On the road

to breaking the
energy curve
A key building block
for a Net Zero future

ericsson.com/
breaking-the-energy-curve
2 Ericsson | On the road to breaking the energy curve

Foreword

Every sector of society must work together

to reach the global climate and energy At Ericsson we have both the
challenge. Net Zero is the North Star of experience and foresight to plan,
climate action and focusing on mobile deploy and operate networks
networks will be key for the industry. that prioritize performance and
Breaking the energy curve of mobile generate energy and cost savings.
networks will have the benefit of reducing
energy use, costs and environmental impact.
Ericsson has set the ambition to be 5G is here, and while the expansion
Net Zero across our value chain by 2040. of this connectivity continues and
The first major milestone in 2030 will be the benefits of an energy conscious,
to reduce supply chain and portfolio in future-proof portfolio are evident, it’s
use emissions by 50% and be Net Zero in simply not enough on its own just to
Fredrik Jejdling our own activities at the same time. dramatically reduce energy consumption
Executive Vice President Since our last report, where we for entire mobile networks.
and Head of Business Area Networks focused on achieving energy savings in We cannot continue with a ‘business
the transition from 4G to 5G networks, as usual’ approach. We must take
we’ve continued to learn and explore. advantage of wider network changes
We drive sustainability in networks and modernization as opposed to
through improved energy efficiency in our piecemeal swap-outs. We must utilize
solutions and integration of renewable the latest technology advances to enable
energy sources, while increasing energy-saving functionalities to the fullest
performance with more innovation and and consider our energy sources.
accelerating the build of nationwide fully To put it simply, we need to
performing 5G. think differently.
As one example of this in practice, we At Ericsson we have both the experience
have the ambition to reduce the energy and foresight to plan, deploy and operate
consumption of a typical new site by networks that prioritize performance
approximately 40% between 2021 and and generate energy and cost savings.
2025. Together with the transition to By taking a holistic view of network
more renewable energy sources in mobile evolution, expansion and operation, we
networks, the resulting CO2 emissions have advanced our approach to break the
per site can potentially be reduced by upward trajectory of energy consumption in
around 70%. mobile networks.

Figure 1: Breaking the energy curve in mobile networks

Traditional roll-out Breaking the energy curve approach

Total network energy consumption

5G
4G

3G

2G

Network roll-out over time

3 Ericsson | On the road to breaking the energy curve

Enabling
Net Zero

The ICT industry plays a decisive role in

enabling the critical and exponential climate
action demanded today by investors and
regulators, and has the potential to reduce
total global industrial emissions by up to
15%. These demands are spurred on by
intensifying pressures from consumers,
customers and the wider supply chain.
To meet these demands and align
with the global 1.5°C ambition set by the
Paris Agreement, businesses must adopt
a full value-chain approach to setting
climate targets according to a Net Zero
timeline – with a commitment to halve
total emissions by 2030 and reach a Net
Zero state by 2050.

In order to reach Net Zero, it

is important to reduce energy
consumption and break the curve.

Net Zero increased circularity, and reduced footprint

Net Zero describes a state where, for of software, as well as developing supply According to Ericsson’s research,
all greenhouse gas emissions emitted, chain targets and partnerships that quantify mobile networks represent about
the same amount is absorbed from the the business impact on other sectors. Energy 0.2% of global carbon emissions
atmosphere. As the North Star of ICT efficiency will not take us all of the way. and 0.6% of global electricity use.

