skip to main content
10.1145/3575813.3597344acmotherconferencesArticle/Chapter ViewAbstractPublication Pagese-energyConference Proceedingsconference-collections
research-article

Energy vs Privacy: Estimating the Ecological Impact of Federated Learning

Published: 16 June 2023 Publication History

Abstract

The increasing usage of edge devices and stricter data privacy regulations motivate the use of federated learning (FL). At the same time, more and more stakeholders are concerned about the ecological impact of machine learning and its associate network traffic. The current research in FL does not investigate the impact of different network constraints and privacy-enhancing techniques, such as differential privacy, on the network traffic and energy consumption of the clients. Most experiments run either on virtual machines or on one machine with simulated clients. In such environments, it is challenging to measure each client’s network and energy usage. Therefore, we built our "Distributed Edge Device Testbed" (DEDT) and evaluate a convolutional neural network trained on the MNIST data set under different network constraints on DEDT, with differential privacy and with an increasing amount of participating clients. For each experiment, we quantify the network traffic, energy consumption, and training time. The results show the importance of experiments on physically separated nodes and the need to improve software-based power monitoring. The estimated energy consumption deviates by up to 35 % from the measured ones. The accuracy of the estimated network traffic depends on the monitored network interface and gives an error of 18 % for virtual machines in combination with monitoring the Ethernet interface. The training time also increases linearly with the number of participating clients.

