A visual integrated management system based on supercomputing platform

Abstract

The present invention provides a visual integrated management system based on a supercomputing platform, which relates to the field of supercomputing. It includes the following specific steps: S1: multi-layer architecture design 1.1: user interface layer: adopt modern front-end development technology to build a responsive, interactive and friendly user interface; the interface is divided into clear areas according to different functional modules, such as partition management area, node management area, resource management area, job management area and file management area; each area has a specially designed interactive component to realize the corresponding function; 1.2: business logic layer: realize business logic processing based on the back-end programming language; this layer is responsible for receiving the user's operation request on the WEB visualization page, and performing legality verification, business rule processing and data conversion. By providing a convenient operation method for supercomputing platform administrators, users and operation and maintenance personnel, simple operation of the supercomputing platform can be realized through the WEB page, reducing the difficulty of operation and the cost of use.

Description

Technical Field

[0001] The present invention relates to the field of supercomputing technology, and in particular to a visual integrated management system based on a supercomputing platform. Background Art

[0002] In the current supercomputing field, supercomputing platforms based on the Slurm scheduling system are widely used. However, these platforms have significant limitations in operation and management. Traditionally, when using such supercomputing platforms, platform administrators, users, and operation and maintenance personnel mainly rely on logging into the underlying operating system and performing various operations through command lines. This operation method has high technical requirements for operators, requiring them to be proficient in various complex commands and parameters, and has high requirements for the accuracy and standardization of operations. Once an operation error occurs, it may lead to serious consequences such as system failure and data loss.

[0003] In addition, the command line operation mode has a low level of visualization, and it is difficult for users to intuitively understand key information such as the system's operating status, resource allocation, and job execution progress. This not only increases the difficulty of operation, but also reduces work efficiency. For platform administrators, it takes a lot of time and energy to analyze and understand the results of command line output to perform complex tasks such as partition management, node management, and resource management. For users, submitting jobs and monitoring job status have also become very cumbersome, and it is impossible to obtain job execution status and resource usage in a timely manner. For operation and maintenance personnel, when performing system monitoring and troubleshooting, there is a lack of intuitive interfaces and effective tools, making it difficult to quickly locate and solve problems.

[0004] This traditional supercomputing platform operation method has largely limited the wide application and efficient use of supercomputing platforms, and cannot meet the needs of modern scientific computing and data processing for efficient, convenient, and visual management. With the continuous expansion of supercomputing application scenarios and the increasing diversification of user groups, a more convenient, efficient, and visual supercomputing platform management method is urgently needed. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the deficiencies in the prior art, the present invention provides a visual integrated management system based on a supercomputing platform, which solves the problem that serious consequences such as system failure and data loss may occur if an operation error occurs.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: A visual integrated management system based on a supercomputing platform, comprising the following specific steps:

[0009] S1: Multi-layer architecture design

[0010] 1.1: User interface layer: Use modern front-end development technology to build a responsive and interactive user interface; the interface is divided into clear areas according to different functional modules, such as partition management area, node management area, resource management area, job management area and file management area; each area has specially designed interactive components to realize the corresponding functions;

[0011] 1.2: Business logic layer: Implement business logic processing based on the back-end programming language; this layer is responsible for receiving user operation requests on the WEB visualization page, performing legality verification, business rule processing and data conversion;

[0012] 1.3: Data access layer: This layer interacts with the underlying Slurm-based supercomputing platform to read and write data. This layer uses a specific interface protocol to communicate with the supercomputing platform, obtain the platform's real-time data, and convert the user's operation instructions on the WEB visualization page into corresponding operations on the supercomputing platform. At the same time, the data access layer is also responsible for interacting with the database to ensure data consistency and persistence.

[0013] S2: Rich interface design

[0014] 2.1: Integrated command line interface: On the basis of retaining the original command line operation mode, the newly developed functions are also exposed through the CLI method; it not only meets the needs of advanced users who are accustomed to using the command line, but also ensures the compatibility of new functions;

[0015] 2.2: Develop RESTful API: Design a complete set of RESTful APIs for integration and expansion with other systems; external systems can use these APIs to obtain information about the supercomputing platform or operate the supercomputing platform;

[0016] 2.3: Automation and configuration management interface: Create dedicated interfaces and modules for automation and configuration management. Through these interfaces, administrators can set up automated tasks, such as scheduled billing, information alarms, and automatic collection of tenant storage status.

[0017] S3: Detailed implementation of functional modules

[0018] 3.1 Partition management module:

[0019] 3.2 Node management module:

[0020] 3.3 Resource Management Module:

[0021] 3.4 Job Management Module:

[0022] 3.5 File Management Module:

[0023] 3.6 Monitoring module.

[0024] Preferably, the front-end development technology described in S1 is such as HTML5, CSS3 and JavaScript framework (including Vue.js, React.js).

[0025] Preferably, the partition management module implements the operations of adding, deleting, modifying and querying partitions on the WEB visualization page by interacting with the back-end business logic layer and the data access layer;

[0026] The partition management module's partition details viewing function obtains detailed partition data from the supercomputing platform, including a list of nodes associated with the partition, current resource usage, and resource configuration, and displays it to the user in an intuitive manner on the WEB page.

[0027] Preferably, the node management module obtains the node status information in real time, and collects the node memory usage, resource status, CPU load and other data and basic information of the node from the supercomputing platform regularly through the data access layer and lists them in a table on the WEB page;

[0028] The node management module implements the operations of adding, deleting, modifying and checking nodes through interaction with the supercomputing platform; when adding a node, the user inputs the relevant parameters of the node, and the system automatically completes the addition and initialization configuration of the node.

[0029] Preferably, the resource management module supercomputer administrator allocates resources for each tenant and sub-tenant through a resource allocation form on a WEB interface;

[0030] The resource management module also supports dynamic resource adjustment function according to the real-time load of the supercomputing platform and the usage demand changes of the tenants.

[0031] Preferably, the job management module is connected to the job scheduling system, and obtains the job status and detailed information in the job queue in real time through the data access layer;

[0032] The job management module creates a new job through a pop-up job creation interface; the user fills in the job name, selects the job type, specifies the required resources, uploads the job script and other information on the interface.

[0033] Preferably, the file management module is compatible with the permissions of the Linux system, and realizes comprehensive management of files in the job directory on the WEB page;

[0034] The file management module has various file operation functions; when creating a new file, the user can select the file type and perform initial content settings.

[0035] Preferably, the monitoring module displays supercomputing resource pool information in real time through a large data screen; the large data screen adopts advanced visualization technology, using Echarts or Highcharts chart library to display operation status, partition information, node status, and storage usage data;

[0036] The monitoring module data collection is achieved by combining scheduled tasks and real-time event triggering mechanisms; scheduled tasks regularly obtain comprehensive resource data from the supercomputing platform. At the same time, when important system events occur, data collection and updates will be triggered in real time to ensure that the information on the data screen is always kept up to date.

[0037] (III) Beneficial effects

[0038] The present invention provides a visual integrated management system based on a supercomputing platform. It has the following beneficial effects:

[0039] 1. The present invention provides a convenient operation method for supercomputing platform administrators, users and operation and maintenance personnel, and realizes simple operation of the supercomputing platform through WEB pages, thereby reducing the operation difficulty and use cost.

[0040] 2. The present invention integrates the functions of multiple independent platforms, and conveniently enters different platforms through a unified entrance to realize different functions, thereby improving the convenience of operation and overall efficiency.

[0041] 3. The present invention realizes comprehensive WEB visualization page functions, meets the needs of supercomputing platform partition management, node management, resource management, job management and file management, and improves the management level and usability of the supercomputing platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a schematic diagram of the system architecture of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Example:

[0045] like Figure 1 As shown, an embodiment of the present invention provides a visual integrated management system based on a supercomputing platform, comprising the following specific steps:

[0046] S1. Multi-layer architecture design

[0047] 1.1 User interface layer: Modern front-end development technologies such as HTML5, CSS3 and JavaScript frameworks (including Vue.js, React.js, etc.) are used to build a responsive and interactive user interface. The interface is divided into clear areas according to different functional modules, such as partition management area, node management area, resource management area, job management area and file management area. Each area has specially designed interactive components to achieve the corresponding functions. For example, in the job management area, a table is used to display the job list, and each row of job information can be expanded to view the details. At the same time, operation buttons are provided for creating, editing and submitting jobs. Through intuitive WEB visualization pages, complex supercomputing platform operations are simplified into graphical operations, greatly reducing the threshold for use. Whether it is a professional supercomputing administrator or an ordinary scientific researcher or student user, it can be easily used. For example, in file management, users do not need to remember complex Linux commands, and can complete file management through simple clicks and editing operations.

