Cloud Computing Is No Pipe Dream
Tech pundits and practitioners alike have spilled lots of ink to hype cloud computing. They'll encourage you to think of it as IT infrastructure on demand—like plugging into the power grid to get electricity, or turning on a faucet to get water, but getting raw computing power instead. To boot, you'll get access to storage capacity and software-based services—and it can scale infinitely, proponents point out.
If this all sounds too good to be true—like so much cold fusion, the now-debunked tabletop nuclear fusion reactor in a bottle—don't be fooled. This time, getting more for less is real. And with the economy of its currently parlous condition, businesses have never needed cloud computing more.
To understand why, consider the historical perspective:
It has been nearly 15 years since Google (GOOG) Chief Executive Eric Schmidt, then chief technology officer of Sun Microsystems (JAVA), articulated a simple proposition: "When the network becomes as fast as the processor, the computer hollows out and spreads across the network."
Fast, Cheap, and Robust
Well, guess what. Today's networks are able to move data as fast, or faster, than semiconductors can process it. All that late-1990s overinvestment in high-speed fiber-optic networks has given us communications networks that are lightning fast, relatively inexpensive, and remarkably robust—capable of transforming how we access, organize, and share computing resources.
Schmidt's vision is becoming reality. And consumers, for their part, embrace it with gusto. From Facebook apps to Apple's (AAPL) iTunes, from Yahoo! (YHOO) Mail to AOL Instant Messaging, from Flickr photo sharing to YouTube video viewing, cloud services have taken hold. Hundreds of millions of consumers are computing in the cloud, with more users and more offerings added every day. That's just on the PC.
There are nearly 4 billion mobile phones in use around the world, and each phone, especially if it's a smartphone, is a small computer. These handheld machines can't do much without a smart cloud to back them up. When users tap into that cloud, they can suddenly do amazing things. That's one way to explain why just two months after Apple's midsummer launch of its App Store, iPhone and iPod touch owners had downloaded 100 million applications. By Dec. 1 the number had surged to 300 million. Third-party developers have stocked up the store with more than 10,000 paid and free iPhone apps.
Try One On for Size
Now it's time for companies to take note. And the good news is that cloud-based opportunities come in all business sizes—large, midsize, and small.
Consider a big enterprise like The New York Times (NYT). Recently the Times was looking for a way to make its archives—150 years of daily newspaper editions—available online. Digitizing so much material was only half the challenge; the storage requirements were vast. The Times archives included 11 million articles; when digitized, they would consume 1.5 terabytes of memory. The paper tried to manage the process internally, but it proved unrealistic. So it turned to Amazon (AMZN) Web Services, the online retail giant's cloud offering.
Using Amazon's Simple Storage Service (S3), the Times launched a Web application called Times-Machine. The launch took days instead of months, and it cost hundreds of dollars instead of thousands. Users can read PDFs of newspapers from the mid-19th century, zooming in on articles, photos, and even advertising.
Now, consider midsize company Apttus. It's a fast-growing operation that provides software-based contract management systems to corporate clients. Apttus built its prototype using a development environment called Force.com, a cloud service from Safesforce.com (CRM). What would have taken six months of traditional development came to fruition in a "couple of weekends," according to Apttus' founders. Using a third-party platform, they were able to build an enterprise-class system with more functionality than competing offers, while enabling its instant global delivery using cloud servers. After three months of feedback from prospects, Apttus scored its first million-dollar deal. In nine months, the company was profitable. Normally, it would have taken $15 million to $20 million in revenue to reach break-even.
Finally, consider a small outfit like SlideShare. It's a Web startup built to enable users to share slide presentations online. The site quickly became a vibrant community among business professionals, with users uploading decks to share, search, and embed into other presentations. Success created a problem; the more users engaged, the more storage the site required. Like the Times, Slideshare turned to Amazon. Not needing to own and operate servers, firewalls, and security systems, it required little capital to get started. The site's chief technology officer shares his presentation; it's called "Using S3 to Avoid VC."
Millions of Servers in Data Centers
The building blocks of cloud services are data centers, run by tech giants such as Amazon, Google, Microsoft (MSFT), and IBM (IBM), housing hundreds of thousands of servers working together using proprietary operating systems. It's estimated that Google is buying half a million servers a year, while Microsoft is acquiring 200,000. Google alone reportedly has 1.4 million servers operating across its dozen or more data centers today. Moreover, this is not a game that only tech behemoths can play. According to IDC, there are 7,000 data centers in the U.S. alone. The number of servers they house is projected to rise to 15.8 million by 2010.
Gartner (IT) famously estimates 80% of corporate IT dollars go to data center maintenance. If that's true, then companies outsourcing their IT operations to the cloud can focus their technology talent on the challenges unique to their businesses—namely, what really matters to customers. Because cloud environments are shared, collaboration is easier when every application becomes "social" media. That means organizations of all kinds would have new platforms on which to foster cooperation and collaboration, minus the coordination costs typically associated with bringing work and people together. Plus, there's no need to overbuild IT infrastructure for spikes in usage (like seasonality in retail), when you can hand off excess demand to the cloud.
Yes, the vision is still a bit hazy. There's appropriately high anxiety these days about the cloud's data security, consumer privacy, and network downtime issues. There are also plenty of high-level features, such as database management, that cloud services don't yet offer. But these are issues cloud players are addressing. When they do, the cloud will be ready for prime time. In today's moribund economy, it could not come a moment too soon.