Method for Calculating the Uncertainty range of Avoided Primary
Energy Consumption and Environmental Impact applied to Data
Analysis Software Services and Solar Electricity
ANDERS S.G. ANDRAE
Huawei Technologies Sweden AB
Skalholtsgatan 9, 16494 Kista
SWEDEN
Abstract: - The absolute and avoided primary energy consumption (PEC) of Software (SW) services is getting
more attention. However, there is no commonly agreed bottom-up methodology for calculation of the total PEC
of SW services. Life Cycle Assessment (LCA) is a common denominator for most existing methodologies. The
purpose is to test a new simplified methodology which includes the uncertainty and sensitivity. The new
methodology is applied to two illustrative cases: data analysis SW and electricity production. The baseline
results for data analysis SW show that the uncertainty will be quite high at around 30% and the most sensitive
parameters are the production of electricity, the amount of data transfer and the production of the end-user
device.
Moreover - the data bytes transferred from the end-user device per iteration, the PEC per byte data transfer and
the PEC of the production of the end-user device used to access the SW - contribute most to the total
uncertainty. Regarding solar electricity replaced by a proportionate electricity mix, the avoided carbon
emissions in China from 2021 to 2023 were 80±36 million tonnes. Intermediate suppliers to the solar electricity
production systems can claim to have contributed to the avoided emissions according to their contribution ratio.
Key-Words: - avoided impact, data analysis, environmental impact, solar electricity, software, uncertainty, life
cycle assessment.
Received: March 22, 2024. Revised: October 15, 2024. Accepted: November 17, 2024. Published: December 30, 2024.
1 Introduction
For several years there has been a growing interest
to attempt to find the primary energy consumption
(PEC) associated with software (SW) solutions [1]
such as individual SW, packages, middleware, and
operating systems [2]-[26].
SW production impacts can be caused by the use of
electricity in turn used by offices and computers
used by the SW designers [7].
SW systems use energy through the manufacturing
and use of the hardware that they operate on [15].
SW service Life Cycle Assessment (LCA) includes
resources such as terminals, software, networks,
service platforms, and servers. Many SW programs
are currently run in the cloud with data transfer from
data centers via networks to end-user devices.
The LCA methods proposed in [27,28] for cloud
services are less to the point for cloud SW than the
present method which is somewhat more practical.
Most research in this domain uses descriptive
approaches and point out the challenges but recently
more methods have emerged [9,15,16,20]. It is not
clear how the all the input data for such methods can
easily be collected. Here is presented a hands-on
methodology for SW applied to cloud SW services
with an accompanying case study. Moreover,
avoided impact methods including probability
analysis have not been applied to data analysis
software services. The hypothesis is that the data
analysis software service use around 100 Wh PEC
per hour.
Another trend is that Solar electricity has grown to
change the electricity generation system helping
avoid environmental impacts (EI) from other power
International Journal of Environmental Engineering and Development
DOI: 10.37394/232033.2024.2.25
Anders S. G. Andrae
E-ISSN: 2945-1159
283
Volume 2, 2024
sources [29]. However, avoided emission analyses
including probability have not so far been applied to
solar electricity. How much carbon emissions were
avoided in China between 2021 and 2023 in China
due to solar electricity? Here a suggestion of
probable avoided EI from Solar growth in China
from 2021 to 2023 is attempted. The probability
methodology used in the present research is further
described in [30].
2 Materials and Methods
This section describes how the total EI of a specific
SW Service can be calculated with minimal effort
but still explain the important drivers. It also
describes how avoided EI from solar electricity can
be calculated in the interest of intermediate
suppliers and end-user solution providers both.
2.2 Primary energy consumption of software
services
The main PEC of a SW service can be estimated by
adding the end-user device use and production, the
data transfer use and the data centers use stage.
The function of the present SW service is to provide
visualization of data analyses.
The functional unit is “A subsystem enabling
visualization of one analytical iteration by one
employee with a data analysis software.”
2.2.1 End-user device use stage
This stage concerns the use of electrical energy in
the use stage for the end-user device used to access
the SW service. For simplicity one contemporary
laptop (and no other devices) is assumed to be used
to run the data analysis software. The laptop battery
capacity is around 80 Wh electricity [31] and its
uncertainty range is set to 15 Wh. The battery life
when 100% fully charged is assumed to be 8 hours
with uncertainty range 3 hours [32].