[0048] 1.2 Business logic layer: Business logic processing is implemented based on back-end programming languages ​​(such as Java, Ruby, etc.). This layer is responsible for receiving user operation requests on the WEB visualization page, performing legitimacy verification, business rule processing and data conversion. For example, when a user requests to allocate resources to a tenant in the resource management module, the business logic layer verifies the administrator's authority, checks the legitimacy of the tenant information, and calculates the resource allocation plan based on the preset resource allocation algorithm, and then passes the request to the data access layer. Provides a wealth of operation prompts and guidance information. In each functional module, when the user hovers the mouse over the operation button or input box, a detailed prompt message will pop up to explain the function and expected results of the operation. For some complex operations, such as the explanation of the resource allocation algorithm in resource allocation and the job script writing guide in job management, the system also provides special help documents and tutorial links to facilitate user learning and use.

[0049] 1.3 Data access layer: It realizes data reading and writing by interacting with the underlying Slurm-based supercomputing platform. This layer uses specific interface protocols (such as Slurm REST API, SSH) to communicate with the supercomputing platform, obtain the platform's real-time data (such as node status, job information, etc.), and convert the user's operation instructions on the WEB visualization page (such as partition modification, job submission, etc.) into corresponding operations on the supercomputing platform. At the same time, the data access layer is also responsible for interacting with the database (used to store user information, configuration information, etc.) to ensure data consistency and persistence.

[0050] S2. Rich interface design

[0051] 2.1 Integrated Command Line Interface (CLI): While retaining the original command line operation mode, the newly developed functions are also exposed through the CLI. In this way, the needs of advanced users who are accustomed to using the command line are met, and the compatibility of new functions is guaranteed. For example, users can execute specific commands through the command line to start the WEB visualization platform, or directly use new resource allocation instructions in the command line. These instructions will be parsed by the system and perform corresponding operations. The functions of multiple independent platforms are deeply integrated to achieve a one-stop service from supercomputing resource management to system monitoring. Users do not need to switch between different software or platforms, avoiding operation incompatibility and data inconsistency caused by platform differences. This integration not only improves operational efficiency, but also enhances the stability and reliability of the system.

[0052] 2.2 Develop RESTful API: Design a complete set of RESTful APIs for integration and expansion with other systems. External systems can use these APIs to obtain information about the supercomputing platform (such as obtaining job status through GET requests) or perform operations on the supercomputing platform (such as submitting jobs through POST requests). This provides strong support for the openness and scalability of the supercomputing platform. For example, it can be easily integrated with other management systems within the enterprise (such as user authentication systems, network management systems, etc.) and has good integration capabilities with other related systems. Through RESTful APIs and other standard interfaces, the supercomputing visualization platform can easily interact with other information systems within the enterprise for data and business collaboration, achieving more extensive business process automation and information sharing.

[0053] 2.3 Automation and configuration management interface: Create dedicated interfaces and modules for automation and configuration management. Through these interfaces, administrators can set up automated tasks, such as scheduled billing, information alarms, and automatic collection of tenant storage conditions. In terms of configuration management, various platform parameters (such as resource allocation parameters, partition configuration parameters, etc.) can be dynamically adjusted to adapt to different operating environments and business needs.

[0054] S3. Detailed implementation of functional modules

[0055] 3.1 Partition management module:

[0056] 3.1.1 On the WEB visualization page, partitions can be added, deleted, modified, and queried through interaction with the backend business logic layer and the data access layer. When adding a partition, the user enters the partition name, resource quota, and other information on the interface. The system verifies the legitimacy of the input and passes the request to the data access layer, which interacts with the underlying supercomputing platform to complete the partition creation.

[0057] 3.1.2 The function of viewing partition details obtains detailed partition data from the supercomputing platform, including the list of nodes associated with the partition, current resource usage, resource configuration, etc., and displays it to the user in an intuitive way on the WEB page.

