The Sun Shines on ‘The Cloud’

Grazed from The Wall Street Journal. Author: John Bussey.

What’s the real lesson to be learned from Amazon’s AMZN -1.40% cloud computing failure two weeks ago, the one that knocked Netflix NFLX +3.92% and other websites offline?

Far from being a dangerous sign that the cloud is unreliable, it was actually a blip—a painful one—on a trend line that is vectoring upward, much to the benefit of global business. There have been sporadic problems at a range of cloud providers, and each produced worrying headlines: "Can You Trust the Cloud?" "Are We Overdependent on Cloud Services?"

But what we’re seeing isn’t a breakdown. Instead, it’s the rapid expansion of a big new industry that is still in its shakedown phase—finding and fixing problems. This is Henry Ford getting the kinks out of his assembly line or cellphone companies trying to fix the dropped-call problem…

The cloud-computing industry is booming. IDC, a research firm, says public information-technology cloud services world-wide had $16 billion in revenue in 2009. That’s forecast to more than double this year and to hit $73 billion by 2015, adding substantial productivity to business along the way.

"While it’s easy and common to blame the cloud for outages because it’s outside of our control, we found that our overall availability over the past several years has steadily improved," Netflix, the online movie website, said in a statement Friday. This from a company badly walloped June 29 when an Amazon Web Services data center in Virginia lost power.

The cloud is actually getting more reliable, contends Netflix. Sticking with Amazon, says the company’s chief product officer, Neil Hunt, gives Netflix "much larger scale and technical expertise" than it would otherwise have.

EveryBlock, a local news website, was even more forgiving after AWS problems in April 2011 knocked the company offline. "Frankly, we screwed up," the company said in a blog post. "AWS explicitly advises that developers should design a site’s architecture so that it is resilient to occasional failures and outages such as what occurred yesterday, and we did not follow that advice."

"Cloud computing is still very young," says George Reese, chief technology officer of enStratus, which helps companies adapt to the cloud. "Everyone has warts. But if you’re smart about how you adopt cloud computing, then the advantages, the economies of scale, certainly outweigh the warts."

Some companies opt to use the cloud for back-office work, keeping mission-critical functions in house. Others, particularly start-ups, build their entire business on the service.

In either case, companies are aiming to reduce IT costs and add technical capability by outsourcing work to big data centers run by such companies as Amazon, Microsoft, Google and Rackspace. Customers access services over the Internet as if they were in a "cloud," hence the buzzword.

Amazon Web Services has hundreds of thousands of customers and data centers stretching from Japan to Ireland to Brazil and the U.S.

Amazon launched AWS in 2006. Now the largest cloud services provider, it has hundreds of thousands of customers and data centers stretching from Japan to Ireland to Brazil and the U.S.

Bob Igou of Gartner Inc., the research firm, says companies list cost savings as the chief reason for going to the cloud. But they also say maintenance costs and integration problems can affect those savings, and they list data security as the cloud’s chief "challenge."

For additional worries, consider what happened at that Amazon data center in Virginia two weeks ago.

An electrical storm jangled power to the center. Generators kicked in but failed to stabilize the load. Power went off to part of the data center. Then a software bug delayed recovery.

Brandon Wade, CEO of Infostream Group, which runs dating websites, says his company tried to call Amazon for help when service crashed.

His techs couldn’t get through. Amazon’s overall performance, says Mr. Wade, was "inexcusable."

The tech community is split over AWS’s troubles: Some commentators blame customers for failing to build more redundancy into their cloud products by spreading service over multiple data centers or regions. Others say Amazon is at fault for overpromising and underdelivering.

For its part, Amazon apologized for the power outage, posted an explanation of what went wrong, and said it would fix the problems. As for reliability, technology analysts say you’d be hard pressed to find an in-house corporate IT operation that tops AWS’s cloud service. "I’ve never had a client complain about Amazon outages," says Joe Pucciarelli of IDC. "Generally, they say they exceed what they’ve committed to."

"When you look at Amazon, Google, Microsoft, they are more stable and more robust than anything a corporation would be willing to invest in," says Mike Cooke of Booz & Co.

Kay Kinton, a spokeswoman for AWS, says the company commissioned independent research from IDC that examined a small sample of AWS customers and found that four-fifths said their unplanned downtime dropped after they moved operations to the cloud. On average, the downtime fell to 1.50 hours per user per year from 5.40 hours. The firms reported a 70% cost savings from shifting operations to cloud computing, Ms. Kinton says.

But it’s too late to convince Mr. Wade at Infostream. Hurt by a separate snafu at an Amazon data center on June 14, he barely had time to recover before the power problems hit June 29.

So he pulled the plug on AWS. He’s planning to move his dating services to Rackspace, Google or another provider.

Nothing like a little competition to speed an industry’s maturation.