Advanced Computing in the Age of AI | Wednesday, May 1, 2024

Nlyte Optimizes Green Datacenter Design 

<p>Designing a datacenter has to take a lot of optimization into account, from performance and latency to, lately, energy costs and waste. CloudNician Founder and CEO James Grundvig interviewed executives from Nlyte, including CEO Doug Sabella and Marketing VP Mark Harris, on the interest in increasing data center efficiency.</p>

Designing a datacenter has to take a lot of optimization into account, from performance and latency to, lately, energy costs and waste. CloudNician Founder and CEO James Grundvig interviewed executives from Nlyte, including CEO Doug Sabella and Marketing VP Mark Harris, on the interest in increasing data center efficiency.

Possibly somewhat in response to a New York Times series admonishing the environmental inefficiencies of datacenters, new facilities have been designed with that in mind. Reported PUE’s of these new facilities range around 1.3.

According to Harris, however, PUE does not quite measure everything, and what it does measure is difficult to adjust since it only takes place after the facility is built and operational.

"We have found that every data center is so unique, even the largest ones are night and day in comparison," Harris said in discussing datacenter efficiency metrics such as PUE (Power Usage Efficiency). "PUE is a great directional metric, but provides little guidance at actually optimizing inefficiencies. PUE don't address capacity planning within the confines of a data center. It falls short."

The key is to implement software that identifies the optimal facility configuration before the servers and stacks are all laid out. Sabella spoke broadly about Nlyte’s new 7.0 offering assisting to advance many datacenters’ green hopes. "The tech industry has wrestled with how to be Green for some time. Anything that eliminates wasted energy is currently considered Green--which is precisely what Nlyte's offering accomplishes,” Sabella said before going into the specifics of how their technology identifies key efficiency points.

For example, according to Sabella, Nlyte is able to look into specific aspects such as the visualization or cloud layer to determine where the bottlenecks are happening. “Whether it's increasing the amount of work done by servers, or increasing visibility into the cloud or virtualization layers already in use, Nlyte allows the Data Center to run at the lowest cost, with the least amount of energy wasted," Sabella said. This technology could hypothetically be applied to existing data centers running into energy waste issues or planned facilities.

Sabella noted that, for existing centers, the software exists atop a database of efficiency information, which would hypothetically allow them to draw time-dependent models of a facility’s energy usage based on past use cases. “Through our contextual data repository, 7.0 has coordination with scalability and performance. It forecasts where you go from where you have been, accounting for the time-sensitive nature of each."

When it comes to planning new designs, according to Sabella, the Nlyte software would draw upon that database for optimizing layout. "We put these customer-driven inputs into the software that enables superior floor planning, navigating in and around rows of servers, with far fewer through-clicks on user experience,” Sabella said.

Designing datacenters can be quite the engineering challenge when relatively new factors such as environmental impact have to be taken into account. Software such that the Nlyte company champions may be a step in determining the proper layout and advance green IT.

EnterpriseAI