[0058] 3.1 covers all functions from basic to advanced in supercomputing platform management. From partitions, nodes, resources, jobs to file management, each functional module provides detailed and practical functions. For example, in node management, it can not only monitor the node status in real time, but also perform in-depth node performance analysis and fault diagnosis; in job management, in addition to basic job operations, it also supports the management of job templates and complex job dependencies.

[0059] 3.2 Node management module:

[0060] 3.2.1 Obtain node status information in real time. Regularly collect node memory usage, resource status, CPU load and other data from the supercomputing platform through the data access layer, and list the basic information of the node (such as node name, IP address, etc.) in a table on the WEB page.

[0061] 3.2.2 The addition, deletion, modification and query operations of nodes are realized through interaction with the supercomputing platform. When adding a node, the user enters the relevant parameters of the node (such as hardware configuration, network settings, etc.), and the system automatically completes the addition and initialization configuration of the node. When deleting a node, resource migration and verification operations will be performed first to ensure data security before deleting the node from the supercomputing platform.

[0062] Maintenance personnel take measures in advance to optimize system performance and avoid potential failures and performance bottlenecks

[0063] 3.3 Resource Management Module:

[0064] 3.3.1 The supercomputing administrator allocates resources to each tenant and subtenant through the resource allocation form on the WEB interface. The form lists the types of resources that can be allocated (CPU, GPU, memory, core time, number of running jobs, etc.), and the administrator enters the corresponding values ​​and restrictions. The business logic layer calculates the specific allocation plan based on the preset resource allocation algorithm (taking into account tenant priority, historical usage, etc.) and applies the plan to the supercomputing platform through the data access layer.

[0065] 3.3.2 The resource management module also supports dynamic resource adjustment. According to the real-time load of the supercomputing platform and the changes in tenants' usage requirements, administrators can initiate resource adjustment requests through the WEB interface. The system will automatically evaluate the feasibility of the adjustment and complete the dynamic adjustment of resources without affecting other tenants.

[0066] 3.4 Job Management Module:

[0067] 3.4.1 Connect to the job scheduling system and obtain the job status and details in the job queue in real time through the data access layer. On the WEB page, the job information is displayed in a list, including the job name, submitting user, status (such as waiting, running, completed, etc.), required resources, etc. Users can click the job details button to view more detailed job execution information, such as job execution logs, intermediate results, etc.

[0068] 3.4.2 The new job function is implemented through a pop-up job creation interface. The user fills in the job name, selects the job type, specifies the required resources (such as the number of CPU cores, memory size, etc.), and uploads the job script and other information on the interface. The system will generate a job description file based on the user input and submit it to the job scheduling system through the data access layer. The edit job function allows users to modify jobs that have been created but not submitted, including modifying job parameters, changing job scripts, and other operations.

[0069] 3.5 File Management Module:

[0070] 3.5.1 Compatible with Linux system permissions, comprehensive management of files in the job directory can be achieved on the WEB page. By simulating the directory structure and permission model of the Linux file system, users can operate as in the local terminal. For example, when viewing files, the corresponding files and directories are displayed or hidden according to the user's permissions.

[0071] 3.5.2 File operation functions are rich and varied. When creating a new file, users can select the file type (such as text file, binary file, etc.) and initialize the content settings. When deleting a file, permission verification and the Recycle Bin function will be performed (you can set whether to enable it). Renaming and moving files are achieved by dragging and dropping or entering a new name and path. Downloading files supports multiple download methods (such as single file download, batch download, breakpoint resume download, etc.), and copy / paste file operations can be performed between different directories and follow Linux permission rules. For file information display, in addition to displaying basic information such as file owner, permissions, hidden files, etc., it also supports preview functions for common file types (such as text files, image files, code files, etc.), which allows users to quickly view file contents.

[0072] 3.6 Monitoring module:

[0073] 3.6.1 Real-time display of supercomputing resource pool information through the data screen. The data screen uses advanced visualization technology, such as chart libraries such as Echarts or Highcharts, to display data such as job status (including the number of jobs, running jobs, etc.), partition information (partition resource usage ratio, partition load, etc.), node status, storage usage (storage capacity usage, storage usage trend, etc.).

