Research And Implementation Of Task Management Strategy In Fog Workflow System

With the advent of the Internet of Things era,the access of mass end devices brings problems such as high latency and network congestion to traditional cloud computing,so cisco proposed a new computing model,that is,fog computing.The fog computing proposed by cisco in 2012 can effectively solve such problems,but the existence of various complex computing resources in the fog computing environment puts forward higher requirements for the offloading and scheduling of workflow tasks.If the solution is not proper,there will be the problems,such as high energy consumption of end device,high response delay of task,and unguaranteed QoS of users.In addition,there are dependencies between computing tasks in all kinds of workflow in reality,which are generally ignored in the current researches on the fog computing environment.Therefore,how to reasonably optimize the offloading and scheduling schemes of various tasks in the fog computing environment to fully reduce the energy consumption of end device while considering the inter-task dependency is a problem worth studying.Finally,in order to verify the effectiveness of the solution,a workflow system simulation platform in the fog computing environment is needed for verification,but no such simulation platform is available at present.The lack of workflow system simulation platform in fog computing environment will seriously affect the credibility and effectiveness of workflow task management experimental results.Therefore,how to design and develop a workflow system simulation platform in fog computing environment is another problem worth studying.In view of the above problems,this thesis has done the following two aspects of research work.On the one hand,this thesis studies how to reasonably optimize the offloading and scheduling schemes of various workflow tasks in the fog computing environment in order to fully reduce the energy consumption of end device.Firstly,an example is given to illustrate the importance of workflow task dependencies to task management in a fog computing environment.Then the task dependency graph model,energy consumption model and energy consumption transfer model are established according to the dependency relationship.Finally,the graph-based energy consumption optimization task management strategy,named Fog Workflow Task Management Strategy(FWTMS),is proposed.On the other hand,this thesis studies how to design and develop the workflow system simulation platform in the fog computing environment.Firstly,the demand analysis is made according to the preliminary research work,and then the structure of FogWorkflowSim is designed based on the workflow system.The system not only realizes the simulation of fog computing environment and workflow environment,but also integrates the task offloading and task scheduling modules in the task management module.And three kinds of common task management evaluation indexes are built into the evaluation index module.This thesis not only theoretically solves the problem of offloading and scheduling of workflow tasks in the fog computing environment,proposes a graph-based energy optimization task management strategy FWTMS,but also solves the lack of workflow simulation system in the fog computing environment from simulation experiments problems,designs and implements a simulation experiment platform for workflow in a fog computing environment FogWorkflowSim for the first time.This thesis provides an idea and an algorithm simulation tool for the task management of fog workflow system.