We keep our goal of reducing our carbon footprint in mind when planning and operating our data centers. To achieve this goal, we take 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.
The key indicator for energy efficiency at our data centers is the PUE 2008, we have succeeded in reducing the average PUE factor at the T-Systems data centers in Germany from 1.85 to 1.57. Compared to the previous year, the figure in 2018 remained virtually constant (2017: 1.54). The reason for this is the ongoing consolidation of the data centers, where applications from less energy-efficient sites are migrated to new, highly efficient data centers. This increases the PUE value of the decommissioned data centers in the short term. Overall, this process led to a constant PUE value across all data centers in the reporting year – and saw the value shrink further in the medium term.factor. Since
The “DC11@2018” program for improving processes throughout our global data center landscape ran from 2013 to 2018. This project combines physical data center consolidation (i.e., reducing data center space and sites) with logical consolidation (i.e., virtualizing data center infrastructure). The objective of the “DC11@2018” program was to globally consolidate data center capacity at FMO (future mode of operation) data centers using the latest IT technology. The consolidation project was completed by the end of 2018 as planned. Within the scope of the physical consolidation, we decommissioned three old data centers in Germany. At the same time, we optimized the air conditioning at the existing data centers in Biere and Munich and other international data centers, including Barcelona in Spain.
The aim was to achieve an average PUE factor of 1.4 across the Group in all FMO data centers by the end of the project. At the end of 2018, we reached a PUE value of 1.57, meaning that we did not attain our target. This is in part due to the gradual shutdown of old data centers – which become less efficient during decommissioning as less capacity is used – and the negative impact this has on the PUE value. Achieving our target calls for further homogenization of our IT landscape combined with optimum capacity utilization of data center infrastructure, IT hardware, and the software running on the systems. Other factors that have a positive impact on PUE development in a data center are selective cooling and increasing the temperature as far as possible – while simultaneously observing set thresholds. For the year 2020, the latest program plans indicate accumulated reductions in CO2 by up to 51 percent compared to 2012. The PUE value for the network infrastructure is predicted to drop to below 1.4 by 2020.
Our target PUE factor for our highly efficient data center in Biere is 1.3. To achieve this target, it will be necessary to adhere to a prescribed temperature range, achieve a capacity utilization of at least 80 percent, and create an IT landscape that is as homogeneous as possible. We were not yet able to achieve the PUE target value in 2018 because the level of capacity utilization is not yet as high as originally planned. In addition, it was necessary to cool the data center on multiple occasions during the hot summer. As outlined above, the data centers were consolidated as part of “DC11@2018”, with the data center in Biere successfully expanded by May 2018 to make up for lost capacity.