[0074] 3.6.2 Data collection is achieved through a combination of scheduled tasks and real-time event triggering mechanisms. Scheduled tasks regularly obtain comprehensive resource data from the supercomputing platform. At the same time, when important system events occur (such as node failures, job abnormalities, etc.), data collection and updates are triggered in real time to ensure that the information on the data screen is always up to date. The data screen also has data filtering and search functions, and users can view data in a specific range according to their needs, such as viewing the node status of a partition.

[0075] In terms of technical content, the supercomputing visualization system integrates advanced software development technology and supercomputing management concepts. The user interface layer adopts modern design and provides a simple and intuitive operation interface, which allows users to quickly find the required functions. The business logic layer uses efficient algorithms to process user requests to ensure the stability and reliability of the system. The data operation layer uses database management systems and interface technologies to achieve secure data storage and fast interaction. The SLURM-based job scheduling system of the scheduling system layer can intelligently allocate resources, and the hardware resource pool provides powerful computing and storage capabilities.

[0076] In terms of the overall architecture, the layered design makes the functions of each part clear and relatively independent. The user interface layer interacts directly with the user, displaying rich functional options and intuitive visual information. The business logic layer handles complex business logic to ensure system stability. The data operation layer connects the business logic with the underlying system to ensure accurate data transmission. The scheduling system layer and the hardware resource pool constitute the infrastructure, which is responsible for supercomputing task execution and data storage. This architecture has good scalability and maintainability, which facilitates system function expansion and troubleshooting. At the same time, the platform pays attention to data security and user authority management, and ensures data security and user privacy through a variety of security measures.

[0077] 2. User Interface Layer

[0078] In the supercomputing visualization system, the unified portal serves as the entrance to the user interface layer, providing a convenient way for users to enter the supercomputing management platform, visualization platform or data monitoring platform. These three platforms each have unique and critical functions, and together they play a role in the stable operation and efficient use of the supercomputing system.

[0079] 1. Supercomputing Management Platform

[0080] 1.1 Features:

[0081] 1.1.1 Partition management: including creating, deleting, modifying and querying partitions. You can set parameters such as partition name and node queue.

[0082] 1.1.2 Node management: Manage the nodes in the supercomputing system, such as adding, removing, starting, stopping, etc., and monitor the node status.

[0083] 1.1.3 Resource allocation: Rationally allocate and restrict CPU, memory, storage and other resources among different partitions or users to ensure effective use of resources and avoid resource abuse.

[0084] 2 Technical Implementation:

[0085] 2.1 In the back-end development, shell script programming technology is used to accurately splice shell commands that meet functional requirements based on the parameters and operation instructions entered by the user in the interface. The shell commands here are command line instructions based on the Unix / Linux operating system, which can accurately control and operate the underlying system.

[0086] 2.2 A secure connection with the underlying supercomputing management node is achieved through the SSH (Secure Shell) protocol. The SSH protocol is based on public key encryption technology, which ensures the confidentiality and integrity of data transmission. The underlying supercomputing management node usually runs on a high-performance Linux server cluster.

[0087] 2.3 When the connection is established, the spliced ​​shell command is sent to the corresponding underlying supercomputing management node. Scripts with specific functions are deployed on these nodes, such as partition management scripts and node management scripts. These scripts are written in shell language and can parse the received commands and call the corresponding system functions and tools to perform actual operations, such as creating and deleting partitions by calling the system's partition management tool.

[0088] 2. Visualization platform

[0089] 2.1 Functions:

[0090] 2.1.1 Job Management: Users can comprehensively view the detailed status of jobs, including the running progress, resources used, estimated completion time, etc. of the job. New jobs can be created, and parameters such as the job type, required resources, and execution commands can be set during the creation process. Existing jobs can also be edited, such as modifying the job priority, resource allocation, etc.

[0091] 2.1.2 File System: Simulates the common file manager interface, allowing users to easily perform file operations. Highly compatible with the Linux system's permission settings, it determines the display and hiding of files and directories based on user permissions to ensure the security of file operations. It supports basic operations such as uploading, downloading, deleting, and renaming files, and can accurately display and manage file metadata information (such as creation time, modification time, owner, etc.).

[0092] 2.1.3 Integrated CLI: Provides a command line interactive interface for users. Users can directly enter specific commands to perform some complex or specific operations. These commands correspond to visual operations, increasing the flexibility and convenience of operations. The CLI commands entered by users will be parsed and processed, interacting with the background service or underlying system to achieve the corresponding functions.

