Management-oriented data center system and data processing method

Abstract

The invention discloses a management-oriented data center system and a data processing method. The system comprises a 3D visualization module, a resource monitoring module, a machine room management module, an interaction adaptation module, a feedback closed loop module, a system configuration module and a data processing unit. According to the method, the physical layout and the operation state of the data center are visually presented through the 3D visualization technology, and the problem that traditional management information is not visual is solved; the refined resource monitoring and machine room management functions improve the resource allocation and scheduling efficiency; the responsive interactive design and shortcut key support optimize the operation convenience; the feedback closed loop module realizes quick response of user requirements and continuous optimization of the system; a system configuration module and a security mechanism guarantee the stability, expandability and data security of enterprise-level applications, and in addition, through performance optimization strategies such as view cone elimination, assembly lazy loading and virtual rolling, it is ensured that the system can still run smoothly in a large-data-volume scene.

Description

Technical Field

[0001] This invention relates to the field of data center system technology, and in particular to a managed data center system and data processing method. Background Technology

[0002] With the continuous expansion of data center scale and the increase in resident enterprises, the layout and resource allocation of data centers are becoming increasingly complex. Traditional data center management methods have many pain points: 1. The display of physical layout and operational status information is not intuitive, making it difficult for maintenance personnel to quickly locate equipment locations and fault points; 2. Resource usage statistics and enterprise management lack refined means, resulting in low resource allocation efficiency; 3. Data from different modules is not synchronized, leading to delayed operational feedback; 4. User problems and suggestions are difficult to collect and process effectively, resulting in a lack of data support for system optimization; 5. System compatibility, security, and performance stability are also difficult to meet the needs of enterprise-level applications. In view of the above, this application proposes a data center system and data processing method oriented towards management. Summary of the Invention

[0003] Based on the technical problems existing in the background technology, this invention proposes a data center system and data processing method oriented towards management.

[0004] The present invention proposes a data center system for management, including a 3D visualization module, a resource monitoring module, a data center management module, an interactive adaptation module, a feedback closed-loop module, a system configuration module, and a data processing unit;

[0005] The 3D visualization module is built on the WebGL engine of Three.js to construct a 5-story building model. Each floor contains 12 hexagonal ring-shaped computer rooms. It supports two view modes: single-layer focus and overall overview. It realizes interactive navigation with 360° rotation, zoom, and pan. It adopts an enterprise color coding mechanism to assign exclusive identification colors to more than 70 resident enterprises. It displays the computer room operation status, ambient temperature, and disk usage in real time, and supports seamless switching between macro and micro views.

[0006] The resource monitoring module is used to count key indicators such as total number of server racks, server rack utilization rate, number of resident companies, and number of system alarms. It presents resource utilization saturation through a combination of digital cards and progress bars, supports millisecond-level fuzzy search by company name and server room location, and visually displays the resource occupancy of each company on different floors and in different server rooms.

[0007] The data center management module uses a 30×24 grid to restore the physical space of the data center at a 1:1 scale. It uses color to indicate the status of the racks as used, idle, or inactive. It supports batch operation of racks by selecting them, dynamically updating the status of racks, and quickly binding racks to enterprises. It provides preset commands to enable quick layout adjustments and real-time calculation of the overall and regional rack utilization rates of the data center.

[0008] The interactive adaptation module provides a multi-panel draggable and adjustable monitoring dashboard, supports responsive adaptation between PC and tablet, sets up quick switching entry for 1-5 floors and overall view, supports keyboard numeric keypad shortcut operation, and ensures that the 3D view and resource overview data are synchronized when switching floors.

[0009] The feedback loop module provides a structured feedback form containing fields such as name, phone number, feedback content, anonymous submission, and willingness to return for feedback, enabling front-end field validity verification and privacy protection; the administrator side supports paginated display of feedback, statistics on the number of feedbacks today and the number of returned for feedback, and has feedback deletion permission control, status marking, and secondary confirmation operation functions;

[0010] The data processing unit is used to receive data requests from each module, realize data acquisition, cleaning, conversion and synchronization, and support the collaborative operation of the functions of each module.

[0011] The system configuration module is used to maintain basic enterprise information and visual configuration, set monitoring indicator thresholds, adjust 3D view animation parameters, and customize data center structure and rack layout rules. It adopts the RESTful API design specification, implements request retry and timeout handling based on Axios, and achieves functional isolation between administrators and ordinary users through RBAC permission control. It also has a security mechanism for front-end input filtering and secondary confirmation of sensitive operations.

[0012] Preferably, the 3D visualization module uses frustum culling technology to optimize rendering performance, rendering only the cabinets visible in the current view, and supports on-demand loading of floor models. Its specific operating logic is as follows:

[0013] The frustum culling technique first constructs a frustum-shaped frustum based on camera parameters. It converts the local coordinates of the server rack to world coordinates and then to camera coordinates. Then, it uses plane equations to determine whether the center point of the server rack is within the frustum. Only visible server racks are rendered to reduce computation. The on-demand loading of floor models combines user interaction and the coverage of the frustum. It uses formulas to determine whether the target floor needs to be loaded. When loading, it calls lightweight model data and caches it. Floors that are idle or not within the frustum are unloaded. The two work together and rely on a unified spatial coordinate system. The former reduces the loading range of the latter, and the latter reduces the judgment workload of the former, jointly optimizing the 3D view rendering performance and memory usage.

[0014] The conversion formula for transforming the local coordinates of the server rack to world coordinates is as follows: Let the coordinates of the server rack in the local coordinate system of its floor be P. obj =(x obj ,y obj ,z obj The offset of this floor in the global world coordinate system of the data center is... Among them Ox O z For the floor plan position offset, O y Given the floor height offset, the coordinates P of the server rack in the world coordinate system are... world =(x world ,y world ,z world ),satisfy:

[0015] The formula for converting world coordinates to camera coordinates is: Let the coordinates of the cabinet in the world coordinate system be P. world =(x world ,y world ,z world The camera's position in the world coordinate system is C = (x... c ,y c ,z c The camera view matrix is ​​M. view Then, the world coordinates need to be translated to a temporary coordinate system P with the camera as the origin. temp =(x temp ,y temp ,z temp Then, through matrix multiplication, it is converted to coordinates P in the camera coordinate system. view =(x view ,y view ,z view The specific formula is as follows:

[0016] Translate to camera origin P temp =(x world -x c ,y world -y c ,z world -z c );

[0017] Matrix transformation Where the view matrix M view It can be represented as the inverse combination of the camera rotation matrix and the translation matrix. Its core function is to align the world coordinate system to the camera coordinate system with the camera as the origin and the viewing direction as the -Z axis.

