Artificial intelligence (AI) is here to stay. While it’s still far from the capabilities we’ve grown to expect from AI thanks to movies, it’s already quite impressive. And it’s making its way into more and more niches such as data center and edge computing management.
The jobs of IT and data center managers are getting tougher, writes Kim Povlsen, VP of Schneider Electric. Povlsen oversees the digital services and data center software group of the company. He writes on DataCenterDynamics that the IT environments are getting increasingly complex and the demands for them are rising. Thus, managers have to maximize system availability and uptime while maintaining the hybrid IT infrastructure.
Fight data with data
How to manage data centers? With data, of course. Data Center Infrastructure Management (DCIM) systems have evolved quite a bit over the years. Originally, they were very complex and only niche experts could use them properly. This means way too often these systems aren’t used to their full potential. Plus, data center usage is growing a lot not only in the sheer scale and amount of data, but also in use cases.
Staying on top of all of this isn’t easy. So, what’s the solution? Gather even more data. By gathering raw IoT data and centralizing it, managers can get a better understanding of what’s going on in the data center. But why not add another layer on top of all of this? That layer is AI which helps to identify actionable information and patterns, Povlsen says.
AI DCIM systems can also help managers get a better understanding of asset behavior and see what solutions the AI would recommend. Of course, the final decision is up to the humans, but they can now have a lot more actionable information possibly identifying issues that otherwise would have been missed.
Basically, this is a way to use AI to help itself by making sure the infrastructure it uses actually runs properly. It’s also a way for AI to “give back” to the data centers. AI could eventually take over data center management and make it self-driving.
For example, it can even tweak data center cooling to make it more efficient. Google’s DeepMind already did that as an experiment in 2016. It used AI to optimize cooling at Google’s Singapore data center. The results lowered the cooling bill by 40% and improved the Power Utilization Effectiveness by 15%. If AI is so effective in other areas of data center management, the benefits are going to be massive.
Should we be excited or worried?
Having AI take over yet one more thing may sound a bit worrying. For now it’s way too early for something like that, though. Pretty much all AI solutions for data centers require human supervision. It’s entirely possible that this will remain as such for a very long time. So, the jobs of data center managers aren’t threatened. Instead, they will get some much needed help.
Another benefit of these systems will be predictive maintenance. Cloud-based DCIM systems will be able to foresee possible issues and alert managers ahead of time.They could also even go as far as predicting maintenance requests from users and clients before they even send them in. This will lower downtimes even more and increase customer satisfaction as who won’t be happy to receive an email saying an issue was identified and taken care of even before it was noticeable.
So, yes, AI can definitely help with the development and management of data centers, edge computing and cloud services in general. It seems that the development of these features is going in the right direction. And we’re already seeing some initial benefits of that.