0.2
climate action, the ITU (International According to Ericsson’s research, mobile
Telecommunications Union) Net networks represent about 0.2% of the
Zero standard sets a pathway for ICT global carbon emissions, and about 0.6%
businesses to set Net Zero targets and of global electricity use. The demand on

%
strategies, and provides the sector with mobile networks will continue to grow and
the actions needed to achieve this. without action, energy use and related
By following the ITU Net Zero standard, emissions will too. However, the link to
which both integrates the ICT sector increased energy consumption is not so
decarbonization trajectory developed much one of increased subscriptions and
by the ITU and aligns with Ericsson’s traffic growth, but of the deployment of new With the growing demand for mobile
blueprint to breaking the energy curve, frequency bands and network equipment, networks to handle more data than
ICT businesses can reach a critical 2030 reflected in increasing population coverage ever, there is a unique opportunity to
intermediate target by: with multiple mobile generations. This is a combat emissions. While Ericsson works
• r educing and avoiding emissions, critical area where, together, the industry to consistently improve network energy
for example by reducing energy can make a difference. performance, to reach Net Zero, service
consumption providers also need to transition to
• t ransitioning to 100% renewable Championing renewable energy renewable energy – and the action needs
energy sources In order to reach Net Zero, it is important to start now. Service providers around
•a  bating small, residual fossil fuel to reduce energy consumption and break the globe are telling us how important
emissions through permanent the curve. Where energy consumption is energy and sustainability criteria are to
carbon removal in accordance with responsible for 20–40% of operational purchase decisions. Reducing demand –
the ITU standard expenditure (opex) for communications and subsequent consumption – of energy
service providers, Ericsson’s research encourages renewable energy production,
Reaching Net Zero by 2050 also demands shows that if the entire ICT industry and is an important step in simplifying the
action that goes beyond energy efficiency, switched to renewable energy sources transition to renewable sources.
and covers embodied carbon in product for all of its electricity requirements, the
development, content of recycled materials, carbon footprint could be reduced by 80%.
4 Ericsson | On the road to breaking the energy curve

Our evolving approach to

breaking the energy curve

While from a global perspective, mobile

networks have a low impact on electricity Figure 2: Ericsson’s approach to breaking the energy curve
use and carbon emissions; energy use and
costs, as well as the carbon footprint have Sustainable Operate
rightfully become some of our industry’s network evolution intelligently
largest challenges. This is mainly due to the Embrace a holistic view Leverage AI/ML and
need to expand network capacity to meet for network planning and automation to boost
exponential traffic growth. operation energy savings
Every service provider has their own set
of network evolution ambitions. However,
the industry cannot proceed with the same
approach from a decade ago. Globally,
Ericsson has seen service providers hesitate Breaking the
to activate energy-reducing software energy curve
solutions, whether this is due to perceived
complexity or the potential impact on
classic network performance indicators.
Yet to ‘break the energy curve’, we must
challenge how we plan, deploy and run
mobile networks.
Since the previous edition of this report,
5G has been deployed globally in over Expand
200 live networks. As we move towards and modernize
2025, Ericsson believes it is possible to Modernize existing network
scale up 5G, while simultaneously reducing while scaling up 5G
total network energy consumption. We
have streamlined our approach into three
core elements:
Expanding sites with new 5G solutions This can be achieved through a portfolio
Sustainable network evolution often includes deploying new frequency of tools and capabilities including the
Many service providers are entering a bands, which requires adding more latest artificial intelligence and machine
scaling phase as they already run a 5G equipment. To avoid increasing energy learning (AI/ML) – as well as automation
network, and our insights can help while consumption, the installed equipment solutions which can help to reduce
we consistently work to improve the energy needs to be modernized. This is key perceived complexity barriers. Ericsson
performance of our products. to changing the energy consumption is well equipped to support service
Network planning and operations need trajectory in mobile networks. With providers in keeping energy use and
to evolve to cater for achieving business a combined view of investments and carbon emissions at a minimum, with
targets and sustainability ambitions with operating costs, our latest generations of zero-touch problem solving and
the lowest energy consumption possible. multi-band and massive multiple-input predictive energy management.
By applying a work process that prioritizes and multiple output (MIMO) radios, and
a holistic view of all perspectives, including basebands can help service providers
an evaluation of organizational objectives to cut energy consumption and greatly Ericsson recommends that mobile
and network realities, the resulting network improve their energy costs and total cost of networks are planned, built and
evolution plan becomes the foundation to ownership going forward. operated with precision. By working
achieve the desired outcomes. through all three steps of our
Operate intelligently approach, it is possible to break the
Expand and modernize As traffic varies daily, the use of energy curve in mobile networks.
Executing a sustainable network evolution energy-saving applications is fundamental
plan with a stepwise 5G expansion to adjusting the capacity of mobile
provides an opportunity to modernize networks to match demand and to
existing networks with low energy delivering the best user experience
consumption and carbon emissions. with the lowest energy use.
5 Ericsson | On the road to breaking the energy curve

