Google believes machine learning and automation is the key to improving data centres' power useage effectiveness (PUE).
Speaking at the Datacentres Europe 2014 conference in Monaco today, Googe's vice president of data centres, Joe Kava, said machine learning can help organisations break out of the current a "plateau" in PUE.
He said Google had been stuck at a PUE score of around 1.12 before it started experimenting with machine learning, which has resulted in reductions in PUE of about 0.2.
"Anyone with a level of interest and dedication can apply this same learning to the knowledge you already have in your data centre," said Kava, explaining how the machine calculates interactions, and tries to learn variables across 19 different points, in a similar way to speech recognition software.
Variables include server loads, numbers of condenser pumps, air speed and inside air humidity.
But the real challenge now, said Kava, is learning to put the technique to good use across the industry.
"[We need to] ask how we're going to benefit on either efficiency or cost-effectiveness. Because I think you have to be as cost effective as you are efficient," said Kava, adding that the data centre industry is reaching a "critical inflection point" due to the "explosive growth and mass adoption of cloud" making PUE "more important now than ever".
For those interested, Google has published a blog accompanying Kava's keynote today.
Sometimes, the power of the mainframe is the most cost effective answer. Computing's Peter Gothard puts Computing's readers' questions on the future of the mainframe to IBM's Z13 expert Steven Dickens.
This Dummies white paper will help you better understand business process management (BPM)