Research On The Theory And Method Of High Performance Mobile Edge Computing With Cloud-edge-end Collaboration

With the rapid development of real-time Internet of Things(Io T)applications,exponential growth of smart terminals,and increasing computing power of terminals and edge nodes,mobile edge computing technology has emerged.Mobile edge computing can support the implementation of artificial intelligence algorithms such as deep learning and reinforcement learning at the edge of a network by extending computing and storage capabilities from a cloud data center to the edge of the network closer to the data source.This makes it possible to avoid the long network transmission delay when transmitting computing tasks from the edge of a network to a remote data center,and thus meet the demands of real-time Io T applications(e.g.,autonomous driving,drones,and augmented reality).Edge computing technology can also avoid security risks such as privacy disclosure caused by data transmission on public networks and data processing in public computing centers.Therefore,mobile edge computing has become a key technology for the development of next-generation networks.China attaches great importance to5G-based mobile edge computing,and emphasizes the promotion of pilot projects in the applications of mobile edge computing in civilian,commercial,industrial,and military areas.New features such as massive data,ultra-low delay,green computing,and data security also bring many challenges to mobile edge computing:(1)Rapid response to massive tasks.Cloud,edge,and end all are capable of processing tasks,but the capacity of a single node is limited;thus cloud,edge,and end collaborative processing is required;(2)Hierarchical and heterogeneous nature of resources.Wireless bandwidth resources are limited,local resources are insufficient,edge hotspots compete fiercely,cloud resources are remote,and system resources are difficult to effectively manage and utilize;(3)Balance between energy consumption and delay.Task scheduling is affected by energy consumption and also needs to take delay factors into account,which are difficult to coordinate and optimize at the same time;(4)Diversified and hybrid tasks.Tasks are huge in volume and diverse,and their.multiple characteristics such as priority and delay sensitivity are in hybrid form,which is difficult to effectively schedule;(5)Security and efficiency of data transmission.Wireless transmission is susceptible to illegal eavesdropping and attack.The capacity of wireless channels is limited.Data are difficult to transmit securely,reliably,and efficiently.The above challenges have severely restricted the development and application of mobile edge computing.In this paper,we focus on the new features of mobile edge computing and next-generation Io T applications.We combine the resources of mobile terminals,edge servers,and cloud computing centers;study the scheduling of massive computing tasks;reduce delay and energy consumption;improve wireless terminals endurance;and enhance system security and reliability.Our study achieves high performance mobile edge computing,and is of great significance for promoting the development of green edge computing and the industrial application of the real-time Io T.In this paper,we introduce the state of research of mobile edge computing technology,and summarize problems requiring solution in cloud-edge-end collaborative task scheduling.Based on game theory and stability theory,we study edge-end collaborative computation offloading and edge-cloud collaborative task scheduling,and propose cloudedge-end collaborative computing task scheduling focused on massive computing tasks,relaxing delay requirements,imprinting wireless terminal endurance,and balancing energy consumption and delay.Finally,to improve the security and reliability of wireless data transmission,we propose a mobile edge computing data transmission scheme based on physical layer security.The results and innovations of our study are as follows.For the joint decision-making of energy harvesting and computation offloading in the mobile edge computing environment,we build a mathematical model for wireless terminal energy harvesting and task computing.Taking energy and computing power as the constraints,and minimizing the average delay as the objective,we build an energy harvesting computation offloading game model.We theoretically analyze the existence of a Nash Equilibrium and the optimal response of this game,and the stability of the battery energy level of the energy harvesting terminal.On this basis,we propose an optimized energy harvesting algorithm based on Lyapunov drift methods,and an optimal offloading algorithm based on game theory,enabling us to optimize and balance the load of a mobile edge computing system to provide green and sustainable endurance capabilities for energy harvesting terminals and meet the demand for low-latency service for computing tasks.For the problem of collaborative task scheduling between edges and the central cloud,we build a mixed mathematical model of transmission,calculation,and energy consumption of divisible tasks with strict delay requirements and delay-sensitive indivisible tasks.With strict delay requirements and system stability as the constraints,and minimizing energy consumption as the objective,we theoretically analyze the relationship between the end-to-end strict delay guarantee of computing tasks and resource scheduling.We analyze the system stability characteristics based on Lyapunov stability theory,and propose an edge-cloud resource collaborative task scheduling algorithm that combines delay-aware computation offloading and dynamic queue scheduling,with a delay guarantee and minimum energy consumption.This algorithm provides a strict delay guarantee for delay-sensitive computing tasks while minimizing the energy consumption of the edge cloud computing system.Finally,to tackle the problem of data disclosure caused by eavesdroppers during data transmission in mobile edge computing,we construct a secure high-speed cached-aided multi-user relay network communication model with outdated channel state information,and propose two user selection criteria to maximize the channel gain of the direct main link and minimize the channel gain of the direct eavesdropping link.We derive the secrecy outage probability(SOP)of traditional network systems and cache-aided relay systems,and an asymptotic SOP of the two user selection criteria under a high signal-to-noise ratio(SNR)and main-to-eavesdropper ratio(MER).This study provides a theoretical basis for improving the reliability of the relay link and enhancing the security of the system,and ensures the secure transmission of data at the physical layer.