deregulating the telecommunications industry in order to lower prices, the
DOD argued that national security required the end-to-end service that
AT&T provided. It was not until the Pentagon was assured that security
needs would be met that it dropped its opposition to breaking up the
telecommunications giant (Schiller 1981). Given this preference for large,
stable companies, it is not surprising that the government would turn to
AWS to meet some of its cloud-computing needs.
The U.S. government's current move to the cloud is propelled by the
belief that cloud computing must become a central means of meeting its
information-technology needs. In December 2010 the federal Chief Infor-
mation Oficers Council released a plan to reform government information
technology, which included requiring agencies to adopt a “cloud-irst”
policy for new IT deployments. According to the plan, cloud-irst is driven
by three interrelated forces. First, large data centers provide economies
of scale that are necessary to meet the growing needs of the federal gov-
ernment's “computation infrastructure.” For federal IT planners, it is
less expensive to centralize data in a few large centers than to retain it in
local ofices. Second, cloud systems are able to provide almost any type
of computation on demand. It is dificult to predict the type and speed
of processing and analysis that will be needed and the planners side with
those who believe that cloud systems are agile enough to meet their needs,
including those they cannot now anticipate. Finally, the cloud unleashes
unprecedented analytics capability on large data collections. It is clear