References

[1]
David Basin, Søren Debois, and Thomas Hildebrandt. 2018. On Purpose and by Necessity: Compliance Under the GDPR. In Financial Cryptography and Data Security, Sarah Meiklejohn and Kazue Sako (Eds.). Springer Berlin Heidelberg, Berlin, Heidelberg, 20–37.
[2]
Keith Bonawitz, Hubert Eichner, Wolfgang Grieskamp, Dzmitry Huba, Alex Ingerman, Vladimir Ivanov, Chloe Kiddon, Jakub Konečný, Stefano Mazzocchi, H. Brendan McMahan, Timon Van Overveldt, David Petrou, Daniel Ramage, and Jason Roselander. 2019. Towards Federated Learning at Scale: System Design. arxiv:1902.01046 [cs.LG]
[3]
Aurelien Bourdon, Adel Noureddine, Romain Rouvoy, and Lionel Seinturier. 2013. PowerAPI: A Software Library to Monitor the Energy Consumed at the Process-Level. ERCIM News 2013 (2013).
[4]
Hafsa Bousbiat, Roumaysa Bousselidj, Yassine Himeur, Abbes Amira, Faycal Bensaali, Fodil Fadli, Wathiq Mansoor, and Wilfried Elmenreich. 2023. Crossing Roads of Federated Learning and Smart Grids: Overview, Challenges, and Perspectives. arxiv:2304.08602 [cs.LG]
[5]
Christopher Briggs, Zhong Fan, and Peter Andras. 2021. Federated Learning for Short-term Residential Energy Demand Forecasting. arxiv:2105.13325 [cs.LG]
[6]
Pietro Ferrara and Fausto Spoto. 2018. Static Analysis for GDPR Compliance. In ITASEC 2018 - Italian Conference on Cyber Security. CEUR Workshop Proceedings, Milan, Italy, 10. http://ceur-ws.org/Vol-2058/#paper-10
[7]
Yunzhe Guo, Dan Wang, Arun Vishwanath, Cheng Xu, and Qi Li. 2020. Towards Federated Learning for HVAC Analytics: A Measurement Study. In Proceedings of the Eleventh ACM International Conference on Future Energy Systems (Virtual Event, Australia) (e-Energy ’20). Association for Computing Machinery, New York, NY, USA, 68–73. https://doi.org/10.1145/3396851.3397717
[8]
Tzu-Ming Harry Hsu, Hang Qi, and Matthew Brown. 2019. Measuring the Effects of Non-Identical Data Distribution for Federated Visual Classification. arxiv:1909.06335 [cs.LG]
[9]
IEA. 2022. Data Centres and Data Transmission Networks. https://www.iea.org/reports/data-centres-and-data-transmission-networks
[10]
Fabian Kaup, Philip Gottschling, and David Hausheer. 2014. PowerPi: Measuring and modeling the power consumption of the Raspberry Pi. In 39th Annual IEEE Conference on Local Computer Networks. IEEE, Edmonton, Canada, 236–243. https://doi.org/10.1109/LCN.2014.6925777
[11]
Kamar Kesrouani, Houssam Kanso, and Adel Noureddine. 2020. A Preliminary Study of the Energy Impact of Software in Raspberry Pi devices. In 2020 IEEE 29th International Conference on Enabling Technologies: Infrastructure for Collaborative Enterprises (WETICE). IEEE, Bayonne, France, 231–234. https://doi.org/10.1109/WETICE49692.2020.00052
[12]
A. S. Khatouni, M. Trevisan, and D. Giordano. 2019. Data-Driven Emulation of Mobile Access Networks. In 2019 15th International Conference on Network and Service Management (CNSM). IEEE, 1515 South Park Street, Halifax, Nova Scotia, Canada, 1–6. https://doi.org/10.23919/CNSM46954.2019.9012691
[13]
Sangyoon Lee and Dae-Hyun Choi. 2022. Federated Reinforcement Learning for Energy Management of Multiple Smart Homes With Distributed Energy Resources. IEEE Transactions on Industrial Informatics 18, 1 (2022), 488–497. https://doi.org/10.1109/TII.2020.3035451
[14]
Tian Li, Maziar Sanjabi, Ahmad Beirami, and Virginia Smith. 2020. Fair Resource Allocation in Federated Learning. arxiv:1905.10497 [cs.LG]
[15]
Y. Liu, J. J. Q. Yu, J. Kang, D. Niyato, and S. Zhang. 2020. Privacy-Preserving Traffic Flow Prediction: A Federated Learning Approach. IEEE Internet of Things Journal 7, 8 (2020), 7751–7763. https://doi.org/10.1109/JIOT.2020.2991401
[16]
H. Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, and Blaise Agüera y Arcas. 2023. Communication-Efficient Learning of Deep Networks from Decentralized Data. arxiv:1602.05629 [cs.LG]
[17]
D. Moss. 2014. Modelling the network performance of DSL connections using netem.
[18]
State of California Department of Justice. 2018. California Consumer Privacy Act of 2018 [1798.100 - 1798.199.100].
[19]
D. E. O’Leary. 2013. Artificial Intelligence and Big Data. IEEE Intelligent Systems 28, 2 (2013), 96–99. https://doi.org/10.1109/MIS.2013.39
[20]
Aman Priyanshu, Rakshit Naidu, Fatemehsadat Mireshghallah, and Mohammad Malekzadeh. 2021. Efficient Hyperparameter Optimization for Differentially Private Deep Learning.
[21]
Xinchi Qiu, Titouan Parcollet, Daniel J. Beutel, Taner Topal, Akhil Mathur, and Nicholas D. Lane. 2020. Can Federated Learning Save The Planet?https://doi.org/10.48550/ARXIV.2010.06537
[22]
Xinchi Qiu, Titouan Parcollet, Javier Fernandez-Marques, Pedro Porto Buarque de Gusmao, Yan Gao, Daniel J. Beutel, Taner Topal, Akhil Mathur, and Nicholas D. Lane. 2022. A first look into the carbon footprint of federated learning. arxiv:2102.07627 [cs.LG]
[23]
Xidi Qu, Shengling Wang, Qin Hu, and Xiuzhen Cheng. 2021. Proof of Federated Learning: A Novel Energy-Recycling Consensus Algorithm. IEEE Transactions on Parallel and Distributed Systems 32, 8 (2021), 2074–2085. https://doi.org/10.1109/TPDS.2021.3056773
[24]
Sashank Reddi, Zachary Charles, Manzil Zaheer, Zachary Garrett, Keith Rush, Jakub Konečný, Sanjiv Kumar, and H. Brendan McMahan. 2021. Adaptive Federated Optimization. arxiv:2003.00295 [cs.LG]
[25]
Y. Roh, G. Heo, and S. E. Whang. 2021. A Survey on Data Collection for Machine Learning: A Big Data - AI Integration Perspective. IEEE Transactions on Knowledge and Data Engineering 33, 4 (2021), 1328–1347. https://doi.org/10.1109/TKDE.2019.2946162
[26]
Yuris Mulya Saputra, Dinh Thai Hoang, Diep N. Nguyen, Eryk Dutkiewicz, Markus Dominik Mueck, and Srikathyayani Srikanteswara. 2019. Energy Demand Prediction with Federated Learning for Electric Vehicle Networks. In 2019 IEEE Global Communications Conference (GLOBECOM). IEEE, Waikoloa, Hawaii, USA, 1–6. https://doi.org/10.1109/GLOBECOM38437.2019.9013587
[27]
Personal Data Protection Commission Singapore. 2014. Personal Data Protection Act.
[28]
A. Taïk and S. Cherkaoui. 2020. Electrical Load Forecasting Using Edge Computing and Federated Learning. In ICC 2020 - 2020 IEEE International Conference on Communications (ICC). IEEE, Virtual, 1–6. https://doi.org/10.1109/ICC40277.2020.9148937
[29]
Andrew Trask, Emma Bluemke, Ben Garfinkel, Claudia Ghezzou Cuervas-Mons, and Allan Dafoe. 2020. Beyond Privacy Trade-offs with Structured Transparency. https://doi.org/10.48550/ARXIV.2012.08347
[30]
Ye Lin Tun, Kyi Thar, Chu Myaet Thwal, and Choong Seon Hong. 2021. Federated Learning based Energy Demand Prediction with Clustered Aggregation. In 2021 IEEE International Conference on Big Data and Smart Computing (BigComp). IEEE, Jeju Island, Korea, 164–167. https://doi.org/10.1109/BigComp51126.2021.00039
[31]
European Union. 2016. REGULATION (EU) 2016/679 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation).
[32]
Haijin Wang, Caomingzhe Si, and Junhua Zhao. 2021. A Federated Learning Framework for Non-Intrusive Load Monitoring. arxiv:2104.01618 [eess.SP]
[33]
Haijin Wang, Caomingzhe Si, Junhua Zhao, Guolong Liu, and Fushuan Wen. 2021. Fed-NILM: A Federated Learning-based Non-Intrusive Load Monitoring Method for Privacy-Protection. arxiv:2105.11085 [cs.LG]
[34]
Sihua Wang, Mingzhe Chen, Walid Saad, and Changchuan Yin. 2020. Federated Learning for Energy-Efficient Task Computing in Wireless Networks. In ICC 2020 - 2020 IEEE International Conference on Communications (ICC). IEEE, Virtual, 1–6. https://doi.org/10.1109/ICC40277.2020.9148625
[35]
Zhaohui Yang, Mingzhe Chen, Walid Saad, Choong Seon Hong, and Mohammad Shikh-Bahaei. 2021. Energy Efficient Federated Learning Over Wireless Communication Networks. IEEE Transactions on Wireless Communications 20, 3 (2021), 1935–1949. https://doi.org/10.1109/TWC.2020.3037554
[36]
Ashkan Yousefpour, Igor Shilov, Alexandre Sablayrolles, Davide Testuggine, Karthik Prasad, Mani Malek, John Nguyen, Sayan Ghosh, Akash Bharadwaj, Jessica Zhao, Graham Cormode, and Ilya Mironov. 2021. Opacus: User-Friendly Differential Privacy Library in PyTorch. arXiv preprint arXiv:2109.12298 (2021), 18.
[37]
Xinyu Zhou, Jun Zhao, Huimei Han, and Claude Guet. 2022. Joint Optimization of Energy Consumption and Completion Time in Federated Learning. In 2022 IEEE 42nd International Conference on Distributed Computing Systems (ICDCS). IEEE, Bologna, Italy, 1005–1017. https://doi.org/10.1109/ICDCS54860.2022.00101
[38]
X. Zhu, J. Wang, Z. Hong, T. Xia, and J. Xiao. 2019. Federated Learning of Unsegmented Chinese Text Recognition Model. In 2019 IEEE 31st International Conference on Tools with Artificial Intelligence (ICTAI). IEEE, Portland, OR, USA, 1341–1345. https://doi.org/10.1109/ICTAI.2019.00186