[0018] The formula for determining whether the center point of the cabinet is within the view frustum using the plane equation is as follows: Let the plane equation of a certain clipping plane of the view frustum be ax + by + cz + d = 0, where ax, by, cz, and d are the normal vectors of the plane, and the direction of the normal vectors points inward into the view frustum. Let x, y, and z be the coordinates P of the center point of the cabinet in the camera coordinate system. view =(x view ,y view ,z viewSubstituting the coordinates of the center point into the left side of the plane equation, we obtain the determination equation ax. view +by view +cz view +d, if the result of this judgment is ≥0, it means that the center point is located on the side of the clipping plane pointing to the inside of the view frustum. Only when the judgment results of the center point of the cabinet for all 6 clipping planes are ≥0 can it be determined that the cabinet is inside the view frustum and needs to be included in the rendering range.

[0019] The formula for determining whether a target floor needs loading is as follows: Let the target floor be the nth floor, and the height of each floor be H. floor Then the range of the Y-axis of the nth layer in the world coordinate system is [Y n,min ,Y n,max ]=[(n-1)×H floor ,n×H floor Simultaneously, by using the Y-axis coordinates of the upper and lower clipping planes of the view frustum in the world coordinate system, the visible range of the current view frustum along the Y-axis is determined to be [Y]. view,min ,Y view,max If the intersection condition Y is satisfied... n,min ≤Y view,max And Y n,max ≥Y view,min This indicates that there is spatial overlap between the target floor and the visible range of the view frustum, requiring loading to be triggered; otherwise, if there is no overlap, loading is not required.

[0020] Preferably, the system configuration module manages global types uniformly through types / index.ts, achieving full project type coverage for API returns, component Props, and utility functions.

[0021] Preferably, the 3D visualization module has a built-in lightweight model processing unit. The lightweight model processing unit uses the Draco compression algorithm to compress the vertex data of the rack model and converts the texture maps into WebP format, so that the file size of a single-floor 3D model is controlled within 500KB. At the same time, an asynchronous loading component is configured to realize the asynchronous loading of the floor model through the loadAsync method of Three.js, and outputs real-time progress data to the interaction adaptation module during the loading process. The interaction adaptation module displays the loading progress bar on the 3D view interface. After the loading is completed, the view frustum culling pre-judgment process is automatically triggered.

[0022] Preferably, the data processing unit includes a memory leak prevention submodule. When the data center management module performs cabinet status updates and the 3D visualization module unloads floor models, the memory leak prevention submodule automatically calls the Geometry.dispose(), Material.dispose(), and Texture.dispose() methods of Three.js to completely release the geometry, material, and texture resources associated with the model. At the same time, a memory monitoring timer is set to collect browser memory usage data every 30 seconds. When a certain type of resource is detected to occupy more than 50MB of memory for 5 consecutive times, a forced cleanup command is triggered and a log is generated and synchronized to the administrator log panel of the system configuration module.

[0023] This invention also proposes a data processing method for a managed data center system, comprising the following steps:

[0024] S1: Multi-dimensional data collection: The data processing unit collects data center physical structure data and rack resource data through standardized API interfaces, in collaboration with the data center management module, and collects enterprise configuration data in collaboration with the system configuration module and equipment operation status data in collaboration with the resource monitoring module. At the same time, it receives user feedback data transmitted by the feedback loop module. During the collection process, the feedback loop module and the resource monitoring module clean the data through front-end input filtering, remove invalid and malicious data, and then upload it to the data processing unit in a unified manner.

[0025] S2: 3D Scene Construction and Real-time Rendering: The data processing unit pushes the collected physical structure data and enterprise configuration data to the 3D visualization module. The 3D visualization module constructs a 3D stereoscopic model based on the data center scene manager encapsulated by Three.js, and calls the enterprise-specific identification color and rack status color rules preset by the system configuration module. At the same time, the 3D visualization module independently enables frustum culling technology and on-demand loading strategy to optimize rendering performance and achieve smooth display of 3D scenes.

[0026] S3: Dynamic synchronization of resource status: The data processing unit establishes a real-time data synchronization channel through the Axios interceptor to synchronize the device operation status data and resource usage data to the 3D visualization module, resource monitoring module and data center management module, respectively; when the indicators change, the resource monitoring module and 3D visualization module trigger a smooth transition animation to ensure data consistency among the modules.

[0027] S4: Interactive Command Response: The interactive adaptation module receives user input commands for view switching, resource retrieval, rack operation, and configuration modification, and transmits the commands to the data processing unit. The data processing unit parses and processes the commands based on the Pinia hierarchical state model, and synchronously distributes update commands to the corresponding modules: pushing view switching data to the 3D visualization module, pushing search conditions to the resource monitoring module, pushing rack operation commands to the data center management module, and pushing configuration modification requests to the system configuration module. Simultaneously, the interactive adaptation module supports shortcut key responses and drag-and-drop adaptation, improving operational efficiency.

[0028] S5: Feedback Data Closed-Loop Processing: The feedback closed-loop module receives user-submitted feedback data, verifies its validity on the front end, and then transmits it to the data processing unit. The data processing unit anonymizes the feedback data, stores it in the backend database, and synchronizes it to the administrator's statistical dashboard in the feedback closed-loop module. The administrator marks feedback as "processed / pending" and records follow-up results through the feedback closed-loop module, forming a closed-loop management of "submission-review-processing-follow-up," and the operation log is synchronized to the data processing unit for archiving.

[0029] Preferably, the specific logical steps of S2 are as follows:

[0030] S201: The 3D visualization module receives physical structure data and enterprise configuration data pushed by the data processing unit. The physical structure data includes the rack's local coordinates P. obj =(x obj ,y obj ,z obj Floor height H foor The number and layout rules of each server room, enterprise configuration data including the resident enterprise ID and its corresponding unique identifier color, the 3D visualization module sends a request to the system configuration module to call preset parameters, which include rack status color rules and 3D view basic parameters. The rack status color rules include used → blue, idle → gray, not enabled → white. The 3D view basic parameters include the initial camera position CameraPos = (x c ,y c ,z c Vertical field of view f ovY Aspect ratio, near clipping plane distance, far clipping plane distance;

[0031] S202: Data processing unit based on floor height H foor The world offset for the nth layer is generated using the following formula: Among them O y = (n-1)×H floorThe X / Z axis offset is determined by the floor plan position, and the local coordinates of each rack are converted into world coordinates to ensure that the racks are accurately mapped in the 3D scene.

[0032] S203: Based on world coordinate data and hexagonal ring layout rules, the 3D visualization module constructs a 5-story building model using the Mesh class in Three.js. Each floor generates 12 hexagonal server room models, and the server room boundaries are defined by the coordinates of the polygon vertices. Basic materials are assigned to the server room models, and the enterprise ID associated with each server room is recorded. At the same time, the server room management module generates rack models at the corresponding server room world coordinate positions according to a 30×24 grid rule, and assigns material colors to the rack models according to the "enterprise identification color" and "rack status color rules" configured in the system. If the rack has been assigned to enterprise A, then enterprise A's exclusive RGB color is used; if it has not been assigned, the corresponding color is used according to the status.

