Design And Implementation Of Distributed Management And Scheduling System For Heterogeneous Resources

With the accelerated pace of digital transformation and intelligent upgrading of the economy and society,the use of advanced concepts and technologies to improve the information level of mine water inrush perception is one of the hot spots of mine water prevention and control research in the new era.To support the formation of early warning system structure for mine water inrush prevention and control,a new mine information infrastructure based on IoT is essential.However,as the number of resource units accessed in mine IoT continues to increase,the control and scheduling of various resources is becoming more and more complex.The new mine information infrastructure architecture is faced with the following problems: the lack of standardized packaging and access to a large number of heterogeneous information resources and sensors and other equipment resources;the complex hierarchical relationship of various heterogeneous resources,which makes it difficult to model and manage them uniformly;and the lack of means to control and schedule resources on demand.In order to meet the capability requirements of the new mine water inrush perception system,the thesis starts from the control of heterogeneous resources in mine IoT,studies the distributed management and scheduling methods of heterogeneous resources,proposes a framework for distributed management and scheduling system of heterogeneous resources based on microservices and lightweight virtualization container technology,designs and implements a distributed scheduling system that supports standardized encapsulation and access of heterogeneous resources,models multi-level heterogeneous resources with multiple granularities,and supports visualization of heterogeneous resources.To ensure system performance and reliability,and to achieve better results in cluster load balancing,a Kubernetes resource scheduling algorithm based on improved genetic algorithm is applied for container scheduling.The main work of the thesis is summarized as follows.(1)System requirement analysis.According to the new mine information infrastructure heterogeneous resource management and scheduling requirements,it is analyzed in four aspects: standardized encapsulated access to heterogeneous resources,multi-granularity modeling control of system resources,task-oriented process construction method,distributed control of resources and on-demand scheduling,describing the specific interaction process between the system and users,and the functions provided by each module of the system to users.UML modeling is used to provide a detailed description of the functions of resource access,resource modeling,task construction,and task execution in the system,and the non-functional requirements to be possessed by the system are also analyzed.(2)System design and implementation.The overall architecture of the heterogeneous resource distributed management and scheduling system is designed,and the functions of each module are implemented.The overall architecture of the system is drawn,the process and model building of heterogeneous resource packaging and access and the process of task workflow construction and execution are described in layers,and the specific implementation process of each function is described in detail through flowcharts,class diagrams and timing diagrams.A Kubernetes resource scheduling algorithm based on an improved genetic algorithm is used in the system,which is optimized in terms of performance while ensuring the load balancing of the system cluster.Based on the analysis of the requirements of the distributed management and scheduling system for heterogeneous resources,the entities in the system are identified and the detailed design of the database is completed.The physical table structure in the database is described,and the meaning and type of each field in the table is explained.(3)System testing.After completing the coding and deployment of the system,the testing environment of the system was deployed in the laboratory,test cases were designed for each functional module of the system,detailed functional tests were conducted,and whether the response time of each function met the requirements and whether the algorithms used improved the functions.Through the experiments,it was found that the system’s logic of each function is correct,and the response time meets the pre-requirements,which is within the acceptable range for users and can be used normally.At the later stage,the system is online,each function is running normally,the page display is normal,and all the user requirements are satisfied,so it can be put into use.In summary,the heterogeneous resource distributed management and scheduling system is of great significance to the construction of the early warning system structure of mine water inrush prevention and control based on the Internet of Things in mines and provides effective basic technical support.