The duration of one iteration is 20 seconds with
uncertainty range 3 seconds and the global average
electricity production use 2.7 Wh PEC/Wh±0.2
Wh/Wh.
Then the PEC of the use of the end user device is
calculated as:
2.7 Wh PEC/Wh×20 seconds×80 Wh/[8×3600
seconds] = 0.15±0.063 Wh PEC/analytical iteration.
It would be possible to add more devices such as
smartphones and desktops to this section.
2.2.2 End-user manufacturing stage
The laptop manufacturing impact is around 1
million Wh PEC per piece and the uncertainty is 0.5
million Wh [33]. Other assumptions are that
35%±5% of the laptop capacity is used during the
iteration, the lifetime of the laptop is 4±2 years [34],
and the working hours per year are 2000±50 hours.
Then the PEC of the production of the end user
device is calculated as:
1.018 million Wh/laptop×20 seconds×35%/[4
years/laptop×8 hours/day×5 days/week×50
weeks/year×3600 seconds] = 0.247±0.183 Wh
PEC/analytical iteration.
The uncertainty range for manufacturing is very
large due to the variable lifetime of the device and
the EI per device.
Similarly, to section 2.2.1, it would be possible to
add more devices such as smartphones and desktops
to this section.
2.2.3 Data transfer use and manufacturing stages
The analysis results are transmitted via various
networks. The scope for data transfer includes a
fraction of the complete internet infrastructure
including when it is not used but still running.
Supporting infrastructure significantly enables the
operation of software [7]. Such infrastructure could
include:
— Compute resources
— Storage
— Networking equipment
— Memory
— Monitoring
— Idle machines
— Logging
— Scanning
— Build and deploy pipelines
— Testing
— Training Machine Learning models
— Operations
— Backup
— Resources to support redundancy
— Resources to support failover
— End user devices
— Internet of Things-devices
— Edge devices.
International Journal of Environmental Engineering and Development
DOI: 10.37394/232033.2024.2.25
Anders S. G. Andrae
E-ISSN: 2945-1159
284
Volume 2, 2024
This supporting infrastructure is the reason why the
entire Internet electricity intensity is used in this
segment.
The electricity use of data transfer is around 0.223
±0.055 Wh/MegaByte [35]. 0.223 is obtained from
year 2024 in [35] as 2008 TWh divided by 8796
ExaByte divided by 1024 times 1000. On average
around 2±0.5 MegaBytes are transmitted per
iteration.
Then the EI of the data transfer is calculated as:
2.7 Wh PEC/Wh×0.223 Wh/MB×2 MB/iteration =
1.204±0.434 Wh PEC/analytical iteration.
The uncertainty range for the data transfer is very
large mainly due to the Wh/MB and amount MB per
iteration.
Moreover, the data transfer impact might
occasionally be reduced via the design of the SW
solution:
2.7 Wh PEC/Wh×0.223 Wh/MB×0.5 MB/iteration
= 0.3±0.108 Wh PEC/analytical iteration.
2.2.4 Cloud use stage
The data centers play a major role for the SW
services like visualization of data analyses. One
iteration is assumed to require 8±2 virtual Central
Processing Units (CPUs) and 64±4 GigaByte
Memory over 0.5±0.2 seconds. The electricity
consumption is assumed to be 7×10-4±1.4×10-4 Wh/s
per virtual CPU [36] and 1.25×10-4±2.5×10-5 Wh/s
per GB for memories [37]. A virtual Graphical
Processing Unit (GPU) might use more power than
a virtual CPU [38] but this is not explored further.
Then the PEC of the cloud use is calculated as:
2.7 Wh/Wh×0.5 seconds/iteration×[8 virtual
CPU×0.0007 Wh/s/CPU + 64 GB/iteration ×
0.000125 Wh/s/GB] = 0.0184±0.0062 Wh/analytical
iteration.
The time used for each iteration (0.5 seconds)
contributes the most to the uncertainty range for the
cloud use.
2.2.5 Summary
By the building blocks mentioned in sections 2.2.1-
4 it is possible to calculate the PEC from one
iteration and also the avoided PEC by more or less
data transfer.
