Can 5G be sustainable?
Author: Raffaele Sabatino
Introduction
5G is designed to enable enhanced mobile broadband, massive machine-type communications, and ultra-reliable, low-latency communications. This will create significant benefits for use cases in transportation, manufacturing, farming, energy, buildings, entertainment & media, health, and the public sector. Increasing requirements placed on mobile networks in terms of number and type of connected devices, data volumes and supported applications, shall be met by 5G networks.
4G mobile networks do not provide the capabilities to support all of these use cases. The reasons differ from use case to use case and can be related to capacity, latency, or availability requirements. Although some use cases can potentially be realized with communication technologies other than 5G (WiFi-based automated driving, LoRaWAN for precision farming, etc.), there is an advantage in terms of lower costs and less GHG (greenhouse gas) emissions to meeting the requirements of many use cases with just one mobile network technology and infrastructure, as is the case of 5G networks, as opposed to “mixed” technologies. Therefore, assessing the sustainability of 5G networks in general and the GHG abatement potential of this new technology is crucial.
On one hand, many 5G-supported use cases in areas like transportation, manufacturing, farming, energy, buildings, entertainment & media, health and the public sector, and more show a promising potential to reduce GHG emissions through ?flexible work, smart grid, automated driving, precision farming.
On the other hand, the EU civil society is concerned and asks for clarification regarding the societal and environmental impact of the 5G technology ecosystem, focusing in particular on improving the legislative framework, creating new safety standards based on cumulative exposure, improving the local planning procedures by imposing new, specific limits on electromagnetic field (EMF) emissions, and the need to educate populations about the overuse of technology and the electromagnetic field pollution generated by both antennas and any portable device connected to the communications network.
Sustainability is also becoming a business approach. Built on the assumption that developing such strategies fosters company longevity, sustainabilitiy aims to create long-term value by taking into consideration how a given organization operates in the ecological, social, and economic environments.
Network evolution
Since the first commercial mobile network service was launched, four major generations of mobile networks have been rolled out globally, each providing different capabilities for voice or data transmission and related services: 1G, 2G, 3G, 4G, now 5G. The 3rd Generation Partnership Project (3GPP), the industry consortium setting standards for 5G, defines 5G as any system using 5G NR (5G New Radio) software. This definition has come nowadays into general use.
From a merely technical perspective, the data rate and latency (i.e., the delay between sending and receiving) are the key performance indicators of such networks.
In addition, thanks to the boom of IoT (Internet of Things), the number of devices requiring mobile internet connectivity goes far beyond the number of subscribers to mobile networks. “Things” able to transmit and receive data at any time, like household appliances, wearables, car sensors, buildings, plants, robots, and so on, are estimated to reach 75 billion by the end of 2025, against a forecasted 18 billion personal mobile devices worldwide in the same timeframe (Statista). Among them, 1.7 billion 5G subscribers are expected (GSM Association).
Given the increasing use of data-intensive applications such as video streaming, the amount of data transmitted via mobile networks by the end of 2022 is expected to increase to 77.5 exabytes per month globally, a factor of 7 above the level of 2017 (Statista).
In other words, mobile communication has evolved from a voice-only service to a complex, interconnected multi-service environment built on a system that supports a multitude of applications and provides high-speed access to a massive number of subscribers and devices.
While mobile network performance, in terms of data rate and latency, will dramatically increase with 5G, other aspects should be carefully considered, like energy consumption, security and reliability. Because IoT sensors are to be powered by low-capacity batteries, the energy required to transmit must be minimized. Additional stringent requirements derive from the fact that mobile networks will be playing an increasing role in (safety-)critical infrastructures as well as in connecting vehicles or controlling electricity grids. A schematic comparison between mobile network generations is shown in the following table.
As we can see, network performance has increased with each generation. With 1G and 2G, mobile networks could just offer phone calls and transmit none or limited amounts of data. It was only with 3G and 4G that mobile networks could provide Internet access as we experience it today, and the first smartphones came onto the market (Blackberry in 2002, iPhone in 2007), paving the way to new applications like web surfing, audio and video streaming. The extremely high data throughput and very low latency required by 5G cannot be attained only with evolution or modification of the existing 4G network, as such features entail radical changes on base stations and network (core, backhaul) level, comprising several technologies and new approaches like millimetre-wave (mmWave) band, dense deployment of smallcells, D2D (Device-to-Device), M2M (Machine-to-Machine), and massive MIMO (Multiple Input, Multiple Output) with beamforming (s. next figures).
