IT companies for years have talked – or even bragged – about how they “eat their own dog food,” meaning they use the products they make. Personally I never liked that analogy, as it doesn’t sound too appetizing. But the idea of incorporating Internet of Things (IoT) and artificial intelligence (AI) technologies into data center architectures is one I suspect colocation providers will find appetizing indeed. Maybe it’s the influence of working for a French company for many years, but I prefer to think of it as “drinking their own champagne.”
Employing IoT and AI technologies enable colocation providers to dramatically increase their data center efficiency. But it requires more than simply putting sensors on all sorts of devices and components. It means collecting the right data and analyzing it to provide in-depth, actionable information that can improve data center operations and efficiency.
That was the gist of a keynote presentation delivered by my colleague Thierry Chamayou, Schneider Electric’s Regional Vice President EMEA South, at Datacloud Europe, an event in early June in Monaco for data center and cloud leaders. He laid out some of the drivers for IoT trends that are creating data center demand. They range from the amusing (robot cops in Dubai) to the terribly serious, including security and how Gartner predicts that 20% of annual security budgets will be dedicated to IoT solutions by 2020. The proliferation of mobile devices and sensors, all generating data, are likewise significant IoT drivers, he noted.
All this data is also driving the need for edge computing, as is the fact that many IoT applications require low latency, which translates to localized data centers. Regulatory issues and bandwidth requirements likewise drive up demand for localized data centers. Chamayou noted that IDC predicts 40% of IoT data will be stored, processed, analyzed and acted upon at the edge, a trend we’ve covered in previous posts (such as this one from our own colocation provider event).
Dealing with these varied demands presents challenges for colocation and hosting providers, including the need to accelerate their time to market in data center design and build phases, while limiting capital expenses. They also need to lower operating expenses, by both improving operational excellence and reducing energy costs.
This is where drinking your own champagne comes in. Even while IoT data is driving demand for colo providers, it is also the key to improving operational efficiency and lowering energy costs.
Many data center components today can produce data about their health and status that can be used to drive actionable intelligence. The key is to connect all data center components to a network that collects critical data at every level, from sensors to the cloud, and then apply analytics to turn it into meaningful information that can drive actions and decision-making.
IoT technology applied to data centers bridges the IT, mechanical and power room domains. It enables thousands of connections in centralized data centers, but also allows connections with remote edge data centers outfitted with appropriate software. It gives providers real-time monitoring, incident management, analysis and asset utilization data that helps them increase data center resiliency while increasing efficiency.
This sort of data – and, more importantly, knowledge of how your data center is operating – will only become more valuable going forward. The industry will continue to use the data to make incremental improvements in efficiency on the facilities side. But when merged with the same sort of data from the IT stack, the opportunities for step function improvement become possible.
In his keynote, Chamayou provided some highlights of the kind of results providers are getting with this IoT-based approach to data center optimization. Genpact, an Indian company that is itself focused on business process optimization, now spends 75% less time on data center maintenance. St. Luke’s Health System is achieving 99.97% network availability across its distributed data centers. It employs Schneider Electric cooling, UPS, PDUs and racks along with the StruxureWare Data Center Operation and Data Center Expert DCIM tools to effectively manage and control it all.
Green Mountain Data Center in Norway uses Schneider Electric EcoStruxure components to reduce its cooling costs by 30%. It also uses 100% renewable energy and produces zero CO2 emissions. (Check out this video that Chamayou showed during his talk – it’s an amazing story.)
These are just some of the stories we heard at Datacloud Europe. Keep an eye out for more posts from the show: #DatacloudEurope.