olly - Fotolia
IoT data center tech aims to improve operational efficiency
IoT tech can make data center operations data-driven and automated. For a successful implementation, IT managers should evaluate upgrade costs and infrastructure resiliency.
IT departments face a full gamut of challenges, so it seems reasonable to ask if they can benefit from do-more-with-less technology deployments. Internet of things devices and hardware can achieve end-to-end monitoring within a data center. There are many use cases for deploying IoT across infrastructure, but the business requirements and results can vary.
IoT data center monitoring can provide enhanced safety, reliability and cost management. The technology also enables data collection across a variety of locations and end points, said Jason Carolan, chief innovation officer at Flexential, a colocation provider.
"These devices can be deployed broadly as they generally are a lower cost than the wired, difficult-to-install platforms; many are wireless-enabled and can provide data over Wi-Fi or cellular networks, making them deployable anywhere," Carolan said.
With many data centers operating on reduced staff across the industry, even prior to COVID-19, there was a continuous push for greater automation and monitoring of data centers.
This included tools for remote access, management, predictive analytics for maintenance and setups for self-healing infrastructure. Critical to the success of these goals is the ability to monitor, collect and control as many aspects of the data center hardware, said Ed Featherston, distinguished technologist at Cloud Technology Partners Inc., a Hewlett Packard Enterprise company.
Running the data factory
The flow of information is reliant on the data center hardware moving, processing and modifying the raw materials (input data/transactions) into the final product. In manufacturing, IoT devices are at the core of that success.
These devices first provide data on the health of the production line (whether physical or digital) and can supply data to analytics tools to perform predictive analysis of potential problems before they become catastrophic. IoT data center devices can support repair or address problems with minimal human intervention, Featherston said.
Two key factors drive this approach, according to Featherston. The first is cost; the more automation an IT team already uses, the lower the overhead is to upgrade infrastructure. Second is the ability to react and quickly correct issues, which provides high resiliency to the infrastructure and the information/systems.
"IoT is the next generation of what has generally been referred to across industries as operational tech," said Daniel Bizo, principal analyst at 451 Research, part of S&P Global Market Intelligence.
The IoT data center's foundation in DCIM
Much of the discussion around IoT technology flows from traditional data center infrastructure management (DCIM). This is where DCIM maturity is a factor.
"Not all DCIM is created equal and it is not a uniform thing; it is like many other systems that vary in how sophisticated they can be and how sophisticated the customer can be," Bizo said.
Big enterprises typically buy DCIM software or have versions created in-house. In either case, there is usually a need for heavy customization for integration and adaptation to specific infrastructure.
With system management software, IT departments have a foundation for IoT-like capabilities, according to Nitesh Bansal, senior vice president and global head of engineering services at Infosys.
System management software isn't new, but IoT technology provides a new way to manage passive assets such as HVAC systems that help IT teams be more productive and make a data center profitable. Beyond the HVAC systems, possible assets include just about everything else: The facility itself, solar power, cooling and heating, wiring and uninterruptible power supplies (UPSes).
"In the past, many or most of those things were installed based on standard capacity planning, which in turn was based on estimates of peak loads," Bansal said.
Sensor technology opens the door to a more fine-grained understanding and potentially more-efficient operations. Bansal noted that industry studies suggest that 15% of data center costs are energy.
"If you could reduce [energy costs] by a third, that would be more than 4% savings to overall costs," he said.
The enhanced IoT sensor technology can help control costs in three areas of the data center. First, he said, is the of monitoring air flow, heating and cooling, and utilization of space. The same techniques can look at individual aisles, racks, or even have devices to spot potential problems and begin to eliminate wasteful energy use and identify areas of inadequate cooling.
A second area is to look at resiliency systems, especially UPS systems. Those often have multiple large batteries being charged continuously. However, Bansal said, it is not uncommon for some individual batteries to degrade, absorb disproportionate amount of electricity to stay in a charged state and then, perhaps, reduce the performance of the whole system when it must deliver emergency power.
"You might not recognize that problem without IoT," he said.
The third area to look at is the human element. While much of data center monitoring is done remotely, many operations still have people assigned to check LED display lights periodically or make minor changes in settings.
In addition to direct costs in terms of salaries and space needs, this activity affects the equilibrium of HVAC systems as doors are opened and closed and can often be eliminated through the simple addition of display panel video monitoring.
IoT as a business driver
Over the past decade, business models have changed to a more data-driven approach; this is a contrast from infrastructure management workflows that just focused on keeping everything running.
"Now, as a business manager, I want to know exactly what I am paying for, and my customers are asking how energy-efficient I am and whether I can quantify the risk of a data center outage," Bizo said.
Culture is also a factor and depends on how a given organization divides responsibilities. In some organizations, according to Bizo, the data center is considered to be part of facilities, which remains concerned with physical security, fire suppression and power management, while IT itself focuses more on the data center delivering information and services.
The upshot is that IT management may not be tracking power use and facilities people may not have any insights into the inner workings of IT.
"One of my favorite phrases is 'everything is a tradeoff.' Implementing this kind of IoT environment is a complex, nontrivial task," Featherston said.
There are upfront costs from a design, planning, acquisition and implementation perspective. The payback is entirely dependent on the end goal and how far along an organization is on its IoT data center journey. Featherston said that even improved IoT or DCIM has personnel costs.
"You still need skilled people in place that understand it," because nothing is perfect; even automation can break and fail, he said.