Sustainable network
evolution

Embrace a holistic view of company targets and network

realities to allow for network planning and operation-supporting
business and sustainability ambitions.

To run a sustainable network while

meeting business and operational targets, Figure 3: A holistic approach to meeting business and sustainability targets
taking a holistic view to network evolution
planning is non-negotiable.
Network KPIs User experience
The reality is often that organizations
within a company have different steering
objectives, so a working process that
prioritizes a holistic view of the solutions is
needed. Traditional areas where conflicts
may arise for a service provider are targets
for network KPIs (blue in Figure 3), user Holistic Circular
Net Zero
experience (purple), capital investments view economy
and operational expenditures (yellow)
and energy cost (green).
In most cases, traditional network
KPIs are measured and used as a reward
system for operational departments.
The nature of a good radio network Capex/opex Energy cost
today, based on data delivery, is very
different from networks supporting circuit
switch voice on 2G and 3G networks.
By expanding the focus on network With more service providers comitting to evolution must also consider the utilization
performance to include user experience Net Zero targets as well as focusing on of renewable energy sources such as wind
and its correlation towards energy managing energy costs, the complexity and solar, as well as high-capacity batteries
consumption, we see new opportunities increases. Often, this is combined with for energy storage.
to optimize RAN energy efficiency. increasing circular economy aspirations
Ericsson provides consultative such as prolonged hardware lifetimes. Prioritizing RAN energy efficiency
services to assist in the deployment So, proof points and plans that align The nature of RAN is to provide nationwide
and optimization of energy efficiency with pre-established company targets coverage and service capacity. As the
functionality, while observing savings and are required. coverage and capacity offered by a
balancing potential impact on network Environmental sustainability is also single radio is limited, this is achieved by
KPIs and user experience. a factor for consumers, in addition to deploying thousands of radio sites, which
the traditional “telecom values” such as can often be a factor of 10,000 or more
Holistic solutions for network evolution performance or coverage. when compared to the number of core
Networks have many indicators which can To embrace sustainability targets, network nodes. Therefore, the RAN
observe small deviations in characteristics the traditional areas above must also with its active and passive equipment
that are not even noticeable from an end encompass a new holistic view to reach represents more than 75% of the service
user perspective. If the network engineers the most efficient solution from an overall provider’s network energy consumption.
are only measured on classical network company perspective. This requires that Service providers will continue to prioritize
performance indicators, there will be an the organizational objectives are evaluated RAN energy efficiency as it is the only
inbuilt resistance to activate functionality together to see the full potential from both way to keep energy consumption under
that could reduce energy consumption. business and environmental viewpoints. control, while delivering an excellent
Service providers need a method for making We recommend that network planning user experience.
decisions about running networks in the encompasses all aspects of core, transport Energy metering provides observations
optimal manner, while guiding future and radio access equipment, as well as site of the actual energy consumption inside all
investments that could reduce total cost equipment (including power systems and RAN hardware on a node level and creates
ownership (TCO) over time. energy sources). A sustainable network a good baseline for future actions.
6 Ericsson | On the road to breaking the energy curve

Figure 4: Network realities

Traffic segment (% total data volume) User experience

25% high 50% mid 25% low Poor Fair Good Great

Poor Fair Good Great

5Mbps 10Mbps 20Mbps

4.0s 2.5s 1.5s

Throughput ’at click’

Time-to-content

Service providers can optimize site

build and energy-saving policies
based on a site’s traffic reality in a
given network.