Additionally: new waveform, advance Coordinated Multi Point (CoMP), carrier aggregation, multiple radio access technology (m-RAT), efficient coding techniques, network virtualization, cloud radio access network (c-RAN). These radical changes cross the areas of capacity and speed, latency, spectral efficiency, and massive connectivity-IoT.
Besides high data rates, low latency is a key factor for many 5G scenarios, enabling energy savings and long battery life time at the same time. Pre-5G latency is about 15 ms, which is good for 4G applications, but 5G introduces technologies (tactile internet, two-way real-time gaming, cloud-based applications, augmented reality, etc.) that require latency below 1ms. See some examples in next figure.
This has a major impact on design choices in many areas. Next figure shows how main challenges can be tackled through specific technologies in the 5G ecosystem.
Effects, opportunities, risks
The environmental impact of all the changes due to 5G can be analyzed in terms of GWP (Global Warming Potential), quantified by GHG (Greenhouse Gas) emissions and, more generally, considering the effects on human health, well-being and natural ecosystems. The first goal can be reached in a relatively smooth way, as GHG impact is somewhat easier to measure. For the second impact type, however, it is still difficult to make a final judgment due to the lack of clear, unequivocal scientific indication.
It is convenient to classify the ecological impact of any new technology into direct and indirect effects. These effects have been traditionally analyzed only in the ICT sector, in the industry as well as in the academic world, for a couple of decades. However, the approach for 5G mobile networks can be set in a similar way. In fact, 5G mobile networks have direct effects on GHG emissions as well, because network equipment need to be produced, installed, powered with electricity, and disposed of at the end of service life; and indirect effects, because 5G paves the way to new applications which can potentially be more or less sustainable.
While some of the studies carried out in the industry point out that GHG saving potential of some ICT use cases (indirect effects) is much higher than the GHG footprint of the ICT sector (direct effects), various studies conducted on some other ICT use cases by academic institutions highlight the unfavorable effects on climate protection.
Direct and indirect GHG effects of 5G applications
Direct GHG effects of 5G refer to the emissions caused by producing (production), operating (use), and disposing of the hardware necessary to implement 5G network infrastructures. These emissions constitute the “GHG footprint” of 5G and are, by definition, unfavorable for climate protection. The best method to assess the direct effects of 5G is the ISO-Standard Life Cycle Assessments (LCA) of 5G infrastructures and devices. LCA makes it possible to evaluate the CO2e (CO2 equivalent) impact throughout the lifespan of a product, from the manufacturing to the energy required to power it, and ultimately considers the waste created when the device reaches the end of life and must be disposed. Access, core and transport networks are the infrastructures to be considered, as each of these sectors has its own production and operational peculiarities.
Indirect GHG effects of 5G refer to the consequences caused by the introduction of 5G use cases on other processes and their GHG emissions. These effects can be favorable or unfavorable for climate protection. Categories to be considered are substitution, optimization, induction, and rebound effects. Substitution (e.g., videoconferencing substitutes air travel) and optimization (e.g., a smart thermostat optimizes heating) are positive effects and can lead to a decrease in GHG emissions. Induction effects occur if the use of that technology stimulates additional consumption of goods or services. Rebound effects occur when the improvements in energy efficiency induced by the use case lead to price reductions of a service and boost the same service, or other. For instance, 5G applications in transport might lower the cost of mobility, increase the demand for transport-related activities and causing more energy consumption. Only a systematic, use-case-specific exploration of GHG abatement potentials and mitigation of negative effects can make 5G (included digitalization) neutral or even positive for climate protection. However, given that the potential applications of 5G mobile networks in economic and social systems are practically endless, it is almost impossible to consider all indirect GHG effects of 5G. It is much more reasonable to limit the analysis to some specific 5G use cases and estimate the indirect effects on GHG emissions for them. The best approach is comparative: the use case realized without 5G is assumed as baseline, and the comparison focuses on the same scenario but implemented with 5G.
