Research On Resource Allocation And Task Scheduling Method Based On Cloud And Edge Collaborative Computing Architecture

The pattern of manufacturing industry began to shift from productcentric to user-centric mode,as its manufacturing process is deeply integrated with industrial Internet of Things,edge computing and other technologies.Due to the complex business process of the large-scale manufacturing workshop,the relationship between "people","information system" and "physical system" needs to be coordinated.As a result,traditional resource allocation and task scheduling schemes are prone to load imbalance.The cloud and side collaboration model relies on rich cloud manufacturing scenarios as an effective solution paradigm to realize intelligent manufacturing and strengthen field-level control.It obtains the full life cycle information of physical devices through resource collaboration and data collaboration and manages and mines information in an efficient and low-cost way.To solve the problems of massive heterogeneous data processing and user-customized requirements under the cloud and edge collaborative computing model,this thesis proposes a resource allocation and task scheduling algorithms that can meet the requirements of low-latency,high concurrency and high-throughput.The work mainly includes the following aspects:1.An edge server resource allocation scheme for cloud and edge collaboration is proposed.Firstly,the edge server is divided into IOintensive virtual machines and CPU-intensive virtual machines by decoupling computing from storage;Secondly,for IO-intensive virtual machines,this thesis generates the maximum priority list to uniformly allocate hardware resources in real time according to the dimensional information of the three aspects above,namely the weight value of the type of data uploaded by the terminal device,the size of the pre-allocated space and the actual occupied space corresponding to each IO-intensive virtual machine in the computing node;For CPU-intensive virtual machines,update the resource allocation of CPU-intensive virtual machines based on the relationship between the number of virtual machines and the performance of the physical machine.Finally,the simulation and experiment are performed to compare and analyze the proposed resource allocation algorithms.2.A task scheduling scheme for cloud and edge collaboration is proposed.First,the method merges DAG on cloud server and edge server,then uses the critical path-based segmentation strategy to divide the tasks,effectively improving the accuracy of allocation,and then achieve load balancing through processor allocation.Finally,simulation experiments are performed to compare and analyze the proposed task scheduling algorithms.