Cited By

View all

Index Terms

  1. Energy vs Privacy: Estimating the Ecological Impact of Federated Learning

      Recommendations

      Comments

      Information & Contributors

      Information

      Published In

      cover image ACM Other conferences
      e-Energy '23: Proceedings of the 14th ACM International Conference on Future Energy Systems
      June 2023
      545 pages
      ISBN:9798400700323
      DOI:10.1145/3575813
      Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected].

      Sponsors

      Publisher

      Association for Computing Machinery

      New York, NY, United States

      Publication History

      Published: 16 June 2023

      Permissions

      Request permissions for this article.

      Check for updates

      Author Tags

      1. Distributed Systems
      2. Energy
      3. Federated Learning

      Qualifiers

      • Research-article
      • Research
      • Refereed limited

      Conference

      e-Energy '23
      Sponsor:

      Acceptance Rates

      Overall Acceptance Rate 160 of 446 submissions, 36%

      Contributors

      Other Metrics

      Bibliometrics & Citations

      Bibliometrics

      Article Metrics

      • 0
        Total Citations
      • 177
        Total Downloads
      • Downloads (Last 12 months)64
      • Downloads (Last 6 weeks)3
      Reflects downloads up to 20 Jan 2025

      Other Metrics

      Citations

      Cited By

      View all

      View Options

      Login options

      View options

      PDF

      View or Download as a PDF file.

      PDF

      eReader

      View online with eReader.

      eReader

      HTML Format

      View this article in HTML Format.

      HTML Format

      Media

      Figures

      Other

      Tables

      Share

      Share

      Share this Publication link

      Share on social media