Assumptions for such an analysis about direct and indirect 5G effects are, of course, specific to the country and to the network infrastructure considered. Thus, many factors are involved, that can greatly vary from case to case: energy production, traffic model, operator-specific forecast, demographics, consumption forecasts, a reference year, which technologies will still exist in that year, etc.
Being that 5G is a relatively recent technology, not many of such studies assessing the sustainability of 5G are available, and most of them are quite limited. One of the most complete researches on the matter has been commissioned by the leading Swiss mobile operator Swisscom and refers to direct and indirect GHG effects of 5G implementation in Switzerland by year 2030. Direct effects considered in this analysis, as expected, take into account the impact of 5G according to the implementation plans of Swisscom until 2030. Indirect effects are limited to four selected 5G-enabled use cases and include the footprint of “non-5G” equipment, i.e., systems that are necessary to realize 5G scenarios, although not specifically defined by the 5G standards body.
Direct GHG effects: The ?research mentioned above predicts direct GHG effects of 5G, measured in kg CO2e/Mt CO2e, both per GB of transmitted data and per year. 5G direct effects are compared to those of 2G, 3G, and 4G networks between 2020 and 2030 (where applicable[4]).
The GWP of 5G networks in 2030 per GB of transmitted data is expected to be 44% less than 4G mobile networks in same year and 85% less than than 2-4G mobile networks in 2020. The GWP of 5G networks in 2030 per year of network operation shall be 0.018 Mt CO2e. This is only just 3% of the overall GHG footprint of the ICT sector in Switzerland as projected for 2025 in an earlier study, and the forecast does not substantially change even if 5G networks of other Swiss network operators are included. This means that measures to keep the GHG-footprint of 5G small should not be limited to 5G infrastructures, but rather target the whole ICT sector. Considering the amount of GHG emissions caused by each network technology per year, 5G in 2030 should cause 11% more GHG emissions than 2-4G in 2020 due to the expected mobile data traffic increase between 2020 and 2030. However, comparing an hypothetical 2030 “4G only” scenario, where data volume is totally allocated to 4G, to a “5G only” scenario, where data volume is handled only by 5G, “5G only” would cause roughly 15% less GHG emissions than“4G only”.
The influence of the chosen electricity mix has been investigated as well. Results show that the electricity mix influences the LCA results. As intuitive as it might be, the way energy is produced can determine the GWP of any technology whose infrastructures must be produced and maintained by using that energy, and 5G makes no exception.
This introduces big differencies. For instance, comparing the USA and Switzerland, much of the US electricity production still depends on fossils like coal and natural gas, while Switzerland makes a relatively limited use of these energy sources. All of these different background conditions can drastically change the environmental impact of 5G from one country to another.
Sticking to the Swiss case, some light was shed on the roles played by production and operation in direct GHG effects thanks to a contribution analysis about the four possible networks: 2-4G, 5G in 2020 and 4G, 5G in 2030. This shows that in 2030, for both 5G and 4G, the split of the infrastructure is 57% production to 43% operational. In 2020, however, for networks 2-5G , the ratio is 40% to 60%. Thus, the energy required to operate the network in 2020 is expected to cause more GHG than in 2030, where the influence of production on GHG seems to prevail.
Indirect GHG effects: Among many 5G use cases that are well documented in literature, the following four are the most promising.
Next table provides a summary of the main 5G network capabilities essential to the selected use cases.
领英推荐
A combination of several capabilities within one wireless network standard, rather than single features, provides the main benefit for each use case. Existing mobile networks do not provide the capabilities to support all of these use cases and the reasons vary from case to case; capacity, latency, or availability can be critical. This does not rule out that some use cases could be realized with other technologies than 5G, like WiFi-based AD (Automated Driving), or LoRaWAN for precision farming. Ongoing debates whether AD shall be realized with 5G mobile network or with WiFi technologies seem to confirm this. Nevertheless, a central advantage of 5G over other technologies is that 5G combines the advantages of various technologies (2G-4G, LPWAN, WiFi, fixed networks) and thus can meet the diverse requirements of many use cases.
