At the Electronic Crossroads Once Again: The Myth of the Modern Computer Utility in the United States

The term "computer utility" has been perennially present in the history of computing, though a precise definition of the term has been fleeting, even when it first appeared in the Mid-1960s. The recent resurgence of interest in the remote provision of computing power has brought with it the historical jumble of definitions that make up that term, and "computer utility" is once again being applied to computer architectures though with little justification or exposition given regarding this use. This article attempts to determine whether past and present applications of the term share any commonalities and in doing so questions whether there has ever existed a true "computer utility." To do so, incarnations of the "computer utility" are examined in two time periods--the 1960s and the 1970s--during which time computer time-sharing enjoyed significant prosperity and now as cloud computing has become increasingly popular and prolific.

The Public Utility Model

The term “computer utility” cannot be properly examined without first grounding the notion of “utility” itself. The history of public utilities in the United States is extensive, and a wide array of work has been conducted examining its intricacies.⁶ Some fundamental properties of the public utility can be established, however. In essence, an industry can be considered to be a utility when two characteristics of that industry converge.⁷ The first characteristic is economic, wherein a particular commodity or service can be efficiently provided by monopoly production; the second characteristic is social, wherein “the demand for such commodity or service is sufficiently general so that it has become a common necesslty.”⁸ Traditionally, such a convergence has occurred in industries such as water and electricity, so they serve as useful points of reference.