The total energy use for the baseline product is
1.62±0.477 Wh PEC/analytical iteration.
Moreover, 1.62 Wh PEC per 20 seconds duration of
one iteration 291.6 Wh PEC/hour.
The total energy use of a lower EI design (target
product) is 0.717±0.214 Wh PEC/analytical
iteration.
Potentially avoided PEC from this particular data
analysis SW are:
1.62 Wh/iteration – (1+0.378)×0.717 Wh/iteration =
0.632 Wh±0.381 Wh. The calculation includes a
30% rebound effect, i.e. (0.378×0.717)/(1.62-0.717).
The uncertainty is 60.3% around the mean.
2.3 Solar electricity avoided emissions
Electricity generation is one of the most analyzed
systems in LCA [39].
In China, Solar electricity generation grew by 257.2
TWh (78.6%) between 2021 and 2023 [40,41].
However, it is not clear which source of electricity
this growth replaced for this time period.
Apart from Solar, the predominant sources for
power production in the Chinese market are Coal,
Hydro, Wind, Nuclear, and Gas. The average of
those five is a reasonable assumption for which
technologies Solar replaced between 2021 and 2023.
However, the allocated mix by is difficult to
quantify.
For the present research the function is to provide
electricity.
The functional unit is “A subsystem providing 257.2
TWh electricity in China between 2021 and 2023.”
Anyway, 0.5 kg CO2e/kWh can be used as a rough
approximation if the factor for a certain country
cannot be obtained [42]. 0.12 is assumed as
uncertainty range.
Average solar electricity releases between 37.3 and
72.2 grams CO2e/kWh [43] with an average of
54.75.
Then the avoided EI per functional unit of the Solar
electricity growth is:
252.7 billion kWh × [0.5 kg CO2e/kWh for average
mix – 3.44×0.05475 kg CO2e/kWh for Solar] =
International Journal of Environmental Engineering and Development
DOI: 10.37394/232033.2024.2.25
Anders S. G. Andrae
E-ISSN: 2945-1159
285
Volume 2, 2024
8.016E+10 kg CO2e = 80.16 million tonnes±36.3
million tonnes. The calculation includes a 30%
rebound effect, i.e. (252.7
billion×2.44×0.05475)/(252.7 billion×0.5-252.7
billion×0.05475). The uncertainty is 45.3% around
the mean.
The rebound effect is the potential benefit minus the
actual benefit. To have a rebound effect of 30% for
this system, 2.44 times the impact of the solar
electricity has to be deducted from the potential
avoided emissions.
The approach is also applicable to year by year
estimations of avoided EI.
3 Results
3.1 Data analysis software services
The results are shown in Figures 1, 2 and 3.
Fig. 1. Impact from SW service baseline product.
The total result is 1.62±0.477 Wh PEC/analytical
iteration.
Fig. 2. Impact from SW service target product
The total result is 0.716±0.214 Wh/analytical
iteration.
In Figure 3, the amount of data transferred per
iteration and the amount of Wh used per MB
contribute altogether 90% to the total uncertainty for
the avoided PEC.
Fig. 3. Avoided primary energy consumption from
introducing new data analysis software.
3.2 Solar electricity
Figure 4 shows that the probability is very high that
solar electricity growth helped avoid emissions.
Fig. 4. Avoided emissions from Solar electricity
growth 2021 to 2023 in China.
The EI/kWh uncertainty contributes 72% to the total
uncertainty of 36.3 million tonnes.
4Discussion
There is likely some typical ballpark numbers for
PEC energy consumption for software services e.g.
Joules per GB and per hour. It would be
unreasonable to diverge too much from ≈102
Wh/hour and ≈101 Wh/GB. Anyway, cloud video
streaming software services have several
benchmarks such as [27] at around 30 Wh
electricity/GB and 72 Wh electricity/hour. Another
is Fig. 2 in [10] at 83 Wh PEC/GB and 250 Wh
PEC/hour for streaming, as well as 1440 Wh
PEC/hour [9] and 142 to 1220 Wh PEC/hour [21].
9%
15%
75%
1%
Distribution of data analysis software primary
energy consumption baseline product
End-user device use
stage
End-user
manufacturing stage
Data transfer use and
manufacturing stages
Cloud use stage
21%
34%
42%
3%
Distribution of data analysis software
primary energy consumption target
product End-user device use
stage
End-user
manufacturing stage
Data transfer use and
manufacturing stages
Cloud use stage
0.632
00.2 0.4 0.6 0.8 1
Avoided PEC per iteration in data analysis software
Avoided Primary Energy Consumption (PEC) -Wh
per visualization of one analytical iteration
1.01
0.25
8.02E+10
0.00E+00 5.00E+10 1.00E+11 1.50E+11
Avoided Carbon (kg) from Solar growth in China 2021 to
2023
Avoided kg from Solar growth
1.17E+11
4.38E+10
International Journal of Environmental Engineering and Development
DOI: 10.37394/232033.2024.2.25
Anders S. G. Andrae
E-ISSN: 2945-1159
286
Volume 2, 2024
As shown in Figure 5 the present baseline iteration
(292 Wh PEC/hour) corresponds to around 1.5
GB/hour of video streaming.
Fig. 5. Typical relation between data consumption
and primary energy per hour for software services.
Regarding global annual cloud electricity use, is the
present 0.000125 Wh/s per GB (0.45 W/GB) for
memories consistent with [3] for data centers (0.007
kWh/GB) which led to 400 TWh electricity per
year?
It has been estimated that globally 0.33 ZettaByte
(363 billion GigaByte) of data are generated daily
[37].
As a sanity check for electricity consumption related
to data generation in data centers:
0.45 (J/s)/GB × 3.63E+11 GB data generated
globally/day [37] × 86400 s/day = 1.41E+16 J/day =
3.92 TWh electricity/day leading to 1430 TWh
electricity per year which is a kind of reasonable -
but still much too high - ballpark number. The
reason is that not all data generated use 0.45 W/GB.
This uncertainty is reflected in section 2.2.4 for the
CPUs.
The data transfer energy efficiency is increasing
with every new technology introduced [44]. Such
phenomena could change which parts of the SW
lifecycle are most important from an energy
standpoint.
To this point, Artificial Intelligence data analysis
SW [20] might have a different distribution of the
impacts than shown in Figures 1 and 2. Cloud use
probably has a higher share for AI SW services than
non-AI SW.
Solar electricity helps avoid emissions but so far
there is no agreed method for allocation to different
processes in the ecosystem. Depending on the solar
solution features, PV modules and inverters and
occasionally batteries are necessary for the function
of the Solar Solution. Then the intermediate
manufacturers of PV modules, inverters, batteries
and the Solar system provider itself may claim
shares of the avoided emission. These shares might
be based on different contribution keys and ratios
still to be developed.
Avoided impact scores are very dependent on use
case and geography. The variation of such cases
should be investigated.
5 Conclusions
Avoided PEC and EI calculations are straight-
forward and can be done in a streamlined manner.
Simplified approaches for SW Service impact
evaluations are relatively accurate for early design
stage trade-off potentials.
6 Next steps
It remains to be researched whether the developed
method is generally applicable to SW Services
beyond data analysis.
References:
[1] A. Guldner, R. Bender, C. Calero, G. Fernando, et al.
Development and evaluation of a reference measurement
model for assessing the resource and energy efficiency of
software products and components—Green Software
Measurement Model (GSMM). Future Generation
Computer Systems, Vol. 155, 2024, pp. 402-418.
https://doi.org/10.1016/j.future.2024.01.033
[2] D. Mytton. Assessing the suitability of the
Greenhouse Gas Protocol for calculation of emissions
from public cloud computing workloads. Journal of
Cloud Computing, Vol. 9, No. 1, 2020, p. 45.
https://doi.org/10.1186/s13677-020-00185-8
[3] T. Tabata, and T.Y. Wang. Life cycle assessment of
CO2 emissions of online music and videos streaming in
Japan. Applied Sciences, Vol. 11, No. 9, 2021, p. 3992.
https://doi.org/10.3390/app11093992
[4] M.M. Vopson. Experimental protocol for testing the
mass–energy–information equivalence principle. AIP
Advances, Vol. 12. No. 3, 2022. p. 035311
https://doi.org/10.1063/5.0087175
[5] M. Burgin, R. Mikkilineni. Is information physical
and does it have mass?. Information, Vol. 13, No. 11,
2022, p. 540. https://doi.org/10.3390/info13110540
[6] M.M. Vopson, S. Lepadatu. Second law of
information dynamics. AIP Advances, Vol. 12, No. 7,
2022. p. 075310. https://doi.org/10.1063/5.0100358
[7] F. Alfieri, C. Spiliotopoulos. ICT Task Force study:
Final Report. Publications Office of the European Union,
Luxembourg, 2023, doi:10.2760/486253, JRC133092.
0
100
200
300
400
500
0 1 2 3
Wh PEC/hour [27]
GB per hour
GigaByte and Primary Energy per hour for
software services
Wh PEC/hour [27]
Wh PEC/hour
present case baseline
International Journal of Environmental Engineering and Development
DOI: 10.37394/232033.2024.2.25
Anders S. G. Andrae
E-ISSN: 2945-1159
287
Volume 2, 2024
Available online:
https://publications.jrc.ec.europa.eu/repository/handle/JR
C133092 (accessed 12 December 2024).
[8] ADEME. Methodological standard for the
environmental assessment for Internet Service Provision
(ISP) - General principles for the environmental labelling
of consumer products. 2023. Available online:
https://codde.fr/wp-
content/uploads/2023/05/pcr_internet_services_provision
__english_version.pdf (accessed 12 December 2024).
[9] J. Moulierac, G. Urvoy-Keller, M. Dinuzzi, Z. Ma.
What is the carbon footprint of one hour of video
streaming? (Doctoral dissertation, Université Côte
d'Azur). Available online: https://hal.science/hal-
04069500/ (accessed 12 December 2024).
[10] L.K. Oxenløwe, Q. Saudan, J. Riebesehl, M. Zahidy,
S. Swain. Evaluating Energy Consumption of Internet
Services. IEICE Transactions on Communications, Vol.
E1016.B, No. 11, 2023, pp. 1036-1043.
https://doi.org/10.1587/transcom.2022OBI0001
[11] S. Kruijer, A. Tarara, M. Schulze. Transparency for
Software Climate Impact. 2023. Available online:
https://publication2023.bits-und-baeume.org/redefining-
progress/transparency-for-software-climate-impact/
(accessed 12 December 2024).
[12] J. Manner. Black software—the energy
unsustainability of software systems in the 21st century.
Oxford Open Energy, Vol. 2, 2023, p. oiac011.
https://doi.org/10.1093/ooenergy/oiac011
[13] Z. Ma. Energy metaverse: the conceptual framework
with a review of the state-of-the-art methods and
technologies. Energy Informatics, Vol. 6, No. 1, 2023, p.
42. https://doi.org/10.1186/s42162-023-00297-w
[14] ARCEP. Assessing the digital environmental
footprint in France. 2024. Available online:
https://www.arcep.fr/fileadmin/user_upload/grands_dossi
ers/environnement/ADEME-
Arcep_Study_on_the_digital_environnemental_footprint
_in_France_-_Part_1_and_2.pdf (accessed 12 December
2024).
[15] Greenspector. Greenspector’s methodology - How
Does Greenspector Robustly Measure the Environmental
Impacts of Digital Services? 2024. Available online:
https://greenspector.com/wp-
content/uploads/2024/05/Methodology_Greenspector_ful
l_EN-Version-202405.pdf (accessed 12 December 2024).
[16] M.P. Ajie, M. Stock, D. Beloin-Saint-Pierre, L.R.
Adrianto. Parameterized Life Cycle Assessment Model
for Electronic Devices: Incorporating Internet
Consumption in Environmental Impact Evaluation. 2024.
In 2024 Electronics Goes Green 2024+(EGG) (pp. 1-8).
IEEE. DOI: 10.23919/EGG62010.2024.10631238
[17] Green Software Foundation. Software Carbon
Intensity (SCI) Specification. 2024. Available online:
https://sci.greensoftware.foundation/ (accessed 12
December 2024).
[18] R. Liu, D. Bunke, F. Behrens, J. Gröger.
Environmental Impacts of Digital Infrastructures and
Digital Services: CO2, Resource Consumption,
Substances of Concern and More. 2024. In 2024
Electronics Goes Green 2024+(EGG) (pp. 1-10). IEEE.
