Authors and contributors
IEA (2019), "Tracking Buildings", IEA, Paris https://www.iea.org/reports/tracking-buildings
Global internet traffic has tripled since 2015, and is expected to further double by 2022 to 4.2 zettabytes per year (4.2 trillion gigabytes) (Cisco, 2015; 2018; 2019). The number of mobile internet users is expected to increase from 3.6 billion in 2018 to 5 billion by 2025, while the number of Internet of Things (IoT) connections is expected to triple from 7.5 billion in 2018 to over 25 billion by 2025 (GSM Association, 2019).
These trends are driving exponential growth in demand for data centre and network services.
Most of the world’s Internet Protocol (IP) traffic goes through data centres. Greater connectivity is therefore driving up demand for data centre services and energy use (mostly electricity), with multiplying effects: for every bit of data that travels the network from data centre to end users, another 5 bits of data are transmitted within and among data centres (Cisco, 2016).
Global data centre electricity demand in 2018 was an estimated 198 TWh, or almost 1% of global final demand for electricity (Masanet et al., 2018).
Based on current trends in the efficiency of hardware and data centre infrastructure, global data centre energy demand is projected to decrease slightly to 191 TWh in 2021 (Cisco, 2018; Masanet et al., 2018; Shehabi et al., 2016). This is despite a projected 80% increase in data centre traffic and 50% increase in data centre workloads over the next three years (Cisco, 2018).
Strong growth in demand for data centre services continues to be offset by continued improvements in the efficiency of servers, storage devices, network switches and data centre infrastructure, as well as a shift to much greater shares of cloud and hyperscale data centres. Hyperscale data centres are very efficient, large-scale cloud data centres that run at high capacity, owing in part to virtualisation software that enables data centre operators to deliver higher work output with fewer servers.
The shift away from small, inefficient data centres towards much larger cloud and hyperscale data centres is evident in the shrinking share of data centre infrastructure in total energy demand, given the very low power usage effectiveness (PUE) of large data centres. PUE is a measure of how efficiently a data centre uses energy; the very best hyperscale data centres can have PUE values of around 1.1 (meaning 0.1 kWh used for cooling/power provision for every 1 kWh used for IT equipment).
The nascent IT infrastructure for blockchain and cryptocurrencies is evolving rapidly, and its implications for data centre electricity use are not yet well understood.
Early estimates suggest that the electricity used by Bitcoin miners – one prominent example of emerging blockchain IT infrastructure – may currently amount to around 0.1‑0.3% of global electricity consumption. However, as blockchain applications grow, understanding and managing its energy use implications may become increasingly important for energy analysts and policy makers.
Data networks consumed around 260 TWh globally in 2018, or about 1.1% of total global electricity demand, with mobile networks accounting for two-thirds.
In the near term, the range of possible energy outcomes is wide, hinging largely on data demand growth and the pace of further efficiency improvements.
In a moderate efficiency-improvement scenario of 10% per year (a conservative estimate based on historical improvements), electricity demand could rise nearly 10% to around 280 TWh in 2021.
In contrast, under a high efficiency-improvement scenario of 20% per year (based on rates achieved in well-managed networks with high capacity utilisation), demand could drop by 25% to about 190 TWh.
This range of potential outcomes highlights the importance of policy to drive further efficiency gains.
Several trends are shaping the future of data network electricity use. Global IP traffic increased nearly threefold during 2014‑18, and similar growth is projected for 2018‑22. The nature of data transmission is changing rapidly, with traffic from wireless and mobile devices expected to account for more than 70% of total IP traffic by 2022, up from around half in 2018.
This shift towards greater use of mobile networks may also have significant implications for the energy use of data transmission networks, given the considerably higher electricity intensities (kWh/GB) of mobile networks compared with fixed-line networks at current traffic rates.
While the latest mobile telecommunications technologies are much less energy-intensive than older technologies (e.g. 4G can be more than 50 times more energy efficient than 2G), their higher speeds may allow for greater usage and traffic volumes.
Despite strong growth in data demand and shifts to mobile transmission, data transmission network technologies are rapidly becoming more efficient: fixed-line network energy intensity has halved every two years since 2000 in developed countries (Aslan et al., 2017), and mobile-access network energy efficiency has been improving 10‑20% annually in recent years (Fehske et al., 2011).
