Do you want your datacenter to be greener?

Do you want your data center to be greener?

The European Union wants data centers to be greener. Following the EU publication Shaping Europes Digital Future, Telecommunications and data centers have a significant environmental footprint they are estimated to contribute 5-9 % to global electricity use and 2% of total emissions. Hence, the EU stated with regard to data centers and telecommunications that they can and should become climate neutral by 2030.

Data centers and cloud hosting, the workhorses of the worlds digital infrastructure, contain thousands of compute servers. Overall, servers and cooling systems contribute most to direct electricity use in data centers. They are followed by storage drives and network devices.[1] Since the total number of internet users and also the number of internet-based applications are continuously growing, concerns with regard to data center energy use have risen in the past years. For example, between 2010 and 2018 global data center storage capacity increased by a factor of 25. Additionally, the amount of compute instances (total applications hosted) on servers worldwide increased by a factor of 6.[2]

Moreover, new forms of information services such as artificial intelligence, which are computationally intensive, amplify the growth of power-hungry computing resources. Hence, politics demand data centers to be greener in the future.

We at Xelera provide an approach to reduce energy consumption in data centers dramatically. Take the server rack below as an example. It consists of 7 standard commercial-off-the-shelf servers, each of which is equipped with a highly energy-efficient hardware accelerator card (also a COTS product). The cluster is deployed at a customer site and runs high-performance computing tasks.

The Xelera Suite acceleration software ensures that these tasks are offloaded to the accelerator cards. If we were to match the computing performance of this cluster without accelerators, we would need 52 servers of the same type instead of 7. Besides more space, this results in a significantly higher energy footprint as shown below.

Now remember that the cluster does not run for an hour it runs continuously throughout the year, which amounts to a breath-taking difference of 124584 kWh per year compared to the conventional solution.

To put the impact of this into a practical example, consider a standard baking oven in a modern kitchen: It consumes 0.4 kWh. If you left it on 24/7 throughout a year, it would consume 3504 kWh/year, which is still 35 times less than the power savings of the accelerated server cluster.

Right Xeleras solution saves energy, which could run 35 ovens for a whole year.

The previously mentioned example only included one server rack. However, data centers are not just a single rack, but they consist of a tremendous number of server racks. Therefore, Xelera's acceleration software is able to have an even greater impact on your data center.

Moreover, the ability to reduce energy usage is even greater when using more than one accelerator per server. For example, the servers in our last deployment are able to carry 6 accelerator cards.

Xelera's acceleration software can be used in a variety of use cases.

If you want to make your data center greener, please contact us via or visit our website.

[1] Shehabi, Arman, et al. (2016). United states data center energy usage report. No. LBNL-1005775. Lawrence Berkeley National Lab.(LBNL), Berkeley, CA (United States)

[2] Masanet, Eric, Arman Shehabi, Nuoa Lei, Sarah Smith, and Jonathan Koomey. "Recalibrating global data center energy-use estimates." Science 367, no. 6481 (2020): 984-986

October 23rd, 2020

Do you want your datacenter to be greener?

Data center

The European Union wants data centers to be greener. Following the EU publication “Shaping Europe’s Digital Future”, Telecommunications and data centers have a significant environmental footprint... more
March 10th, 2020

Accelerating Decision Tree-based Predictive Analytics


Gradient boosting frameworks such as XGBoost, LightGBM and CatBoost, as well as Random Forest algorithms are often a part of winning machine learning models in Kaggle competitions ... more
February 25th, 2020

Recognizing Voices with AI

Meeting room

Voice-based digital assistants are on the rise. Xelera Technologies provides an AI module for speech processing systems which distinguishes voices and gives ... more
Visit us on: