David Mytton

Paper notes – An Analysis of Renewable Energy Usage by Mobile Data Network Operators

Published in Data Center Energy, Environment, Paper notes.

Paper notes – An Analysis of Renewable Energy Usage by Mobile Data Network Operators

This is a post in my “Paper Notes” series where I publish my notes as I read papers related to sustainable computing.


Syed, S., Arfeen, A., Uddin, R. & Haider, U. (2021) An Analysis of Renewable Energy Usage by Mobile Data Network Operators. Sustainability. 13 (4), 1886. Available from: doi: 10.3390/su13041886.


The exponential growth in mobile data traffic has resulted in massive energy usage and therefore has increased the carbon footprint of the Internet. Data network operators have taken significant initiatives to mitigate the negative impacts of carbon emissions (CE). Renewable Energy Sources (RES) have emerged as the most promising way to reduce carbon emissions. This article presents the role of renewable energy (RE) in minimizing the environmental impacts of mobile data communications for achieving a greener environment. In this article, an analysis of some selected mobile data network operators’ energy consumption (EC) has been presented. Based on the current statistics of different mobile network operators, the future energy values are estimated. These estimations of carbon emissions are based on the predicted data traffic in the coming years and the percentage consumption of energy from renewable sources by the network operators. The analysis presented in this article would be helpful to develop and implement energy policies that accelerate the process of increasing the renewable shares in total energy requirements. Incrementing the share of renewable energy in total energy requirements can be a way forward to reach Goal 7 of the United Nations Sustainable Development Goals (SDGs).

General notes

  • This paper aggregates data from the sustainability reports of Vodafone, Telefonica and Telstra, examines their annual data rates, total energy consumption and renewable energy purchases, then makes predictions for how renewable energy will affect carbon emissions over the 6 years to 2019-2025.
  • The authors believe this is an important topic because of the recent growth in internet data volumes, the growth in the usage of mobile devices and the correlation between data usage and energy consumption. They make reference to a claim that mobile base stations make up 70% of the energy consumption of mobile networks and link carbon emissions with the generation of the electricity being consumed.
  • It is true that electricity generation causes carbon emissions because the majority of global electricity is generated using fossil-fuels (only 5% renewables as of 2019). However, the variability of the grid mix means that carbon intensity changes on a regular basis. The mobile operators included in this paper are global organisations and so their electricity carbon intensity will vary significantly by region. Linking total electricity consumption to total carbon emissions is a historical calculation based on known carbon emissions factors – it is much more difficult to predict future carbon emissions. The grid mix cannot easily be predicted in advance. New renewables projects are always in development. Energy efficiency may improve. The weather changes. I see no evidence that these factors have been considered by the predictions in this paper (nor noted as a limitation).
  • The quality of the language in the manuscript is poor – there are numerous typos, many repeated sentences, and it is quite difficult to follow in places. English language ability should not be a barrier to publishing scientific research so I wonder why the publisher has not provided editing and proofreading prior to publication.
  • The workings are missing. There is no way to verify the calculations presented in the manuscript because they are not provided. I cannot find any supplementary information such as the graph sources, parameterised equations, or the regression analysis that the results are based upon. The equations included in the manuscript are not fully explained.

Specific comments

  • The “related work” section seems like a missed opportunity – it is just a description of other papers about renewable energy rather than a proper literature review explaining why those papers are relevant.
  • The authors state:

The operators’ goal is to save the environment and reduce carbon emissions by applying sustainable energy usage techniques.

but is that true? That might be the marketing goal but what about simply reducing energy costs? This paper is not a study of operator sustainability strategies so I would be cautious about attributing intention.

  • The selection of these three mobile operators appears arbitrary. The authors say:

The reason for selecting these three operators is that Vodafone, Telstra and Telefonica are among the world’s top ten telecommunication networks. These companies have subscribers from all over the world. Some of the leading telecommunication companies are working in numerous countries. However, their significant share is from one country, like AT&T, which generates more than 90% of its capital share from the USA only. The reason for the inclusion of Telstra is to check the status of relatively smaller network operators regarding energy efficiency. The secondary reason for considering these three network operators is that the operators’ most current and comprehensive data is also available, through which analysis can be performed.

but these do not seem like robust inclusion criteria. Just looking at a Wikipedia article on the largest mobile operators by number of subscribers shows many other operators that have not been included. Why has Telstra been picked to represent a “smaller network operator”. I appreciate the difficulty of finding sufficient data to analyse operators – increasing transparency is one of my own research themes – but the paper does not define what “top” means.

  • The description of how the percentage of renewable energy is calculated says “using the TREND function”. What is the trend function? Presumably the linear trend function in Excel, but that is not explained and the values used are missing.
Figure 5b and 5c.
  • Figure 5 shows carbon emissions reductions, current and forecast, and all three show a sudden increase from 2019 to 2020. This is interesting because it is the end of the actual values (2019) and start of the projected values (2020). It is most pronounced in Figure 5c. The paper says:

From the analysis, it is noticed that initially, the share of renewable energy usage is likely to decline as compared to its current use, due to the enormous increase in data usage. Despite the initial reduction, the use of renewable energy sources is estimated to increase afterward.

However, this is not really an explanation, just a description. It is not clear what the explanation for this jump is.

  • In the discussion, this doesn’t make sense:

Though all these mentioned mobile data operators have experienced incremental data traffic, the electricity consumption has not been changed or differs negligibly. The reason for unchanged electricity is the use of renewable energy by these mobile data network operators.

Perhaps this is a drafting error and they mean that despite the increased data traffic, the carbon emissions do not change because of the use of renewable energy. Using renewable energy will have no impact on electricity consumption.


There is value in analysing the sustainability reporting of major mobile operators because of the significant increase in mobile traffic over the past few years. Despite expected future growth of mobile data traffic, mobile technology has received only limited research attention.

Unfortunately, this paper is a missed opportunity. Equation parameterisation and workings have not been provided so the results cannot be verified. The methods have not been adequately explained and the discussion makes some strange statements that do not make sense. Some of the results presented have characteristics that are unexplained. The manuscript also needs significant language editing.