[0033] S204: The 3D visualization module calculates the width and height dimensions of the near / far clipping planes based on camera parameters and determines the shape of the view frustum using a formula. The formula used is...

[0034]

[0035] S205: Set the rack's world coordinates P world Convert the coordinates to the camera coordinate system and use the plane equation to determine whether the center point of the cabinet is inside the view frustum;

[0036] S206: If the user is in the "Overall View" view, the 3D visualization module calculates the visible range of the view frustum along the Y-axis in the world coordinate system using camera parameters. view,min ,Y view,max If in "Single-Floor Focus" view, the visible range is directly the Y-axis range of the current floor. For the nth floor, its Y-axis range is first calculated using a formula.

[0037] [Y n,min ,Y n,max ]=[(n-1)×H floor ,n×H floor Then, the intersection formula is used to determine whether loading is needed. The formula is as follows:

[0038]

[0039] S207: For floors determined to be "need to be loaded", the 3D visualization module loads a lightweight model through Three.js's GLTFLoader, and caches it to the "floor model cache pool" after loading; for floors that are "not loaded", only the basic location information is retained, and the model data is not loaded for the time being.

[0040] S208: The 3D visualization module renders the visible cabinets in the rendering queue in the order of "closest to the camera and furthest away", and calls Three.js's WebGLRenderer to output the 3D scene to the monitoring dashboard of the interactive adaptation module.

[0041] S209: The data processing unit synchronizes the equipment operation status data to the 3D visualization module in real time. The 3D visualization module immediately updates the rendering status of the corresponding computer room / rack and triggers a smooth transition animation to ensure that the 3D scene is synchronized with the actual status in real time.

[0042] Preferably, the specific logical steps of S4 are as follows:

[0043] S401: Receive user operations, including clicking the "Floor Switch" button, entering resource search keywords, selecting cabinet operations, and pressing shortcut keys, and encapsulate the instructions into a "type + parameter" format and transmit them to the data processing unit;

[0044] S402: After receiving the instruction, the data processing unit first calls the RBAC permission rules of the system configuration module to verify the permissions of the instruction initiator. If it is a "configuration modification" instruction and the initiator is a regular user, it directly returns a "insufficient permissions" response, and the interaction adaptation module pops up a window to prompt the user. If the permissions are compliant, it enters the instruction parsing stage.

[0045] S403: The data processing unit loads the Pinia hierarchical state model, which contains three types of state pools. Each state pool is bound to the instruction parsing logic. The global state pool is used to store cross-module shared data, the module state pool is used to store the configuration specific to each module, and the page state pool is used to store the current page interaction data. The data processing unit extracts the associated data from the corresponding state pool according to the instruction type to complete the instruction parameter completion and logic parsing.

[0046] S404: The data processing unit pushes a "view switching parameter package" to the 3D visualization module, which includes the target floor ID, floor world offset and view mode. The 3D visualization module switches to the 2nd floor 3D view based on the parameter package, and updates the view frustum range simultaneously. At the same time, the interaction adaptation module updates the highlighted state of the "floor control panel" and records the frequency of shortcut key usage.

[0047] S405: The data processing unit pushes a "search condition package" to the resource monitoring module, containing the keyword "Company A", search fields, and matching rules. Based on the fuzzy matching rules, the resource monitoring module traverses the list of enterprise resource data and filters out target enterprises whose names contain 'A'. The number of search results N is counted according to the formula:

[0048] The resource monitoring module displays the search results in a table format and triggers a lightweight animation, while simultaneously caching the results to the Pinia module's state pool.

[0049] S406: The data processing unit pushes a "rack operation parameter package" to the data center management module, which includes the operation type, a list of target rack IDs, and the target enterprise ID. The data center management module updates the rack-enterprise association table, changes the target rack status from "idle" to "used," and synchronizes the data center utilization rate. At the same time, the 3D visualization module receives status update notifications from the data center management module and switches the target rack color to Enterprise A's exclusive identification color to ensure that the 3D view is consistent with the management data;

[0050] S407: The data processing unit pushes a "configuration modification parameter package" to the system configuration module, which includes the modification type, parameter name, old value, and new value. The system configuration module updates the "monitoring indicator threshold table", changes the temperature alarm threshold to 32℃, and generates a configuration modification log simultaneously. At the same time, the system configuration module pushes a "threshold update notification" to the resource monitoring module through the data processing unit. The resource monitoring module immediately uses the new threshold to determine the alarm status.

[0051] S408: After each module executes the command, it returns the "execution result" to the interaction adaptation module. The interaction adaptation module provides visual or status feedback to the user. The interaction adaptation module counts the frequency of shortcut key usage. If the usage frequency of a certain shortcut key is higher than that of the default shortcut key for 7 consecutive days, it automatically prompts the administrator whether to set it as the new default shortcut key. When dragging and adapting, the interaction adaptation module uses the "grid snapping algorithm" to ensure that the panel position is aligned, avoid panel overlap, and improve the ease of operation.

[0052] Compared with existing technologies, the beneficial effects of this invention are:

[0053] 1. Construct a three-dimensional model of the data center through the 3D visualization module, and combine the enterprise's exclusive identification color and the cabinet status color rules to intuitively present the physical layout and equipment operation status. It supports seamless switching between "macro" and "micro" views, and maintenance personnel can quickly locate equipment locations and fault points without relying on traditional 2D drawings, thus improving the efficiency of fault diagnosis.

[0054] 2. The resource monitoring module provides real-time statistics on key indicators such as rack utilization and enterprise resource usage, and supports fuzzy search by enterprise / data center; the data center management module enables batch rack operations and precise allocation, and quantifies resource usage using the utilization rate calculation formula. Managers can optimize allocation strategies based on the data to reduce resource idleness and improve resource utilization.

[0055] 3. The feedback loop module collects user questions and suggestions through structured forms. Administrators can statistically analyze feedback data and track processing progress, forming a closed loop of "submission-review-processing-follow-up". The data processing unit regularly generates feedback analysis reports to provide data support for system parameter adjustments and promote continuous system optimization.

[0056] 4. Through the real-time synchronization channel between the data processing unit and Axios, the data of modules such as 3D visualization, resource monitoring, and data center management are updated in a coordinated manner. Changes in indicators trigger smooth animation feedback, avoiding the problem of "delayed data update after operation" in traditional management and ensuring consistency between operation and data status.

[0057] 5. The interoperability module supports responsive adaptation between PC and tablet, meeting the needs of multi-device use; the system configuration module uses RBAC permission control to isolate the functions of administrators and ordinary users, and combines front-end input filtering and secondary confirmation of sensitive operations to prevent risks; the 3D visualization module reduces system resource consumption and ensures the stable operation of enterprise-level applications through view frustum culling, on-demand floor loading, and lightweight model processing.