We have long been using energy Any network consisting of many physical could be called “time to content” since it
performance to measure energy efficiency. sites follows a similar distribution of represents the time taken for the user to
By exploring the relation between “wanted traffic load per site. Ericsson recommends reach their desired content. Ericsson has
performance” and “energy consumption”, categorizing sites (or sectors) in three developed specific observation counters
we can observe deviations on cluster or traffic segments: for ease of data extraction. When a site
site level. When different activities are • the red traffic segment contributes reports DL speeds below predefined
executed in the cluster, we can observe how to 25% of total traffic volume in the thresholds for downloads of the most
energy performance is changing over time. sites with the highest load of the common applications, we grade the site
observed network (or sector) to determine improvement
Site traffic volume and user experience • the blue traffic segment includes all areas. This offers an overall view of
Each site has a special role in a network sites that provide the next 50% of total current network status, and helps guide
cluster. To provide a good understanding traffic volume (red and blue together the operational department to reducing
of the network and how each site relates identify sites that bring 75% of energy consumption.
to the overall network performance, the network data volume) Via the traffic split (Figure 4,
recommendation is to observe the network • the remaining 25% of the network left picture) we can find sites with
status during the most loaded periods to data volume belongs to the green significantly lower traffic volumes
ensure that potential capacity bottlenecks traffic segment with low traffic load (green sites). Typically, 50-70% of the
are identified. If no internal process is in sites belong to the green segment. If
place to observe network reality, Ericsson Network realities explored these sites are correlated with user
recommends to apply the following two Traffic load varies across nodes in the experience observations (Figure 4, right
types of observations: traffic volume network and over time. Energy savings are picture), we can see sites that have great
per site, and user experience per site. achieved by proper dimensioning of radios performance in busy hours, indicating
Together they can assist service providers and deployed frequency bands for example, that the site is over-deployed when it
in finding valuable insights. By using as well as using software to scale power comes to activated frequency bands and
an easy-to-understand view, service consumption with traffic. An important deployed radio units. This is one example
providers can define strategies for network result from this study is to identify “the most of how to identify sites where extended
deployment and operation, such as to valuable sites” (Figure 4 – left, red and blue energy savings settings can be applied,
recommend how to apply energy-saving traffic segments), and ensure these operate such as putting parts of the hardware in
software functionality with better precision. with efficient network equipment. sleep mode or turning off radios.
The service providers also needs to With a sustainable network evolution
understand how the network performance plan in place, the network observations
is experienced by the end users. The most can be used to ensure traffic performance
important aspect from a user perspective is and define different operational strategies
to have a reliable minimum performance. for energy savings across the network
Ericsson’s studies show that a good clusters and sites.
proxy for user experience is to observe In summary, upgrading the network
downlink (DL) speed when the network evolution plan is the foundation of the
is mostly loaded. The observations approach to breaking the energy curve.
 7 Ericsson | On the road to breaking the energy curve

Expand and
modernize

Effective modernization of existing networks is

essential when scaling 5G to reduce the mobile
network total energy consumption.

To reduce energy consumption when include all costs from their company, one multi-standard baseband 6631.
scaling up 5G, modernizing the existing including opex from energy cost and site The modernized site will have 60%
equipment is vital. Multi-band technology rental, and pair this with the modernization higher capacity with 50% lower energy
allows the functionality of several radio of the existing infrastructure on site. consumption, half the radio weight and
units to be combined into a single physical one-third of the of the size, while adding
unit for existing frequency bands. The Adding more spectrum through an additional FDD frequency band. It will
energy efficiency, size and weight can be modernization also support 5G on existing frequency
improved by integrating more components Modernization can both help to add bands while 3G is removed.
into the same radio. both spectrum and capacity, as well The transition from 4G to 5G involves
With the new generation of multi-band as reducing energy consumption. a significant increase in RAN equipment
radios, we can add frequency bands while Advancements in Ericsson Silicon allow processing requirements. Energy efficiency
reducing the amount of radio units and Ericsson to combine multiple frequency is top of our agenda and – together
energy consumption. Since site lease bands into a single remote radio unit. with increased capacity demands – it
contracts are often stipulated by site In this example, the Radio 6626 guides development of new components,
footprint (number of units, volume and enables one radio unit to replace products and features. For example, the
weight), a modernization to more compact six single-band radios. purpose-built Ericsson Silicon processing
and energy-efficient site builds can reduce The legacy site to the left in Figure 5, hardware is designed to meet these
TCO significantly. However, the current has nine radios using three frequency performance demands. It plays a key
challenge is that it is not always the same bands. Combined with the three role in creating high-performing and
service providers’ budgets that control the basebands, the site supports 2G, 3G lightweight products and has increased
network TCO. Therefore, we recommend and 4G. This configuration is replaced energy efficiency by a factor of 10 from
service providers improve the ability to by two Radio 6626 in combination with 2016 to 2022.

