Nicholas Carr is a prominent commentator on the tech biz scene, a former executive editor of the Harvard Business Review and proprietor of a blog called Rough Type [www.roughtype.com] that manages the hat trick of being thoughtful, entertaining and a tad acerbic.
In 2004, he published “Does IT Matter?,” which argued that information technology was rapidly losing its one-time status as a source of competitive advantage for business. In Mr. Carr’s view, IT has ceased to matter in the sense that it has become so ubiquitous that the playing field is now level, and a bank, an airline, an auto company is unlikely to gain an edge by glomming onto the next jazzy idea.
The book stirred a lively debate — CIOs and CTOs were not persuaded and said so loudly, and the dispute has continued since.
In “The Big Switch,” Mr. Carr takes his line of thought a step further. Drawing heavily on the historical experience with electricity, and the changeover from localized, area- or factory-specific generators to massive central facilities with distribution over wires, he says that computing is on the same general path. Individual businesses, rather than running their own IT centers, will increasingly rely on central computing centers.
Mr. Carr grants some of the problems. Computing requires continuing interactions; it is not the same as an electrical grid, which has done its job when it delivers standardized power to the plug, leaving the entire specific application to the customer. Businesses are wary of committing their vital data and apps to the Internet cloud, and so on.
Nonetheless, he thinks the efficiencies, including the benefits of not needing an internal IT department, are too great to be denied, and that the trend is ineluctable.
It is a strong case, and events are making it stronger, as Google and Microsoft build out data centers, Amazon actually offers a variety of computing utility, Sun sells its data-center-in-box and on-line services such as Salesforce make headway.
There are some counters that Mr. Carr doesn’t pay much attention to that that will not be omitted by skeptics. One goes to the heart of the original thesis about IT; there may still be competitive advantage to be wrung from proprietary IT, as is demonstrated by the exemplar of utility computing, which is Google. Google’s competitive advantages lie in its proprietary search algorithms, and in the proprietary tweaks it has made to the open source Linux operating system. It is not entirely clear that similar strategies are totally useless in other enterprises. As Mr. Carr points out, 30 years ago, American Airlines gained substantial advantage from computerized reservation systems.
In addition, as storage and processing costs decline, the benefits of avoiding them lessen. To the ire of programming purists, part of Bill Gates’ genius was to realize that economy in code writing was worth little in a Moores’ Law world. There is also the nanosecond problem. An optical bit moves only about a foot in a nanosecond, which means that sending it a thousand miles takes 5/millionths of a second, an eternity in compuworld, so the hope of a totally dumb terminal performing high powered tasks at a distance is vain. Significant localized processing and storage power must always exist, which reduces potential gains. If the resource to be economized is the time of IT professionals, then intermediate approaches may be possible, with computing power distributed and its administration centralized.
Finally, there is the ever-present specter of government. For utility computing to work, businesses must be sure that they can buy the services they need. A continuing threat that the FCC or the Antitrust Division reserves the right to impose net neutrality, unbundling or some other cockamamie theory that would inhibit investment in connections or force carriers to treat business traffic on a par with P2P file sharing or YouTube is a serious deterrent here.
Questions aside, Mr. Carr makes an arresting case, and this is where my reservations about book come in. Instead of delving in detail into implications of utility computing for business and the tech world, the book takes a sharp angle off into a discussion of familiar themes of Web 2.0. The idea of cloud computing, in which websites draw on eclectic sources, the mash-up culture, user-generated content, and all the accompanying issues of privacy, data collection, the demise of classic business models and the problem of supporting for content creation.
It is not that Mr. Carr’s comments on these themes are uninteresting — Mr. Carr is always interesting. He analogizes much of Web 2.0 to a variety of “sharecropping” in which the company supplies the tools, the consumers supply the work, and the company collects the entire harvest. He punctures the assumptions of many Internet devotees that this is a new sphere immune to the claims of the physical world, noting the power of the Internet as a mechanism of control. “In reality [on the Internet], not only is it known that you’re a dog, but it’s probably known what breed you are, your age, where you live, and the kind of treat you prefer.”
But the Web 2.0 version of “the cloud” — services to consumers and user interactions delivered via data centers — is not the same phenomenon as Carr’s utility computing hypothesis, and I wish he had stuck more closely to the latter, exploring the objections and hybrid possibilities, and examining in more detail the implications for business. For example, he comments, but only briefly, on the fact that massive computing power can be placed at the beck of any small businessperson. This has considerable import for industrial organization generally, and is worth more exploration.
Oh well. It is one of Mr. Carr’s strengths as a writer that one finds oneself arguing with him, and one of the great things about the computer age is that a published book marks the beginning of a discussion, not its closure. So log on to Rough Type, where the debate will undoubtedly continue.
James V. DeLong is special counsel at Kamlet Shepherd Reichert, LLP and vice president and senior analyst at the Convergence Law Institute LLC in Washington, D.C.
Please read our comment policy before commenting.