發展結合物聯雲霧計算平台與異質生產設備之智慧化資安技術暨攻防演練場域驗證

The Study of Key Information Security Techniques and of Offensive and Defensive Verifications on Intelligent IoT Cloud-edge Computing Platform and Heterogamous Manufacturing Equipment

陳彥霖(Yen-Lin Chen) ； 楊士萱(Shih-Hsuan Yang) ； 陳金聖(Chin-Sheng Chen) ； 何昭慶(Chao-Ching Ho) ； 林志哲(Chih-Jer Lin) ； 王正豪(Jeng-Haur Wang) ； 曾釋鋒(Shih-Feng Tseng) ； 張世豪(Shih-Hao Chang) ； 劉佳明(Chia-Ming Liu) ； 余兆偉(Chao-Wei Yu)

《科儀新知》 234期 (2023/03) Pp. 28-43

摘要

本整合型研究將以工業4.0之發展情境中之製造工廠為例並加入IEC62443系列標準之資訊技術系統的安全標準為基礎，建構一個基於智動機電平台之智慧工廠，並將於工廠中可能存在的資安事件為樣本，並在臺北科技大學機械工程系自動加工機械工廠作為攻防驗證與IEC62443-2-4認證的試驗場域，為臺灣企業對於智動機電的實現與轉型提供一個完整的解決方案與示範樣本。在研究主軸上可分為三個總體研究目標：「IT（Information Technology）-應用層」、「CT（Communication Technology）-傳輸層」與「OT（Operation Technology）-感知層」。本研究預期工廠產線的加工設備運作資訊以及生產資料會由「OT-感知層」感測器在設備上進行來進行相關機電感測單機的擷取資料整合，並且再根據資料種類透過「CT-傳輸層」之閘道器實現異質網路將資料上傳至「OT-感知層」的「邊緣」端，以進行初步運算與分析管控；「邊緣」端主要工作為透過「OT-感知層」設備取得的影像與感測資訊，並進行資料的標示辨識與分析運算，然後再將分析後的模型透過高效率加密編碼的方式傳送至「IT-應用層」區域的霧端平台上。當「霧端」蒐集到資料後，會進行更進一步的學習與整合，並且透過學習衍生出AI的決策分析，最後再由「OT-感知層」針對相關工具機進行加工參數的最佳化決策，已進行加工設備的生產調整；由於此層資料甚為珍貴，故在資安部分，本研究會透過混沌加密演算法與3DES及傳輸層安全性協定（TLS）來確保資料傳遞的安全性，防止在交換資料時受到竊聽及篡改。最後經由「霧端」分析與學習後的模型及決策會與「OT-感知層」進行辨識模型與AI決策的模型更新與強化，最後再將相關控制指令傳送至「實體」設備上進行相關控制數據的修正，使得產線生產上得以最佳化最有效率，使得形成一個「IT-應用層」、「CT-傳輸層」與「OT-感知層」之間相輔相成的智動機電資安系統。

並列摘要

This study takes the manufacturing field in the development context of Industry 4.0 as an example and integrates key points in IEC62443 standards. We plan to build up a smart factory based on a smart-motor-electrical platform which takes the possible security incidents in the factory as a practical example and adopt the smart manufacturing factory of National Taipei University of Technology as a POC site for offensive and defensive verification on information security issues and gets the IEC62443-2-4 certification. The study includes three overall research goals: "IT-application layer", "CT transmission layer", and "OT-aware layer". This plan expects that the sensor machine inductive sensing "OT-sensing layer" will capture and integrate the operation information as well as production data of the factory production line, then is realized through the "CT transmission layer" gateway according to obtained data. The heterogeneous network uploads data to the "Edge" end of the "OT-sensing layer" for preliminary learning, analysis and control. The "edge" end mainly works with the image and sensor information obtained through the "OT-sensing layer" equipment. In addition, the proposed system will perform data label identification and analysis calculations, and then transmit the analyzed model to the fog terminal platform in the "IT-application layer" area through high-efficiency encryption and encoding. When the "Fog End" collects the data, it will conduct further learning and integration then derive the AI decision-making analytics. Because the manufacturing data is very precious, in the information security part, this plan will apply chaotic encryption algorithms, 3DES and the Transport Layer security (TLS) to ensure the security of data transmission and prevent eavesdropping and tampering when exchanging data. Finally, the model and decision after the analysis and learning of the "fog end" which will be updated and strengthened with the "OT-sensing layer" for the identification model and AI decision model. The relevant control commands will be sent to the "physical" device for relevant control. The correction of the data makes the production line optimized and most efficient, which means that the "IT-application layer", "CT transmission layer", and "OT-sensing layer" become a perfect cooperated intelligent electrical security system.