Research On Computing Offloading Strategy In Cloud And Fog Collaboration

As more and more compute-intensive applications execute on mobile terminals and data traffic grows explosively,mobile communication networks need to find a new way to meet the increasing demand for services.Cloud computing and mobile communication networks are combined to form a new computing model,namely mobile cloud computing,which utilizes the powerful computing power and rich resources of cloud computing to solve the shortcomings of mobile terminal computing,storage and battery energy consumption.As more applications are offloaded to the cloud,the offloading distance to the cloud is larger and the likelihood of network congestion increases as the access data increases,The tasks can not be completed within the time-delay tolerance range of applications with time requirements.In order to solve these problems,fog computing is proposed as an extension of cloud computing.The fog computing system is composed of various types of computers,mobile terminal users,switches and other devices with relatively weak computing energy.Computing offloading is the main service delivery mode of mobile terminals and cloud or fog.It solves the limitations of mobile terminals due to their own computing power and storage space.It takes a lot of time and energy to process computationally intensive applications.The mobile terminal offloads the computing task to the cloud or the fog,which has energy cost and time cost,and needs to propose a suitable offloading strategy to maximize the revenue obtained by the mobile terminal.At present,research on cloud collaboration usually has only one fog node and one cloud node.However,in practical applications,there is not only one fog node around the mobile terminal user.This paper mainly studies the computational offloading technology of cloud collaboration,how to use these fog nodes efficiently,and further reduce the delay and energy consumption.The main research work and innovations in this paper are as follows:(1)A cloud-fog collaborative computing offload strategy based on Markov decision process is proposed.Considering the offloading to the fog node,the number of available fog nodes in the vicinity is sensed,and the offloading task is divided into mutually independent computing task groups,which are respectively transmitted to the fog node,so that the transmission delay and the calculation delay are further reduced.The Markov decision model is established and the optimal strategy is obtained through iteration.Finally,the algorithm is simulated to prove that the algorithm not only reduces energy consumption and delay,also maximizes total revenue.(2)An offloading strategy based on artificial fish swarm algorithm is proposed.Under the network multi-cloud,multi-fog and multi-mobile terminals,the delay and energy consumption required by multi-users to perform offloading tasks are analyzed,the total energy consumption of computing tasks is minimized,and the penalty function is introduced to optimize the constraint items.The problem turns into an unconstrained target model.Then the objective function is optimized by improving the artificial fish swarm algorithm.Finally,the algorithm is simulated to prove that the algorithm can effectively reduce the energy consumption in the time delay tolerance,and prove the convergence of the algorithm and analyzes the factors affecting convergence.