Furthermore, should each use case be realized with alternative technologies, the number of communication networks operated and managed in parallel would increase, resulting in more administrative complexity, higher costs, and worse environmental impact. To avoid this, it is essential to establish a cross-sectoral dialogue aiming to realize a communication infrastructure which meets the requirements while minimizing the amount of network hardware and electricity which would be necessary to build and operate several networks in parallel. This process should also foresee a coordinated phase-out of previous network generations once they are obsoleted by 5G.
The Swiss research has determined the indirect effects of the four use cases mentioned above for the year 2030. Following factors have been considered: GHG abatement levers (for instance, increase of fuel efficiency of road transport vehicles through AD); baseline emissions, i.e., the prospective emissions caused in 2030 with current patterns before the use case was realized (for instance,?expected GHG emissions due to transport with no significant adoption of AD); level of adoption of the use case in 2030, i.e., the share of the population that will adopt the use case (for instance, share of AD vehicles in 2030); impact on GHG emissions per unit of adoption of the use case (for instance, fuel savings due to AD); expected rebound effect (for instance, increase in vehicle-miles travelled due to cost reductions of car transport).
Three scenarios have been considered: pessimistic (use case adoption rate taken from lower data range), expected (use case adoption rate taken from average data range), optimistic (use case adoption rate taken from higher data range). An abatement potential of the GHG emissions between 0.1 and 2.1Mt CO2e (Switzerland, 2030) for these use cases could be established, according to the three scenarios, with best potentials for flexible work, smart grids and precision farming (s. next figure).
Being the expected adoption of the four use cases by 2030 still relatively low, specifically for AD, the theoretical potential (maximum adoption, no rebound) of the use cases to contribute to climate protection is even higher.
Indirect GHG effects of non-5G equipment: The use cases rely on additional ICT equipment besides 5G infrastructure. For instance, smart grids require smart meters, precision farming requires specific sensors, etc. Therefore, GHG emissions caused by production and operation of this additional “non-5G” equipment must be considered. The Swiss analysis mentioned before estimates a footprint of such equipment between 0.03 Mt CO2e/year and 0.16 Mt CO2e/year, according to the three hypothetical scenarios. The more “optimistic” the use cases assumptions, the bigger the GHG footprint, because more use case-specific non-5G equipment is needed (s. next figure).
Flexible work and smart grid have the largest footprint because their adoption rate in 2030 is assumed to be higher than the adoption rate of AD and precision farming and because households and commuters in Switzerland are more numerous than farms. It must also be made clear that the Swiss analysis could not take into account all of the non-5G ICT equipment (fixed networks, data centers, ICT equipment installed in road infrastructure, and so on) needed to implement these use cases.
Overall results: Next figure provides an overview of the overall results of the Swiss study mentioned.
Results of such analyses should be interpreted with some care because of several uncertainty factors, among others the assumption of an hypothetical baseline scenario (business-as-usual GHG emissions in 2030) and the fact that one specific 5G network is considered. GHG abatement potential of use cases selected, and their rebound effects, imply a certain amount of uncertainty too. First of all, because it is almost impossible to include in the footprint estimation all of the ICT-equipment necessary for each use case considered; furthermore, 5G supports many other use cases with both increasing and decreasing effects on GHG emissions. There is also e dependency on the assumed use cases adoption rates, for a certain target year: should the assumption be low, GHG impact (negative or positive) over the time horizon considered might be in reality significantly higher. It is worthy, finally, to stress the fact that many other use cases may exist which provide either substantial GHG abatement potentials, or even contribute to increasing GHG emissions. Use cases of either type can always be created by innovation, most probably even after the 5G infrastructure will be fully rolled out.
Other effects of 5G
Most research about 5G sustainability focus on GWP, although it would be methodologically possible to include other environmental impact categories beyond global warming potential. Apart from GHG emissions and energy consumption, 5G can impact environment in many other ways. At least the following factors should be considered: human health, constructions, deforestation.
WHO (World health Organization) established the International Electromagnetic Fields (EMF) Project in 1996, aiming to investigate the health impact of exposure to electric and magnetic fields in the frequency range 0-300 GHz and to advise national authorities on EMF radiation protection. On the potential health risks from 5G, WHO says that no adverse health effect has been causally linked with exposure to wireless technologies, even though these conclusions are drawn from studies performed across the entire radio spectrum and, so far only few studies have been carried out at the frequencies to be used by 5G. WHO considers tissue heating as the main mechanism of interaction between radiofrequency fields and human body, stating that RF exposure levels from current technologies result in negligible temperature rise in the human body. As the frequency increases, there is less penetration into the body tissues and absorption of the energy becomes more confined to the surface of the body, basically skin and eye. Provided that the overall exposure remains below international guidelines, no consequences for public health are anticipated by WHO. However, pushed by many public initiatives and growing concern about the dangers caused by radiations at 5G frequencies, WHO is currently conducting a health risk assessment from exposure to radiofrequencies, covering the entire radiofrequency range, including 5G, to be published by 2022. Scientific evidence related to potential health risks from 5G exposure shall be reviewed by WHO as the new technology is deployed, and as more public health-related data become available. WHO is committed to further research into the possible long-term health impacts of all aspects of mobile-telecommunications.
Consequences in the field of constructions should also be considered. Higher 5G frequencies necessitate high density of end-points compared to previous generations. Moreover, the costs for building tower stations are high when compared to other networks, and could generate more constructions necessary to realize 5G coverage, which might have a negative impact on the environment, causing destruction of trees, nature, and animals. Further factors potentially boosting new constructions are use cases like Smart Cities, because they rely on navigation systems, AD and smart buildings, causing high density of sensors and users, particularly in areas like stadiums and transit terminals. The fact that antennas are being moved from towers to street level and in general more buildings need to be built can have aesthetic implications. Furthermore, 5G greatly relies on optical fiber infrastructure to achieve the required data transmission speeds. Along with the higher equipment density needed, this can create additional challenges in performing upgrades on existing buildings that weren’t originally designed to accommodate 5G infrastructures, hiding negative consequences for the environment. All this exposes 5G to rebounds effects that are difficult to foresee: more brand new buidings might be constructed, many buildings might need refurbishing works, and there is the challenge of the materials used. To what extent these potentially negative environmental factors could be mitigated, if not over-compensated, by virtuous consequences from the use cases implemented is an open point.
5G can also have a huge influence on the constructions business itself, due to 5G-enabled applications like robots, drones, autonomous cranes, loaders and trucks, 3D-printing robots, combined with other new technologies such as Artificial Intelligence (AI), Augmented Reality (AR), and Virtual Reality (VR), requiring massive data volumes, super low latency and enormous capacity to provide seamless and scalable actions. Considered from ecological perspectives, all this holds a potential rebound effect, because it might induce more construction through higher speed and lower cost. On the other hand, it could also have a positive influence on the environment. Potential advantages of these 5G-driven applications are:
Because 5G works best with line-of-sight communications, there is the risk of bad environmental practices like cutting threes - especially high ones?- and, in general, that elimination of natural elements is encouraged, because considered as “obstacles”. These consequences might get even heavier, should 5G technologies be massively extended to rural areas, where nature is more present.
Another potential problem is that pre-5G phones are not compatible with 5G technology. Thus, consumers will need to replace their current 4G phone, even a perfectly working one, with a 5G-compatible one, generating e-waste. To cope with this issue, systemic measures should be taken, including device buy-back programs and plans for the decommissioned devices beyond recycling centres.
Summing up
5G has for sure an impact on the GHG emissions. However, it seems reasonable to expect that it will cause, in the next 10 years, less direct GHG emissions per unit of transmitted data than today’s mobile networks. Country-specific factors such as energy mix, traffic models, as well as a timespan considered can greatly influence the forecast. Therefore, countries should do their best to produce much more green energy and drastically move away from energy production based on fossils like coal and natural gas.
Indirect effects of many 5G-enabled use cases are favorable, and their GHG reduction potential is far bigger than the GHG footprint of ICT equipment, 5G-specific or not, required to enable those use cases. ?Thus, if 5G network capabilities are systematically utilized to support use cases with high GHG abatement potentials, significant GHG abatements can be realized in the years to come.
