Everyone is familiar with Amazon’s (NASDAQ:AMZN) Web-based retail business, the e-readers and tablets it sells and its online media offerings. What tends to get a little less publicity is AWS, or Amazon Web Services, the division that provides the hardware and infrastructure powering some of the Web’s most popular sites and applications.
It’s easy to visualize a company like Netflix — with 27 million customers in North America and accessible through all sorts of devices — as running a massive data center, with thousands of servers, backup generators, miles of fiber and hundreds of technicians dedicated to keeping things humming 24/7. Not so. Netflix instead turned to cloud computing, entrusting its video streaming business to Amazon.
Perhaps this is even more surprising: Even though Netflix reportedly shells out extra for maximum redundancy measures (so that one Amazon data center going down shouldn’t take Netflix down altogether) and for extra bandwidth (to prevent spikes in viewership from slowing the network), the company has been knocked offline by Amazon cloud outages multiple times. Besides the Christmas Eve incident, Netflix went dark in June when lightning hit an Amazon data center in Virginia, and many of its customers experienced service disruptions in August, thanks to another Amazon cloud outage.
The June outage exposed one of the dangers inherent in cloud computing: multiple services relying on a single data center. The pitch for AWS is that Amazon invests in the data center infrastructure, and a myriad of smaller companies that couldn’t normally afford such high-tech solutions pay a monthly fee for AWS to host their websites. They share elements of the infrastructure located at that data center.
However, when that AWS center goes down, it’s not just a headache for Amazon, it can take dozens or hundreds of websites with it, including big ones like Netflix. That June outage also took out Quora and Facebook’s (NASDAQ:FB) popular Instagram photo-sharing service, Pinterest. In April 2011, another major Amazon outage brought down FourSquare, Reddit and others.
Here’s another potential danger inherent in having multiple businesses hosted through a cloud computing provider: security. Between physical security measures, state-of-the-art network security and on-site experts, an AWS data center is undoubtedly safer from cyber-criminals than most corporate networks.
However, cloud computing means a lot of data (including sensitive personal and financial info) is housed in a single location. This offers what a Homeland Security official described in InformationAge as a “rich target for hackers.” While hacking AWS may be tough, the potential payoff is huge. And if it were to ever happen, the damage could be extremely widespread.
An AWS data center failure hits more than just websites and Web-based apps. Many companies are also storing their data in the cloud. An outage means they can’t access their data, leading to business interruptions.
Steve Wozniak (the other Apple Steve) was quoted in The Register, trash talking cloud services just as Apple (NASDAQ:AAPL) was ramping up its own iCloud offering. Wozniak is concerned about the potential for disaster when data storage and processing are cloud-based, as well as the implications for data ownership when information is not stored locally.
Despite the inherent risks, the many benefits of cloud computing continue to draw customers, especially among start-ups that couldn’t otherwise afford to build their own data centers. There’s still buzz around cloud computing as a cost-saving measure for cash-strapped IT departments, too, as companies struggle to cut operational costs while they recover from the recession. As The New York Times points out, some very big corporations use AWS, including Pfizer (NYSE:PFE) — hardly a typical dot-com start-up.
In its Q3 earnings press release, Amazon noted that its cloud computing services are now used by 1,500 educational institutions and over 300 government agencies.
Amazon doesn’t report AWS revenue separately, but GigaOM thinks it might be good for $1.5 billion in revenue in 2012. Of course, revenue doesn’t translate directly into profit. Data centers are expensive to build and operate, and competition keeps prices low. The same GigaOM piece notes that in Q1, Amazon’s CFO mentioned $800 million to $900 million in capital spending at least partially tied to expanding AWS infrastructure.
The competition that’s snapping at Amazon’s heels? Besides thousands of small ISPs that offer cloud computing packages, some very big players are also in the game. Google (NASDAQ:GOOG) is another company that has leveraged its own considerable experience and infrastructure to push into cloud services, and it’s going after Amazon in a big way with aggressive expansion and price cuts.
Rackspace (NYSE:RAX) claims its 80,000 servers are hosting data and services for 197,000 businesses. Microsoft’s (NASDAQ:MSFT) Azure is only a few years old, but it’s seeing rapid growth in capacity and customers, while IBM (NYSE:IBM) is pushing its SmartCloud service for enterprise customers.
So, the future looks good for cloud computing (just have a look at this Forbes article highlighting series of forecasts predicting huge growth rates in cloud server and application growth to 2016). Amazon is in the lead now and has the opportunity to capitalize on its position and make cloud computing an even bigger part of its revenue mix.
The trick will be overcoming competition, commoditization, service outages (especially among high-profile clients like Netflix), steep capital spending and the risk of a large-scale hack to turn those big revenues into big profits.
As of this writing, Brad Moon didn’t own any securities mentioned here.