Figure 5: Site expansion and modernization – adding FDD spectrum

From legacy sites To next generation sites

Single-band 2T2R FDD radios Multi-band Radio 6626 FDD

Capacity is increased by 60%

3x3x 1x

Add FDD spectrum 3x2T2R 3x2T2R

B8B20 B1B3

2G DUG 2G
3G DUW Energy use is reduced by 50% 4G
4G DUS 5G Baseband 6631

Capacity x1.6
Capacity 1 Energy consumption 50%
Energy consumption 100% (kWh)
(kWh) Spectrum [MHz] 800, 900, 1800,
Spectrum [MHz] 900, 1800, 2100 2100
8 Ericsson | On the road to breaking the energy curve

Figure 6: Site expansion and modernization – adding 5G mid-band and FDD spectrum

From legacy sites To next generation sites

Single-band 2T2R FDD radios Multi-band Radio 6646 FDD
Multi-band Radio 6626 FDD
Massive MIMO AIR 3268

Capacity increases more than 10 times

3x 1x 3x
Add TDD mid-band

Add FDD mid-band 4G

2G DUG
3G DUW 5G
4G DUS RP 6651

Energy use is reduced by 37%

Capacity 1 Capacity >10x

Energy consumption 100% Energy consumption 63%
(kWh) (kWh)
Output power 400W Output power 680W
per sector per sector
Spectrum [MHz] 800, 900, 1800, Spectrum [MHz] 700, 800, 900,
2100 1800, 2100, 3500

During recent years, Ericsson has taken Choosing the right segment while multi-band and multi-sector radios,
significant steps to improving the energy scaling 5G mid-band will ensure the and three Massive-MIMO radios. In this
efficiency of Massive MIMO radios by up best user experience with optimal example the radio 6626 is a 2-band,
to 50% and our multi-band radios by up to energy consumption. 3-sector radio that replace 6 single band
20% compared to the earlier generations. It is also important to use software radios to provide 4G capacity. Radio
At the same time, our basebands are capabilities to avoid adding more 6646 is a 3-band, 3-sector radio which
30–60% more energy efficient than new hardware: will provide a coverage layer for 4G while
competitors when delivering the same • Carrier Aggregation expands mid-band adding one new FDD 5G band. Finally, the
capacity. Together, this sets us on a good coverage with the help of low band. AIR 3268, an ultra-lightweight Massive
path towards significantly reducing the By moving data to a more energy-efficient MIMO radio, is for the new 5G TDD
power consumed per delivered GB. 5G band where service providers often frequency band. The modernized site will
The new network performance have a 100 MHz carrier, the energy per have 5 radios and more than 10 times
demands require increased spectral transferred bit can be reduced by higher capacity with about one-third lower
efficiency and flexibility. Massive a factor of 10 while improving energy consumption. This forward-looking
MIMO technology allows service user experience. site configuration will significantly boost
providers to deliver a powerful 5G • Innovative solutions, like both 5G coverage and mid-band (3.5GHz)
experience nationwide. Ericsson Spectrum Sharing, offer performance in an energy-efficient way
The Ericsson Massive MIMO radio service providers a way to deploy 5G and scale up standalone deployment.
solutions are divided into three segments: on existing 4G band without sacrificing
• Capacity covers the most performance, by reusing pre-deployed An efficient mobile network requires fewer
capacity-demanding sites and hardware – enabling nationwide 5G sites and less energy. Automated, optimized
provides superior performance in all coverage from day one. networks across fewer sites lower the
deployments. environmental impact and cost less to run.
• Coverage targets deployments with Adding mid-band Massive MIMO Commissioning, maintaining
large inter-site distances, such as and modernizing FDD band and upgrading sites traditionally
suburban or rural areas with a smaller Adding 5G Massive MIMO on required engineers to visit the site
spread of users in the vertical domain. mid-band does not need to increase to work on it, with the obvious carbon
• Compact covers sites where there the energy consumption. footprint impact of transportation.
are constraints on the deployment. The legacy site to the left of Figure 6 Increased automation for virtual drive
It prioritizes TCO and ensures that has 12 radios and supports 4 frequency tests, monitoring, root-cause analysis,
mechanical properties are in line with bands with 2G, 3G, and 4G. This and remote upgrades significantly reduces
site constraints, such as size and weight. configuration is upgraded to two the need to spend time on site.
9 Ericsson | On the road to breaking the energy curve