[0093] 2 Technical Implementation:

[0094] 2.1. The job management module uses network communication technology, such as socket programming based on TCP / IP protocol, to establish a stable connection with the business logic layer. The business logic layer is responsible for interacting with the underlying job scheduling system, which is a scheduling system based on a distributed computing framework (such as Hadoop YARN, Slurm, etc.), which manages the allocation and execution of jobs on supercomputing resources. Through the interaction of this multi-layer architecture, job information is obtained in real time, and after obtaining it, data parsing and processing technology is used to convert the original data into a format that can be displayed by the user interface, such as converting the job status code into a text description that is easy for users to understand.

[0095] 2.2. File system module Based on the application programming interface technology of the file system, these APIs can be standard file operation APIs provided by the operating system (such as read, write, open, close and other functions in the Linux system), or specific APIs developed for distributed file systems (such as Ceph, HDFS, etc.). These APIs simulate the functions of common file managers. When interacting with the underlying storage system, network storage protocols (such as variants or optimized versions of protocols such as NFS and CIFS) are used to ensure efficient transmission of data in the network. At the same time, according to the Linux permission model, when a user operates a file, the system will query the permission information of the file in the underlying storage system (such as the file owner, the group to which it belongs, read, write and execute permissions, etc.) through the permission check module to determine whether the user has the permission to perform the corresponding operation.

[0096] 2.3. CLI integration uses command line parser technology to perform syntax analysis and parameter extraction on CLI commands entered by users. The parsed commands interact with the business logic layer or the underlying system through interfaces, which are implemented based on technologies such as RESTful API and RPC (remote procedure call). According to the type and parameters of the command, the corresponding functional module or system service is called to complete the operation. For example, if it is a CLI command to query job information, it will trigger interaction with the job management module to obtain and return relevant information.

[0097] 3. Data monitoring platform

[0098] 1Function:

[0099] 1.1 Comprehensively collect various key data of the supercomputing system through HTTP or JDBC, including job status, resource usage and partition load.

[0100] 1.2 The collected data is displayed in the form of intuitive and easy-to-understand visual charts on the system monitoring interface. These charts are of various types, such as line charts to show resource usage trends, bar charts to compare the status of different nodes, and pie charts to show resource allocation ratios, etc., which are convenient for users to monitor the status of the supercomputing platform in real time.

[0101] 2 Technical Implementation:

[0102] 2.1 Data Collection Environment Data acquisition is achieved through three methods, namely HTTP, JDBC or file import. The data sources are mainly Slurm REST API, Slurm underlying database and supercomputing management platform database. For HTTP method, it is used to interact with Slurm REST API, send requests according to API specifications and receive job-related data, such as job status, resource allocation and other information. JDBC method is used to connect Slurm underlying database and supercomputing management platform database, and accurately obtain required data, including hardware resource allocation information, historical job data, etc. by writing query statements that conform to the database structure. The file import method can process files in Excel and CSV formats, and plays a role when there are some pre-stored data files or when data needs to be supplemented from the outside, ensuring the diversity of data sources, thereby ensuring the comprehensiveness of data collection. At the same time, a timer mechanism is used to trigger data collection tasks at a certain time interval (such as every 5 minutes) to ensure the timeliness and accuracy of data.

[0103] 2.2 The data visualization phase is to first use data processing algorithms to clean, organize and analyze the raw data after obtaining the data. For example, filter or mark abnormal data (such as unreasonable high resource utilization data). Then, use visualization chart libraries (such as ECharts, Hi ghcharts, etc.) in programming languages ​​such as JavaScript to convert the processed data into visualization charts. These chart libraries provide a wealth of chart types and configuration options. By writing JavaScript code, data is bound to chart elements, and the style, title, axis labels and other attributes of the chart are set. Finally, the generated visualization chart is displayed on the system monitoring interface built on HTML5 and CSS, providing users with a clear and intuitive view of the operation of the supercomputing platform.