Mobile networks are shifting rapidly away from older networks towards more efficient 4G (and eventually 5G). By 2022, 4G and 5G are expected to carry a combined 83% of mobile traffic, compared with less than 1% for 2G (Cisco, 2019).
Demand for data centre and network services is expected to continue to grow strongly, but how this affects energy use will continue to be determined largely by the pace of energy efficiency gains.
Government policies, as well as action and commitments by data centre and network operators, will be essential to driving further efficiency improvements to moderate overall ICT energy use.
With data demand growth expected to be strongest in the Asia-Pacific region and North America, continued efforts to make data centres and networks in these regions more efficient is critical. Huang and Masanet (2015) offer a summary of best practices and how to calculate savings for incentives programmes.
Governments and network operators could be instrumental in implementing policies and programmes to improve the energy efficiency of data transmission networks.
Actions could include accelerating the phase-out of energy-intensive legacy networks, implementing network device energy efficiency standards, improving metrics and incentives for efficient network operations, and supporting international technology protocols.
Improving data collection on ICT and their energy-use characteristics can help inform energy analysis and policy making.
For example, the US Energy Information Administration collects data on connected devices in homes (RECS) and commercial buildings (CBECS), as well as on servers in data centres (CBECS).
Companies and industries could agree to voluntary efficiency and CO2 emissions targets.
For example, the European Union and the United States have adopted voluntary agreements with companies to improve the efficiency of connected set-top boxes. Verizon has announced public efficiency goals for its network operations, for which it annually publishes indicators (kWh/GB) and progress.
Demand for data centre and data transmission network services is expected to continue to grow strongly over the next decade. Innovation will be critical to ensuring that energy efficiency gains continue to keep overall energy demand in check.
Global internet protocol (IP) traffic is increasing rapidly, and is expected to triple by 2022. This traffic is increasingly shifting to wireless and mobile: wireless and mobile devices expected to account for more than 70% of traffic by 2022, up from around half in 2018.
This shift toward greater use of mobile networks may have significant implications for energy use, given the considerably higher electricity intensities (kWh/GB) of mobile networks compared with fixed-line networks at current traffic rates.
Demand for data centre services is expected to continue to grow strongly after 2020, and data centre energy use will continue to be largely determined by the pace of energy efficiency gains. While the continued shift to efficient cloud and hyperscale data centres will reduce the energy intensity of data centre services, applying artificial intelligence (AI) and machine learning to tap further efficiency gains may become increasingly important.
- Aslan, J. et al. (2017), "Electricity Intensity of Internet Data Transmission: Untangling the Estimates", Journal of Industrial Ecology, https://doi.org/10.1111/jiec.12630.
- Cisco (2015), "The History and Future of Internet Traffic", https://blogs.cisco.com/sp/the-history-and-future-of-internet-traffic.
- Cisco (2016), "Cisco Global Cloud Index 2015-2020 - Cisco Knowledge Network Session".
- Cisco (2018), Cisco Global Cloud Index: Forecast and Methodology, 2016–2021, https://www.cisco.com/c/en/us/solutions/collateral/service-provider/global-cloud-index-gci/white-paper-c11-738085.pdf.
- Cisco (2019), Cisco Visual Networking Index : Global Mobile Data Traffic Forecast Update , 2017 – 2022, https://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/white-paper-c11-738429.pdf.
- Fehske, A. et al. (2011), "The global footprint of mobile communications: The ecological and economic perspective", IEEE Communications Magazine, Vol. 49/8, pp. 55–62, https://doi.org/10.1109/MCOM.2011.5978416.
- GSM Association (2019), The Mobile Economy, , https://www.gsmaintelligence.com/research/?file=b9a6e6202ee1d5f787cfebb95d3639c5&download.
- Huang, R. and Masanet, E. (2015), "Data Center IT Efficiency Measures", The Uniform Methods Project: Methods for Determining Energy Efficiency Savings for Specific Measures, https://www.nrel.gov/docs/fy17osti/68576.pdf.
- Masanet, E. R. et al. (2018), Global Data Center Energy Use: Distribution, Composition, and Near-Term Outlook, Evanston, IL.
- Shehabi, A. et al. (2016), United States Data Center Energy Usage Report, Berkeley, California, https://doi.org/LBNL-1005775.
Eric Masanet, Northwestern University