Operate
intelligently

Leverage AI/ML and automation for maximized

traffic performance of deployed hardware with
minimized energy use.

With mobile networks being deployed to example, might complement grid electricity.
meet expected peak traffic demands for The energy behavior of these infrastructures Customer case
the next three to five years, more capacity would change based on load and
than needed will be provided for most traffic trends. Ericsson and Indosat:
hours of the day. Variation in traffic load is To manage mobile networks Building a greener
inherent in mobile networks, such as day efficiently, we must capitalize on the fact
and night differences or short variations that different regions, clusters and sites
network
down to the millisecond level. Our need unique energy-saving protocols. This Challenge
energy-reducing software solutions make is where data, AI and ML will streamline • Indosat wanted to reduce emissions
use of these load variations and allows different automation schemes to transform while increasing data traffic
the power consumption to vary up to 97% network operations. • To achieve a more sustainable
between full traffic and no traffic (Figure 7). We believe there are three important network by aggressive LTE spectrum
Energy-saving functions and pillars that intelligent operations towards migration and Ericsson Radio
features are available for all radio access sustainable networks stand on today: System modernization
technologies and enable large savings in • maximizing the power of data to
networks today. boost automation Solution
Regarding functions that act on • working holistically on energy-saving • Energy-saving features decreased
millisecond range, our recommendation is actions whilst keeping the user overall consumption
to always have them activated. Features experience at optimal levels • Hardware, software and related
deactivating larger shares of components • establishing platforms focusing services were upgraded to build a
or equipment, such as antenna branches on sustainable operations through modular RAN
or cells, enable the largest savings when predictiveness, automation and • Ericsson Energy Infrastructure
optimized on an individual cell level. orchestration Operations (EIO) was used to manage
However, we also observe a hesitation to all energy-related assets efficiently


deploy these solutions in networks today
due to the potential impact on network Figure 7: Portfolio-wide innovations Impact
performance indicators or perceived • The solution achieved the highest
complexity to manage them effectively. Full traffic spectral efficiency in all of Jakarta and
To take full advantage of energy-saving 100% power produced more traffic on fewer sites
functionalities, we must challenge the way consumption than competitors
we operate mobile networks. To achieve • There were energy savings of
the wanted savings with optimized user Up to 97% around 20-30% compared to
experience, careful management of less power the old equipment
network performance and energy saving consumption • Energy bills were reduced additionally
actions is required. at low traffic by up to 3.6% and CO2 emissions by
A key issue in solving the energy Typical traffic 4% on the sites where the EIO solution
efficiency puzzle in mobile networks is 30-40% power is deployed
to have an end-to-end view of the entire consumption
network and its elements, including both
passive and active site infrastructure. • Micro sleep Tx
Passive infrastructure represents the • Deep Sleep
support ecosystems which keep radio sites • Traffic-aware
up and running, including batteries, power network services
supply units and climate control units. operations
There is also the complexity of different Zero traffic hours • Predictive
power sources being connected into the 3-10% power Energy
passive infrastructure, where utilization consumption Management
of diesel generators or solar panels, for
10 Ericsson | On the road to breaking the energy curve

