ETRI Journal

  • 제46권2호
  • /
  • Pages.205-217
  • /
  • 2024
  • /
  • 1225-6463(pISSN)
  • /
  • 2233-7326(eISSN)
한국전자통신연구원 (Electronics and Telecommunications Research Institute)
권호 표지
DOI QR코드

DOI QR Code

Bi-LSTM model with time distribution for bandwidth prediction in mobile networks

  • Hyeonji Lee (Department of Computer Engineering, Changwon National University) ;
  • Yoohwa Kang (Network Research Division, Electronics and Telecommunications Research Institute) ;
  • Minju Gwak (Department of Computer Engineering, Changwon National University) ;
  • Donghyeok An (Department of Computer Engineering, Changwon National University)
  • 투고 : 2022.12.16
  • 심사 : 2023.03.27
  • 발행 : 2024.04.20
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

We propose a bandwidth prediction approach based on deep learning. The approach is intended to accurately predict the bandwidth of various types of mobile networks. We first use a machine learning technique, namely, the gradient boosting algorithm, to recognize the connected mobile network. Second, we apply a handover detection algorithm based on network recognition to account for vertical handover that causes the bandwidth variance. Third, as the communication performance offered by 3G, 4G, and 5G networks varies, we suggest a bidirectional long short-term memory model with time distribution for bandwidth prediction per network. To increase the prediction accuracy, pretraining and fine-tuning are applied for each type of network. We use a dataset collected at University College Cork for network recognition, handover detection, and bandwidth prediction. The performance evaluation indicates that the handover detection algorithm achieves 88.5% accuracy, and the bandwidth prediction model achieves a high accuracy, with a root-mean-square error of only 2.12%.

키워드

과제정보

This work was supported by the Institute of Information & Communications Technology Planning & Evaluation (IITP) under grant funded by the Korean government (MSIT) (No. 2020-0-00974, Development of Ultra-reliable and Low-Latency 5G+ Core Network and TSN Switch Technologies).