[0104] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A visual integrated management system based on a supercomputing platform, characterized in that it includes the following specific steps: S1: Multi-layer architecture design 1.1: User interface layer: Use modern front-end development technology to build a responsive and interactive user interface; the interface is divided into clear areas according to different functional modules, such as partition management area, node management area, resource management area, job management area and file management area; each area has specially designed interactive components to realize the corresponding functions; 1.2: Business logic layer: Implement business logic processing based on the back-end programming language; this layer is responsible for receiving user operation requests on the WEB visualization page, performing legality verification, business rule processing and data conversion; 1.3: Data access layer: This layer interacts with the underlying Slurm-based supercomputing platform to read and write data. This layer uses a specific interface protocol to communicate with the supercomputing platform, obtains the platform's real-time data, and converts the user's operation instructions on the WEB visualization page into corresponding operations on the supercomputing platform. At the same time, the data access layer is also responsible for interacting with the database to ensure data consistency and persistence. S2: Rich interface design 2.1: Integrated command line interface: On the basis of retaining the original command line operation mode, the newly developed functions are also exposed through the CLI method; it not only meets the needs of advanced users who are accustomed to using the command line, but also ensures the compatibility of new functions; 2.2: Develop RESTful API: Design a complete set of RESTful APIs for integration and expansion with other systems; external systems can use these APIs to obtain information about the supercomputing platform or operate the supercomputing platform; 2.3: Automation and configuration management interface: Create dedicated interfaces and modules for automation and configuration management. Through these interfaces, administrators can set up automated tasks, such as scheduled billing, information alarms, and automatic collection of tenant storage status. S3: Detailed implementation of functional modules 3.1 Partition management module: 3.2 Node management module: 3.3 Resource Management Module: 3.4 Job Management Module: 3.5 File Management Module: 3.6 Monitoring module.

2. According to claim 1, a visual integrated management system based on a supercomputing platform is characterized in that: S1 mentioned front-end development technologies such as HTML5, CSS3 and JavaScript frameworks (including Vue.js and React.js).

3. According to claim 1, a visual integrated management system based on a supercomputing platform is characterized in that: The partition management module implements the adding, deleting, modifying and querying operations of partitions on the WEB visualization page by interacting with the back-end business logic layer and the data access layer; The partition management module's partition details viewing function obtains detailed partition data from the supercomputing platform, including a list of nodes associated with the partition, current resource usage, and resource configuration, and displays it to the user in an intuitive manner on the WEB page.

4. The visual integrated management system based on a supercomputing platform according to claim 1, characterized in that: The node management module obtains the node status information in real time, and regularly collects the node's memory usage, resource status, CPU load and other data from the supercomputing platform through the data access layer, and lists the basic information of the node in a table on the WEB page; The node management module implements the operations of adding, deleting, modifying and checking nodes through interaction with the supercomputing platform; when adding a node, the user inputs the relevant parameters of the node, and the system automatically completes the addition and initialization configuration of the node.

5. The visual integrated management system based on a supercomputing platform according to claim 1, characterized in that: The resource management module supercomputing administrator allocates resources to each tenant and sub-tenant through a resource allocation form on the WEB interface; The resource management module also supports dynamic resource adjustment function according to the real-time load of the supercomputing platform and the usage demand changes of the tenants.

6. The visual integrated management system based on a supercomputing platform according to claim 1, characterized in that: The job management module connects to the job scheduling system and obtains the job status and detailed information in the job queue in real time through the data access layer; The job management module creates a new job through a pop-up job creation interface; the user fills in the job name, selects the job type, specifies the required resources, uploads the job script and other information on the interface.

7. The visual integrated management system based on a supercomputing platform according to claim 1, characterized in that: The file management module is compatible with the Linux system permissions and implements comprehensive management of files in the job directory on the WEB page; The file management module has various file operation functions; when creating a new file, the user can select the file type and perform initial content settings.

8. The visual integrated management system based on a supercomputing platform according to claim 1, characterized in that: The monitoring module displays supercomputing resource pool information in real time through a large data screen; the large data screen adopts advanced visualization technology and uses Echarts or Highcharts chart library to display operation status, partition information, node status, and storage usage data; The monitoring module data collection is achieved by combining scheduled tasks and real-time event triggering mechanisms; scheduled tasks regularly obtain comprehensive resource data from the supercomputing platform. At the same time, when important system events occur, data collection and updates will be triggered in real time to ensure that the information on the data screen is always kept up to date.

Images