Preserving Power in the Cloud

Arthur Cole

In the old days, data infrastructure would go down from time to time and the affected business would suffer lost productivity and lost revenue until service was restored. Now, thanks to the cloud, data infrastructure goes down and multiple businesses, even entire sectors of the economy, shut down.

Nothing takes out a data center like a loss of power, however, which is why many cloud providers are paying critical attention to both primary and backup energy supplies.

This point was made in dramatic fashion earlier this month when Amazon lost power at a key Virginia data center that supports a large portion of its Elastic Compute Cloud (EC2) service. The crash knocked out a number of top clients, including Quora, Hipchat and Salesforce's Heroku service. According to the company, the problem originated at a pair of utility substations, which shut down due to a cable fault. Then two backup systems failed, one due to a defective cooling fan and another from an incorrectly configured distribution circuit. While many observers were quick to point out that the mishaps in the backup infrastructure could have been spotted and corrected under regular systems checks, the fact is that no amount of safeguarding is foolproof, meaning the cloud is no more or less immune to disruption than any other data environment.

Still, the growing reliance on cloud computing is seen as a much-needed boon to UPS providers, particularly in areas of the world where data infrastructure is being built from scratch. Frost & Sullivan, for example, says the Asia-Pacific region could see revenues jump from $1.46 billion last year to more than $2 billion by 2017 as both private industry and governments turn to the cloud to keep data costs under control. It also doesn't hurt that much of the region is geographically unstable, as last year's Fukushima earthquake in Japan demonstrated, making utility-based energy supplies inherently unstable.

Firms like Active Power are already poised to take advantage of this trend. The company is slated to deploy a 1000 kVA version of its CleanSource UPS system at a massive co-location facility in China later this year. The project is part of a $300 billion program by the Chinese government to boost IT capabilities over the next five years. This year alone, the country is looking to expand its cloud infrastructure some 30 percent.

At the same time, enterprises are taking a closer look at energy management through tools like Eaton Corp.'s Intelligent Power Manager. The system works with VMware's vSphere High Availability (HA) mode to provide the ability to remotely shut down host servers that are threatened by power loss. The intent is to protect unsaved data and preserve overall data integrity by offering a controlled power down rather than a sudden interruption of service. Through integration with VMware environments, Power Manager enables visibility and monitoring of physical and virtual servers, UPSes, PDUs and other power systems through the vCenter Server platform.

As I mentioned, no system is foolproof. Fools are just too smart. Ultimately, data reliability will come down to careful monitoring of primary and backup power systems, but it will also call for highly advanced, dynamic data networks that enable load balancing on a massive scale should significant portions of the data environment suddenly vanish.

Add Comment      Leave a comment on this blog post

Post a comment





(Maximum characters: 1200). You have 1200 characters left.




Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.