A recent 451 Research quarterly Datacenter Knowledgebase (DCKB) report noted that there are now over 4,700 collocation and wholesale data center facilities in the world. With cloud computing pushing the need for data center space and every organization, no matter the size, depending more and more on data, the data center market is set to continue to grow. Whether it is on-premises, managed by a service provider or supporting a public cloud infrastructure, the data center market is here to stay.
With greater processing power and the demands of instant access to applications like customer interfaces, sales platforms and stockholding updates, comes greater power consumption and uptime demand. In short, data centers are becoming increasingly expensive to run.
Enter artificial intelligence (AI) – technology that enable machines to process data that would otherwise require human intelligence, and again, at a cost.
AI could be a means to curb future costs by radically reducing data center energy consumption, improving uptime and minimizing the requirement for human intervention. It offers this potential without compromizing performance, as the technology will enable data centers to run more efficiently.
Today’s technology is characterized by constant change, so it’s inevitable that data centers need to keep pace with increasingly complex IT infrastructures and the functional scope needed for smooth operations. Rather than be seen as a challenge, change should be seen as creating myriad exciting possibilities to leverage new technologies such as AI, for enhanced ‘brain’ functionality and efficiency.
In short, AI allows machines to learn by themselves, drawing conclusions from data interpretations, and implementing such conclusions automatically, without human intervention. This means that businesses may start optimizing resource management without necessarily throwing more manpower at it, with the concomitant increase in the required skill set and the high price tag with which it usually comes.
AI can be used in a multitude of roles within the enterprise. Its core role is still automation – that of systems, decision-making and execution. Over and above what we can call ‘collective robot behaviour’, that is connecting various systems within the enterprise and using these to perform functions, in real time, it can also be used in a predictive role.
This could involve companies using it for long-term planning. Many research groups within enterprises use it to predict the implications of business decisions before they are implemented, and the data is then used to forecast what environmental factors may present themselves during manufacturing processes.
In the same vein, the technology could be used to predict how markets may react to business decisions by collating information for a multitude of sources and data feeds, analyse these on the fly and forecast a likely market reaction. It’s similar to what the industry terms ‘game theory’, and is a remarkably accurate and indispensible tool with a wide range of applications, from political campaigns to the roll-out of new products.
AI and the data center
Not only are data centers a crucial part of housing and processing the data generated by AI, they can also directly benefit. AI can now successfully be deployed, using the same collective-behaviour capabilities, to work in conjunction with data center infrastructure management technologies to check power consumption and cooling, system status and capacity.
AI can check power consumption and cooling, system status and capacity.
One of the earliest examples of this is Google’s adoption of AI technology to enhance the productivity of its data centers. Implementing this technology in 2014, Google was able to use machine learning to reduce the amount of energy used for cooling by up to 40 percent
The reduction of power consumption experienced by Google will save the company hundreds of millions of pounds based on the large number of data centers it runs, but this cost reduction would make an impact on a business of any size. Through the critical analysis and automated decision-making it is feasible to significantly improve the efficiency of any data center and the reduction in cost could then be passed on to customer if running a data center that hosts many organizations.
With margins cut to the bone, AI might well be what businesses need to gain a critical advantage over the competition through their optimized resource management.