Method and device for generating dynamically switchable heat dissipation strategy based on OpenBMC (Open Baseboard Management Controller)

Abstract

The invention discloses a method and device for generating a dynamically switchable heat dissipation strategy based on an OpenBMC, and relates to the technical field of server intelligent heat dissipation regulation and control, and the method comprises the steps: firstly constructing a configuration library comprising optimal heat dissipation strategies corresponding to various devices and models and Sensor thresholds; the configuration library predefines a series of tested heat dissipation strategies and sensor thresholds according to factors such as hardware characteristics, power consumption and heat dissipation capability of equipment. And according to the equipment type, model and the like acquired by the equipment information acquisition module, searching corresponding configuration in the heat dissipation strategy and Sensor threshold configuration library. According to the method, corresponding thresholds and strategies are quickly and accurately matched by using technologies such as a hash table and database query, and the heat dissipation strategies are adjusted in real time according to parameters such as temperature provided by the sensor so as to meet different heat dissipation requirements generated by different devices and models, so that the technical problems of sensor threshold stiffness and non-agility in heat dissipation strategy switching in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent heat dissipation control of servers, and in particular to a method and device for generating a heat dissipation strategy that can be dynamically switched based on OpenBMC. Background Art

[0002] In the current localization process, different manufacturers design the working environment temperature of the same equipment such as SSD, PSU, PCIE, etc., which leads to different temperature alarms and heat dissipation requirements. The current technology cannot adjust these temperature alarm thresholds and heat dissipation strategies according to different manufacturers or models of equipment in real time and take effect in real time when the BMC is running. This is mainly reflected in the following two points:

[0003] Rigid sensor thresholds: In current technologies, sensor threshold settings are often fixed. Even if they can be manually modified, the server BMC or functional module must be restarted to take effect. Dynamic adjustments cannot be made based on the real-time status of the device or changes in external pluggable devices (such as SSDs, PSUs, etc.). Because the operating temperature environments of devices from different manufacturers or different models of devices from the same manufacturer are different, the over-temperature alarm threshold settings are different, which will not only cause the corresponding temperature sensor to have false alarms, but also cause problems such as overheating or excessive heat dissipation of the device.

[0004] The switching of cooling strategies is not agile: The switching mechanism of cooling strategies is not fast and intelligent enough, and lacks flexibility. Because the working performance points of devices from different manufacturers or different models of devices from the same manufacturer are different, the requirements for cooling are also different. If the cooling strategy cannot be adjusted in real time and cannot quickly respond to the optimal performance working temperature points for different devices, it will affect the stable operation and performance of the equipment, and may also cause problems such as equipment downtime and data loss. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a method and device for generating a dynamically switchable heat dissipation strategy based on OpenBMC. The following technical solutions are adopted:

[0006] The method for generating a dynamically switchable cooling strategy based on OpenBMC includes the following steps:

[0007] Step 1: The dynamic strategy module builds a heat dissipation strategy and sensor threshold configuration library, which contains the optimal heat dissipation strategy configuration data and corresponding sensor thresholds for various devices and models;

[0008] Step 2: When a new device is connected to the system, the sensor threshold and cooling strategy switching interface definition are added, the information acquisition module obtains the parameters of the newly connected device, and interacts with the dynamic strategy module to exchange device parameters;

[0009] Step 3: The dynamic strategy module matches the corresponding optimal cooling strategy configuration data and the corresponding sensor threshold in the cooling strategy and sensor threshold configuration library based on the device parameters;

[0010] Step 4. If the match is successful, the dynamic strategy module exchanges the optimal cooling strategy configuration data and the corresponding sensor threshold with the newly connected device based on the DBUS protocol;

[0011] Step 5: The newly connected device executes the optimal heat dissipation strategy configuration data and the corresponding sensor threshold to dissipate heat.

[0012] By adopting the above technical solution, a configuration library containing the optimal cooling strategies and sensor thresholds corresponding to a variety of devices and models is constructed. The configuration library predefines a series of tested cooling strategies and sensor thresholds based on the hardware characteristics, power consumption, cooling capacity and other factors of the device. According to the device type, model and other information obtained by the device information acquisition module, the corresponding configuration is found in the cooling strategy and sensor threshold configuration library. Using hash tables, database queries or configuration file parsing and other technologies, the corresponding thresholds and strategies are quickly and accurately matched. According to the temperature and other parameters provided by the sensor, and according to the cooling strategy adjusted in real time by Dynamic-policy, the speed of the internal fan of the server is adjusted in real time to cope with the different cooling requirements generated by different devices and models. The FSC module provides an interface for modifying the cooling strategy, including pid, setpoint and linear speed regulation strategies, so as to solve the technical problems of rigid sensor thresholds and inflexible switching of cooling strategies in the existing technology.

[0013] Optionally, if the corresponding device parameters are not matched in step 4, the following method is used to generate the optimal cooling strategy configuration data and the corresponding sensor threshold;

[0014] Step 41, extracting multiple related features that have an important impact on the formulation of the heat dissipation strategy from the device parameters;

[0015] Step 42, inputting multiple associated features into the trained large model, and the large model outputs a preliminary heat dissipation strategy;

[0016] Step 43, optimize the preliminary heat dissipation strategy to output the final heat dissipation strategy and sensor threshold, and continue to execute steps 4 and 5.

[0017] Optionally, during the execution of the cooling strategy in step 5, the operating status of the device is continuously interacted with. If it is determined that the cooling-related status of the device does not meet expectations or the device has an abnormal situation, the cooling strategy and sensor threshold are adjusted and the cooling effect is verified again until the cooling expectations are met.

[0018] Optionally, the optimal cooling strategy and corresponding sensor threshold of the new device after optimization and verification are added to the cooling strategy and sensor threshold configuration library.

[0019] