| With the development of modern network technologies such as software-defined networks,network function virtualization,and mobile edge computing,the network architecture and service capabilities of wireless networks are undergoing revolutionary changes.In the existing 4G wireless system,the backhaul used to connect the wireless access network and the core network are facing the problem of heavy traffic,and the backhaul is becoming a bottleneck of the entire cellular system.Wireless system design goals need to include ultra-high reliability and ultra-low latency.As an architecture that provides path diversity and fast network reconfiguration,Mesh networks will likely significantly improve the efficiency of backhaul to meet these needs.After a detailed literature review,the first research content of this thesis explores the ultra-low latency edge network architecture.The thesis integrates the Mesh networks and SDN into an edge network architecture to efficiently support real-time communication applications with tight latency requirements.The edge network consists of a number of moderately sized Mesh components that are geographically adjacent.The Mesh components use a virtual switching mechanism internally to achieve an almost constant backhaul delay.The design uses the shortest path routing and non-rejected network traffic control mechanisms.Then the second study is a joint communication and computing design.The thesis considers a beyond 5G edge network architecture for applications such as autopilot,which are computationally intensive and delay sensitive applications.On the premise of meeting the user's timely experience and battery life expectancy,mobile users can offload complex algorithms and large amounts of data to Cloudlet servers deployed near the Mesh components.Computer simulation experiments show that the joint communication and computational design can improve the energy efficiency of the system under the constraint of time delay by using the SCA algorithm. |
PDF Full Text Request