INDUSTRY INTELLIGENCE POWERED BY THE DCA INDUSTRY INTELLIGENCE POWERED BY THE DCA
Embracing AI to build the next generation of DCIM
As the demand for data and digital services is set to drive exponential growth in the need for data centres , it ’ s important to embrace technologies that are on hand to deal with such rising demand . Parvez Alam Kazi is the Head of Product at Smartia , and suggests openly embracing Artificial Intelligence in order to build the next generation of DCIM and support the industry .
The proliferation of hyperscale
data centres over the past decade and their impact of global energy consumption have been topics of constant debate . Contrary to popular belief , as per an article recently published in Science magazine , while the computing power at data centres increased 550 % from 2010 to 2018 , the energy consumption has only increased by 6 % during the same period . Nevertheless , as per another article in Nature magazine , data centres currently are estimated to be using almost 200 terawatt hours ( TWh ) of electricity per year ; that ’ s a staggering 1 % of all global energy consumption , potentially more than the energy consumption of some countries . Interestingly , that proportion back in 2010 was still 1 %.
That demonstrates a conscious effort by data centre organisations to build more energy-efficient data centres . Hyperscale data centres , by design , are highly efficient by virtue of their sheer scale and also in part due to use of virtualisation technology that enables higher computing power despite using fewer servers . Additionally , the big players like Google , Amazon and Microsoft have resorted to the use of Artificial Intelligence ( AI ) to improve energy efficiency and data centre operations . For example , starting in 2014 , Google pioneered the use of AI to develop cooling control systems that are now delivering 30 % in annual energy savings consistently . In 2019 , Google ’ s data centres were able to achieve an annual average power usage effectiveness ( PUE ) value of 1.10 , that being a record low . However , the industry average stood considerably higher at 1.67 . In principle , using AI to reduce energy consumption and optimise operations shouldn ’ t be a privilege limited to the larger players . The benefits of AI should be democratised .
Traditional data centre infrastructure management ( DCIM ) tools , which enable data centre operators to optimise systems and manage resources , have stuck to the standard ‘ measure , monitor and manage ’ motto for far too long . For example , DCIM tools generally provide the ability to measure and monitor real-time PUE as a flagship feature . However , as discussed later , using AI to model and contextualise PUE brings with it a whole new set of possibilities . Next-generation DCIM tools ( DCIM 2.0 ), which not only build on top of years of domain knowledge , but also openly embrace AI-powered features , have the potential to revolutionise how data centres will operate in the future .
AI-enhanced DCIM 2.0 can augment solutions to a lot of existing use cases . Taking Google ’ s example again , Machine www . intelligentdatacentres . com Issue 20
21