Please use this identifier to cite or link to this item:

Title: 以卷積神經網路優化5G時代下智慧家庭的服務流量分類
Using CNN to Optimize Traffic Classification for Smart Homes in 5G era
Authors: 蔡宗諺
Tsai, Tsung-Yen
Contributors: 張宏慶
Jang, Hung-Chin
Tsai, Tsung-Yen
Keywords: 第五代行動通訊網路
Smart Home
Traffic Classification
Deep Learning
Software Defined Networking (SDN)
Date: 2021
Issue Date: 2021-11-01 12:00:30 (UTC+8)
Abstract: 近年來,隨著物聯網及人工智慧技術的迅速發展與進步,愈來愈多業者將住宅結合新興科技打造智慧家庭,以提升住戶的生活品質。因此在未來的智慧家庭中,許多類似5G三大應用場景特性的服務將應用於不同種類的智慧裝置,智慧家庭的整體網路流量必然大量增加,使智慧家庭中的網路流量管理成為值得深入探討的議題。由於5G時代的網路流量大幅增加與網路加密技術的廣泛使用,無法輕易從大多數網路應用服務中解密流量取得資訊,更無法透過傳統的網路流量分類方法將各類服務流量進行分類,加以發送到對應的應用類別進行管理。
本論文藉由軟體定義網路技術模擬多租戶的智慧家庭環境,依據3GPP LTE QoS Class Identifier (QCI)表,篩選出適用於未來智慧家庭類別的服務,模擬不同類別的智慧家庭服務流量,並利用卷積神經網路對網路流量進行分類。透過本論文,ISP業者能依分類好的服務類別,設定頻寬比例並配置到對應的服務類別,達到有效提升QoS及使用者QoE的目的。實驗結果顯示,CNN模型對智慧家庭模擬流量的分類精準度,透過調整後的參數組合與設定大小為1500 bytes的Payload輸入,能有最佳的分類準確率86.5%,相較一般神經網路模型準確率提升了6.5%。
In recent years, with the rapid development and progress of the Internet of Things and artificial intelligence technologies, more builders have combined housing with emerging technologies to create smart homes to improve the quality of life of residents. Therefore, there are many services similar to the three major application scenarios of 5G will be applied to different types of smart devices in the future. The network traffic of the smart home will increase significantly, network traffic management in the smart home became a significant topic to be explored. Due to the substantial increase in network traffic in the 5G era and the widespread use of network encryption technology, it is impossible to decrypt traffic easily and classify various traffic through traditional network traffic classification methods.
In this research, we use SDN technology to simulate a multi-tenant smart home environment. According to the 3GPP LTE QCI table, we select services suitable for smart home categories to simulate different types of smart home service traffic, and using CNN to classify network traffic. Through this research, ISP operators can set the bandwidth ratio according to the classified service category and configure it to the corresponding service category, so as to achieve the purpose of effectively improving QoS and QoE. The experimental results show that the CNN model has the best classification accuracy rate of 86.5% through the adjusted parameter combination and setting the payload input with a set size of 1500 bytes, compared with the general neural network model. The accuracy rate has increased by 6.5%.
Reference: [1] W. Chen, X. Fan and L. Chen, "A CNN-based Packet Classification of eMBB, mMTC and URLLC Applications for 5G," International Conference on Intelligent Computing and its Emerging Applications (ICEA), Tainan, Taiwan, pp. 140-145, 2019.
[2] W. -J. Eom, Y. -J. Song, C. -H. Park, J. -K. Kim, G. -H. Kim and Y. -Z. Cho, "Network Traffic Classification Using Ensemble Learning in Software-Defined Networks," 2021 International Conference on Artificial Intelligence in Information and Communication (ICAIIC), pp. 089-092,2021
[3] H. Y. He, Z. Guo Yang and X. N. Chen, "PERT: Payload Encoding Representation from Transformer for Encrypted Traffic Classification," 2020 ITU Kaleidoscope: Industry-Driven Digital Transformation (ITU K), pp. 1-8, 2020
[4] H. Jang and J. Lin, "SDN based QoS aware bandwidth management framework of ISP for smart homes," 2017 IEEE SmartWorld, Ubiquitous Intelligence & Computing, Advanced & Trusted Computed, Scalable Computing & Communications, Cloud & Big Data Computing, Internet of People and Smart City Innovation (SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/SCI), pp. 1-6, 2017.
[5] H. Lim, J. Kim, J. Heo, K. Kim, Y. Hong, and Y. Han, "Packet-based Network Traffic Classification Using Deep Learning," 2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC), Okinawa, Japan, pp. 46-51, 2019.
[6] E. Nazarenko, V. Varkentin and A. Minbaleev, "Application for Traffic Classification Using Machine Learning Algorithms," 2020 International Conference Quality Management, Transport and Information Security, Information Technologies (IT&QM&IS), pp. 269-273,2020
[7] S. Rezaei and X. Liu, "Deep Learning for Encrypted Traffic Classification:
An Overview," IEEE Communications Magazine, vol. 57, no. 5, pp. 76-81, May 2019.
[8] W. Wang, M. Zhu, J. Wang, X. Zeng and Z. Yang, "End-to-end encrypted traffic classification with one-dimensional convolution neural networks," 2017 IEEE International Conference on Intelligence and Security Informatics (ISI), Beijing, China, pp. 43-48, 2017.
[9] SDN Three Tier Architecture,,
retrieved date: 2021/2/10.
[10], retrieved date: 2021/2/10.
retrieved date: 2021/2/10.
[12] J. H. Cox et al., "Advancing Software-Defined Networks: A Survey," in IEEE Access, vol. 5, pp. 25487-25526, 2017.
[13] ITU-R Recommendation M.2083-0, "IMT Vision – Framework and overall. objectives of the future development of IMT for 2020 and beyond," Sep. 2015,!!PDF-E.pdf, retrieved date: 2021/2/10.
[14] Jan, Qasim and Shahzad Latif. “IP MULTIMEDIA SUBSYSTEM (IMS) SECURITY MODEL.” (2013), retrieved date: 2021/2/10.
[15] 3GPP TS 23.203 Standard, "Policy and Charging Control Architecture," retrieved date: 2020/11/10.
[16] Network Traffic Classification using Deep Learning,,retrieved date: 2020/11/10.
[17] Global Smart Home Market Still Growing,, retrieved date: 2020/11/10.
[18], retrieved date: 2021/2/10.
retrieved date: 2021/2/10.
[20] ITU vision on 5G usage scenarios. [Online]. Available:!!PDF-E.pdf, retrieved date: 2021/2/10.
[21], retrieved date: 2021/2/10.
[22], retrieved date: 2021/2/10.
Description: 碩士
Source URI:
Data Type: thesis
Appears in Collections:[資訊科學系] 學位論文

Files in This Item:

File Description SizeFormat
320501.pdf7998KbAdobe PDF0View/Open

All items in 學術集成 are protected by copyright, with all rights reserved.

社群 sharing