Using less electricity by making our network operations more energy efficient is a central element of our climate protection target. We take different approaches for the fixed network, mobile communications and data centers. All of these approaches, however, include the use of innovative ICT solutions for load-based control of network performance and energy consumption.
Telekom operates its own fixed-line and mobile communications networks in Europe and the US. Much of our energy demand comes 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 approaches to further reducing our energy needs:
- We update our network infrastructure through measures such as switching over to IP technology and installing highly efficient systems to guarantee the availability of electricity, systematically removing technology we no longer need
- We optimize energy supply and energy conversion through improved energy storage facilities
- We use more energy-efficient technology for lighting, monitoring and cooling in our technical network facilities
Our internal energy service provider, Power & Air Solutions, who is responsible for energy management at most of Telekom's companies in Germany, plays a key role in these activities. Power & Air Solutions has been using an ISO 50001-compliant energy management system since 2013.
Telekom was among the companies that signed the EU Broadband Code of Conduct and the EU Data Centre Code of Conduct. By signing these documents, Telekom has committed itself to continuing to effectively improve the energy efficiency of its products and services.
Measuring and controlling energy requirements
The PUE factor (Power Usage Effectiveness), which we record annually, is an important indicator that we use to manage our climate protection measures. Our goal for operation of our fixed-line network in Germany is to reduce our PUE factor to 1.4 by 2020. By doing this, we will be able to compensate for the increased energy requirements associated with growing amounts of data and new service features.
More computing capacity – less energy consumption
There is a lot of room to reduce energy consumption at data centers. We tap these possibilities systematically each time we build new data centers or update old ones. In these activities, we focus on the following:
- Optimizing heating, ventilation and cooling technology
- Energy-efficient components
- Optimizing stand-by features
- More efficient archiving and storage
In 2014 we began operations at a new, large data center at the Biere location in Magdeburg. We are setting the bar at this data center in all of these areas with innovative technology.
Biere: Role model for energy efficiency
The new cloud data center in Biere is the largest of its kind in Germany as well as one of the largest in Europe. It has room for 30,000 servers and its PUE factor of 1.3 represents a significant milestone in the achievement of our climate target. Its innovative cooling concept alone cuts its total energy requirements by around one-third relative to comparable data centers. We received the Gold LEED certification (LEED: Leadership in Energy and Environmental Design) for the building's energy-efficient, environmental design. Only ten data centers in the world have this certification.
We have fitted out an existing data center in Magdeburg, not far from Biere, almost identically. The two data centers work together closely as "twins." Data is stored at both data centers at all times so that is is always available, even if there is a problem at one of the data centers. Outsourcing data to the cloud also requires the highest security standards. That is why the building complexes are systematically partitioned; the strictest security measures protect data from unauthorized access.
Optimizing and consolidating data centers: Less is more
We are also making efforts to reduce CO2 emissions in the way we operate and design our data centers. The approach taken by Group subsidiary T-Systems basically falls into two phases: optimization measures at the different data center sites and global optimization spanning the entire data center landscape worldwide.
The first phase was conducted from 2008 to 2013 by thoroughly optimizing current data center sites. 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. In this context, upgrading our IT systems and migrating the application environments to this new technology helped us to improve energy efficiency considerably.
The PUE factor serves as an indicator for improvements in energy efficiency. Between 2008 and 2014, we saw a positive change in the average PUE factor of the T-Systems data centers from 1.85 to 1.64 thanks to the measures described above. Despite the overall positive trend, we saw a slight increase in the PUE level in 2014 compared to the previous year. This is due to the fact that, while energy consumption by the servers declined, the energy required for non-ICT components remained unchanged.
The second phase began in 2013 and involved combining the physical consolidation of data centers (i.e., reducing the amount of data center space and the number of sites) with logical consolidation (i.e., virtualization). 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. We expect these measures to massively reduce CO2 emissions. Current program plans estimate a cumulated CO2 reduction of up to 51 percent by 2020 based on the 2012 figures.
Winning the Code of Conduct Data Centre Award 2014
We continue to improve the efficiency of our data centers to help us achieve our climate protection target. In May 2014, T-Systems was presented the European Union's Data Centre Energy Efficiency Award for its success with these measures in 2013. Our submission included publishing the energy values of all of our data centers and submitting a list of improvement measures.
Power-save mode: Smart and economical
Network control is a major lever when it comes to making broadband services more environmentally friendly and efficient for our customers. We collaborated with a network equipment manufacturer to develop an innovative operating mode (L2 mode) for the fixed-line network. This mode makes it possible to flexibly adjust transmission capacity to the actual data volume in several switch levels. We can use this mode to significantly reduce energy consumption in access networks during low traffic periods. We planned to conduct a field test in 2014. The quality controls at the lab, however, have not yet been completed. We have tested the L2 mode in live network operation, but without customer participation. Depending on the results, we should be able to start the field test in 2015. Once the field test has been completed, we are planning to gradually put the new mode into live operation.
The diagram on the right shows the method with two switch levels, which already allows a rough approximation to the curve representing actual data traffic (depicted here in magenta). Analyses have shown that the enhanced L2 mode will make it possible to avoid large amounts of CO2 emissions and save operating costs.
Generating heat and power under a single roof
In addition to traditional power grids, we are also increasingly using combined heat and power units (CHP) to provide power that we generate ourselves to our network nodes (Internet distribution points) in a way that is efficient and environmentally friendly. These units convert around 90 percent of the energy into usable energy (electricity and heat), whereas electricity generated via conventional means and sourced from the German power grid only comprises round 40 percent usable energy. Using CHP units has enabled us to significantly reduce the CO2 emissions of our network operations. We added four new CHP units with electrical power output ranging from 50 kW to 250 kW in 2014.
Technological and efficient ways to use CHP units
CPU units basically cover the electricity and heating needs of our network nodes via thermal energy. We have improved load distribution by introducing a new control approach in order to tone down excessive electricity needs (peak loads) during times of exceedingly high data traffic volumes. However, the German Renewable Energy Act (EEG), which was passed in 2014, made conditions for operators of environmentally friendly systems like these more challenging. Operators are now subject to a 40 percent EEG fee on self-generated energy from CHP units used in their own operations.
CO2 emissions reduced by 8,687 metric tons
In February 2015, Telekom Deutschland was operating its own CHPs at a total of 27 network nodes. In 2014, these units generated around 33 GWh of electricity. Because the units are very efficient, they emitted 8,687 tons less CO2 compared to power taken from public grids.
Photovoltaics for mobile base stations
T-Mobile Austria is testing different solutions to support power supply to base stations with photovoltaics and wind energy at five mobile base stations in its Green Base Station project. The main focus of the tests is whether and under which circumstances the systems are economically viable.
Photovoltaic modules with specially developed holders are attached directly to the base station masts and small wind turbines are attached to the tops of the masts. Any excess energy produced by the plants is stored on-site in batteries to cover any power outages. With the help of a new control system, the stored energy can also be used to power the stations on days with little sunshine or wind. Test findings from 2014 show that up to 40 percent of the energy needed to power the base stations can be produced by renewable sources. The use of renewable energy generated by photovoltaics has successfully passed the functional tests.
Apart from functionality, the goal is to develop an efficient approach for installing additional plants. There are still a few issues to be addressed.
Because we are now able to monitor the system over a longer period of time, we can also use information regarding the behavior of the systems to optimize them. Tests on the small wind turbine were not extended due to heavy vibrations of the mast. We are analyzing the use of other types of wind turbines for additional tests instead.