| 摘要: | 隨著行動雲端、巨量資料與寬頻網路技術快速演進，促使智慧聯網應用的快速發展，萬物互聯的物聯網時代來臨，所牽涉資料的移動、儲存、處理、分析需求，既有的雲端服務模式已無法因應服務龐大且分散的用戶需求。此外，傳統電信網路配置和管理的方法，也無法滿足即時資料處理和服務管理的需求。為了解決傳統網路的發展瓶頸，軟體定義網路（Software Defined Network, SDN）以及網路功能虛擬化（Network Functions Virtualization, NFV）技術的興起，將現今複雜的網路架構轉變成虛擬化、可程式化與標準化的開放架構，不僅帶動ICT產業的發展變革，也催生了邊緣運算（Edge Computing）、軟體定義廣域網路（SD-WAN）甚至是機器學習（Machine Learning）跨領域技術應用的結合，為未來5G微營運商（Micro Operator）建構智慧城市聯網應用的發展基礎。本計畫分三年進行，第一年度主要研究目標是邊緣網路訊務快篩機制並應用機器學習技術分類訊務流量，可提供資料預先處理，降低訊務流量與強化通訊安全。第二年延續第一年的成果，針對邊緣網路訊務流量變動進行控制與管理，提供按需（On-Demand）的邊緣運算服務並搭配SD-WAN廣域網路技術即時判斷最佳傳輸路徑，進行連外網路的即時分配傳遞，可提昇QoS服務品質與傳輸效能。第三年將整合第一、二年度的成果，以多雲邊緣（Cloudy Edge）協同運作與自動化管理為研究重點，活化跨邊緣網路中微營運商資料中心資訊共享，實現雲霧運算自動化管理的服務模式，期能作為未來5G微營運商微型電信服務管理的參考依據。此外，計畫主持人深耕網路服務管理技術多年，發表國內外論文兩百餘篇、擁有二十餘件專利，參加國際競賽更是屢獲肯定。計畫主持人自2009年便開始SDN/OpenFlow領域之研發，有執行開發型產學合作計畫之實績，成果已獨家技術移轉予廠商，且榮獲科技部頒發海報展示傑出獎。研發過程中使用NTT開發的Ryu SDNFramework 開放源碼發現系統上的BUG，並被NTT證實，且提供在3.7版本中修改。加之於2015與2017通訊大賽-SDN創新應用競賽中榮獲『兩次冠軍』，也是歷來唯一榮獲兩次冠軍之研究團隊，顯見主持人之研究實務成果已受產業界的肯定。 ;With the rapid development of mobile, cloud, big data and broadband network technologies, the increasing of networking applications has driven the era of Internet of Things (IoT). The IoT system included mobility, storage, processing and analysis of data, the existing cloud services has been unable to satisfy the large and scattered services. In addition, traditional telecom network configuration and management methods cannot meet the requirements of real-time data processing and service management. In order to solve the bottleneck of traditional networks, the emergence of Software Defined Network (SDN) and Network Functions Virtualization (NFV) technology turns the complicated network architecture into a virtual and programmable network. The SDN/NFV not only brings significant change the development of ICT, but also spawns the combination of Edge Computing, SD-WAN and even Machine Learning. The development of these emerging technologies also forms the basis for the future development of smart city networking applications for the 5G micro-operators.This project is over three years. In the first year, the goal is utilizing SDN/NFV technologies to quickly inspect the important traffic information and the machine learning technology is used to classify the traffic flow that provide data pre-processing, reduce traffic flow and enhance communications security in the edge network. The second year continues the results of the first year. To control and manage the traffic changes on the edge network, provide On-Demand edge computing services and match the SD-WAN WAN technology to decide the optimal transmission path and transmit the real-time traffic of the external network to improve the QoS and transmission efficiency. The third year will integrate the results of the first and second year that focus on the Cloudy-Edge collaborative operation and automated management. Study on resource sharing among micro data centers across different edge networks and realize the cloud-fog computing architecture to automate service management in future 5G operator micro-telecom services. In addition, the project host studied the network service management technology for many years, received awards in many international competitions, published two hundred of papers, and has more than 20 patents. Since the beginning of SDN/OpenFlow research in 2009, the team has a lot of experience in implementation of cooperation plans between industry and school. The results have been transferred to vendors. They also won a Poster Award of Excellence issued by Ministry of Science and Technology. The team issued a bug report of the Controller Ryu confirmed by NTT and corrected in version 3.7, and also obtained twice champion in 2015 and 2017 Communications Contest of SDN Innovative Application. It is obvious that the host’s research ability and achievement has been affirmed by industrial circles. |
