As companies continue to build edge computing data centers at a furious pace to support Internet of Things (IoT) applications and other distributed computing efforts, they often run into challenges with traditional room-based cooling systems. For high-density IT requirements (including micro data centers), it’s time to take a good look at in-row cooling as an alternative.
The traditional room method for data center cooling is to use down-flow perimeter cooling units that distribute cold air under a raised floor while consuming return air (warm air) at the top of each AC unit. As the hot air moves from the rear of the IT racks to the tops of the AC units, it mixes and is no longer hot which decreases the efficiency, capacity and predictability of the AC units. This often leads to hot spots, AC over sizing and unpredictable environments especially with a higher rack power consumption.
With in-row cooling, the AC units are close coupled with the IT racks ensuring they consume the warmest return air possible. This approach allows for low to high density racks while also increasing AC efficiency and providing a very predictable data center.
In-row corrects typical air flow issues in data centers because the AC units consume air directly from the hot aisle before the hot air rises and mixes. The hot air is cooled and then provided directly to the cold aisle ensuring the racks consume cool air.
The latest generation of in-row cooling offers at least five advantages vs. room-based systems, including better reliability and serviceability as well as improved energy efficiency. What’s more, in-row cooling systems take up far less space, leaving more room for IT equipment. And they offer greater flexibility in terms of where the condensing unit is located, making them suitable for almost any location.
In this article, we’ll go through each of the five advantages of in-row cooling in turn.
1. Improved reliability
With traditional room-based cooling, if any component fails it can have a devastating effect on the entire cooling system. For example, most room-based systems have only two fans. If one of them fails, you lose 50 percent of the cooling capacity until it can be repaired. If certain other components fail, you may lose cooling entirely.
The latest in-row cooling systems, on the other hand, have dual power supplies and hot-swappable versions of the components that are most susceptible to failure. The units may have as many as eight fans, for example, all hot swappable, so you can pull one out and slide in another without taking the cooling system out of service.
2. Improved serviceability
Room-based cooling systems typically have all critical components located inside the data center. That means if any component needs servicing,you have a technician inside the data center white space — which makes any IT group nervous and may cause downtime. Should a compressor fail, for example, the technicians may be on-site for a good long while.
Compressors and variable frequency drives are the components that typically need the most service. The compressor for the latest in-row cooling systems are placed outdoors, outside of the data center white space. When it needs service, it all happens outside the data center, away from sensitive IT equipment.
3. Increased energy efficiency
Energy efficiency is another area where in-row cooling bests room-based systems. Room-based systems expel cool air into the entire room, where it mixes with the hot air coming out of the rear of IT racks. That means not all the cool air is directed to that IT equipment, where it’s needed.
In-row systems sit in the very racks where the IT gear is located, and the cool air is directed right to the IT equipment. Containment systems collect the warm air the IT gear expels and delivers it outside, so the cool and warm air never mix. In practice, some 85 to 95 percent of the cool air an in-row system produces goes to cooling IT equipment, whereas only about 70 to 80 percent does for room-based systems.
That means it takes less energy for the in-row system to do its job. In fact, the latest in-row systems have an energy efficiency rating (EER) of 3.1 (up from 2.8 or 3 in previous generations). EER is a ratio of output cooling (in Btu/hr.) to input electrical power (in watts) at a certain outside temperature and humidity. A room-based system, by contrast, typically has an EER of around 2.8.
4. More space
In-row cooling systems do their job in a much smaller footprint vs. room-based systems, leaving more room in the data center for IT equipment. That can be an especially important issue in an edge or micro data center, which by nature are small to begin with.
In a 100 square-foot data center, for example, with the latest in-row cooling system, about 90 to 95 percent of the space would be available for IT equipment. With room-based cooling, you’d need at least 20 percent of the space for cooling equipment.
5. More flexibility
Finally, the latest generation of in-row cooling systems offer a distinct advantage vs. earlier generations in terms of where you can put the condensing unit. With the second-generation Schneider Electric UniflairTM Direct Expansion InRow Cooling system, for example, you can place the condenser up to 120 meters away, making it suitable even for high-rise buildings.
A fresh approach to cooling at the edge
Uniflair is also highly scalable, with 30 to 100 percent modulating cooling capacity. A single unit, for example, can ramp from 8kW to 30kW without any additional investment, thanks to new electronically communicated fans and variable frequency drives (VFDs). In short, Uniflair can keep up with the ever-changing capacity of an edge data center.
It’s innovation like Uniflair that attracted us to work with Schneider Electric in the first place, and we’re excited to be able to offer customers this new cooling line that offers so many advantages over traditional room-based systems. To learn more about how WWT and Schneider Electric can solve your edge computing cooling or other needs, contact the Facilities Team or request a workshop or Cooling Briefing.