| With the rapid development and popularization of mobile Internet and Internet of Things,People have entered the era of Internet of Everything.With the rapid increase in the number of network edge devices,mass data is generated by the perception layer of the Internet of Things,which leads to a sharp increase in the load of the cloud computing network,resulting in a long network delay;Mobile edge computing is the key technology to improving the user experience of 5G network in the future,which is close to the data source,so that it can effectively reduce the network transmission delay.Edge computing will sink the business to the edge of the network,provide computing services and storage services,and provide task migration platform for users.Task migration technology can transfer complex tasks on mobile devices to remote edge server through wireless networks,rely on the rich computing resources of remote servers,complete computing tasks,and return the results to mobile devices,so as to solve the problems of inadequate computing power and limited battery capacity of mobile terminals.At present,the existing task migration strategy is to make the migration decision under the premise that the migration service node has been established.It does not take into account the scenarios when the multi service nodes are available,and most of the migration strategies do not take into account the complex situation of the multi dependency relationship within the mobile terminal application.The application scenarios of these strategies are relatively simple and cannot give full play to the characteristics and advantages of mobile edge computing.On the basis of the learning of cloud computing and mobile edge computing,this paper studies the two directions of task migration:the selection of task migration destination and how to optimize the migration according to the target of energy consumption optimization.First of all,from the perspective of reducing the delay and improving the resource utilization equilibrium,the issue of "where to migrate" is solved under the scenario of multi service nodes.A task migration destination selection strategy based on similarity between supply and demand and dynamic price model is proposed.The optimal migration location is selected in the mobile edge computing environment.The task is migrated to the optimal edge micro data center to improve the resource utilization and task throughput.The experimental results under the CloudAnalyst retrofit simulation platform show that this selection algorithm can effectively reduce the network delay and response time,and improve resource utilization and task throughput.Secondly,due to the current battery development speed of mobile devices is far behind the development speed of its processor and memory,it has become an important factor constraining the development of mobile terminal equipment,unable to meet the energy consumption requirement of emerging Internet of Things applications.Therefore,to solve the problem of "how to realize the low energy migration of complex dependency application",from the perspective of reducing energy consumption,based on the determination of task migration destination,for mobile terminal scenarios with complex multi-dependencies,a fine-grained directed acyclic graph task partition model is established based on the characteristics of mobile edge computing,and the relationships among the divided subtasks are analyzed to construct the minimization energy consumption problem under the execution time limit,then uses the genetic algorithm to find the optimal solution,and obtains the result of the migration decision for each sub task,that is,the optimal energy-saving migration plan for the entire mobile terminal application.The experimental results show that the fine-grained task migration strategy proposed in this paper makes full use of the advantages of mobile edge computing,and can effectively reduce the energy consumption of mobile terminal devices on the premise of meeting the task execution delay. |
PDF Full Text Request