Cardiff University

Managing the cooling cost and decreasing energy consumption at Cardiff University's HPC and general IT-use data center was made easier through StruxureWare Data Center Operation: Energy Efficiency.

University CTO Hugh Beedie explains how DCIM analytics enabled a process of continual improvements to PUE which halved the data center power bill.

Play Video 3:21 minutes

Quick Overview

  • House and protect

    patient data
  • Centralized

    data center
  • Managed

    customer data
  • Shared

    space and services
  • Economy

    of scale


"Schneider Electric Data Center Expert software gives us the measurement numbers we need to evaluate the energy efficiency in our data center. The Data Center Operations us both real-time data – with down to minute-to-minute variation and historical data so I can look back over the last day, week, or month, to see the PUE performance."

"Data Center Expert software...saves us a load of money...Just to put the numbers in context, that data center runs about 250,000 pounds of electricity per year to run and by having a good PUE going from 1.7 down to 1.2, I’m saving 125,000 pounds of electricity every year."

Chief Technology Officer

  • Simultaneous running of IT and supercomputer loads requires an efficiently controlled environment.
  • We need to see water temperatures; leaving, entering the chillers, how fast chillers are operating, how much the compressor pum is operating, an dall of those associated numbers.


"We’ve been through three main phases in the data center. In 2007, we designed and installed a new environmentally efficient data center. Our PUE numbers were around 1.5.

A few years later, another round of changes came along to add a whole load more of supercomputers. So we added more racks and chillers, but we could see that it was gradually getting less and less efficient over time and our PUE had crept up to maybe 1.7.

At the last stage of the upgrade, we made a case to the university to invest a half a million pounds based on the fact that we could decrease energy consumption with much more efficient cooling. Our case said we could deliver a PUE better than 1.4, with a payback time of around 8-to-10 years. In fact, we achieved a PUE of nearly 1.2, which meant the payback period was well below 5 years."

Chief Technology Officer

  • With Data Center Expert, we have the mesaurement numbers to evaluate energy efficiency in our data center.
  • Data Center Operations provides us with real-time data - with down to the minute variation and historical data.
  • Real-time graphs allow us to discover special problems like that chiller running is wasting £5,000 of electricity tonight.

Product Solution

•  Real-time data

•  Historical data – last day, week or month

•  Improved PUE results

•  Controlled data center environment

•  Significant electricity savings

StruxureWare for Data Centers Value

DCIM analytics enabled a process of continual improvements to PUE which halved their data center power bill.

Environmentally Efficient

Data center

"In the past, general IT services and supercomputers existed in quite different environments, but with virtualizing techniques, we've now found that the infrastructure to support general IT and supercomputers is converging super computing tasks and the load put on them can vary enormously.

If all the supercomputers happen to be running complex calculations t the same time, then that load would go up to 400-450 kilowatts. That means you need a very good controlled environment to make sure it's running efficiently."

Chief Technology Officer

Lowered PUE

Results in reduced energy consumption and cost savings

When Cardiff University installed the StruxureWare modules, Data Center Expert and Data Center Operations, they expected to get a PUE of around 1.4, with a payback time of around 8-10 years.

They were pleasantly surprised when the resulting PUE was close to 1.2, with a payback period of well below 5 years. Are they happy with the results...needless to say, "The proof is in the pudding."