DOI: 10.23919/EGG62010.2024.10631234
[19] Sustainable Web Design. Estimating Digital
Emissions - What is the Sustainable Web Design Model?
2024. Available online:
https://sustainablewebdesign.org/estimating-digital-
emissions/ (accessed 12 December 2024).
[20] A. Berthelot, E. Caron, M. Jay, L. Lefèvre.
Estimating the environmental impact of Generative-AI
services using an LCA-based methodology. Procedia
CIRP, Vol. 122, 2024. pp. 707-712.
https://doi.org/10.1016/j.procir.2024.01.098
[21] Y. Omidi. Estimating the greenhouse gas emissions
of daily internet traffic of end-users in the Netherlands
(Bachelor's thesis, University of Twente). 2024.
Available online: https://essay.utwente.nl/100943/
(accessed 12 December 2024).
[22] É. Guégain, T. Simon, A. Rahier, R. Rouvoy.
Managing Uncertainties in ICT Services Life Cycle
Assessment using Fuzzy Logic. ICT4S'24 - 10th
International Conference on ICT for Sustainability, Jun
2024, Stockhlom, Sweden. ⟨hal-04532041⟩, Available
online: https://hal.science/hal-04532041/ (accessed 12
December 2024).
[23] J. Gröger, F. Behrens, R. Liu, D. Bunke. Life cycle
assessment of digital services. 2024. Available online:
https://ecodigit.de/fileadmin/user_upload/Method_sketch
_LCA_Software.pdf (accessed 12 December 2024).
[24] L. Yang, S.T. Ni, Y. Wang, A. Yu, J.A. Lee, P. Hui.
Interoperability of the Metaverse: A Digital Ecosystem
Perspective Review. 2024. arXiv preprint
arXiv:2403.05205.
https://doi.org/10.48550/arXiv.2403.05205
[25] R. Istrate, V. Tulus, R.N. Grass, L. Vanbever, W.J.
Stark, G. Guillén-Gosálbez. The environmental
sustainability of digital content consumption. Nature
Communications, Vol. 15, No. 1, 2024. p. 3724.
https://doi.org/10.1038/s41467-024-47621-w
[26] G. Guidi, F. Dominici, J. Gilmour et al.
Environmental Burden of United States Data Centers in
the Artificial Intelligence Era. 2024. arXiv preprint
arXiv:2411.09786.
https://doi.org/10.48550/arXiv.2411.09786
[27] N. Ullrich, F.M. Piontek, C. Herrmann, A. Saraev,
T. Viere. Estimating the resource intensity of the Internet:
A meta-model to account for cloud-based services in
LCA. Procedia CIRP, Vol. 105, 2022, pp. 80-85.
https://doi.org/10.1016/j.procir.2022.02.014
[28] A.S.G. Andrae. Method for Calculating the Avoided
Impact of Specific Information and Communication
Technology Services. International Journal of
Environmental Engineering and Development, Vol. 2,
2024, pp. 73-87. DOI: 10.37394/232033.2024.2.7
[29] K. Ukoba, K.O. Yoro, O. Eterigho-Ikelegbe, C.
Ibegbulam, and T.C. Jen. Adaptation of solar power in
the Global south: Prospects, challenges and opportunities.
Heliyon, Vol. 10, No.7, 2024, e28009. DOI:
10.1016/j.heliyon.2024.e28009
[30] A.S.G Andrae. Method for Uncertainty and
Probability Estimation of Avoided Impacts from
International Journal of Environmental Engineering and Development
DOI: 10.37394/232033.2024.2.25
Anders S. G. Andrae
E-ISSN: 2945-1159
288
Volume 2, 2024
Information and Communication Technology Solutions.
International Journal of Recent Engineering Science
(IJRES), Vol. 11, No.5, 2024, pp. 103-108. DOI:
https://doi.org/10.14445/23497157/IJRES-V11I5P110
[31] Y.H. Chang, C.H. Chu, H.W. Lin. Improvement in
Laptop Heat Dissipation with Taguchi Method.