1. Maximizing the power of data to 2. Working holistically on As an example, the Ericsson Predictive
boost automation energy-saving actions whilst keeping Cell Energy Management solution reduces
Automation is a critical part of the digital the user experience at optimal levels energy spend at site level based on an
network transformation. The active The introduction of AI and automation AI/ML and automation architecture. The
network elements are already able to into the management of mobile networks predictive engine identifies which RAN
report power consumption alongside other will harness the power of data generated cells that can have energy-saving actions
performance data over the management across the network infrastructure and applied, such as cell locking, deep sleep
interface. However, the energy enable autonomous optimization. activation or RAN feature configuration,
optimization of passive infrastructure is Network orchestration towards energy without impacting network quality and
often overlooked, as energy spent on this efficiency based on changing traffic and user experience. Multiple criteria sets and
equipment is hard to measure and control. usage profiles while providing optimal energy-saving thresholds can be deployed
To maximize the power of automation, user experience will be the critical for daytime vs. nighttime and different
we must capture data and digitalize balance; intelligent operations will network clusters.
the complete site ecosystem enabling be the differentiator. The solution complements existing RAN
intelligent measurement and control. features with customized energy-saving
Deployment of site controllers, smart Predictive RAN Energy Optimization actions, reducing an additional 2-8% on
sensors and enclosures on passive Different criteria sets and business logic network level. Applied in an Asian Tier-1
infrastructure will be key, enabling remote can be implemented in radio units and network with over 350 million subscribers,
measurement and control of power cells, based on traffic and utilization the savings potential amounted to
systems, cooling systems and renewable predictions, combined with user experience USD 15 million per year and 100 kilotonnes
energy sources. The combined data observations. Orchestrating energy-saving of CO2 emissions across the full network.
will enable correlations to be identified actions must involve considering different Another example is the Ericsson
between site load and traffic trends traffic profiles with geographical realities, AI-Powered MIMO sleep mode feature.
versus the energy behavior of passive which are deployed on a spread of radio It adds an additional ~5% of energy
infrastructure elements, where holistic access technologies and frequency bands. savings on on top of manual settings for
energy-saving actions can be defined and This is essential in providing coverage and the same feature. However, the biggest
deployed per site. capacity in an energy-efficient manner. value comes from an increased acceptance

Figure 8: AI/ML solutions and their applications to the network

Predictive RAN Energy Optimization Predictive Infrastructure Energy Optimization

AI/ML for RAN performance with energy optimization AI/ML to operate each site’s infrastructure

Site Energy Profiling Power Source Optimization

AI/ML to enable anomoly detection and AI/ML for site power optimization and
predictive operations ROI improvement
11 Ericsson | On the road to breaking the energy curve

and use of the MIMO sleep mode feature

through the simplified activation since Figure 9: Control loops for RAN automation
many operational departments wouldn’t
otherwise prioritize the manual work
needed to fine tune the parameter settings. Weeks

rApps on EIAP
Predictive Infrastructure
Energy Optimization
Days
Having digitalized the site infrastructure,
we can utilize the same principles as
in RAN optimization to manage many
critical elements in the site in a more Seconds
efficient manner. We can run climate

