Server production method and system based on industrial internet

Abstract

The invention relates to the technical field of control systems, and provides a server production method and system based on an industrial internet. Acquiring hardware parameters, environment parameters and production parameters of the server through the intelligent sensor; performing filtering processing on the hardware parameters, the environment parameters and the production parameters to generate a structured space-time matrix; performing nonlinear autoregression processing on the space-time matrix, constructing a space-time coupled partial differential equation, and outputting a space-time distribution model of the server load through the partial differential equation; based on resource allocation attributes corresponding to the spatial-temporal distribution model, a server function is generated and solved, and a resource allocation strategy is obtained; and generating a control instruction based on the resource allocation strategy, and sending the control instruction to related production equipment. Perceiving, decision making, execution and optimization collaboration are achieved, and the resource utilization rate, the system energy efficiency and the flexibility of production control are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of control system technology, and more specifically, to a server production method and system based on the Industrial Internet. Background Technology

[0002] The application of server production methods based on the Industrial Internet stems from the dual drivers of digital transformation in the manufacturing industry and the demand for high-performance computing. Currently, technologies such as cloud computing, big data, artificial intelligence, and industrial control systems are developing rapidly, and the scale of data centers is growing exponentially. As a core infrastructure, the production efficiency and operational energy efficiency of servers directly affect enterprise costs and service quality.

[0003] Traditional server production models have many pain points, such as static resource allocation, which relies on manual experience to configure resources such as the central processing unit (CPU) and memory. This makes it difficult to adapt to dynamic load changes and can easily lead to overload or idle resources. Therefore, related technologies have insufficient flexibility in resource allocation during server production. Summary of the Invention

[0004] This application provides a server production method and system based on the Industrial Internet, which can at least partially solve the problem of insufficient flexibility in resource allocation during the server production process.

[0005] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.

[0006] According to one aspect of this application, a server production method based on the Industrial Internet is provided, comprising: acquiring server hardware parameters, environmental parameters, and production parameters through intelligent sensors; filtering the hardware parameters, environmental parameters, and production parameters to generate a structured spatiotemporal matrix; performing nonlinear autoregressive processing on the spatiotemporal matrix to construct a spatiotemporally coupled partial differential equation, and outputting a spatiotemporal distribution model of server load through the partial differential equation; generating and solving a server function based on the resource allocation attributes corresponding to the spatiotemporal distribution model to obtain a resource allocation strategy; generating control commands based on the resource allocation strategy, and sending the control commands to relevant production equipment.

[0007] In this application, based on the aforementioned scheme, the step of filtering the hardware parameters, environmental parameters, and production parameters to generate a structured spatiotemporal matrix includes: generating a prior state estimate for the current moment based on the hardware parameters, environmental parameters, and production parameters; determining the prior error covariance matrix for the current moment based on the prior error covariance matrix for the previous moment, and determining the posterior parameters using the prior error covariance matrix for the current moment; determining the posterior state estimate for the current moment based on the prior state estimate and the posterior parameters; and outputting the spatiotemporal matrix corresponding to the target dataset based on the posterior state estimate for the current moment.

[0008] In this application, based on the aforementioned scheme, generating the prior state estimate at the current moment based on the hardware parameters, the environmental parameters, and the production parameters includes: generating the prior state estimate at time k based on the hardware parameters, the environmental parameters, and the production parameters. for:

[0009]

[0010] Wherein, the state transition matrix A is derived from the thermodynamic model of the server, B represents the control input matrix, and u k This refers to the control parameters of the hardware device.

[0011] In this application, based on the aforementioned scheme, determining the prior error covariance matrix at the current time based on the prior error covariance matrix at the previous time, and determining the posterior parameters using the prior error covariance matrix at the current time, includes:

[0012] Based on the prior error covariance matrix corresponding to the previous time step, determine the prior error covariance matrix P corresponding to the current time step. k|k-1 for:

[0013] P k | k-1 =AP k-1 | k-1 A T +Q

[0014] Among them, P k-1 | k-1 Let Q represent the prior error covariance matrix corresponding to the previous time step, Q represent the process noise covariance, and T represent the transpose operation.

[0015] Based on the prior error covariance matrix P corresponding to the current time k|k-1 Determine the posterior parameter K k for:

[0016] K k =P k |k-1 H T HP k | k-1 H T +R) -1

[0017] Where H represents the observation matrix, which maps parameters such as temperature and vibration to sensor readings, and R represents the sensor noise covariance.

[0018] In this application, based on the aforementioned scheme, the step of performing nonlinear autoregressive processing on the spatiotemporal matrix to construct a spatiotemporally coupled partial differential equation, and outputting a spatiotemporal distribution model of server load through the partial differential equation, includes: performing nonlinear autoregressive processing on the spatiotemporal matrix to construct a spatiotemporally coupled partial differential equation; and discretizing and solving the partial differential equation using the finite difference method to output a spatiotemporal distribution model of server load.

[0019] In this application, based on the aforementioned scheme, the step of generating and solving a server function based on the resource allocation attributes corresponding to the spatiotemporal distribution model to obtain a resource allocation strategy includes: generating a server function based on the resource allocation attributes and dual variables corresponding to the spatiotemporal distribution model; and iteratively solving the server function using the gradient method to obtain a resource allocation strategy.

[0020] In this application, based on the aforementioned scheme, the step of generating control instructions based on the resource allocation strategy and sending the control instructions to the relevant production equipment includes: converting the resource allocation strategy into control instructions based on a preset mapping rule; and sending the control instructions to the relevant production equipment.

[0021] According to one aspect of this application, an industrial internet-based server production system is provided, comprising:

[0022] The acquisition module is used to acquire the server's hardware parameters, environmental parameters, and production parameters through smart sensors;

[0023] The filtering module is used to filter the hardware parameters, environmental parameters, and production parameters to generate a structured spatiotemporal matrix.

[0024] The model module is used to perform nonlinear autoregressive processing on the spatiotemporal matrix, construct spatiotemporally coupled partial differential equations, and output the spatiotemporal distribution model of server load through the partial differential equations.

[0025] The strategy module is used to generate and solve server functions based on the resource allocation attributes corresponding to the spatiotemporal distribution model to obtain the resource allocation strategy.

[0026] The instruction module is used to generate control instructions based on the resource allocation strategy and send the control instructions to the relevant production equipment.

[0027] According to one aspect of this application, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements the server production method based on the Industrial Internet as described in the above embodiments.

[0028] According to one aspect of this application, an electronic device is provided, comprising: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the industrial internet-based server production method as described in the above embodiments.

[0029] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the industrial internet-based server manufacturing method provided in the various optional implementations described above.

[0030] In the technical solution of this application, intelligent sensors comprehensively collect hardware, environmental, and production parameters, solving the problems of lag and incompleteness in traditional manual data collection; filtering processes suppress noise and correct missing values, generating a structured spatiotemporal matrix to provide a unified input for modeling; nonlinear autoregression and partial differential equations couple the spatiotemporal propagation law of the load, breaking through the limitations of linear models and accurately predicting load distribution; dynamic matching of resources and loads through multi-objective constraints avoids idleness or overload; control commands are transmitted and executed via a reliable protocol, and closed-loop feedback continuously optimizes the strategy. Overall, it achieves coordinated perception, decision-making, execution, and optimization, significantly improving resource utilization, system energy efficiency, and the flexibility of production control.

[0031] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0032] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0033] Figure 1The flowchart illustrating a server production method based on the Industrial Internet is shown in one embodiment of this application.

[0034] Figure 2 The flowchart illustrating the generation of a spatiotemporal matrix is ​​shown in one embodiment of this application.

[0035] Figure 3 The illustration shows a schematic diagram of a server production system based on the Industrial Internet of Things in one embodiment of this application.

[0036] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0037] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art.

[0038] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.

[0039] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0040] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0041] The implementation details of the technical solution of this application are described below:

[0042] Figure 1 A flowchart illustrating a server manufacturing method based on the Industrial Internet according to an embodiment of this application is shown. (Refer to...) Figure 1 As shown, this server production method based on the Industrial Internet includes at least steps S110 to S150, which are described in detail below:

[0043] The S110 acquires server hardware parameters, environmental parameters, and production parameters through smart sensors.

[0044] In one embodiment of this application, the intelligent sensor network covers the entire lifecycle of the server, and hardware parameter acquisition is achieved through embedded sensors and protocol interfaces. In this embodiment, the intelligent sensors embed IoT chips or sensor network chips. Specifically, during server production, static parameters such as the number of CPU cores and memory bandwidth are directly read through the data transmission interface, while dynamic data such as temperature and voltage are captured in real time by the transmission tool. Environmental parameter monitoring uses distributed temperature and humidity sensors and triaxial vibration sensors, covering key locations such as racks and hard disk arrays, and data is uploaded at a preset frequency through the data transmission protocol. Production parameters are obtained by parsing historical logs through the technology stack, extracting time-series information such as fault codes and maintenance records, forming a raw data stream covering hardware status, environmental conditions, and production records.

[0045] By predicting and updating the iteratively corrected state estimate, the state transition matrix A is derived from the server's thermodynamic model. The control input matrix B relates to adjustment parameters such as the cooling fan speed, and the observation matrix H maps parameters such as temperature and vibration to sensor readings. The filtering process uses the sensor noise covariance calibrated by the variance method to adjust the Kalman gain, balancing the weights of predicted and observed values. Actual tests show that this filtering method reduces noise variance by 12%, improves the signal-to-noise ratio to 28dB, and achieves a temporal resolution of 10ms for the output dataset. The spatial dimension includes 24 hardware and environmental features, directly driving subsequent dynamic load modeling and resource optimization.

[0046] Specifically, in the process of generating the state transition matrix A, based on the dynamic characteristics of the server's thermodynamic model, a set of differential equations describing thermodynamic state variables such as CPU temperature and ambient temperature is first established using the law of conservation of energy. For example, the rate of change of CPU temperature is expressed as a linear combination of its own temperature, cooling fan speed, and ambient temperature. Subsequently, the nonlinear thermodynamic equations are linearized and transformed into a standard linear state-space model, where the elements of the state transition matrix A correspond to the coupling coefficients between state variables (such as thermal resistance and thermal capacity parameters). Finally, the continuous-time model is transformed into a discrete form using a discretization method to obtain the state transition matrix A used in Kalman filtering. Its element values ​​are determined through server aging experiments or theoretical calculations, reflecting the intensity of thermal interaction between various components.

[0047] The above process ensures the comprehensiveness of the data source by covering three types of parameters: hardware, environment, and production through intelligent sensors; it suppresses noise and corrects missing values ​​by using Kalman filtering and standardized anomaly detection, thereby improving data reliability; and it transforms discrete data into a spatiotemporal matrix through spatiotemporal alignment and structured processing, providing a unified input format for subsequent modeling. This layer solves the problems of lag and incompleteness in traditional manual data collection, enabling accurate perception of dynamic loads.

[0048] S120, the hardware parameters, environmental parameters, and production parameters are filtered to generate a structured spatiotemporal matrix.

[0049] In one embodiment of this application, the edge computing node performs core preprocessing tasks. For missing values ​​caused by sensor disconnection, a regression model is used to predict and fill in the missing values, with a preset number of sample points set in the neighborhood. Outliers with measured values ​​exceeding the mean are removed using a standardization method to suppress environmental interference noise. Time synchronization uses a data transmission protocol to calibrate the clocks of each sensor, ensuring that the synchronization accuracy of timestamps from multiple data sources is less than a set threshold. A mapping relationship between server hardware parameters and sensor spatial locations is established, forming a spatial index table. Finally, the cleaned and aligned data is compressed into a structured spatiotemporal matrix according to the time series, providing a unified input format for subsequent analysis.

[0050] like Figure 2 As shown, in one embodiment of this application, the hardware parameters, environmental parameters, and production parameters are filtered to generate a structured spatiotemporal matrix, including:

[0051] S210, Based on the hardware parameters, the environmental parameters, and the production parameters, generate a priori state estimate for the current moment;

[0052] S220, based on the prior error covariance matrix corresponding to the previous time step, determine the prior error covariance matrix corresponding to the current time step, and determine the posterior parameters through the prior error covariance matrix corresponding to the current time step;

[0053] S230, Based on the prior state estimate at the current moment and the posterior parameter, determine the posterior state estimate at the current moment;

[0054] S240, based on the posterior state estimate at the current moment, output the spatiotemporal matrix corresponding to the target dataset.

[0055] In this embodiment, based on the hardware parameters, environmental parameters, and production parameters, a priori state estimate at time k is generated by predicting and updating the iteratively corrected state estimate. for:

[0056]

[0057] Here, the state transition matrix A is derived from the server's thermodynamic model, and B represents the control input matrix, which is associated with adjustment variables such as the cooling fan speed. u represents the prior state estimate corresponding to the previous time step. k This represents the control parameters of the hardware device. Specifically, the state transition matrix A is derived from the server's thermodynamic model, and the control input matrix B is generated by adjusting parameters such as the cooling fan speed.

[0058] Simultaneously, based on the prior error covariance matrix corresponding to the previous time step, the prior error covariance matrix P corresponding to the current time step is determined. k|k-1 for:

[0059] P k | k-1 =AP k-1 | k-1 A T +Q

[0060] Among them, P k-1 | k-1 Let represent the prior error covariance matrix corresponding to the previous time step, Q represent the process noise covariance, and T represent the transpose operation. The prior error covariance matrix is ​​calculated as described above to quantify the uncertainty of the prediction.

[0061] Then, based on the prior error covariance matrix P corresponding to the current time... k|k-1 Determine the posterior parameter K k for:

[0062] K k =P k | k-1 H T HP k | k-1 H T +R) -1

[0063] Where H represents the observation matrix, which maps parameters such as temperature and vibration to sensor readings, and R represents the sensor noise covariance.

[0064] Then, based on prior state estimation and posterior parameter K k The posterior state estimate at time k is determined as follows:

[0065]

[0066] Among them, z k This represents the sensor observation value at time k, such as the CPU temperature read by an infrared thermometer at time k.

[0067] By calculating the posterior state estimate at the current moment to balance the weights of predicted and observed values, the noise variance of the final output is reduced, generating a high-signal-to-noise ratio, high-quality dataset. Optionally, this dataset is output in the form of a spatiotemporal matrix, with the time dimension covering a preset sampling interval and the spatial dimension containing a preset number of hardware and environmental features, directly driving subsequent dynamic load modeling and resource optimization.

[0068] The above process, through edge computing nodes, controls data preprocessing latency within a preset time, meeting the response requirements of the server production line. It achieves the transformation from multi-source heterogeneous data to high-quality spatiotemporal datasets, providing a data foundation for industrial internet server production methods.

[0069] S130, perform nonlinear autoregressive processing on the spatiotemporal matrix to construct a spatiotemporally coupled partial differential equation, and output the spatiotemporal distribution model of the server load through the partial differential equation.

[0070] In one embodiment of this application, nonlinear autoregressive processing extracts deep features from the spatiotemporal matrix through a network. The network employs a dual hidden layer structure. The input layer receives multi-scale spatiotemporal features decomposed by wavelet transform, such as low-frequency approximation components and high-frequency detail components. The hidden layer captures nonlinear relationships through activation functions, and the output layer generates load prediction values. The spatiotemporally coupled partial differential equation is constructed by combining diffusion and logistic growth terms. The diffusion term simulates the propagation effect of load in physical space (e.g., between racks), the logistic growth term characterizes the self-increase trend of load over time, and the external driving term integrates the real-time impact of environmental sensor data on the load. The partial differential equation is discretized and solved using the finite difference method, with the spatial term using central difference. Parameter calibration relies on thermal simulation experiments and fitting with historical load data. The final output load spatiotemporal distribution model can dynamically reflect load changes in various regions of the server with temporal resolution and millimeter-level spatial resolution, providing a precise basis for resource scheduling.

[0071] In one embodiment of this application, the spatiotemporal matrix is ​​subjected to nonlinear autoregressive processing to construct a spatiotemporally coupled partial differential equation, and a spatiotemporal distribution model of the server load is output through the partial differential equation, including:

[0072] The spatiotemporal matrix is ​​subjected to nonlinear autoregressive processing to construct a spatiotemporally coupled partial differential equation;

[0073] The partial differential equation is discretized using the finite difference method to solve the problem, and the spatiotemporal distribution model of the server load is output.

[0074] The technical solution of this application is based on the structured spatiotemporal matrix output from the aforementioned steps, which includes multiple spatiotemporal features such as CPU temperature, vibration amplitude, and memory bandwidth. First, a spatiotemporal grid covering the physical space of the server (e.g., rack, CPU core location) is constructed, and the time dimension is divided according to preset time intervals. The spatiotemporal matrix is ​​then initialized into a prediction model using a nonlinear autoregressive network. The network employs a dual hidden layer structure; the input layer connects to multi-source sensor data, and the hidden layer extracts multi-scale information of the spatiotemporal features through wavelet transform, providing a data-driven framework for load dynamic modeling.

[0075] In this embodiment of the application, the spatiotemporal matrix is ​​subjected to nonlinear autoregressive processing to construct a spatiotemporally coupled partial differential equation as follows:

[0076]

[0077] Where α, β, and γ represent the diffusion coefficient, logical growth coefficient, and external driving coefficient, respectively. L represents the Laplace operator; L(t,x) represents the spatiotemporal distribution model of server load, i.e., the load value at time t and spatial location x. max This indicates the maximum load, determined by the server hardware specifications; w i D represents the contribution of the i-th sensor feature to the server load. i (t,x) represents the i-th sensor feature in the structured spatiotemporal matrix, such as the measurement of CPU temperature at time t and location x, where i and n represent the identifier and total number of sensor features, respectively.

[0078] After discretization using the finite difference method, and combined with server physical characteristics for parameter calibration, the diffusion coefficient, determined by thermal simulation experiments, describes the spatial diffusion rate of the load. The logical growth coefficient, fitted based on historical load data, characterizes the nonlinear trend of load self-growth. The external driving coefficient quantifies the influence weight of environmental data on the load through correlation analysis. The original data is decomposed into low-frequency approximations and high-frequency details, improving the model's ability to capture sudden load changes.

[0079] Optionally, the partial differential equations are discretized using the finite difference method, and multi-scale features are extracted by combining wavelet transform to generate a high-precision spatiotemporal load distribution model. The spatiotemporal distribution model of the server load is output, providing a dynamic basis for resource scheduling and forming a complete closed loop from data acquisition to load prediction to resource optimization.

[0080] The above process extracts multi-scale spatiotemporal features through a nonlinear autoregressive network combined with wavelet transform, capturing both local fluctuations and global trends in the load. Partial differential equations, through the coupling of diffusion and logistic growth terms, characterize the propagation patterns of the load in both physical space and time, overcoming the limitations of traditional linear models in representing complex dynamics. This layer transforms the raw data into a predictable load distribution model, providing a basis for resource scheduling decisions.

[0081] S140, Based on the resource allocation attributes corresponding to the spatiotemporal distribution model, generate and solve the server function to obtain the resource allocation strategy.

[0082] In one embodiment of this application, the core of the process of generating a resource allocation strategy based on a spatiotemporal distribution model lies in transforming dynamic load characteristics into an executable resource scheduling scheme. The system first extracts key parameters from the load spatiotemporal distribution model, such as the peak load time window, load diffusion rate, and environmental interference coefficient for each server node. It then constructs a resource-load mapping relationship by combining this with a hardware resource list, which includes information such as the number of CPU cores, memory bandwidth, and storage speed. The multi-objective optimization problem is decomposed into a main problem and subproblems using the Lagrange relaxation algorithm. The main problem focuses on minimizing latency and energy consumption, while the subproblems handle resource capacity constraints. The two are iteratively coupled using the subgradient method. In specific implementation, an initial dual variable is set to represent the resource scarcity level. Each iteration adjusts the CPU allocation ratio and memory quota until the server function converges to a stable value.

[0083] In one embodiment of this application, a server function is generated and solved based on the resource allocation attributes corresponding to the spatiotemporal distribution model to obtain a resource allocation strategy, including:

[0084] Based on the resource allocation attributes and dual variables corresponding to the spatiotemporal distribution model, a server function is generated.

[0085] The server function is solved iteratively using the gradient method to obtain the resource allocation strategy.

[0086] In this embodiment, the spatiotemporal matrix of server load in the spatiotemporal distribution model is used as the core input, and a resource pool is constructed by combining server hardware configurations, such as the number of CPU cores and memory capacity. First, the load data is spatiotemporally aligned, mapping the spatiotemporal matrix to physical resource locations, such as rack coordinates, to generate a task priority queue. The server function is designed based on a weighted sum of latency, energy consumption, and cost in resource allocation attributes. The weighting coefficients are set according to the business scenario; specifically, the weighting coefficient for real-time tasks is 0.6, for energy-sensitive scenarios it is 0.3, and for cost-priority scenarios it is 0.1. The resource allocation matrix A is initialized as a zero matrix, representing an unallocated state.

[0087] The constraint violation is incorporated into the server function as a penalty term. The server function L(A,u) is constructed based on the dual variables as follows:

[0088]

[0089] Where A represents the set of allocation strategies, μ represents the set of dual variables; λ1, λ2, and λ3 represent the weight coefficients of real-time tasks, energy-sensitive scenarios, and cost-priority scenarios, respectively, and E yan E nen E che These represent the linear relationships between task latency, energy consumption, and cost consumption and resource usage, respectively; μ i Let A represent the dual variable. i,j C represents the allocation strategy for resource type i to task j. i This represents the configuration parameters of the server hardware, where i and j represent the resource type and task identifier, respectively.

[0090] Optionally, the initial value of the dual variable is set to 0.1 and dynamically adjusted according to resource scarcity. For example, when CPU resources are scarce, the CPU's dual variable automatically increases, forcing the algorithm to prioritize satisfying CPU constraints. This method reduces the computational complexity of the problem and significantly improves solution efficiency. Through the above process, with the dual variable μ fixed... i Then, the minimum value of the server function with respect to A is found to obtain the optimal allocation strategy under the current dual variable, and the dual variable is updated according to the constraint violation.

[0091] Specifically, after fixing the dual variables, the optimal allocation decision variables under the current dual variables are obtained by calculating the minimum value of the server function with respect to the decision variables. Here, the decision variables represent the amount of resource type i allocated to task j at time t and location x, optimized using the Lagrange relaxation algorithm. Subsequently, the dual variables are updated based on the constraint violation amounts, generating new dual variables. for:

[0092]

[0093] in, represents the original dual variable, and 'a' represents the step size, which can take the value 0.05; C represents the optimal allocation strategy for assigning resource type i to task j. i This represents the configuration parameters of server hardware resource i. The iterative process continues until the rate of change of the objective function value is less than a set threshold or the set number of iterations is reached, ultimately outputting the resource allocation strategy.

[0094] After generating the server function, a subgradient method is used to iteratively solve the server function, generating the optimal resource allocation strategy. Through this process, the resource allocation strategy achieves multi-objective balance, reducing latency and energy consumption in real-time task scenarios, improving CPU utilization in cloud server cluster testing, and shortening task completion time. The output resource allocation strategy directly drives the resource scheduling module, dynamically adjusting parameters such as the number of CPU cores and memory bandwidth, forming a complete closed loop from load prediction to resource optimization. While ensuring computational efficiency, this achieves efficient, energy-saving, and low-cost resource management in industrial internet server production.

[0095] The above process, executed in a distributed manner by edge computing nodes, enables rapid solution of the strategy, ensuring millisecond-level response. Multi-objective constraints are transformed into a dual problem, and the weights of latency, energy consumption, and cost are iteratively balanced using the subgradient method. The final output is a resource allocation table incorporating both time and spatial dimensions (rack-level positioning), avoiding local optima in single-objective optimization. This drives production equipment to call upon computing, storage, and network resources on demand, solving the problems of resource idleness and overload in traditional static allocation, and achieving precise matching between dynamic load and resource supply.

[0096] S150, generate control instructions based on the resource allocation strategy, and send the control instructions to the relevant production equipment.

[0097] In one embodiment of this application, the core of the process of generating control commands based on resource allocation strategies and sending them to production equipment lies in transforming abstract resource parameters into specific operations that the equipment can execute. The system first parses the resource allocation table to determine the CPU frequency, memory quota, and storage bandwidth required by each server node at a specific time. Then, it maps the parameters to device commands using a preset rule base: for example, CPU frequency adjustment commands need to be converted into performance status control commands, and memory allocation generates a non-uniform memory access configuration script. After the commands are generated, the industrial internet platform packages the messages using a message transmission protocol and pushes them to edge nodes through a lightweight publish-subscribe model. High-frequency control commands (such as cooling fan speed adjustment) are sent directly to the server's baseboard management controller, while low-frequency strategy commands (such as storage resource allocation) are transmitted to the storage array.

[0098] Optionally, during transmission, the integrity of instructions is verified through digital signatures, and a quality of service (QoS) level is set to ensure that critical instructions are transmitted with priority. Upon receiving an instruction, the device executes it immediately; for example, the server adjusts the frequency based on CPU instructions and simultaneously transmits actual power consumption data back via an interface. If a device times out (e.g., no response for more than 500ms), a redundancy mechanism is automatically triggered, forwarding the instruction to a backup device. Finally, the execution results of all devices are aggregated in the monitoring module, forming a closed-loop feedback loop that dynamically adjusts resource allocation strategies to ensure the continuous and efficient operation of the production environment.

[0099] In one embodiment of this application, generating control commands based on the resource allocation strategy and sending the control commands to the relevant production equipment includes:

[0100] Based on preset mapping rules, the resource allocation strategy is converted into control commands;

[0101] The control commands are sent to the relevant production equipment.

[0102] The abstract resource allocation strategy translates resource parameters (such as the number of CPU cores and memory bandwidth) into specific executable instructions for the device. For example, if the strategy requires a server's CPU utilization to reach 80%, the system will use preset mapping rules to convert the utilization target into a frequency adjustment instruction, such as adjusting from 2.4GHz to 3.2GHz, and simultaneously generate a memory allocation table to specify the memory quota for each process. This process must be considered in conjunction with the device's hardware specifications to ensure that the instructions are feasible within physical limitations.

[0103] Multi-device collaboration is achieved through an industrial internet platform. The system uses data transmission protocols to package instructions into lightweight messages and push them to the edge computing nodes of the corresponding devices. For example, an instruction to adjust the cooling fan speed is sent directly to the server's baseboard management controller, while storage resource allocation instructions are transmitted to the storage array via the iSCSI protocol. During transmission, the platform verifies the integrity of the instructions and ensures security through digital signatures to prevent unauthorized operations.

[0104] Upon receiving a command, the device immediately executes it and reports its status. For example, the server adjusts its performance based on the CPU frequency command and simultaneously transmits actual power consumption data back via an interface. The system monitors execution deviations in real time. If a device times out (e.g., exceeding 500ms), a redundancy mechanism is triggered, forwarding the command to a backup device. Ultimately, the actual resource usage of all devices is aggregated in the data acquisition module of step 1, forming a closed-loop optimization to ensure that resource allocation strategies continuously adapt to changes in the production environment.

[0105] The above process transforms abstract strategies into device commands through a pre-defined rule base. Message transmission protocols and quality of service mechanisms ensure the reliability and real-time performance of command transmission. Redundant device switching and execution result feedback form a closed loop, dynamically adjusting resource allocation strategies. This layer ensures that theoretical strategies can be implemented and that the model is continuously optimized through actual operational data, forming a complete closed loop of perception, decision-making, execution, and optimization.

[0106] This application's technical solution acquires server hardware, environmental, and production parameters using intelligent sensors; filters these parameters to generate a structured spatiotemporal matrix; applies nonlinear autoregressive processing to the spatiotemporal matrix to construct spatiotemporally coupled partial differential equations, outputting a spatiotemporal distribution model of the server load; generates and solves a server function based on the resource allocation attributes corresponding to the spatiotemporal distribution model to obtain a resource allocation strategy; generates control commands based on the resource allocation strategy and sends these commands to relevant production equipment. The intelligent sensors comprehensively collect hardware, environmental, and production parameters, solving the lag and incompleteness problems of traditional manual data collection; filtering suppresses noise and corrects missing values, generating a structured spatiotemporal matrix to provide a unified input for modeling; nonlinear autoregression and partial differential equations couple the spatiotemporal propagation law of the load, overcoming the limitations of linear models and accurately predicting load distribution; multi-objective constraints dynamically match resources and loads, avoiding idleness or overload; control commands are transmitted and executed via a reliable protocol, with closed-loop feedback continuously optimizing the strategy. It enables overall coordination of perception, decision-making, execution, and optimization, significantly improving resource utilization, system energy efficiency, and production control flexibility.

[0107] The following describes embodiments of the industrial internet-based server production system of this application, which can be used to execute the industrial internet-based server production method in the above embodiments of this application. It is understood that the industrial internet-based server production system can be a computer program (including program code) running on a computer device; for example, the industrial internet-based server production system is an application software. This industrial internet-based server production system can be used to execute the corresponding steps in the method provided in the embodiments of this application. For details not disclosed in the embodiments of the industrial internet-based server production system of this application, please refer to the embodiments of the industrial internet-based server production method described above.

[0108] Figure 3 A block diagram of an Industrial Internet-based server production system according to an embodiment of this application is shown.

[0109] Reference Figure 3 As shown, an industrial internet-based server production system according to an embodiment of this application includes:

[0110] The acquisition module 310 is used to acquire the server's hardware parameters, environmental parameters, and production parameters through intelligent sensors;

[0111] The filtering module 320 is used to filter the hardware parameters, the environmental parameters, and the production parameters to generate a structured spatiotemporal matrix.

[0112] Model module 330 is used to perform nonlinear autoregressive processing on the spatiotemporal matrix, construct spatiotemporally coupled partial differential equations, and output a spatiotemporal distribution model of server load through the partial differential equations.

[0113] Strategy module 340 is used to generate and solve server functions based on the resource allocation attributes corresponding to the spatiotemporal distribution model to obtain resource allocation strategies;

[0114] The instruction module 350 is used to generate control instructions based on the resource allocation strategy and send the control instructions to the relevant production equipment.

[0115] In this application, based on the aforementioned scheme, the filtering module includes:

[0116] The estimation unit is used to generate a priori state estimate for the current moment based on the hardware parameters, the environmental parameters, and the production parameters.

[0117] The first determining unit is used to determine the prior error covariance matrix corresponding to the current time based on the prior error covariance matrix corresponding to the previous time, and to determine the posterior parameters through the prior error covariance matrix corresponding to the current time.

[0118] The second determining unit is used to determine the posterior state estimate at the current time based on the prior state estimate at the current time and the posterior parameter.

[0119] The output unit is used to output the spatiotemporal matrix corresponding to the target dataset based on the posterior state estimate at the current time.

[0120] In this application, based on the aforementioned scheme, the estimation unit includes:

[0121] The generation subunit is used to: generate a prior state estimate at time k based on the hardware parameters, the environmental parameters, and the production parameters. for:

[0122]

[0123] Wherein, the state transition matrix A is derived from the thermodynamic model of the server, B represents the control input matrix, and u k This refers to the control parameters of the hardware device.

[0124] In this application, based on the aforementioned scheme, the first determining unit includes:

[0125] The prior subunit is used to determine the prior error covariance matrix P at the current time based on the prior error covariance matrix at the previous time step. k|k-1 for:

[0126] P k |k-1 =AP k-1 | k-1 A T +Q

[0127] Among them, P k-1|k-1 Let Q represent the prior error covariance matrix corresponding to the previous time step, Q represent the process noise covariance, and T represent the transpose operation.

[0128] The posterior subunit is used to determine the prior error covariance matrix P corresponding to the current time step. k|k-1 Determine the posterior parameter K k for:

[0129] K k =P k | k-1 H T HP k | k-1 H T +R) -1

[0130] Where H represents the observation matrix, which maps parameters such as temperature and vibration to sensor readings, and R represents the sensor noise covariance.

[0131] In this application, based on the foregoing scheme, the model module includes:

[0132] The construction unit is used to perform nonlinear autoregressive processing on the spatiotemporal matrix to construct spatiotemporally coupled partial differential equations;

[0133] The solution unit is used to discretize and solve the partial differential equation using the finite difference method, and output the spatiotemporal distribution model of the server load.

[0134] In this application, based on the foregoing scheme, the strategy module includes:

[0135] The function unit is used to generate server functions based on the resource allocation attributes and dual variables corresponding to the spatiotemporal distribution model.

[0136] The strategy unit is used to iteratively solve the server function using the gradient method to obtain the resource allocation strategy.

[0137] In this application, based on the foregoing scheme, the instruction module includes:

[0138] The conversion unit is used to convert the resource allocation strategy into control commands based on preset mapping rules;

[0139] The control unit is used to send the control commands to the relevant production equipment.

[0140] In this technical solution, intelligent sensors comprehensively collect hardware, environmental, and production parameters, solving the problems of lag and incompleteness in traditional manual data collection; filtering processes suppress noise and correct missing values, generating a structured spatiotemporal matrix to provide a unified input for modeling; nonlinear autoregression and partial differential equations couple the spatiotemporal propagation law of the load, breaking through the limitations of linear models and accurately predicting load distribution; dynamic matching of resources and loads through multi-objective constraints avoids idleness or overload; control commands are transmitted and executed via a reliable protocol, with closed-loop feedback continuously optimizing the strategy. Overall, it achieves coordinated perception, decision-making, execution, and optimization, significantly improving resource utilization, system energy efficiency, and the flexibility of production control.

[0141] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown.

[0142] It should be noted that the computer system of the electronic device in this embodiment is only an example and should not impose any limitations on the function and scope of use of the embodiments of this application.

[0143] In this embodiment, the computer system includes a central processing unit 401, which can perform various appropriate actions and processes based on a program stored in a read-only memory 402 or a program loaded from a storage section 408 into a random access memory 403, such as executing the server production method based on the Industrial Internet described in the above embodiment. The random access memory 403 also stores various programs and data required for system operation. The central processing unit 401, the read-only memory 402, and the random access memory 403 are interconnected via a bus 404. An input / output interface 405 is also connected to the bus 404.

[0144] The following components are connected to the input / output interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the input / output interface 405 as needed. A removable medium 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 410 as needed so that computer programs read from it can be installed into the storage section 408 as needed.

[0145] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit 401, it performs various functions defined in the system of this application.

[0146] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. The transmitted data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0147] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0148] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0149] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in the various alternative implementations described above.

[0150] In another aspect, this application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to implement the server production method based on the Industrial Internet described in the above embodiments.

[0151] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0152] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the method according to the embodiments of this application.

[0153] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0154] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A server production method based on the Industrial Internet, characterized in that, include: The server's hardware parameters, environmental parameters, and production parameters are acquired through smart sensors. The hardware parameters, environmental parameters, and production parameters are filtered to generate a structured spatiotemporal matrix; The spatiotemporal matrix is ​​subjected to nonlinear autoregressive processing to construct a spatiotemporally coupled partial differential equation, and the spatiotemporal distribution model of the server load is output through the partial differential equation. Based on the resource allocation attributes corresponding to the spatiotemporal distribution model, a server function is generated and solved to obtain the resource allocation strategy. Control commands are generated based on the resource allocation strategy and then sent to the relevant production equipment.

2. The server production method based on the Industrial Internet according to claim 1, characterized in that, The hardware parameters, environmental parameters, and production parameters are filtered to generate a structured spatiotemporal matrix, including: Based on the hardware parameters, environmental parameters, and production parameters, a priori state estimate for the current moment is generated. Based on the prior error covariance matrix corresponding to the previous time step, determine the prior error covariance matrix corresponding to the current time step, and determine the posterior parameters through the prior error covariance matrix corresponding to the current time step. Based on the prior state estimate at the current moment and the posterior parameters, determine the posterior state estimate at the current moment; Based on the posterior state estimate at the current moment, the spatiotemporal matrix corresponding to the target dataset is output.

3. The server production method based on the Industrial Internet according to claim 2, characterized in that, Based on the hardware parameters, environmental parameters, and production parameters, a priori state estimate for the current moment is generated, including: Based on the hardware parameters, environmental parameters, and production parameters, a priori state estimate for time k is generated. for: Here, the state transition matrix A is derived from the server's thermodynamic model, and B represents the control input matrix. u represents the prior state estimate corresponding to the previous time step. k This refers to the control parameters of the hardware device.

4. The server production method based on the Industrial Internet according to claim 3, characterized in that, Based on the prior error covariance matrix corresponding to the previous time step, the prior error covariance matrix corresponding to the current time step is determined. The posterior parameters are then determined using the prior error covariance matrix corresponding to the current time step, including: Based on the prior error covariance matrix corresponding to the previous time step, determine the prior error covariance matrix P corresponding to the current time step. k|k-1 for: P k | k-1 =AP k-1 | k-1 From T +Q Among them, P k-1|k-1 Let Q represent the prior error covariance matrix corresponding to the previous time step, Q represent the process noise covariance, and T represent the transpose operation. Based on the prior error covariance matrix P corresponding to the current time k|k-1 Determine the posterior parameter K k for: K k =P k|k-1 H T (HP k|k-1 H T +R) -1 Where H represents the observation matrix and R represents the sensor noise covariance.

5. The server production method based on the Industrial Internet according to claim 1, characterized in that, The spatiotemporal matrix is ​​subjected to nonlinear autoregressive processing to construct a spatiotemporally coupled partial differential equation. The spatiotemporal distribution model of the server load is then output through this partial differential equation, including: The spatiotemporal matrix is ​​subjected to nonlinear autoregressive processing to construct a spatiotemporally coupled partial differential equation; The partial differential equation is discretized using the finite difference method to solve the problem, and the spatiotemporal distribution model of the server load is output.

6. The server production method based on the Industrial Internet according to claim 1, characterized in that, Based on the resource allocation attributes corresponding to the spatiotemporal distribution model, a server function is generated and solved to obtain a resource allocation strategy, including: Based on the resource allocation attributes and dual variables corresponding to the spatiotemporal distribution model, a server function is generated. The server function is solved iteratively using the gradient method to obtain the resource allocation strategy.

7. The server production method based on the Industrial Internet according to claim 1, characterized in that, Based on the resource allocation strategy, control commands are generated and sent to the relevant production equipment, including: Based on preset mapping rules, the resource allocation strategy is converted into control commands; The control commands are sent to the relevant production equipment.

8. A server production system based on the Industrial Internet, characterized in that, include: The acquisition module is used to acquire the server's hardware parameters, environmental parameters, and production parameters through smart sensors; The filtering module is used to filter the hardware parameters, environmental parameters, and production parameters to generate a structured spatiotemporal matrix. The model module is used to perform nonlinear autoregressive processing on the spatiotemporal matrix, construct spatiotemporally coupled partial differential equations, and output the spatiotemporal distribution model of server load through the partial differential equations. The strategy module is used to generate and solve server functions based on the resource allocation attributes corresponding to the spatiotemporal distribution model to obtain the resource allocation strategy. The instruction module is used to generate control instructions based on the resource allocation strategy and send the control instructions to the relevant production equipment.

9. The server production system based on the Industrial Internet according to claim 8, characterized in that, The filtering module includes: The estimation unit is used to generate a priori state estimate for the current moment based on the hardware parameters, the environmental parameters, and the production parameters. The first determining unit is used to determine the prior error covariance matrix corresponding to the current time based on the prior error covariance matrix corresponding to the previous time, and to determine the posterior parameters through the prior error covariance matrix corresponding to the current time. The second determining unit is used to determine the posterior state estimate at the current time based on the prior state estimate at the current time and the posterior parameter. The output unit is used to output the spatiotemporal matrix corresponding to the target dataset based on the posterior state estimate at the current time.

10. The server production system based on the Industrial Internet according to claim 9, characterized in that, The estimation unit includes: The generation subunit is used to generate a prior state estimate at time k based on the hardware parameters, the environmental parameters, and the production parameters. for: Here, the state transition matrix A is derived from the server's thermodynamic model, and B represents the control input matrix. u represents the prior state estimate corresponding to the previous time step. k This refers to the control parameters of the hardware device.

Images