[0058] This invention uses 3D visualization technology to intuitively present the physical layout and operational status of a data center, solving the problem of unintuitive traditional management information. Its refined resource monitoring and data center management functions improve resource allocation and scheduling efficiency. Responsive interactive design and shortcut key support optimize ease of operation. A feedback loop module enables rapid response to user needs and continuous system optimization. System configuration modules and security mechanisms ensure the stability, scalability, and data security of enterprise-level applications. Furthermore, performance optimization strategies such as view frustum culling, component lazy loading, and virtual scrolling ensure smooth system operation even under large data volume scenarios, fully meeting the technical standards and functional requirements of modern data center management. Attached Figure Description

[0059] Figure 1 This is a block diagram of a management-oriented data center system proposed in this invention;

[0060] Figure 2 This is a flowchart of a data processing method for a managed data center system proposed in this invention. Detailed Implementation

[0061] The present invention will be further explained below with reference to specific embodiments.

[0062] Example

[0063] Reference Figure 1-2This embodiment proposes a management-oriented data center system, including a 3D visualization module, a resource monitoring module, a data center management module, an interactive adaptation module, a feedback closed-loop module, a system configuration module, and a data processing unit.

[0064] The 3D visualization module uses the WebGL engine of Three.js to build a 5-story building model. Each floor contains 12 hexagonal ring-shaped server rooms. It supports two view modes: single-layer focus and overall overview. It enables interactive navigation with 360° rotation, zoom, and pan. It adopts an enterprise color coding mechanism to assign exclusive identification colors to 70+ resident enterprises. It displays the server room operating status, ambient temperature, and disk usage in real time, and supports seamless switching between macro and micro views.

[0065] The 3D visualization module uses frustum culling technology to optimize rendering performance, rendering only the racks visible in the current view, and supports on-demand loading of floor models. Its specific operating logic is as follows:

[0066] The frustum culling technique first constructs a frustum-shaped frustum based on camera parameters. It converts the local coordinates of the server rack to world coordinates and then to camera coordinates. Then, it uses plane equations to determine whether the center point of the server rack is within the frustum. Only visible server racks are rendered to reduce computation. The on-demand loading of floor models combines user interaction and the coverage of the frustum. It uses formulas to determine whether the target floor needs to be loaded. When loading, it calls lightweight model data and caches it. Floors that are idle or not within the frustum are unloaded. The two work together and rely on a unified spatial coordinate system. The former reduces the loading range of the latter, and the latter reduces the judgment workload of the former, jointly optimizing the 3D view rendering performance and memory usage.

[0067] The conversion formula for transforming the local coordinates of the server rack to world coordinates is as follows: Let the coordinates of the server rack in the local coordinate system of its floor be P. obj =(x obj ,y obj ,z obj The offset of this floor in the global world coordinate system of the data center is... Among them O x O z For the floor plan position offset, O y Given the floor height offset, the coordinates P of the server rack in the world coordinate system are... world =(x world ,y world ,z world ),satisfy:

[0068] The formula for converting world coordinates to camera coordinates is: Let the coordinates of the cabinet in the world coordinate system be P. world =(x world ,yworld ,z world The camera's position in the world coordinate system is C = (x... c ,y c ,z c The camera view matrix is ​​M. view Then, the world coordinates need to be translated to a temporary coordinate system P with the camera as the origin. temp =(x temp ,y temp ,z temp Then, through matrix multiplication, it is converted to coordinates P in the camera coordinate system. view =(x view ,y view ,z view The specific formula is as follows:

[0069] Translate to camera origin P temp =(x world -x c ,y world -y c ,z world -z c );

[0070] Matrix transformation Where the view matrix M view It can be represented as the inverse combination of the camera rotation matrix and the translation matrix. Its core function is to align the world coordinate system to the camera coordinate system with the camera as the origin and the viewing direction as the -Z axis.

[0071] The formula for determining whether the center point of the cabinet is within the view frustum using the plane equation is as follows: Let the plane equation of a certain clipping plane of the view frustum be ax + by + cz + d = 0, where ax, by, cz, and d are the normal vectors of the plane, and the direction of the normal vectors points inward into the view frustum. Let x, y, and z be the coordinates P of the center point of the cabinet in the camera coordinate system. view =(x view ,y view ,z view Substituting the coordinates of the center point into the left side of the plane equation, we obtain the determination equation ax. view +by view +cz view +d, if the result of this judgment is ≥0, it means that the center point is located on the side of the clipping plane pointing to the inside of the view frustum. Only when the judgment results of the center point of the cabinet for all 6 clipping planes are ≥0 can it be determined that the cabinet is inside the view frustum and needs to be included in the rendering range.

[0072] The formula for determining whether a target floor needs loading is as follows: Let the target floor be the nth floor, and the height of each floor be H. floor Then the range of the Y-axis of the nth layer in the world coordinate system is [Y n,min ,Yn,max ]=[(n-1)×H floor ,n×H floor Simultaneously, by using the Y-axis coordinates of the upper and lower clipping planes of the view frustum in the world coordinate system, the visible range of the current view frustum along the Y-axis is determined to be [Y]. view,min ,Y view,max If the intersection condition Y is satisfied... n,min ≤Y view,max And Y n,max ≥Y view,min This indicates that there is spatial overlap between the target floor and the visible range of the view frustum, requiring loading to be triggered; otherwise, if there is no overlap, loading is not required.

[0073] The 3D visualization module has a built-in lightweight model processing unit. This unit uses the Draco compression algorithm to compress the vertex data of the rack model and converts the texture maps to WebP format, keeping the file size of a single-floor 3D model within 500KB. It also configures an asynchronous loading component, which uses the loadAsync method of Three.js to load the floor model asynchronously and outputs real-time progress data to the interaction adaptation module during the loading process. The interaction adaptation module displays the loading progress bar in the 3D view interface, and automatically triggers the view frustum culling pre-judgment process after loading is completed.

[0074] The resource monitoring module is used to count key indicators such as total number of server racks, server rack utilization rate, number of resident companies, and number of system alarms. It presents resource utilization saturation through a combination of digital cards and progress bars, supports millisecond-level fuzzy search by company name and server room location, and visually displays the resource occupancy of each company on different floors and in different server rooms.

[0075] The data center management module uses a 30×24 grid to restore the physical space of the data center at a 1:1 scale. It uses color to indicate the status of the racks as used, idle, or inactive. It supports batch operation of racks by selecting them, dynamic updates of rack status, quick binding of racks to enterprises, and provides preset commands to enable quick layout adjustments. It also calculates the overall and regional rack utilization rates of the data center in real time.

[0076] The interactive adaptation module provides a multi-panel draggable and adjustable monitoring dashboard, supports responsive adaptation between PC and tablet, sets up quick switching entry for floors 1-5 and the overall view, supports keyboard numeric keypad shortcut operation, and ensures that the 3D view and resource overview data are synchronized when switching floors.

