Can Machine Learning Prevent Downtime?

Posted by James Ball on Oct 26, 2018 9:00:00 AM
It's Friday again, which means it's time for another installment of the Data Center News Roundup. This week, we look at the promising news that machine learning might help data center operators prevent downtime; the rise of hyperconvergence; how sea water can generate electricity and other futuristic tidbits from the news.  

AdobeStock_170135489

Adobe

Data Center News Roundup for Friday, October 26, 2018

Hyperconvergence to Replace Traditional Hardware?

BizTech Magazine makes the bold assertion this week that “hyperconvergence, software-defined data centers and other low-touch approaches will likely replace traditional hardware.” While not going as far as saying “the data center is dead,” the piece cites the rising trends of Internet of Things, edge computing and SaaS to point to an evolution in the way data centers operate. The story focuses its attention on private, rather than traditional data centers, saying the former are already using “cloudlike management and methodologies” to make this prediction about hyperconvergence. Check out the full piece here.

 

It’s the Motion in the Ocean

We’ve previously discussed Microsoft’s Project Natic, an underwater data center situated off the coast of Scotland. Now there’s word of a 1.25 megawatt data center from Ocean Energy that will come online in 2019, powered by wave motion. At 125 feet long, the hydrokinetic data center vessel is named the OE Bouy and it will weigh 826 tons, Network World reports. Plans are to deploy the vessel off Hawaii, bringing the twofold benefits of ocean energy and ocean cooling all in one. Word is, rent’s pretty cheap and the neighbors are quiet, if not a little curious! Read the story here.  

 

Toward a More Sustainable Future

Given the fact that data centers currently use as much electricity as 6 million houses globally, the University of Arizona and Microsoft have teamed up to make data centers “better neighbors” to their communities, energy-wise, AZPM reports. The primary questions this joint venture will aim to answer revolve around sustainability and being environmentally-friendly. Dubbed the “Cloud Infrastructure Renewal Center,” the project brings together architects, optical scientists, engineers and others to help reduce waste heat. Learn more here.

 

Machine Learning May Prevent Downtime

Finally, this week, software startup AdeptDC also has its eyes on the future. The future it envisions is one where software will go beyond managing cooling and power by tapping into machine learning to “holistically optimize for efficiency, troubleshoot, issue incident alerts, and prevent equipment failures by identifying anomalies,” according to Data Center Knowledge. Pretty heady, futuristic stuff, but if it the pilot projects are any indication, the software could one day predict – and prevent – massive outages. Read it here.

 

 Data Center Fit Up

Topics: News