참고문헌

  1. Percentage of mobile device website traffic worldwide from 1st quarter 2015 to 2nd quarter 2022. https://www.statista.com/statistics/277125/share-of-website-traffic-coming-from-mobiledevices/ Accessed: 2022-10-06.
  2. Consumer internet data traffic worldwide by application category from 2016 to 2022. https://www.statista.com/statistics/454951/mobile-data-traffic-worldwide-by-application-category/ Accessed: 2022-10-06.
  3. J. Kua, G. Armitage, and P. Branch, A survey of rate adaptation techniques for dynamic adaptive streaming over http, IEEE Commun. Surv. Tutorials 19 (2017), no. 3, 1842-1866.
  4. S. E. Co, 5G vision, white paper, 2015.
  5. GSM Association, The mobile economy 2022, 2022.
  6. E. Ko, D. An, I. Yeom, and H. Yoon, Congestion control for sudden bandwidth changes in TCP, Int. J. Commun. Syst. 25 (2012), no. 12, 1550-1567.
  7. P. Li, X. Jiang, G. Jin, Y. Yu, and Z. Xie, ALSTM: An attention-based LSTM model for multi-scenario bandwidth prediction, (IEEE 27th International Conference on Parallel and Distributed Systems (ICPADS), Beijing, China), 2021, pp. 98-105.
  8. S. Goutam and S. Unnikrishnan, Decision for vertical handover based on naive Bayes algorithm, (International Conference on Advances in Computing, Communication and Control (ICAC3), Mumbai, India), 2019, pp. 1-6.
  9. Z.-H. Huang, Y.-L. Hsu, P.-K. Chang, and M.-J. Tsai, Efficient handover algorithm in 5G networks using deep learning, (Globecom 2020-2020 IEEE Global Communications Conference, Taipei, Taiwna), 2020, pp. 1-6.
  10. J. Lima, A. Medeiros, E. Aguiar, V. A. D. S. Junior, and T. Guerra, User-level handover decision making based on machine learning approaches, J. Commun. Inform. Syst. 37 (2022), no. 1, 104-108.
  11. N. Han, S. Qiao, D. Liu, P. Ding, Y. Zhang, X. Xiong, M. Wang, and L. A. Gutierrez, Handover detection approach based on trajectory data mining techniques, J. Eng. 2018 (2018), no. 16, 1534-1537.
  12. T. Huang and J. Subhlok, Fast pattern-based throughput prediction for TCP bulk transfers, (CCGRID 2005. IEEE International Symposium on Cluster Computing and the Grid, Cardiff, UK), 2005, pp. 410-417.
  13. J.-H. Hwang and C. Yoo, Formula-based TCP throughput prediction with available bandwidth, IEEE Commun. Lett. 14 (2010), no. 4, 363-365.
  14. M. Mirza, J. Sommers, P. Barford, and X. Zhu, A machine learning approach to TCP throughput prediction, IEEE/ACM Trans. Netw. 18 (2010), no. 4, 1026-1039.
  15. J. Padhye, V. Firoiu, D. F. Towsley, and J. F. Kurose, Modeling TCP reno performance: a simple model and its empirical validation, IEEE/ACM Trans. Netw. 8 (2000), no. 2, 133-145.
  16. N. Parvez, A. Mahanti, and C. Williamson, An analytic throughput model for TCP new reno, IEEE/ACM Trans. Netw. 18 (2009), no. 2, 448-461.
  17. L. Mei, R. Hu, H. Cao, Y. Liu, Z. Han, F. Li, and J. Li, Realtime mobile bandwidth prediction using LSTM neural network and Bayesian fusion, Comput. Netw. 182 (2020), 107515.
  18. B. Wei, W. Kawakami, K. Kanai, J. Katto, and S. Wang, TRUST: a TCP throughput prediction method in mobile networks, (IEEE Global Communications Conference (GLOBECOM, Abu Dhabi, United Arab Emirateds), 2018, pp. 1-6.
  19. N. Nayakwadi and R. Fatima, Machine learning based handover execution algorithm for heterogeneous wireless networks, (Fifth International Conference on Research in Computational Intelligence and Communication Networks (ICRCICN), Bangalore, India), 2020, pp. 54-58.
  20. D. Yun, W. Liu, C. Q. Wu, N. S. V. Rao, and R. Kettimuthu, Performance prediction of big data transfer through experimental analysis and machine learning, (IFIP Networking Conference (Networking), Paris, France), 2020, pp. 181-189.
  21. UCC MISL 5G dataset. https://github.com/uccmisl/5Gdataset Accessed: 2022-10-25.
  22. D. Raca, D. Leahy, C. J. Sreenan, and J. J. Quinlan, Beyond throughput, the next generation: a 5g dataset with channel and context metrics, (Proceedings of the 11th ACM multimedia systems conference, Istanbul, Turkey), 2020, pp. 303-308.
  23. G-NetTrack Pro. https://gyokovsolutions.com/manual-gnettrack/ Accessed: 2023-3-13.
  24. N. S. Keskar, D. Mudigere, J. Nocedal, M. Smelyanskiy, and P. T. P. Tang, On large-batch training for deep learning: generalization gap and sharp minima, arXiv preprint, 2016. https://doi.org/10.48550/arXiv.1609.04836
  25. D. Masters and C. Luschi, Revisiting small batch training for deep neural networks, arXiv Preprint, 2018. https://doi.org/10.48550/arXiv.1804.07612
  26. 3GPP TS 36.214 evolved universal terrestrial radio access (e-utra); physical layer; measurements (version 14.2.0 release 14), 2017. 3GPP.
  27. J. Um, I. Kim, and S. Park, Implementation of platform for long-term evolution cell perspective resource utilization analysis, ETRI J. 43 (2021), no. 2, 232-245.