Eliminating downtime and preventing disasters from stopping business operations in their tracks is one of the calling cards of quality data center and colocation facilities. Data center design has come a long way over the past several years, with extremely low downtime rates like the mythical five nines reliability well within reach. However, downtime still can and does happen. When it does, it’s likely to come with high cost consequences.
Businesses today use their data center and colocation facilities under the assumption that availability will be constant and consistent. Since this is true such a high percentage of the time, it’s easy to build upon this supposition and create layers of operative capacities that wholly rely on 100 percent availability and could prove disastrous if availability doesn’t meet this mark in practice. This is particularly true for enterprises that choose to keep their data center needs on premises or in house, as they’re likely to be beholden to other business issues, such as budgetary, personnel or time management concerns, that can supersede reliability management efforts. Divided attention is less likely in a third party colocation provider, as their scrutiny and resources are wholly focused on maintaining or resuming operation of their tenants’ core operations needs and avoiding downtime altogether.
Data Center Finance: Downtime Losses
Data centers do go dark, and the causes range from the obvious to the outrageous. A recent report by VirtualHosting examined some of the more pernicious – and downright surprising – cases of data center downtime. In times of extended downtime, the financial losses are likely to be substantial – on average, organizations lose $138,000 per hour of data center downtime – a rate 38 percent costlier in 2012 than it was in 2010. Average hourly loss further breaks down by data center performance tiers – the average best-in-class company loses about $101,600 per hour of downtime, while industry average enterprises lose $161,700 per hour. The typical companies lose more per hour due to a greater overall length of disruption time, while most best-in-class outages are quickly amended.
Overall, companies lose $26.5 billion in revenue each year due to downtime. It’s an issue that affects top companies as well as lesser-known ones – 59 percent of Fortune 500 companies experience at least 1.6 hours of downtime every week. The company also engaged in some admittedly unlikely speculation to show just how directly the world’s economic fortunes are tied to data center usage. For instance, if every data center in the world went dark simultaneously, the world would lose $69 trillion every hour. Yes, that’s “trillion,” with a “t”.
However, data center usage and investment continues to rise at an exponential rate. Global IP traffic is expected to hit 1.0 zettabytes annually by 2015 and 1.4 ZB per year by 2017. Cloud colocation demand rose 63 percent in 2012, while traffic that flows between data centers and users or between two separate colocation facilities continues to increase.
The Culprits of Data Center Downtime
The most common cause of all data center outages and stoppages is an unsurprising one – human error. Accounting for 73 percent of all downtime, human error can be attributed to several root causes, including poor training, a lack of operational governance and insufficient maintenance practices. However, the data center’s No. 1 enemy could be a seemingly innocuous one – squirrels. These wily creatures account for a high percentage of cable damage and can easily short-circuit significant amounts of data center operating capacity.
Other common data center downtime contributors include natural disasters (including hurricanes, floods, lightning and security exposures), cyber terrorism and hacking, and poor data center design.
Are Businesses Prepared?
Although many enterprises admit that significant suspension of operations would be devastating to their productivity and profitability, many aren’t doing everything they can to prevent a catastrophe. While the 99.999 percent availability standard seems to offer the most reliability possible in practice, one in 10 companies stated that they need even greater reliability. Eighty-seven percent of businesses admitted that data loss resulting from a data center crash would be damaging to their business, but 56 percent of enterprises in North America lack a sufficient disaster recovery policy. Overall, these statistics indicate that high data center availability and capacity to prevent any downtime are chief considerations for organizations to consider in their data center research.