Research On Intelligent Optical Device Design Software Based On Distributed Cluster

With the gradual maturity of high-precision etching technology,the processing technology of micro/nano-optical devices has reached nano-level precision,and the design of micro/nano-optical devices is developing in the direction of intelligence,miniaturization,rapidity and high integration.Traditional optical device design software only relies on local single-node server computing,and limited resources cannot meet the requirements of high-precision optical device design,and cannot achieve multi-user,multi-task and parallel optical device design scenarios.Therefore,this paper combines the increasingly mature cloud computing and cloud native technologies to develop an optical device intelligent design software that can make full use of distributed cluster computing resources.This thesis first analyzes the requirements of optical device design software,and divides the software system into front-end interface module,back-end control module,computing module,data persistence module and cloud platform module based on business requirements.While showing the overall architecture of the system,the adopted technical architecture is introduced,including SpringBoot architecture,Docker container technology,Kubernetes container orchestration system,MySQL database and CephFS distributed file system.Starting from the database table structure design and software interface design,the design concept of the optical device intelligent design software is expounded.After that,each module of the software system is specifically implemented according to the outline design of the optical device intelligent design software.While realizing the optical device intelligent design software and the optical device design cloud platform,this thesis also proposes a container scheduling strategy KSFC(Kubernetes Scheduling For Computing)based on the multicriteria decision analysis algorithm for the design characteristics of optical devices.Compared with other container scheduling strategies,the simulation results show that in the optical device design scenario,KSFC has better scheduling performance,which can improve the resource utilization rate of the optical device design cloud platform and improve the design efficiency of optical devices.Finally,the main contributions of this thesis are summarized,and the optimization space and research direction of future work are pointed out.