To paraphrase an old joke, in the future, data centers will be run by one human and one dog. The human’s role will be to feed the dog; the dog’s role will be to make sure the human does not touch any of the equipment.
The demands from artificial intelligence (AI), machine learning (ML), the Internet of Things, 5G technology and for more computing power continues to exponentially increase the number of servers and other devices in the average data center. It also impacts the complexity and time necessary to plan and manage these systems. There is also another driver and that is edge computing. As companies lay out their plans to deploy many small computer nodes close to where data is generated, they quickly run up against the problem of how to operate this disparate infrastructure in an economical way.
While technology has advanced to allow an individual system administrator to manage thousands of devices, some believe we are headed for a wall. The situation is made even worse by what some call the “grey tsunami” of expected staff retirements. Is it time to fully embrace remote managed and unstaffed data center environments, optimized for machines, not humans?
The Evolution of Unstaffed Data Centers
This idea of unstaffed (also called lights-out) data centers is not new, but it is certainly evolving. The COVID-19 crisis will undoubtedly lead to greater considerations given to unstaffed solutions as data centers will themselves be led by software and other innovations. Data Center Infrastructure Management (DCIM) specialists such as Schneider Electric and Vertiv, and edge specialists like EdgeConnex and Vapor IO, are all making significant contributions in this innovative trend.
However, according to a report completed by 451 Research, capacity planning and power monitoring are the top two benefits data center managers want out of their DCIM solution. Unfortunately, 25 per cent of managers surveyed, reported that their DCIM tool failed and 10 per cent completely abandoned their DCIM project altogether.
Also, there is concern that not having staff on-hand is risky, especially in emergency situations. According to the Uptime Institute, one to two qualified staff are needed on-site at all times to support the safe operation of Tier III or Tier IV facilities.
The technology creating the need for more remotely managed data centers is also a part of the solution. Adding AI and ML to DCIM solutions has been met with great success. Schneider Electric reported late last year that they were collecting enough operational data from customer devices to start rolling out viable predictive features. When data centers include 200,000 or more devices, the devices generate billions of rows of data. This data is organized into data lakes allowing data scientists to build and test complex algorithms. By deploying and testing the models at scale, their accuracy can be enhanced which in turn, changes the way services are predicted and managed.
How TMGcore’s Technology is Innovative
There is little doubt that autonomous data centers of the future will be managed by software with robotics playing a greater role. One novel innovation comes from TMGcore whose technology uses a robotics system to remove failed servers and replace them with new ones. TMGcore’s technology also includes a two-phase cooling system in which servers are immersed in coolant fluid that boils off as the chips generate heat, removing the heat from the rack as it changes from liquid to vapor. A closed-loop water system is used to condense the fluid and return it to the tank. The robotic arm can latch onto a server, lift it out of the immersion tank, grab a replacement server and lift it back into the correct place. While the direct liquid immersion captures more than 90 per cent of the IT heat load, it makes no contribution to cooling for human operators.
The Benefits of an Unstaffed Data Center
If, as some people speculate, we view data centers as complex and integrated machines rather that real estate facilities with people, then they can be designed and optimized for IT, rather than human operators. For example, IT equipment can generally operate at temperature higher than would be comfortable for humans to work in. Rack heights and aisle width typically designed to the limits of human interaction could be designed specifically for robots. This in turn can lower data center construction costs and the cost of operations. Arguably, the facilities would be safer as the risk of human electrocution or other health and safety risks would be eliminated. Finally, since human error is one of the top causes of data center downtime, autonomous data centers could improve this important performance metric.
The Bottom Line
There are a number of impediments that need to be addressed before unstaffed data centers become more widely adopted. The technology needs to be more broadly proven and the perception of risks by data center operators and accreditation authorities needs to change. No question though, these advances are coming.
For more information on Cushman & Wakefield’s Global Data Center Advisory Group, contact us.