Two risks can counteract this GHG abatement potential, that is rebound effects and non-5G-specific ICT equipment needed to make the 5G use cases work, like laptops for flexible work, sensors for precision farming, smart meters for smart grids, and so on. 5G networks as such will be responsible for a small fraction of the footprint of the whole ICT sector, therefore in order to make 5G sustainable and reduce GHG emissions, the GHG footprint of the ICT sector should be kept small. ?Because 5G use cases also need additional ICT equipment, as mentioned above, it is important to take in consideration all ICT infrastructure and equipment necessary for 5G.
A Switzerland-specific analysis, under considerations of some methodological limitations, shows that significant ecological advantages in 2030 can be attributed to the use cases flexible work, smart grid, and precision farming. This speaks for a world where business travel and commuting are massively reduced by virtual collaboration, renewable energy share in the electricity supply system is very high, use of agricultural inputs such as fertilizers is reduced, and productivity in livestock farming is increased.
The GHG abatements enabled by “virtuous” use cases should be encouraged. For instance, “green” use cases in transportation, energy, and farming sectors should be unleashed by creating conditions that are conducive to flexible work models, new mobility services, renewable energy sources, and forms of farming that target the reduction of CO2e (i.e. not only CO2, but also emissions from CH4, N2O, an other dangerous gases). Not making it happen would mean accepting the risk that reductions will remain smaller than the footprint of ICT equipment needed to implement the 5G use cases, and rebound effects might compensate, if not overcompensate for the possible GHG emission reductions.
While short term economic advantages of 5G seem to be clear, social awareness should be raised about the 5G-enabled use cases which hold highest GHG abatement potential and environmental advantages, and these should be pushed. This is only possible if mobile companies worldwide do integrate in their 5G implementation plans a serious impact analysis of 5G on the environment.
Long-term environmental consequences and risks are not limited to GHG emissions; general effects of 5G must also be understood and possibly mitigated as well. The technology needed to power the new 5G network, based on the adoption of millimeter waves instead of radio waves, the new use of small cells, MIMO, new capabilities, and the significant higher amount of devices to be produced will not only significantly change how mobile devices are used and influence energy consumption, but might also have detrimental consequences for human health and natural ecosystems.
However, the environmental impact of 5G on personal health and nature in general is still being studied by biologists and environmental scientists, and many points are still under discussion. There have been several theories, hoaxes, and controversies, though it is still unclear if there is a long-term environmental impact of 5G deployment and how severe it might be. It is important that public information material is made available by world organizations like WHO and that dialogue among scientists, governments, and the public is promoted to increase understanding around health and 5G mobile communications.
Construction of new, additional infrastructures necessary to accommodate 5G and realize 5G coverage might also have negative consequences on the environment, including aesthetic impacts. Mitigation is possible: on one hand, through co-location with non-5G infrastructures whenever possible; on the other hand, new construction should also plan for spare capacity to accommodate future technologies. In the construction sector, 5G also entails use cases (for instance, giant 3D printing) which have a green footprint and contribute to positive practices: safer work environment, innovation, and more efficient passive homes, as well as carbon-negative buildings. However, considered from ecological perspectives, all this also holds a potential rebound effect, because it might induce more non-ecological construction through higher speed and lower cost. To what extent these potentially negative environmental factors could be mitigated, if not over-compensated, by virtuous consequences from the use cases implemented is an open point.
The implementation of 5G might encourage bad environmental practices like cutting threes, elimination of natural elements, and generation of e-waste. Systemic measures should also be taken in this area, like buy-back and recycling programs for non-5G compatible devices.
--------
About the Author
Raffaele Sabatino is Senior Consultant specializing in telecommunications at mm1, the Consultancy for Connected Business.?
--------
[1] Providing service continuity upon changing the point of attachment of a mobile user in the case of same radio access network technology is horizontal handover; providing service continuity upon changing the point of attachment of a mobile user in the case of different radio access technology is vertical handover.
[2] CS = Circuit Switched; PS = Packet Switched
[3] Heterogeneous network. HetNet indicates the use of multiple types of access nodes in a wireless network, to offer wireless coverage in an environment with a wide variety of wireless coverage zones. A HetNet is a network with complex interoperation between macrocell, small cell, and in some cases WiFi network elements used together to provide a mosaic of coverage, with handoff capability between network elements.
[4] 2G and 3G are assumed to be dismissed by year 2030, in the specific case