- Reducing energy requirements
- PUE factor of 1.4 in the fixed network by 2020
- Consolidating data centers
- Migration to IP
- PUE factor
- Energy consumption
Deutsche Telekom operates its own fixed-line and mobile communications networks in Europe and the United States. Most of our energy requirements come from operating this network infrastructure. In the interest of our customers, we continue to increase the capacity and performance of our networks so that we can handle growing amounts of data and increasing demands on the speed and quality of data transmission. This always goes hand-in-hand with higher energy requirements. We pursue three different approaches to further reduce our energy needs:
- We are modernizing our network infrastructure by migrating over to IP technology, installing highly efficient plants to ensure availability of power and consistently removing technology that we no longer need.
- We optimize energy supply and conversion, e.g., by improving power converters.
- We use more energy-efficient technology for lighting, monitoring and above all cooling our facilities.
Our internal energy service provider, Power & Air Solutions , which is responsible for energy management in Germany, plays a key role in these activities. Power & Air Solutions has been ISO 50001 certified since 2013 for their energy efficient approach.
Measuring and controlling energy consumption
The PUE factor (Power Usage Effectiveness) is measured annually and is an important performance indicator for managing our climate protection measures. Our goal for our fixed network in Germany is to reduce our PUE factor to 1.4 by 2020, allowing us to compensate for increasing energy requirements due to growing volumes of data and new features.
Research on efficient network technologies
We constantly work to increase the energy efficiency of our network infrastructure. In 2015, we launched the EARNI project research project together with partners from industry (Alcatel-Lucent) and science (Fraunhofer Heinrich Hertz Institute). The project's goal is to improve network capacity utilization and energy efficiency by using multipurpose technologies. We suffered a setback in a comparable project in March 2015 when we decided to cancel the project after a field test had shown that introducing the technology for load-adaptive network operation would only be feasible with substantial effort and considerable losses in terms of reliability/availability.
We also keep our goal of reducing our carbon footprint in mind when planning and operating our data centers. To achieve this goal, T-Systems takes a two-step approach: It starts with optimizing energy consumption at each data center site and then continues with improving processes throughout the global data center landscape.
Between 2008 and 2013 the focus was on optimizing existing data center space. Some examples include updating the cooling systems and installing cold aisle containment to control the flow of cooling air as needed and minimize cold air waste in the IT areas. Advancements in IT technology also made it possible to considerably reduce energy consumption. In addition to these fundamental technical optimization measures, we are always fine-tuning our climate control technology; this in combination with optimum data center capacity utilization enables us to continue to improve efficiency.
The PUE factor serves as an indicator for improvements in energy efficiency. We were able to reduce the average PUE factor at the T-Systems data centers from 1.85 to 1.63 between 2008 and 2015. Despite these favorable overall results, we recorded fluctuations in the PUE factor during the reporting period compared to the previous year. This is due to the fact that, while the energy required for non-ICT components remained unchanged, computer energy consumption fluctuated in the wake of the consolidation process.
The second phase has been ongoing since 2013. This phase combines physical data center consolidation (i.e., reducing data center space and sites) with logical consolidation (i.e., virtualizing data center infrastructure). The DC11@2018 program is working to globally consolidate data center sites with the latest IT technology to a few FMO (future mode of operation) data centers. The target average PUE factor at the FMO data centers will be 1.4 once the program has been completed. This requires a homogeneous IT landscape combined with optimum capacity utilization of data center infrastructure, IT hardware and the software running on the systems. We expect these measures to bring about another massive reduction in CO2 emissions. Current program plans estimate a cumulated CO2 reduction of up to 51 percent by 2020 based on the 2012 figures.
Operating our networks and data centers accounts for the majority of our energy consumption. Through various measures, we intend to reduce our energy consumption and improve the energy efficiency of our infrastructure, moving us closer to our climate target. We initiated several projects in 2015 including the following:
Creating an energy dashboard
During the reporting period we developed a new web-based energy application, the Energy Dashboard. The app makes it possible to monitor approximately 8,000 fixed-network nodes and 20,000 mobile base stations. This gives all Power & Air Solutions employees who monitor the energy consumption levels of our systems a quick overview – at any time and from anywhere. The app includes a notification system in case large consumption points exceed certain threshold values. This lets us identify possible causes early on and take counteractive measures. The app also makes it possible to identify and selectively control central network nodes with high traffic volumes.
Improved cooling technology for data centers
In 2014, we launched operations at our cloud data center in Biere. The cooling concept developed in-house cuts total energy requirements by around one-third relative to comparable data centers. But we do not limit our efforts to new data centers. We also strive to cut back on energy consumption in our current network infrastructure. That is why we tested two new ways to cool data centers and switching centers in 2015:
- controlling ventilators already in place with the help of 3D room temperature measurement
- new heat removal systems.
We will use the test findings to further optimize our current cooling systems.
We are always expanding and optimizing our networks so that we can offer services that are demand-oriented and energy-efficient. Whenever technically feasible, we are increasingly employing system technology based on multipurpose hardware . We joined forces with Alcatel-Lucent and the Fraunhofer Heinrich Hertz Institute in 2015 to launch the Energy-Aware Network Infrastructure (EARNI) project. The aim of this project is to find out whether using new multipurpose systems has a favorable impact on network energy efficiency.
For the research project, we used an actual use case to calculate energy requirements and compared the results with conventional technology. The results of this comparison show that, with careful planning and network design, the capacity of multipurpose systems can be much better utilized and therefore promise better energy efficiency compared to conventional technology. The system also permits load-adaptive operation.
Multipurpose hardware brings other advantages as well. The technology makes it possible to incorporate controllers to control local energy supply systems. Cogeneration plants and photovoltaic systems can be used to provide power to the telecommunications network. What's more, this equipment can be controlled with the same hardware that is used for the network technology. The EARNI project also analyzed a use case pertaining to the control of power supply.
We are increasingly using combined heat and power (CHP) plants in addition to the conventional power grid. They are an efficient, environmentally friendly way to provide power to our network nodes: CHP plants convert around 90 percent of the energy into usable energy (electricity and heat); in contrast, conventional electricity generation in the German power grid has an efficiency of only around 40 percent.
By using these plants, we can considerably reduce the CO2 emissions caused by network operation. Power & Air Solutions was operating its own CHP plants at a total of 27 network nodes in February 2016. These units generate around 7.9 GWh of electricity each year. Due to the high efficiency of these units, they generated 2,700 metric tons less CO2 than energy taken from public power grids.
More units planned
We commissioned two additional CHP units to be installed in Berlin and Frankfurt am Main in 2015. We also looked into how to increase the efficiency of our current CHP plants. Based on the findings of our analysis, we designed a new energy concept for one of our sites in Berlin. The concept should be in place by the end of 2016.
CHP units basically cover the electricity and heating needs of our network nodes via thermal energy. To tone down excessive electricity needs (peak loads) during times of exceedingly high data traffic volumes, we have improved load distribution by introducing a new control approach.