Research On Task Scheduling Optimization Of Workflow System In Fog Computing Environment

In order to alleviate the huge pressure faced by cloud data centers when they process massive Internet data,such as excessive communication load,unpredictable delay and insufficient support for the mobility of users,a new computing mode called"fog computing"was proposed by Cisco.In fog computing,fog nodes can realize low-latency interaction at the edge of the network and reduce the response delay effectively.Cloud data centers can provide centralized processing resources with stronger computing capacity and larger storage space,so that it can provide rapid computing.With fog computing commercialization gradually expand,fog node service with its infrastructure of the advantages of low price,simple to deploy and maintain attracted the attention of the user.The use of only fog nodes resources can save task execution cost to a large extent,but it needs to pay a lot of time cost,which affects the user's service experience.Collaborative use of fog-cloud multi-layer resources to provide services can reduce the cost without affecting the service experience.Workflow system can manage complex fog-cloud multi-layer resource allocation and optimize task scheduling according to users' needs.How to reasonably use fog-cloud multi-layer resource to provide services satisfying the Quality of Service(QoS)requirements for tasks of workflow system is worth studying.Due to the mobility of the user terminal in the actual scenario and the limited range of services that the fog node can provide,the change of user's location will cause the change of the fog node that can be connected to.To some extent,this limits the range of fog nodes that can be used in fog computing.Although task transmission to fog node can get response quickly,it is slow and inefficient in execution.Although user location change will affect the request fog node resource,it has no impact on the request cloud resource.In the process of user movement,collaborative use of fog-cloud resource to perform tasks can ensure that tasks can timely request to the resource.On the one hand,this thesis studies how to utilize the fog-cloud multi-layer resources to provide services satisfying the QoS requirements for tasks in the fog computing.Combined with Particle Swarm Optimization(PSO)and Min-Min algorithm,this thesis propose a Cost-effective Time-constrained task scheduling Strategy using fog-cloud multi-layer resources in Fog computing,denoted as CTSF.On the other hand,considering that the user location change will limit the scope of connecting to the fog node when the task requests resources,it will have an impact on the QoS of task.Based on the CTSF strategy,this thesis propose a Cost-effective Time-constrained task scheduling Strategy using fog-cloud multi-layer resources in Location-aware Fog computing,denoted as CTSLF.The main work of this thesis is as follows:1.Since the use of single layer fog resources cannot provide users with services that meet the requirements of low delay and low cost at the same time,this thesis studies how to coordinate the use of fog-cloud multi-layer computing resources in task scheduling of workflow system.In task scheduling,PSO algorithm is used to optimize the scheduling scheme and the Min-Min algorithm is used to solve the resource conflict problem in the scheduling scheme.In this thesis,the task scheduling algorithms is combined with the time model and cost model of task execution,and proposes a CTSF strategy.The experiment compares and analyzes the proposed strategy from the aspects of fitness value,execution cost,execution time,computation time and communication time,and the results show that the proposed scheduling strategy can produce an optimal scheduling scheme that meets the deadline constraint and reduce costs.2.Due to the mobility of the user terminal in the actual scenario,this thesis further considers how to use fog-cloud collaborative computing resources to provide services satisfying the QoS requirements for tasks in the process of user movement and proposes CTSLF strategy based on the study of CTSF strategy.In CTSLF strategy,the location of the user request resources determines which fog nodes can be connected to the task.Under the influence of the precursor task in the workflow,the time when the same task requests resources is different in different scheduling schemes,so the thesis designs the method of how to locate the users' location when the task requests the resources and how to determine the scope of service resources that the task can connect.The CTSLF strategy was compared and analyzed in terms of fitness value,execution cost,execution time,computation time and communication time in the experiment.The results show that the CTSLF strategy can generate scheduling schemes that meet the deadline constraints and optimize costs.Because the services provided by the single layer fog resources for users cannot meet the users'QoS requirements(such as low delay,low cost,etc.),this thesis study how to manage the fog-cloud multi-layer resources to provide tasks with services satisfying QoS requirements.Based on the PSO algorithm and the Min-Min algorithm,the thesis designs a CTSF strategy that can generate the scheduling scheme satisfying the deadline constraints and optimizing costs.On this basis,considering the user's mobility in the actual scenario,this thesis studies how to manage the fog-cloud multi-layer resources to provide tasks with services satisfying QoS requirements in the case of user movement,and CTSLF is designed based on CTSF strategy.The effectiveness of two strategies in this thesis is illustrated through experiments finally.