The downtime experienced by Google on May 14, the latest in a series of temporary shutdowns experienced by cloud-based service companies throughout 2009, evokes one of those questions of burning importance to the enterprise: Are cloud-based services truly ready to meet business needs that require virtually continuous uptime?
Google's occasional shutdowns reliably bring a great deal of media attention, as with the February incident that took down Gmail in the United States and the United Kingdom for about 2.5 hours. A few months before that, in August 2008, Google Gmail and Google Apps underwent 15 hours of downtime.
In addition to the public cloud sphere, private cloud computing has found itself confronted with similar breakdowns. In March, the early test release of Microsoft Azure, an enterprise-capable cloud platform designed to eventually compete against Google Apps, underwent a 22-hour outage that left users unable to access its applications.
To the casual observer, such outages might suggest that cloud computing, despite the hype, is not quite ready, so to speak, for business prime time.
"We're still in the early days of exactly how cloud computing is being defined and delivered, and also what companies are expecting out of these services," Charles King, an analyst with Pund-IT Research, told eWEEK in an interview after Microsoft Azure experienced its temporary shutdown.
If a company wanted to increase the reliability of its private cloud-based services to the much-vaunted "five nines" (99.999) level, and thus lower its potential downtime to near zero, then King suggested the company would have to pay for it-and potentially pass those costs along to its customers. Within that context, particularly in the midst of a global recession, most companies may opt to extend 99.99 percent availability, and trust that their customers will live with the slightly increased risk of downtime.
Whether small and midsize businesses utilize private clouds or build their businesses on public cloud-based services such as Google Apps, their tolerance for downtime is almost certainly higher than that of the enterprise, which generally demands virtually zero unexpected shutdowns.
"If your revenue is based on being able to stay in contact with people, and you have an outage, it can build to significant levels of damage quickly, so your tolerances are tight," Rob Enderle, an analyst with the Enderle Group, said in an interview.
"The outages that Google experiences [don't] happen in a well-run enterprise," Enderle added. "I'm not sure Google gets the enterprise; even with Microsoft, it took them bringing in employees from places like IBM before they understood it. Google has not yet gone through that process, even with a CEO coming out of Sun-it requires a fairly large infusion of people who get the enterprise."