Tag: data traffic

With big data traffic, comes big energy costs

In 2009, mobile phone networks carried 91 Petabytes (that’s 9.1×1016 bytes) of data per month. Ten years later, mobile networks are estimated to carry around 30 Exabytes of data per month. That’s 30 x1019 bytes or 15 billion HD movies, or an average of 2.5 movies for the 6 billion smartphone users on the planet. This represents an increase of over 330-fold in data traffic. This rapid increase in data carried by mobile networks is projected to slow, but even a reduction in CAGR to 30% a year would see volumes reaching over 130 Exabytes a month in 5 years.

This increase in data travelling over mobile networks reflects the increasingly data-heavy applications running on mobile devices as well as the increasing penetration of smartphones in many developing markets. Although enterprises and public sector firms will also drive demand for mobile data, this is minor compared to consumer user demand. As mobile device penetration rates continue to increase and mobile device owners adopt more data heavy applications such as video streaming and immersive experiences, growth in volumes will continue far past 2024 and we could easily still see a 20-fold increase over current levels over the next 10 years.

In many ways this is exciting, as more computing power reaches the hands of more people around the world delivering applications that help billions of people in their daily lives. But this comes with a caveat – there is an input inherent in delivering this data traffic: the energy needed for running the network infrastructure.

The electrical energy required to power networks represents a cost to operators, but it also represents CO2 emissions arising from burning fossil fuels to power the network (either directly from local dedicated generators or through the power grid). Greenhouse gas emissions therefore also risk increasing significantly as a result of data growth, particularly in countries heavily dependent on fossil fuels for their electricity production. Previously, this has been managed by the fact that mobile networks have been optimised to support larger amounts of data with a similar topology in terms of infrastructure, if slightly higher energy needs and costs. However, as spectrum is a limited resource, continued growth of mobile traffic over current LTE networks would quickly lead to densification – an increased amount of antenna and network infrastructure – by some estimates this would be an increase of 160% by 2025.

Even with improvements in hardware performance, growth in mobile data over LTE networks would result in significant growth in energy consumption which represents a significant source of emissions. This is at odds with the goals that operators have set for themselves in terms of greenhouse gas emissions and risks breaching the standards to which (consumer, public and private sector) customers are increasingly holding their suppliers. Investors are also factoring in the higher risk profile associated with companies with high carbon emission exposure5. Finally, this also matters to employees – particularly younger ones with the digital skills-sets that operators are looking to attract.

The key question posed by this dilemma then is: How should mobile network operators deal with this rapid rate of growth in data and the associated energy consumption and CO2 emissions?

In this report, we focus on one part of the answer – accelerating the adoption of more energy efficient 5G technologies and associated operational practices. A faster roll-out of 5G networks is a key weapon in operators’ arsenal of measures for de-coupling energy costs and carbon emissions arising from data growth.

 

Scope of this report

Network operators need to mitigate the ever-increasing energy costs and carbon footprint of their networks resulting from the forecast data growth in some way. There are six ways that the accelerated adoption of 5G can do this:

1. Direct curtailment of energy consumption in mobile access networks through the better energy “performance” of 5G network equipment and operational practices relative to 4G.
2. Direct curtailment of energy consumption in 5G core networks through the better energy “performance” of network equipment and operational practices relative to 4G core networks.
3. Reduced energy consumption by devices (particularly smartphones and IoT devices).
4. Decarbonising the grid: indirectly enabling lower levels of national carbon emissions from electricity generation through 5G supported “smart-grid” applications, increasing the proportion contributed by renewables and improving wider efficiencies in distribution and non-renewables generation.
5. Indirectly improving energy efficiency across all sectors through reducing waste and improving operations. Reduced emissions are largely a by-product of improved productivity and process efficiencies.
6. Reducing carbon emissions from travel through reducing the number of journeys (e.g. remote monitoring and management, virtual meetings) and reducing the emissions per journey.

Areas where 5G could impact global carbon emissions

Source: STL Partners

In this report, we focus on the first two – the management of energy consumption via increasing the carbon performance of the network (expressed as a reduction in the tonnes of CO2 per TB of data transmitted). While we see significant potential upside in “de-carbonising the grid”, in enabling greater energy efficiencies and reducing waste across the economy, these are not in operators’ direct control. They are also more challenging to estimate. We would recommend this form part of a future study.

For nearly all operators, over 90% of the direct energy usage of network operators is accounted for via electricity drawn from the grid to service their own networks. Limiting the growth of this would represent the largest direct reduction in future greenhouse gas production for a mobile network.

Four scenarios modelled

We have modelled our analysis around four 5G scenarios. We treat each of these scenarios differently in our model and scenario assumptions vary by country type.

No 5G roll out

No 5G roll out assumes that there is no roll out of 5G radio access or core technologies. This scenario shows emissions growth well below the growth in data volumes. This is because we anticipate reductions from lower-carbon power generation (e.g. renewables) and significant performance improvements in 4G core and access networking. These are discussed at length in this report.

Slow 5G rollout

A slow roll out of 5G would see the most delayed launch dates for 5G (between 2021 and 2024) and assumes that 5G accounts for the lowest share of data volumes over time (10-25% by 2025 and 60- 80% by 2030).

Medium 5G rollout

A medium roll out of 5G can be considered a base-case. This would see an average launch date of 2019-2022. There would see a significant volume of data running over 5G – up to 60% by 2025 and 85% by 2030. Implicitly, we would expect decommissioning of 2G or 3G (or even 4G) networks with spectrum re-farming to 5G.

Table of Contents

• Preface
• Executive Summary
• Introduction
  • Scope of this report
• Four scenarios modelled
• Our findings
  • Faster 5G roll out could reduce cumulative carbon emissions by 0.5 billion tonnes of CO2 globally by 2030
• How will accelerating 5G roll-out reduce carbon emissions from mobile networks?
  • 5G technologies as drivers of sustainability
• Country level findings: Uneven distribution of carbon savings
• Conclusions and Recommendations
  • Operators
  • Regulators and other national authorities
  • Tower and power suppliers
  • Technology providers
• Methodology
  • Projections
  • Scenarios
  • Country level differences
  • Other Assumptions
• Appendix

Table of Figures

• Figure 1: Faster 5G roll-out would have a material impact on greenhouse emissions
• Figure 2: Areas where 5G could impact global carbon emissions
• Figure 3: Cumulative reduction in emissions under different roll-out scenarios
• Figure 4: Projected CO2 emissions from mobile networks under 4 scenarios
• Figure 5: Where do emissions reductions come from
• Figure 6: Access technologies’ evolving energy performance
• Figure 7: Carbon intensity of different countries used in modelling emissions
• Figure 8: Potential reduction in emissions from fast roll-out of 5G against carbon intensity of grid
• Figure 9: Top 30 Countries by potential reduction in emissions from fast 5G roll-out
• Figure 10: STL’s carbon emissions methodology