隨著網路的快速發展，越來越多網路設備連上網路，而每個設備服務又皆有其各自的網路傳輸之需求，若以傳統的網路架構進行網路傳輸，是無法妥善地針對不同的網路傳輸需求去進行資源的分配的，此時便需要透過網路切片(Network slicing) 的技術來實現差異化的網路傳輸服務。網路切片能將一條共用之網路線段切分成數個彼此獨立的邏輯拓樸，其藉由對各個切片進行客製化的網路設定，並令每個服務透過相應的網路切片進行傳輸，以此來針對不同的網路需求進行最佳的網路資源分配。 本研究針對網路切片資源分配問題提出顧客/服務混合式網路切片架構，其透過並存兩種傳統式網路切片架構：服務式網路切片架構(Service based slicing) 與顧客式網路切片架構(Customer based slicing) 來進行網路的資源管理，並結合本論文所提出之 Superior Decision for Resource Allocation(SDRA) 機制來對企業用戶之過往網路使用行為，透過回歸分析進行分析及預測，來改善傳統網路切片架構中資源使用率低落以及網路壅塞的問題。根據本論文之實驗數據顯示，透過本論文之機制進行網路資源之分配及管理，能有效改善傳統網路切片架構 58.7% 的網路資源使用率，以及 97.5% 的網路效能，其與同樣透過網路切片進行資源分配之相關研究相比，本論文提供了更佳且更為穩定的網路資源分配方案。 ;The rise of 5th generation mobile networks (5G) has driven the development of the Internet of Things (IOT) technology. When there are a large number of IOT devices connected to the Internet, and each service has their own network requirement, but the traditional network architecture cannot properly allocate resources for different network requirements. In this case, it needs to be implemented through the technology of Network slicing. Network slicing can divide a shared network into several independent logical topologies, by customizing the network settings for each slice and allowing each service to transmit through the corresponding network slice, to give the best allocation scheme of network resources for different network services. This study proposes a customer/service hybrid based network slicing architecture for network slicing resource allocation. It coexists with two traditional network slicing architectures: Service based slicing and customer based network slicing architecture to carry out network resource management. And this paper used hybrid based slicing architecture combined with the Superior Decision for Resource Allocation (SDRA) mechanism proposed in this paper to solve the network usage issue and network congestion in traditional network slicing architectures through regression analysis and prediction. According to the experiment result of this paper, Through the mechanism of this paper to allocate and manage network resources can effectively improve the network resource utilization rate of 58.7% of the traditional network slicing architecture and the network performance of 97.5%, and compared with the other researches on resource allocation through network slicing, the mechanism proposed in this paper provides a better resource allocation solution.
