New service identifies efficiency gains

Want to save up to 20 percent on your data center’s electricity bills? And while you’re at it: potentially lower your environmental impact and improve IT performance? You might want to consider a DCEE (Data Center Efficiency Evaluation) – a new joint offering from Atos and Siemens.

Can you hear that giant sucking sound, deep within your office building? It could be your data center – the heart of your organization’s IT infrastructure – running up its massive utility bill.


Data centers are notoriously hungry for power. They swallow 10 to 50 times as much energy per unit of floor space as a typical office does, says a report by the United States Office of Energy Efficiency and Renewable Energy. Indeed, the same report finds that data centers consume 2 percent of all electricity – globally, it even amounts to 3 percent of the total electricity consumption.


Of course with our growing demand for computing (think smartphones, tablets, the Internet of Things), consumption is set to keep climbing. Various market analysts predict that data centers will for the foreseeable future grow by about 10 percent every year.

Another factor to consider in this respect is that the infrastructure regulations and standards to which data centers are built tend to improve continuously, offering further potential for optimizing operation expenditures (OPEX) over the course of a data center’s life cycle. In addition, IT hardware efficiency, performance and utilization of assets change over time and require continuous optimization. 


Checking in with the latest technology and infrastructure standards appears even more appealing keeping in mind that nearly half of the life cycle costs of a data center are energy costs. However, most of the energy services on the market today are reactive in nature. Overall transparency is missing, and the most efficient investment options might be hard to identify. 

Tracking energy waste

That is where the joint efficiency evaluation approach by Atos and Siemens comes in to help you identify areas with optimization potential to improve the overall data center efficiency. Most existing data centers, especially the original ones built 15 to 20 years ago, have a high potential of optimizing their energy efficiency. You can order up a DCEE that will track down where energy is going astray and recommend ways how to optimize. 


The main outcome is the identification of efficiency gains in the range of up to 20 percent, such as exchanging inefficient gear, optimizing thermal parameters, or improving power management with software and analytics tools. 


But further fringe benefits also include: 

  • ensuring that your energy bills are actually correct (with ever-more complex tariff structures, an average of 1 out of 18 bills are wrong);
  • more transparency and awareness to define goals and achieve targets with the powerful cloud-based Navigator platform;
  • reductions in liability for electricity taxes and fees, which in many countries are rising;
  • a reduction of the physical server footprint and power requirements by making servers virtual;
  • a decrease in downtime and crashes that can threaten data integrity (and annoy users);
  • declines in your organization’s carbon footprint and other environmental impacts associated with power generation.
There is still a lot of potential for further improvement.
Heike Ludwig, Senior Data Center Consultant at Siemens

Double-breasted approach

The efficiency gains are to be had in two areas. One is in Atos’s field of expertise: IT infrastructure. The other is in Siemens’s field of expertise: facility infrastructure. 


The holistic end-to-end approach from Atos and Siemens helps to identify, on the one hand, where changes in the facility infrastructure can offer greater benefits in the IT space and, on the other hand, where IT can open doors for optimizing the facility infrastructure. 


Both areas, says Heike Ludwig, a Senior Data Center Consultant at Siemens, have vast energy needs and, thus, carry a huge potential for energy savings.


On the IT side, common problems include out-of-date hardware, under-optimized storage or configurations, obsolete networking and under-use of virtualization and the cloud. On the facility side, the trouble usually starts with the lack of an energy management system, and it goes on to suboptimal white-space layouts, oversized cooling and inscrutable flows of energy within the data center.


Beyond that, there is the simple ticking of the equipment clock. IT equipment has a life cycle of around three years, facility infrastructure a life cycle of around 15 years. Anything older tends to be, well, out of date and should be replaced with new technology.

Optimizing power usage effectiveness

Newer generations of equipment almost always offer improved efficiency. Cooling systems, for example, are designed to make more use of ambient air. Operating temperatures need not be as frosty as in the past. Controls are more accurate and more precise, and they deliver more information about operations, so that those in turn can be better optimized.

How much better? In Germany, the national average PUE* for data centers fell from 2.2 in 2008 to 1.8 in 2015, and it is expected to hit 1.65 by 2020.


Good, says Ludwig, but she adds that there is still a lot of potential for further improvement. Internet data center operators have reported achieving overall PUEs of as low as 1.12 in favorable geographic locations and are striving to optimize the efficiency and lower the PUE even further.

How does a Data Center Efficiency Evaluation work?

DCEEs make sense, Ludwig says, for data centers as small as 50 fully built racks and, of course, anything larger. It involves two to three analysts from Atos and Siemens, who typically spend one day on-site at the data center. They then spend another two to three days compiling preliminary, broad recommendations for improvement, or five to seven days to write a detailed report. 


At the end of a DCEE, common recommendations for IT can be: optimization of data storage, swap-outs of inefficient servers, or virtualization of servers, and moving low-risk data and apps to the cloud. For the facility, solutions typically call for a revamp of cooling, installation of modern energy controls and sometimes for a rethink of electricity supply. It can even lead customers to generate some or all of their own electricity.


Whatever the combination, it’s likely to turn that sucking sound into a quiet hum.

* Power usage effectiveness (PUE) is the common metric to measure the energy efficiency of data centers. It is calculated by the ratio of a data center’s total energy use to its IT energy use.


Author: Eric Johnson, business journalist based in Switzerland.

Picture credits: Siemens AG

Subscribe to our Newsletter

Stay up to date at all times: everything you need to know about electrification, automation, and digitalization.