[0077] The feedback loop module provides a structured feedback form with fields for name, phone number, feedback content, anonymous submission, and willingness to follow up, enabling front-end field validity verification and privacy protection; the administrator side supports paginated display of feedback, statistics on the number of feedbacks today and the number of follow-up visits received, and has feedback deletion permission control, status marking, and secondary confirmation operation functions;

[0078] The data processing unit is used to receive data requests from various modules, realize data acquisition, cleaning, transformation and synchronization, and support the coordinated operation of various module functions;

[0079] The data processing unit includes a memory leak prevention submodule. When the data center management module updates the rack status or the 3D visualization module unloads the floor model, the memory leak prevention submodule automatically calls the Geometry.dispose(), Material.dispose(), and Texture.dispose() methods of Three.js to completely release the geometry, material, and texture resources associated with the model. At the same time, a memory monitoring timer is set to collect browser memory usage data every 30 seconds. When a certain type of resource is detected to occupy more than 50MB of memory for 5 consecutive times, a forced cleanup command is triggered and a log is generated and synchronized to the administrator log panel of the system configuration module.

[0080] The system configuration module is used to maintain basic enterprise information and visual configuration, set monitoring indicator thresholds, adjust 3D view animation parameters, and customize data center structure and rack layout rules. It adopts the RESTful API design specification, implements request retry and timeout handling based on Axios, and achieves functional isolation between administrators and ordinary users through RBAC access control. It also has a security mechanism for front-end input filtering and secondary confirmation of sensitive operations.

[0081] The system configuration module manages global types uniformly through types / index.ts, achieving full project type coverage for API returns, component props, and utility functions.

[0082] This embodiment proposes a data processing method for a managed data center system, including the following steps:

[0083] S1: Multi-dimensional data collection: The data processing unit collects data center physical structure data and rack resource data through standardized API interfaces, in collaboration with the data center management module, and collects enterprise configuration data in collaboration with the system configuration module and equipment operation status data in collaboration with the resource monitoring module. At the same time, it receives user feedback data transmitted by the feedback loop module. During the collection process, the feedback loop module and the resource monitoring module clean the data through front-end input filtering, remove invalid and malicious data, and then upload it to the data processing unit in a unified manner.

[0084] S2: 3D Scene Construction and Real-time Rendering: The data processing unit pushes the collected physical structure data and enterprise configuration data to the 3D visualization module. The 3D visualization module constructs a 3D stereoscopic model based on the data center scene manager encapsulated by Three.js, and calls the enterprise-specific identification color and rack status color rules preset by the system configuration module. At the same time, the 3D visualization module independently enables frustum culling technology and on-demand loading strategy to optimize rendering performance and achieve smooth display of 3D scenes.

[0085] The specific logical steps are as follows:

[0086] S201: The 3D visualization module receives physical structure data and enterprise configuration data pushed by the data processing unit. The physical structure data includes the rack's local coordinates P. obj =(x obj ,y obj ,z obj Floor height H foor The number and layout rules of each server room, enterprise configuration data including the resident enterprise ID and its corresponding unique identifier color, the 3D visualization module sends a request to the system configuration module to call preset parameters, which include rack status color rules and 3D view basic parameters. The rack status color rules include used → blue, idle → gray, not enabled → white. The 3D view basic parameters include the initial camera position CameraPos = (x c ,y c ,z c Vertical field of view f ovY Aspect ratio, near clipping plane distance, far clipping plane distance;

[0087] S202: Data processing unit based on floor height H foor The world offset for the nth layer is generated using the following formula: Among them O y = (n-1)×H floor The X / Z axis offset is determined by the floor plan position, and the local coordinates of each rack are converted into world coordinates to ensure that the racks are accurately mapped in the 3D scene.

[0088] S203: Based on world coordinate data and hexagonal ring layout rules, the 3D visualization module constructs a 5-story building model using the Mesh class in Three.js. Each floor generates 12 hexagonal server room models, and the server room boundaries are defined by the coordinates of the polygon vertices. Basic materials are assigned to the server room models, and the enterprise ID associated with each server room is recorded. At the same time, the server room management module generates rack models at the corresponding server room world coordinate positions according to a 30×24 grid rule, and assigns material colors to the rack models according to the "enterprise identification color" and "rack status color rules" configured in the system. If the rack has been assigned to enterprise A, then enterprise A's exclusive RGB color is used; if it has not been assigned, the corresponding color is used according to the status.

[0089] S204: The 3D visualization module calculates the width and height dimensions of the near / far clipping planes based on camera parameters and determines the shape of the view frustum using a formula. The formula used is...

[0090]

[0091] S205: Set the rack's world coordinates P world Convert the coordinates to the camera coordinate system and use the plane equation to determine whether the center point of the cabinet is inside the view frustum;

[0092] S206: If the user is in the "Overall View" view, the 3D visualization module calculates the visible range of the view frustum along the Y-axis in the world coordinate system using camera parameters. view,min ,Y view,max If in "Single-Floor Focus" view, the visible range is directly the Y-axis range of the current floor. For the nth floor, its Y-axis range is first calculated using a formula.

[0093] [Y n,min ,Y n,max ]=[(n-1)×H floor ,n×H floor Then, the intersection formula is used to determine whether loading is needed. The formula is as follows:

[0094]

[0095] S207: For floors determined to be "need to be loaded", the 3D visualization module loads a lightweight model through Three.js's GLTFLoader, and caches it to the "floor model cache pool" after loading; for floors that are "not loaded", only the basic location information is retained, and the model data is not loaded for the time being.

[0096] S208: The 3D visualization module renders the visible cabinets in the rendering queue in the order of "closest to the camera and furthest away", and calls Three.js's WebGLRenderer to output the 3D scene to the monitoring dashboard of the interactive adaptation module.

[0097] S209: The data processing unit synchronizes the equipment operating status data to the 3D visualization module in real time. The 3D visualization module immediately updates the rendering status of the corresponding computer room / rack and triggers a smooth transition animation to ensure that the 3D scene is synchronized with the actual status in real time.

[0098] S3: Dynamic synchronization of resource status: The data processing unit establishes a real-time data synchronization channel through the Axios interceptor to synchronize the device operation status data and resource usage data to the 3D visualization module, resource monitoring module and data center management module, respectively; when the indicators change, the resource monitoring module and 3D visualization module trigger a smooth transition animation to ensure data consistency among the modules.

[0099] S4: Interactive Command Response: The interactive adaptation module receives user input commands for view switching, resource retrieval, rack operation, and configuration modification, and transmits the commands to the data processing unit. The data processing unit parses and processes the commands based on the Pinia hierarchical state model, and synchronously distributes update commands to the corresponding modules: pushing view switching data to the 3D visualization module, pushing search conditions to the resource monitoring module, pushing rack operation commands to the data center management module, and pushing configuration modification requests to the system configuration module. Simultaneously, the interactive adaptation module supports shortcut key responses and drag-and-drop adaptation, improving operational efficiency.

