It took a bunch of clouds to disrupt the cloud.
In addition to knocking out power to millions of customers over the weekend, massive storms along the eastern seaboard also managed to disrupt power to Amazon’s facilities in northern Virginia, affecting its cloud services. Netflix, Instagram and Pinterest all experienced outages, along with smaller companies that also rely on Amazon’s cloud infrastructure to provide services.
By the morning of July 2, the AWS Service Health Dashboard indicated that all aspects of the service were operating normally, including the facilities in northern Virginia. The other companies managed to restore their respective services over the weekend.
The disruption illustrates two concepts related to cloud computing: first, even the hardiest clouds are vulnerable to extraordinary “real world” events, and second, any company relying on the cloud needs to be aware of that fact.
There are ways, of course, to install more redundancy in cloud offerings, such as supporting multiple data centers in geographically diverse locations. However, this can prove an expensive step, whether you’re an enterprise building your own infrastructure from the ground-up or a startup paying another company to host your data in its cloud.
Indeed, a March survey by cloud-services firms StorageCraft and Symform found that, out of 600 companies queried about their storage habits, two thirds felt that the costs associated with cloud or backup systems as a “problem.” Only a third of respondents relied on the cloud as a primary or secondary data backup.
The question about cloud redundancy is ever more important, considering the number of companies embracing cloud services as a way to save costs and make their core products more flexible. McKinsey & Company recently estimated that the combined global public and private cloud services market would grow from $11 billion in 2010 to somewhere between $65 billion and $85 billion by 2015.
Other studies have reinforced the notion of a rapidly growing cloud; a recent Parallels study put the current size of the small- and medium-size business market for cloud services in the United States at $15.1 billion (with an additional $1.6 billion for non-employer businesses), predicting that number would grow to $24.3 billion by 2014.
Along with the rise of the cloud, though, has come the perception that downtime is simply a way of life. Indeed, outages among higher-profile cloud providers such as Amazon and Google have raised short-term ire among clients, but that’s translated into seemingly little impact on overall cloud adoption.
“Firms that have shifted their infrastructure over to cloud services suddenly realize that maybe they haven’t diversified their risk enough,” analyst Roger Kay wrote in a July 2 blog posting. “Amazon has a nice front-end program that allows a customer to choose where its servers are hosted. In theory, a firm would pick geographic diversification that balanced risk. But a large user of even well distributed services like Netflix is likely to be affected anyway, as it cannot diversify away risk entirely.”
Image: Steshkin Yevgeniy/Shutterstock.com