Although still a relatively recent phenomenon, cloud computing’s impact across industries is already proving to be broad and significant.

Cloud computing is affecting the competitive dynamics of the hardware, software, and consulting industries. In the past, firms seeking to increase computing capacity invested heavily in expensive, high margin server hardware, creating a huge market for computer manufacturers. But now hardware firms find these markets may be threatened by the cloud. The trend shifting from hardware to services is evident in IBM’s quarterly numbers. The firm recently reported its overall earnings were up 12 percent, even though hardware sales were off by 20 percent.^[446] What made up the difference? The growth of Big Blue’s services business. IBM is particularly well positioned to take advantage of the shift to services because it employs more technology consultants than any other firm in the world, while most of its competitors are forced to partner to offer something comparable. Consulting firm Capgemini’s partnership to offer cloud services through Amazon is one such example.

The shift to cloud computing also alters the margin structure for many in the computing industry. While Moore’s Law has made servers cheap, deploying SaaS and operating a commercial cloud is still very expensive—much more so than simply making additional copies of conventional, packaged software. Microsoft surprised Wall Street when it announced it would need to pour at least two billion dollars more than analysts expected into the year’s server farmserver farmA massive network of computer servers running software to coordinate their collective use. Server farms provide the infrastructure backbone to SaaS and hardware cloud efforts, as well as many large-scale Internet services. capital spending. The firm’s stock—among the world’s most widely held—sank 11 percent in a day.^[447] As a result, many portfolio managers started paying closer attention to the business implications of the cloud.

Cloud computing can accelerate innovation and therefore changes the desired skills mix and job outlook for IS workers. If cloud computing customers spend less on expensive infrastructure investments, they potentially have more money to reinvest in strategic efforts and innovation. IT careers may change, too. Demand for nonstrategic skills like hardware operations and maintenance are likely to decrease. Organizations will need more business-focused technologists who intimately understand a firm’s competitive environment, and can create systems that add value and differentiate the firm from its competition.^[448] While these tech jobs require more business training, they’re also likely to be more durable and less likely to be outsourced to a third party with a limited understanding of the firm.

By lowering the cost to access powerful systems and software, barriers to entry also decrease. Firms need to think about the strategic advantages they can create, even as technology is easily duplicated. This trend means the potential for more new entrants across industries, and since startups can do more with less, it’s also influencing entrepreneurship and venture capital. The CTO of SlideShare, a startup that launched using Amazon’s S3 storage cloud, offers a presentation on his firm’s site labeled “Using S3 to Avoid VC.” Similarly, the CEO of online payments startup Zuora claims to have saved between half a million and one million dollars by using cloud computing: “We have no servers, we run the entire business in the cloud.”^[449] And the sophistication of these tools lowers development time. Enterprise firm Apttus claims it was able to perform the equivalent of six months of development in a couple of weekends by using cloud services. The firm scored its first million-dollar deal in three months, and was break-even in nine months, a ramp-up time that would have been unheard of, had they needed to plan, purchase, and deploy their own data center, and create from scratch the Web services that were provided by its cloud vendor.^[450]

So What’s It Take to Run This Thing?

In the countryside surrounding the Columbia River in the Pacific Northwest, potato farms are yielding to server farms. Turns out the area is tailor made for creating the kinds of massive data installations that form the building blocks of cloud computing. The land is cheap, the region’s hydroelectric power costs a fraction of Silicon Valley rates, and the area is served by ultrafast fiber-optic connections. Even the area’s mild temperatures cut cooling costs.

Most major players in cloud computing have server farms in the region, each with thousands of processors humming away simultaneously. Microsoft’s Quincy, Washington, facility is as big as ten American football fields and has nearly six hundred miles of wiring, 1.5 metric tons of battery backup, and three miles of chiller piping to keep things cool. Storage is big enough to store 6.75 trillion photos. Just a short drive away, Yahoo has two facilities on fifty acres, including one that runs at a zero carbon footprint. Google has a thirty-acre site sprawled across former farmland in The Dalles, Oregon. The Google site includes two massive buildings, with a third on the way. And in Boardman, Oregon, Amazon has a three building petabyte palace that sports its own ten-megawatt electrical substation.^[451]

While U.S. activity has been particularly intense in the Pacific Northwest, server farms that support cloud computing are popping up from Shanghai to São Paulo. Not only does a diverse infrastructure offer a degree of fault tolerance and disaster recovery (Oregon down? Shift to North Carolina), the myriad of national laws and industry-specific regulatory environments may require some firms to keep data within a specific country or region. To meet the challenge, cloud vendors are racing to deploy infrastructure worldwide and allowing customers to select regional availability zones for their cloud computing needs.

The build-out race has become so intense that many firms have developed rapid-deployment server farm modules that are preconfigured and packed inside shipping containers. Some of these units contain as many as three thousand servers each. Just drop the containers on site, link to power, water, and telecom, and presto—you’ve got yourself a data center. More than two hundred containers can be used on a single site. One Microsoft VP claimed the configuration has cut the time to open a data center to just a few days, claiming Microsoft’s San Antonio facility was operational in less time than it took a local western wear firm to deliver her custom-made cowboy boots!^[452] Microsoft’s Dublin-based fourth generation data center will be built entirely of containers—no walls or roof—using the outside air for much of the cooling.^[453]

Figure 10.5. 

This Sun server-packed container is designed for rapid data center deployment.

While firms are buying less hardware, cloud vendors have turned out to be the computing industry’s best customers. Amazon has spent well over two billion dollars on its cloud infrastructure. Google reportedly has 1.4 million servers operating across three dozen data centers.^[454] Demonstrating it won’t be outdone, Microsoft plans to build as many as twenty server farms, at costs of up to one billion dollars each.^[455] Look for the clouds to pop up in unexpected places. Microsoft has scouted locations in Siberia, while Google has applied to patent a method for floating data centers on an offshore platform powered by wave motions.^[456]