[0100] The specific logical steps are as follows:

[0101] S401: Receive user operations, including clicking the "Floor Switch" button, entering resource search keywords, selecting cabinet operations, and pressing shortcut keys, and encapsulate the instructions into a "type + parameter" format and transmit them to the data processing unit;

[0102] S402: After receiving the instruction, the data processing unit first calls the RBAC permission rules of the system configuration module to verify the permissions of the instruction initiator. If it is a "configuration modification" instruction and the initiator is a regular user, it directly returns a "insufficient permissions" response, and the interaction adaptation module pops up a window to prompt the user. If the permissions are compliant, it enters the instruction parsing stage.

[0103] S403: The data processing unit loads the Pinia hierarchical state model, which contains three types of state pools. Each state pool is bound to the instruction parsing logic. The global state pool is used to store cross-module shared data, the module state pool is used to store the configuration specific to each module, and the page state pool is used to store the current page interaction data. The data processing unit extracts the associated data from the corresponding state pool according to the instruction type to complete the instruction parameter completion and logic parsing.

[0104] S404: The data processing unit pushes a "view switching parameter package" to the 3D visualization module, which includes the target floor ID, floor world offset and view mode. The 3D visualization module switches to the 2nd floor 3D view based on the parameter package, and updates the view frustum range simultaneously. At the same time, the interaction adaptation module updates the highlighted state of the "floor control panel" and records the frequency of shortcut key usage.

[0105] S405: The data processing unit pushes a "search condition package" to the resource monitoring module, containing the keyword "Company A", search fields, and matching rules. Based on the fuzzy matching rules, the resource monitoring module traverses the list of enterprise resource data and filters out target enterprises whose names contain 'A'. The number of search results N is counted according to the formula:

[0106] The resource monitoring module displays the search results in a table format and triggers a lightweight animation, while simultaneously caching the results to the Pinia module's state pool.

[0107] S406: The data processing unit pushes a "rack operation parameter package" to the data center management module, which includes the operation type, a list of target rack IDs, and the target enterprise ID. The data center management module updates the rack-enterprise association table, changes the target rack status from "idle" to "used," and synchronizes the data center utilization rate. At the same time, the 3D visualization module receives status update notifications from the data center management module and switches the target rack color to Enterprise A's exclusive identification color to ensure that the 3D view is consistent with the management data;

[0108] S407: The data processing unit pushes a "configuration modification parameter package" to the system configuration module, which includes the modification type, parameter name, old value, and new value. The system configuration module updates the "monitoring indicator threshold table", changes the temperature alarm threshold to 32℃, and generates a configuration modification log simultaneously. At the same time, the system configuration module pushes a "threshold update notification" to the resource monitoring module through the data processing unit. The resource monitoring module immediately uses the new threshold to determine the alarm status.

[0109] S408: After each module executes the command, it returns the "execution result" to the interaction adaptation module. The interaction adaptation module provides visual or status feedback to the user. The interaction adaptation module counts the frequency of shortcut key usage. If the usage frequency of a certain shortcut key is higher than that of the default shortcut key for 7 consecutive days, it will automatically prompt the administrator whether to set it as the new default shortcut key. When dragging and adapting, the interaction adaptation module uses the "grid snapping algorithm" to ensure that the panel position is aligned, avoid panel overlap, and improve the ease of operation.

[0110] S5: Feedback Data Closed-Loop Processing: The feedback closed-loop module receives user-submitted feedback data, verifies its validity on the front end, and then transmits it to the data processing unit. The data processing unit anonymizes the feedback data, stores it in the backend database, and synchronizes it to the administrator's statistical dashboard in the feedback closed-loop module. The administrator marks feedback as "processed / pending" and records follow-up results through the feedback closed-loop module, forming a closed-loop management of "submission-review-processing-follow-up," and the operation log is synchronized to the data processing unit for archiving.

[0111] This embodiment uses 3D visualization technology to intuitively present the physical layout and operational status of the data center, solving the problem of unintuitive traditional management information. Refined resource monitoring and data center management functions improve resource allocation and scheduling efficiency. Responsive interactive design and shortcut key support optimize ease of operation. The feedback loop module enables rapid response to user needs and continuous system optimization. The system configuration module and security mechanisms ensure the stability, scalability, and data security of enterprise-level applications. Furthermore, performance optimization strategies such as view frustum culling, component lazy loading, and virtual scrolling ensure smooth system operation even in scenarios with large data volumes, fully meeting the technical standards and functional requirements of modern data center management.

[0112] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A management-oriented data center system, characterized in that, It includes a 3D visualization module, a resource monitoring module, a data center management module, an interactive adaptation module, a feedback loop module, a system configuration module, and a data processing unit; The 3D visualization module is built on the WebGL engine of Three.js to construct a 5-story building model. Each floor contains 12 hexagonal ring-shaped computer rooms. It supports two view modes: single-layer focus and overall overview. It realizes interactive navigation with 360° rotation, zoom, and pan. It adopts an enterprise color coding mechanism to assign exclusive identification colors to more than 70 resident enterprises. It displays the computer room operation status, ambient temperature, and disk usage in real time, and supports seamless switching between macro and micro views. The resource monitoring module is used to count key indicators such as total number of server racks, server rack utilization rate, number of resident companies, and number of system alarms. It presents resource utilization saturation through a combination of digital cards and progress bars, supports millisecond-level fuzzy search by company name and server room location, and visually displays the resource occupancy of each company on different floors and in different server rooms. The data center management module uses a 30×24 grid to restore the physical space of the data center at a 1:1 scale. It uses color to indicate the status of the racks as used, idle, or inactive. It supports batch operation of racks by selecting them, dynamically updating the status of racks, and quickly binding racks to enterprises. It provides preset commands to enable quick layout adjustments and real-time calculation of the overall and regional rack utilization rates of the data center. The interactive adaptation module provides a multi-panel draggable and adjustable monitoring dashboard, supports responsive adaptation between PC and tablet, sets up quick switching entry for 1-5 floors and overall view, supports keyboard numeric keypad shortcut operation, and ensures that the 3D view and resource overview data are synchronized when switching floors. The feedback loop module provides a structured feedback form containing fields such as name, phone number, feedback content, anonymous submission, and willingness to return for feedback, enabling front-end field validity verification and privacy protection; the administrator side supports paginated display of feedback, statistics on the number of feedbacks today and the number of returned for feedback, and has feedback deletion permission control, status marking, and secondary confirmation operation functions; The data processing unit is used to receive data requests from each module, realize data acquisition, cleaning, conversion and synchronization, and support the collaborative operation of the functions of each module. The system configuration module is used to maintain basic enterprise information and visual configuration, set monitoring indicator thresholds, adjust 3D view animation parameters, and customize data center structure and rack layout rules. It adopts the RESTful API design specification, implements request retry and timeout handling based on Axios, and achieves functional isolation between administrators and ordinary users through RBAC permission control. It also has a security mechanism for front-end input filtering and secondary confirmation of sensitive operations.