Electronics, Vol. 13, No.5, 2024, p. 882.
https://doi.org/10.3390/electronics13050882
[32] K. Kunal, N. Fuhr, A. Papenmeier. Product Reviews
as Source for Extracting Product Information: Lessons
Learned. eCom’24: ACM SIGIR Workshop on
eCommerce, Washington, District of Columbia, United
States. 2024. Available online: https://sigir-
ecom.github.io/eCom24Papers/paper_23.pdf (accessed
12 December 2024)
[33] H. André, M.L. Söderman, A. Nordelöf. Resource
and environmental impacts of using second-hand laptop
computers: A case study of commercial reuse. Waste
Management, Vol. 88, 2019, pp. 268-279
https://doi.org/10.1016/j.wasman.2019.03.050
[34] M. Hektoen. Modelling the environmental benefit of
a lifetime extension on a laptop to facilitate pro-
environmental behaviour (Master's thesis, NTNU), 2016.
Available online: https://ntnuopen.ntnu.no/ntnu-
xmlui/bitstream/handle/11250/2401244/15764_FULLTE
XT.pdf?sequence=1&isAllowed=y (accessed 12
December 2024)
[35] A.S.G. Andrae. New perspectives on internet
electricity use in 2030. Engineering and Applied Science
Letters, Vol. 3, No. 2, 2020, pp. 19-31.
doi:10.30538/psrp-easl2020.0038
[36] R. Liu, Y. Wu, K. Zhao et al. Online Resource
Provisioning and Batch Scheduling for AIoT Inference
Serving in an XPU Edge Cloud. IEEE Transactions on
Emerging Topics in Computing, 2024, pp. 1-16. DOI:
10.1109/TETC.2024.3403874
[37] H.S. Lella, R. Chattaraj, S. Chimalakonda and M.
Kurra. Towards Comprehending Energy Consumption of
Database Management Systems-A Tool and Empirical
Study. In Proceedings of EASE '24: 28th International
Conference on Evaluation and Assessment in Software
Engineering 2024, pp. 272-281.
https://doi.org/10.1145/3661167.3661174
[38] A. Andrae. Towards hundred Thousand-fold
improvement in energy performance for the coming
ronnabyte era? International Journal Of Advanced
Research in Engineering & Management (IJAREM), Vol.
10, No. 4, 2024, pp. 1-13. Available online:
http://www.ijarem.org/papers/v10-i4/1.IJAREM-
G7320.pdf (accessed 12 December 2024)
[39] M.V. Barros, R. Salvador, C.M. Piekarski et al.. Life
cycle assessment of electricity generation: a review of the
characteristics of existing literature. The International
Journal of Life Cycle Assessment, Vol. 25, 2020, pp. 36-
54. https://doi.org/10.1007/s11367-019-01652-4
[40] Energy Institute. Statistical Review of World Energy
2023. Available online:
https://www.energyinst.org/statistical-review (accessed
12 December 2024)
[41] Energy Institute. Statistical Review of World Energy
2024. Available online:
https://www.energyinst.org/statistical-review (accessed
12 December 2024)
[42] N. Wang, Y. Ren, T. Zhu, F. Meng et al. Life cycle
carbon emission modelling of coal-fired power: Chinese
case. Energy, Vol. 162, 2018, pp. 841-852.
https://doi.org/10.1016/j.energy.2018.08.054
[43] L. Wang, T. Qiu, M. Zhang, Q. Cao et al. Carbon
emissions and reduction performance of photovoltaic
systems in China. Renewable and Sustainable Energy
Reviews, Vol. 200, 2024, p. 114603.
https://doi.org/10.1016/j.rser.2024.114603
[44] A. Seidel, N. May, E. Guenther, F. Ellinger.
Scenario-based analysis of the carbon mitigation potential
of 6G-enabled 3D videoconferencing in 2030. Telematics
and Informatics, Vol. 64, 2021, p. 101686.
https://doi.org/10.1016/j.tele.2021.101686
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
The author contributed in the present research, at all
stages from the formulation of the problem to the
final findings and solution.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The author has no conflict of interest to declare that
is relevant to the content of this article.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
International Journal of Environmental Engineering and Development
DOI: 10.37394/232033.2024.2.25
Anders S. G. Andrae
E-ISSN: 2945-1159
289
Volume 2, 2024