Functions in RAN
control units on the site with AI/ML that
dynamically adjust cooling settings based Milliseconds
on temperature trends and heat-resisting
factors. We can turn off power supply units
based on traffic load and utilization trends
synergized with RAN energy optimization Microseconds
actions. Similar applications can optimize
diesel generator run hours in relation to
other available energy sources.
Through our SMO, the Ericsson Intelligent
Site Energy Profiling Automation Platform (EIAP), we have Customer case
With the gathered end-to-end network extended the rApp - RAN automation
data, it is possible to create correlations applications - concept to also automate Deutsche Telekom (DT)
based on traffic load and infrastructure existing, purpose-built 4G and 5G RAN and Ericsson driving
performance that can be used to networks, in addition to Open RAN and
benchmark the energy efficiency Cloud RAN. towards sustainable 5G
of similar sites and predict anomalies. We are accelerating our development
The Ericsson Node/Radio Power of rApps that increase operational Challenge
Efficiency Map solution measures energy efficiency and network performance in Identify and validate both cost and
efficiency per site and analyzes each the areas of Network Evolution, Network energy-efficient solutions based on
site against dimensioning, hardware, Deployment, Network Optimization and consumption and increased usage of
software and radio frequency for a detailed Network Healing. These rApps enable renewable energy sources
understanding of any inefficiency’s root cognitive, AI/ML-based, powerful
cause. Our Energy Profiling uses a predictive automation applications with large Solution
model to identify anomalies based on trend energy and carbon emission reductions • Addition of a wind turbine to the
analysis of AC and DC energy meter data potential. Our solutions cover fast and Dittenheim site, which was already
coming from site controllers. slow control loop scenarios with use part-powered by solar panel energy
cases that can be delivered as a radio • Integration of additional energy
Power Source Optimization feature, a network service, rApp or sources such as fuel cells will soon
ML algorithms help increase utilization combination of them. The objective is to replace the need for emergency
of high-capacity lithium-ion batteries increase efficiency and performance by diesel generators
or renewable energy sources instead executing AI in both control loops, where
of relying on grid or fuel-based we can have centralized automation Impact
energy sources. capabilities to enable complex decisions • The turbine can provide up to five
For example, the Ericsson Infrastructure with rApps through predictive models, kilowatts of additional power, as a
Operations Power Source Optimization complemented with distributed second renewable energy
solution uses AI/ML to predict the best automation functions at RAN level power source
available energy source on the site based where implementations can be done in • Initial tests showed that on windy
on cost, load and energy source availability real time. (See Figure 9). days, more renewable energy
trends including site battery autonomy As an increasing amount of data could be generated than was
detection. This reduces grid usage during and measurements come from consumed by site operations
expensive tariff zones, diesel generator run telecom networks, AI and automation
hours, and improves return on investment expand the potential for high-impact,
for renewable sources. energy-saving and carbon emissions
reduction opportunities. The ability to
3. Establishing platforms focusing analyze vast amounts of data relating to
on sustainable operations through traffic patterns, real-time demand, and
predictiveness, automation and network-resource availability allows for
orchestration quick, automated decisions to be taken
Going forward, usage of common Service to enable a number of use cases to create
Management and Orchestration (SMO) needed efficiencies. Continuous innovation
platforms with sets of powerful on applications and business logics
sustainability automation applications using AI and automation will enable the
will emerge. sustainable networks of the future.
About Ericsson Ericsson enables communications service providers to
capture the full value of connectivity. The company’s
portfolio spans the following business areas: Networks,
Cloud Software and Services, Enterprise Wireless Solutions,
Global Communications Platform, and Technologies and
New Businesses. It is designed to help our customers go
digital, increase efficiency and find new revenue streams.
Ericsson’s innovation investments have delivered the
benefits of mobility and mobile broadband to billions
of people globally. Ericsson stock is listed on Nasdaq
Stockholm and on Nasdaq New York.
www.ericsson.com

Ericsson The content of this document is subject GFTL-22:001268 Uen

SE-164 80 Stockholm, Sweden to revision without notice due to © Ericsson 2022
Telephone +46 10 719 0000 continued progress in methodology,
www.ericsson.com design and manufacturing. Ericsson
shall have no liability for any error or
damage of any kind resulting from the
use of this document