Research On Resource Scheduling Of Satellite Network Based On Edge Computing

The mobile Internet provides highly reliable and low-latency network services to users around the world across multiple areas,and its development is closely associated with the world’s progress.The rapid growth of data traffic and the newer iterations of smart devices have resulted in more applications for users,and with that comes an extremely large amount of data to be processed by computer systems.Therefore,cloud computing centers require more powerful capabilities to satisfy the demands,which pose significant challenges for communication networks.Mobile Edge Computing(MEC)can alleviate the pressure on links and cloud computing centers by deploying nodes with computing,storage,and network resources to transfer partial capabilities of cloud computing centers to sites closer to users.Restricted by technology and costs,terrestrial communication facilities fail to provide global coverage,while satellites can provide wider coverage and reliable network connections that are not subject to natural disasters and other similar impacts.With the rapid development of satellite technology in recent years,it enables the largescale deployment of fully functional satellite communication systems in the future to cover the shortage of terrestrial cellular communication systems,which will become an important component of communication systems in the future.As a key component to support future mobile communications,low earth orbit(LEO)satellite networks can also incorporate the advantages of mobile edge computing by deploying computing and storage resources on satellites to enhance the service capability of satellite nodes and provide computing services to users.Accordingly,the following innovative work has been carried out in this article for LEO satellite networks,with the main research efforts stated as follows.1.Design and realization of LEO satellite network platform based on edge computing:In the present study,an LEO satellite network platform with edge computing capability is proposed and designed based on virtualization technology and mobile edge computing system architecture.The platform achieves virtualized satellite nodes based on K3S,third-party applications are deployed in the platform using microservices to provide edge computing services externally.This platform deploys computing and storage resources,achieves resources scheduling,models the characteristics of the low-orbit satellite edge computing system,sets the parameters of the low-orbit satellite constellation,achieves task migration and computation offloading strategy.Given the mobility of satellites,i.e.,the ever-changing location of satellites,a user access and switching method is introduced to the platform.2.Task migration strategy based on task attributes:In light of the absence of high-quality services in certain cases due to resource and mobility constraints of LEO satellites,this paper investigates the task migration strategy in satellite edge computing scenarios and proposes a task migration strategy based on task attributes.Considering the computing resource situation and task attributes of satellite nodes,in the event of poor completion of user tasks by the access satellite,satellite resources are allocated to the tasks proposed by the user through decision-making,and the user’s tasks are migrated to a satellite with more computing power near the access satellite or the next access satellite for processing,which can effectively reduce the end-to-end delay,improve the user experience and the resource utilization efficiency of the system.Compared with migration strategy based on random,the task migration strategy based on task attributes reduces the delay by relatively 18%and improves the success rate by relatively 12%.3.Computation offloading of satellite edge computing systems:To address the need to improve system efficiency associated with the diversification of user tasks and user requirements,a reasonable offloading strategy can be adopted for divisible computational tasks to boost the efficiency of task processing and the resource utilization of satellite systems.In this paper,a computation offloading strategy based on an improved positive feedback genetic algorithm is proposed for the computation offloading issues of user tasks.Specifically,the task information and the information of the satellite nodes are collected,and an improved positive feedback genetic algorithm is used to optimize the feasible solution during the iterative process with the goal of the shortest user end-to-end delay to yield a better offloading decision,followed by the final offloading of the user task.The algorithm proposed in this paper can effectively reduce the latency generated by the computational task versus the reference method.Compared with computation offloading strategy based on random,computation offloading strategy based on an improved positive feedback genetic algorithm reduces the delay by relatively 26%.