2. The data center system for management as described in claim 1, characterized in that, The 3D visualization module uses frustum culling technology to optimize rendering performance, rendering only the cabinets visible in the current view, and supports on-demand loading of floor models. Its specific operating logic is as follows: The frustum culling technique first constructs a frustum-shaped frustum based on camera parameters. It converts the local coordinates of the server rack to world coordinates and then to camera coordinates. Then, it uses plane equations to determine whether the center point of the server rack is within the frustum. Only visible server racks are rendered to reduce computation. The on-demand loading of floor models combines user interaction and the coverage of the frustum. It uses formulas to determine whether the target floor needs to be loaded. When loading, it calls lightweight model data and caches it. Floors that are idle or not within the frustum are unloaded. The two work together and rely on a unified spatial coordinate system. The former reduces the loading range of the latter, and the latter reduces the judgment workload of the former, jointly optimizing the 3D view rendering performance and memory usage. The conversion formula for transforming the local coordinates of the server rack to world coordinates is as follows: Let the coordinates of the server rack in the local coordinate system of its floor be P. obj =(x obj ,y obj ,z obj The offset of this floor in the global world coordinate system of the data center is... Among them O x O z For the floor plan position offset, O y Given the floor height offset, the coordinates P of the server rack in the world coordinate system are... world =(x world ,y world ,z world ),satisfy: The formula for converting world coordinates to camera coordinates is: Let the coordinates of the cabinet in the world coordinate system be P. world =(x world ,y world ,z world The camera's position in the world coordinate system is C = (x... c ,y c ,z c The camera view matrix is ​​M. view Then, the world coordinates need to be translated to a temporary coordinate system P with the camera as the origin. temp =(x temp ,y temp ,z temp Then, through matrix multiplication, it is converted to coordinates P in the camera coordinate system. view =(x view ,y view ,z view The specific formula is as follows: Translate to camera origin P temp =(x world -x c ,y world -y c ,z world -z c ); Matrix transformation Where the view matrix M view It can be represented as the inverse combination of the camera rotation matrix and the translation matrix. Its core function is to align the world coordinate system to the camera coordinate system with the camera as the origin and the viewing direction as the -Z axis. The formula for determining whether the center point of the cabinet is within the view frustum using the plane equation is as follows: Let the plane equation of a certain clipping plane of the view frustum be ax + by + cz + d = 0, where ax, by, cz, and d are the normal vectors of the plane, and the direction of the normal vectors points inward into the view frustum. Let x, y, and z be the coordinates P of the center point of the cabinet in the camera coordinate system. view =(x view ,y view ,z view Substituting the coordinates of the center point into the left side of the plane equation, we obtain the determination equation ax. view +by view +cz view +d, if the result of this judgment is ≥0, it means that the center point is located on the side of the clipping plane pointing to the inside of the view frustum. Only when the judgment results of the center point of the cabinet for all 6 clipping planes are ≥0 can it be determined that the cabinet is inside the view frustum and needs to be included in the rendering range. The formula for determining whether a target floor needs loading is as follows: Let the target floor be the nth floor, and the height of each floor be H. floor Then the range of the Y-axis of the nth layer in the world coordinate system is [Y n,min ,Y n,max ]=[(n-1)×H floor ,n×H floor Simultaneously, by using the Y-axis coordinates of the upper and lower clipping planes of the view frustum in the world coordinate system, the visible range of the current view frustum along the Y-axis is determined to be [Y]. view,min ,Y view,max If the intersection condition Y is satisfied... n,min ≤Y view,max And Y n,max ≥Y view,min This indicates that there is spatial overlap between the target floor and the visible range of the view frustum, requiring loading to be triggered; otherwise, if there is no overlap, loading is not required.

3. A management-oriented data center system according to claim 1, characterized in that, The system configuration module manages global types uniformly through types / index.ts, achieving full project type coverage for API returns, component props, and utility functions.

4. A management-oriented data center system according to claim 2, characterized in that, The 3D visualization module has a built-in lightweight model processing unit. This unit uses the Draco compression algorithm to compress the vertex data of the rack model and converts the texture maps to WebP format, keeping the file size of a single-floor 3D model within 500KB. It also features an asynchronous loading component that uses the loadAsync method of Three.js to load the floor model asynchronously. During loading, it outputs real-time progress data to the interaction adaptation module, which displays a loading progress bar in the 3D view interface. Once loading is complete, it automatically triggers a frustum culling pre-judgment process.

5. A management-oriented data center system according to claim 1, characterized in that, The data processing unit includes a memory leak prevention submodule. When the data center management module updates the rack status or the 3D visualization module unloads the floor model, the memory leak prevention submodule automatically calls the Geometry.dispose(), Material.dispose(), and Texture.dispose() methods of Three.js to completely release the geometry, material, and texture resources associated with the model. At the same time, a memory monitoring timer is set to collect browser memory usage data every 30 seconds. When a certain type of resource is detected to occupy more than 50MB of memory for 5 consecutive times, a forced cleanup command is triggered and a log is generated and synchronized to the administrator log panel of the system configuration module.

6. A data processing method for a management-oriented data center system as described in claim 1, characterized in that, Includes the following steps: S1: Multi-dimensional data collection: The data processing unit collects data center physical structure data and rack resource data through standardized API interfaces, in collaboration with the data center management module, and collects enterprise configuration data in collaboration with the system configuration module and equipment operation status data in collaboration with the resource monitoring module. At the same time, it receives user feedback data transmitted by the feedback loop module. During the collection process, the feedback loop module and the resource monitoring module clean the data through front-end input filtering, remove invalid and malicious data, and then upload it to the data processing unit in a unified manner. S2: 3D Scene Construction and Real-time Rendering: The data processing unit pushes the collected physical structure data and enterprise configuration data to the 3D visualization module. The 3D visualization module constructs a 3D stereoscopic model based on the data center scene manager encapsulated by Three.js, and calls the enterprise-specific identification color and rack status color rules preset by the system configuration module. At the same time, the 3D visualization module independently enables frustum culling technology and on-demand loading strategy to optimize rendering performance and achieve smooth display of 3D scenes. S3: Dynamic synchronization of resource status: The data processing unit establishes a real-time data synchronization channel through the Axios interceptor to synchronize the device operation status data and resource usage data to the 3D visualization module, resource monitoring module and data center management module, respectively; when the indicators change, the resource monitoring module and 3D visualization module trigger a smooth transition animation to ensure data consistency among the modules. S4: Interactive command response: The interactive adaptation module receives user input commands for view switching, resource retrieval, rack operation, and configuration modification, and transmits the commands to the data processing unit. The data processing unit parses and processes instructions based on the Pinia hierarchical state model, and synchronously distributes update instructions to the corresponding modules: pushes view switching data to the 3D visualization module, pushes search conditions to the resource monitoring module, pushes rack operation instructions to the data center management module, and pushes configuration modification requests to the system configuration module; at the same time, the interaction adaptation module supports shortcut key response and drag-and-drop adaptation to improve operation efficiency. S5: Feedback Data Closed-Loop Processing: The feedback closed-loop module receives user-submitted feedback data, verifies its validity on the front end, and then transmits it to the data processing unit. The data processing unit anonymizes the feedback data, stores it in the backend database, and synchronizes it to the administrator's statistical dashboard in the feedback closed-loop module. The administrator marks feedback as "processed / pending" and records follow-up results through the feedback closed-loop module, forming a closed-loop management of "submission-review-processing-follow-up," and the operation log is synchronized to the data processing unit for archiving.

