Fake Bandwidth Classification in Internet Networks Using Convolutional Neural Networks
Keywords:
Fake Bandwidth; Convolutional Neural Network; Internet Service ProviderAbstract
The rising incidence of corruption in bandwidth procurement within the Indonesian ISP sector has attracted considerable public attention. This phenomenon can be mathematically quantified when users receive bandwidth that does not correspond to their contractual agreements, resulting in significant financial losses due to service quality falling substantially below what is paid for. The concept of "fake bandwidth" was introduced by Azriel Christian Nurcahyo during 2023–2024, referring to situations where allocated bandwidth is not fully delivered actual performance may be reduced by 20–30%, or even exceed 50%. Empirical verification remains challenging as it requires capable hardware and real-time analytical computation. This motivated investigation using Convolutional Neural Networks (CNN), selected for its processing speed advantages over GRU or LSTM, and its capability for continuous real-time computation over 1–2 weeks without degrading server performance. The model was implemented at the University of Technology Sarawak (UTS), Malaysia, for objective evaluation prior to deployment in Indonesia. Testing employed training-testing ratios of 30%, 50%, and 70%, with continuous data processing over 10–12 days using a symmetrical 100 Mbps configuration. Ground Truth results indicated: Fake 23.39%, Genuine 53.85%, No Heavy 6.41%, and Unclassified 16.35%. CNN classification identified: Fake 23.25%, Genuine 52.75%, No Heavy 7.12%, and Unclassified 16.88%. The CNN achieved an error rate of 3.07% with classification accuracy of 96.93% from 1,300,100 test samples. The novelty of this research lies in demonstrating that even within a major university network, a fake bandwidth rate of approximately 23% persists. The system remains operational and under active development to enhance efficiency and adapt to various router types.
References
[1] R. Narwal, N. Panwar, and P. Yadav, “Computer networking: Fundamentals, models, and security considerations,” International Journal of Advanced Research in Science, Communication and Technology, 2024. https://doi.org/10.48175/ijarsct-17621
[2] M. Cüppers, “Quality of service (QoS) based wireless network for mobile Ad Hoc networks,” I-Manager’s Journal on Wireless Communication Networks, vol. 10, no. 2, p. 20, 2022. https://doi.org/10.26634/jwcn.10.2.18934
[3] A. C. Nurcahyo, A. T. H. Yong, and A. F. Atanda, “Classification of simulated fake bandwidth data using LSTM,” TEPIAN, vol. 5, no. 3, pp. 35–47, 2024. https://doi.org/10.51967/tepian.v5i3.3106
[4] A. C. Nurcahyo, T. H. Yong, and A. F. Atanda, “Optimisation of network logs for fake bandwidth classification using CNN,” TEPIAN, vol. 6, no. 2, pp. 85–96, 2025. https://doi.org/10.51967/tepian.v6i2.3260
[5] A. C. Nurcahyo, H. Y. Ting, and A. F. Atanda, “Network log implementation for GRU based bandwidth classification,” Journal of Computers and Digital Business, vol. 4, no. 2, pp. 76–89, 2025. https://doi.org/10.56427/jcbd.v4i2.763
[6] S. Ali and H. S. Bakhtiar, “Audit forensik dan bukti digital dalam mengungkap kasus korupsi BTS Kominfo 2023,” Intellektika, vol. 3, no. 1, pp. 115–125, 2024. https://doi.org/10.59841/intellektika.v3i1.2036
[7] N. Ahmad, S. Alfira, and C. Paripurna, “Corruption in digital transformation: A case study of the impact of misappropriation of Kominfo 4G BTS project funds on e-government governance in Indonesia,” n.d. Available: https://jpabdimas.idjournal.eu/index.php/soshum/article/view/4022
[8] Y. Xiao, M. Varvello, and A. Kuzmanovic, “Monetizing spare bandwidth: The case of distributed VPNs,” in Measurement and Modeling of Computer Systems, 2022. https://doi.org/10.1145/3489048.3530966
[9] S. Sasikumar et al., “Unlocking complex patterns through the power of deep learning for advanced data analysis and prediction,” pp. 621–625, 2024. https://doi.org/10.1109/cybercom63683.2024.10803222
[10] I. Ferjani, M. S. Hidri, and A. Frihida, “Dynamic knowledge expansion: Real-time text classification with deep convolutional neural networks,” 2024. https://doi.org/10.21203/rs.3.rs-3826991/v1
[11] A. Jenefa and V. Nair, “A robust deep learning-based approach for network traffic classification using CNNs and RNNs,” in IEEE Conference on Systems, Process and Control, pp. 106–110, 2023. https://doi.org/10.1109/ICSPC57692.2023.10125858
[12] Y. Zhang and Z. Wang, “Feature engineering and model optimization based classification method for network intrusion detection,” Applied Sciences, 2023. https://doi.org/10.3390/app13169363
[13] P. Maxwell, E. Alhajjar, and N. D. Bastian, “Intelligent feature engineering for cybersecurity,” in International Conference on Big Data, pp. 5005–5011, 2019. https://doi.org/10.1109/BIGDATA47090.2019.9006122
[14] C. Tripp et al., “An empirical deep dive into deep learning’s driving dynamics,” arXiv, 2022. https://doi.org/10.48550/arXiv.2207.12547
[15] S. Chinta, “Advancements in deep learning architectures: A comparative study of performance metrics and applications in real-world scenarios,” SSRN Electronic Journal, 2025. https://doi.org/10.2139/ssrn.5046551
[16] X. Ma, G. Chen, H. Zhang, and T. Qin, “Towards hierarchically extracting network topologies,” pp. 386–392, 2024. https://doi.org/10.1109/nana63151.2024.00071
[17] N. Nurdianto, M. L. Singgih, and I. K. Gunarta, “Building composite performance index for broadband internet customer experience,” pp. 669–675, 2024. https://doi.org/10.1109/isct62336.2024.10791161
[18] D. C. Marcu and C. Grava, “The importance of data quality in training a deep convolutional neural network,” in International Conference on Engineering of Modern Electric Systems, pp. 1–4, 2023. https://doi.org/10.1109/EMES58375.2023.10171785
[19] F. Yu et al., “A survey of large-scale deep learning serving system optimization: Challenges and opportunities,” arXiv, 2021. Available: https://arxiv.org/abs/2111.14247
[20] F. Yu et al., “Characterizing and understanding deep neural network batching systems on GPUs,” BenchCouncil Transactions on Benchmarks, Standards and Evaluations, 2023. https://doi.org/10.1016/j.tbench.2024.100151
[21] O. Matsiievskyi et al., “Application of neural networks to optimize distributed computing in cloud and edge environments,” pp. 1–5, 2025. https://doi.org/10.1109/sist61657.2025.11139218
[22] I. Ferjani, M. S. Hidri, and A. Frihida, “Dynamic knowledge expansion: Real-time text classification with deep convolutional neural networks,” 2024. https://doi.org/10.21203/rs.3.rs-3826991/v1
[23] B. Özkilbaç, I. Y. Ozbek, and T. Karacali, “Real-time fixed-point hardware accelerator of convolutional neural network on FPGA based,” in International Conference on Computing and Information, pp. 1–5, 2022. https://doi.org/10.1109/icci54321.2022.9756093
[24] M. Z. Hussain et al., “A fine-tuned convolutional neural network model for accurate Alzheimer’s disease classification,” Scientific Reports, vol. 15, 11616, 2025. https://doi.org/10.1038/s41598-025-86635-2
[25] N. Jayakrishna and N. N. Prasanth, “A hybrid deep learning model for detection and mitigation of DDoS attacks in VANETs,” Scientific Reports, vol. 15, 34170, 2025. https://doi.org/10.1038/s41598-025-15215-1
[26] R. R. Singh, H. Fida, A. Kumar, and A. Kumar, “Leveraging CNNs for accurate weather prediction: A comparative analysis of custom and transfer learning models,” in Innovative Computing and Communications (ICICC 2025), LNNS, vol. 1431, Springer, Singapore, 2025. https://doi.org/10.1007/978-981-96-6681-2_39
[27] S. Sinha and Y. M. Lee, “Challenges with developing and deploying AI models and applications in industrial systems,” Discover Artificial Intelligence, vol. 4, p. 55, 2024. https://doi.org/10.1007/s44163-024-00151-2
[28] B. Roth et al., “Specification overfitting in artificial intelligence,” Artificial Intelligence Review, vol. 58, p. 35, 2025. https://doi.org/10.1007/s10462-024-11040-6
[29] B. Russell, “Understanding overfitting in AI and its impact on cybersecurity,” World Journal of Advanced Research and Reviews, vol. 27, no. 2, pp. 361–372, 2025. https://doi.org/10.30574/wjarr.2025.27.2.2859
[30] Z. Huang, N. Ma, S. Wang, and Y. Peng, “GPU computing performance analysis on matrix multiplication,” The Journal of Engineering, vol. 2019, no. 23, pp. 9043–9048, 2019. https://doi.org/10.1049/joe.2018.9178
[31] Y. Hong and D. Kim, “Performance and efficiency gains of NPU-based servers over GPUs for AI model inference,” Systems, vol. 13, no. 9, p. 797, 2025. https://doi.org/10.3390/systems13090797
