Outage calls attention to Amazon’s web dominance
Company’s Internet services unit has become essential to smooth operation of popular websites
SEATTLE— There have been many opportunities to consider how crucial Amazon has become to the smooth operation of the Internet over the past few years.
The most recent involved a case of fat fingers. That event occurred late last month when an Amazon employee entered an incorrect set of commands on a computer, unintentionally knocking out a large set of servers in an Amazon data centre in Northern Virginia and, with it, an array of online services from other companies.
Among the many consequences of the shutdown: Users of the business messaging service Slack couldn’t upload files. Photos on the technology news site the Verge didn’t display. Quora, a popular question-and-answer site, couldn’t be reached.
It was a rare fumble for Amazon in cloud computing, in which companies pay to run their online applications in data centres operated by big providers.
Its computing business, Amazon Web Services, hauled in $12.2 billion (U.S.) in revenue last year from customers ranging from Netflix to the CIA.
Amazon’s service interruption, though not nearly bad enough to spark a panic among customers, was viewed by some as a moment for reflection.
“It really is a wake-up call to enterprises,” said Craig McLuckie, a former Google technologist who is now chief executive of Heptio, a startup that makes software tools for more easily moving applications between clouds. “They certainly need to understand to what level of dependency they have on a single provider.”
Lydia Leong, an analyst at Gartner, did not expect what transpired to cause a rush of companies to shift their cloud computing away from Amazon. She said it was highly unusual for companies to run the same application in more than one cloud because it introduced new complexities and costs.
To avoid being knocked offline, companies can set up their applications to run in multiple regions, across the company’s network of data centres. A few years ago, Hearst, the media company, was exclusively using Amazon for cloud computing, but later added services from Microsoft and Google, said Philip R. Wiser, Hearst’s chief technology officer. Because it’s relatively early in the emergence of cloud computing, Wiser said he believed it was important to spread one’s bets around.
“Being multi-cloud as a company and having the skill and understanding of how to move between investments — that is an asset,” he said.