7. The data processing method for a management-oriented data center system according to claim 6, characterized in that, The specific logical steps of S2 are as follows: S201: The 3D visualization module receives physical structure data and enterprise configuration data pushed by the data processing unit. The physical structure data includes the rack's local coordinates P. obj =(x obj ,y obj ,z obj Floor height H foor The number and layout rules of each server room, enterprise configuration data including the resident enterprise ID and its corresponding unique identifier color, the 3D visualization module sends a request to the system configuration module to call preset parameters, which include rack status color rules and 3D view basic parameters. The rack status color rules include used → blue, idle → gray, not enabled → white. The 3D view basic parameters include the initial camera position CameraPos = (x c ,y c ,z c Vertical field of view f ovY Aspect ratio, near clipping plane distance, far clipping plane distance; S202: Data processing unit based on floor height H foor The world offset for the nth layer is generated using the following formula: Among them O y = (n-1)×H floor The X / Z axis offset is determined by the floor plan position, and the local coordinates of each rack are converted into world coordinates to ensure that the racks are accurately mapped in the 3D scene. S203: Based on world coordinate data and hexagonal ring layout rules, the 3D visualization module constructs a 5-story building model using the Mesh class in Three.js. Each floor generates 12 hexagonal server room models, and the server room boundaries are defined by the coordinates of the polygon vertices. Basic materials are assigned to the server room models, and the enterprise ID associated with each server room is recorded. At the same time, the server room management module generates rack models at the corresponding server room world coordinate positions according to a 30×24 grid rule, and assigns material colors to the rack models according to the "enterprise identification color" and "rack status color rules" configured in the system. If the rack has been assigned to enterprise A, then enterprise A's exclusive RGB color is used; if it has not been assigned, the corresponding color is used according to the status. S204: The 3D visualization module calculates the width and height dimensions of the near / far clipping planes based on camera parameters and determines the shape of the view frustum using a formula. The formula used is... S205: Set the rack's world coordinates P world Convert the coordinates to the camera coordinate system and use the plane equation to determine whether the center point of the cabinet is inside the view frustum; S206: If the user is in the "Overall View" view, the 3D visualization module calculates the visible range of the view frustum along the Y-axis in the world coordinate system using camera parameters. view,min ,Y view,max If in "Single-Layer Focus" view, the visible range is directly defined by the Y-axis range of the current floor. For the nth floor, its Y-axis range is first calculated using a formula. n,min ,Y n,max ]=[(n-1)×H floor ,n×H floor Then, the intersection formula is used to determine whether loading is needed. The formula is as follows: S207: For floors determined to be "need to be loaded", the 3D visualization module loads a lightweight model through Three.js's GLTFLoader, and caches it to the "floor model cache pool" after loading; for floors that are "not loaded", only the basic location information is retained, and the model data is not loaded for the time being. S208: The 3D visualization module renders the visible cabinets in the rendering queue in the order of "from closest to furthest from the camera", and calls Three.js's WebGLRenderer to output the 3D scene to the monitoring dashboard of the interactive adaptation module. S209: The data processing unit synchronizes the equipment operation status data to the 3D visualization module in real time. The 3D visualization module immediately updates the rendering status of the corresponding computer room / rack and triggers a smooth transition animation to ensure that the 3D scene is synchronized with the actual status in real time.

8. The data processing method for a management-oriented data center system according to claim 6, characterized in that, The specific logical steps of S4 are as follows: S401: Receive user operations, including clicking the "Floor Switch" button, entering resource search keywords, selecting cabinet operations, and pressing shortcut keys, and encapsulate the instructions into "type + parameter" format and transmit them to the data processing unit; S402: After receiving the instruction, the data processing unit first calls the RBAC permission rules of the system configuration module to verify the permissions of the instruction initiator. If it is a "configuration modification" instruction and the initiator is a regular user, it directly returns a "insufficient permissions" response, and the interaction adaptation module pops up a window to prompt the user. If the permissions are compliant, it enters the instruction parsing stage. S403: The data processing unit loads the Pinia hierarchical state model, which contains three types of state pools. Each state pool is bound to the instruction parsing logic. The global state pool is used to store cross-module shared data, the module state pool is used to store the configuration specific to each module, and the page state pool is used to store the current page interaction data. The data processing unit extracts the associated data from the corresponding state pool according to the instruction type to complete the instruction parameter completion and logic parsing. S404: The data processing unit pushes a "view switching parameter package" to the 3D visualization module, which includes the target floor ID, floor world offset and view mode. The 3D visualization module switches to the 2nd floor 3D view based on the parameter package, and updates the view frustum range simultaneously. At the same time, the interaction adaptation module updates the highlighted state of the "floor control panel" and records the frequency of shortcut key usage. S405: The data processing unit pushes a "search condition package" to the resource monitoring module, containing the keyword "Company A", search fields, and matching rules. Based on the fuzzy matching rules, the resource monitoring module traverses the list of enterprise resource data and filters out target enterprises whose names contain 'A'. The number of search results N is counted according to the formula: The resource monitoring module displays the search results in a table format and triggers a lightweight animation, while simultaneously caching the results to the Pinia module's state pool. S406: The data processing unit pushes a "rack operation parameter package" to the data center management module, which includes the operation type, a list of target rack IDs, and the target enterprise ID. The data center management module updates the rack-enterprise association table, changes the target rack status from "idle" to "used," and synchronizes the data center utilization rate. At the same time, the 3D visualization module receives status update notifications from the data center management module and switches the target rack color to Enterprise A's exclusive identification color to ensure that the 3D view is consistent with the management data; S407: The data processing unit pushes a "configuration modification parameter package" to the system configuration module, which includes the modification type, parameter name, old value, and new value. The system configuration module updates the "monitoring index threshold table", changes the temperature alarm threshold to 32℃, and generates a configuration modification log simultaneously. At the same time, the system configuration module pushes a "threshold update notification" to the resource monitoring module through the data processing unit. The resource monitoring module immediately uses the new threshold to determine the alarm status. S408: After each module executes the command, it returns the "execution result" to the interaction adaptation module. The interaction adaptation module provides visual or status feedback to the user. The interaction adaptation module counts the frequency of shortcut key usage. If the usage frequency of a certain shortcut key is higher than that of the default shortcut key for 7 consecutive days, it automatically prompts the administrator whether to set it as the new default shortcut key. When dragging and adapting, the interaction adaptation module uses the "grid snapping algorithm" to ensure that the panel position is aligned, avoid panel overlap, and improve the ease of operation.

Images