I spent the early part of this week in a new world, International Telecoms Week (ITW). While many of the attendees have been, and continue to be customers of ours, it was a conference far different from the more facility-oriented conferences I am used to. Being a fish somewhat out of water I took the opportunity to learn as much as I could about upcoming trends in the data center space from those on the bleeding edge of IT instead of mechanical or facility design. Most of the time I felt like I was drinking from a fire hose but when I zoom out two large trends effecting thermal management design became abundantly clear. Large scale deployments of edge facilities will be necessary as new computing technologies further interconnect with our daily lives, and loads become increasingly dense.
Trends in Thermal Control
As digital transformation becomes a top priority in 2018, the Internet of Things will continue to be a trend to watch. IOT-enabled communications, where time-to-result matters, will continue to take center stage. We can’t increase the speed of light (at least at the moment) which means there is a fundamental limit to how quickly computations can be performed through the internet. This may not be a big deal for checking your bank account or scrolling through articles but what about self-driving cars? Real-time virtual surgery? Robotics on deep water oil-rigs? The proliferation of the internet of things into critical, life threatening applications will force latency to a minimum and the only way to do this is to locate some of the computing as close to the device as possible i.e. the edge. There are many companies looking to serve to edge with modular or containerized data center solutions – one metric to differentiate these is energy efficiency.
Solutions in Edge Computing
Free-cooling is essential to having a low PUE but free-cooling is not one-size-fits all. Some companies looking to take an early lead in the edge market are designing their solutions to interchangeably connect to different methods of free-cooling depending on site location. For example, an edge facility in Florida would not be able to take advantage of water-side free-cooling which could be an excellent option in Colorado.
The second big takeaway from ITW is that loads are getting more and more dense; dense to the point where Thermal Management providers are going to need to design new methods for cooling racks hitting 50KW+. This is where my two big lessons from ITW converge. As more and more devices connect to the internet, more and more edge facilities will need to be deployed, and as the edge proliferates into more and more remote locations, dense computing equipment will be required to enable installation and successful operation.
If you are a looking to build your footprint at the edge or work in an organization building containerized solutions — be sure that your thermal management partner is looking at how the impact site location and density will have effect on your designs. At Data Aire, we look for opportunities to optimize standard systems for specific applications when working with our partners; our production processes enables rapid deployment of flexible and customized systems — in terms of design and scale which will allow us to support our partners as they expand their network edge.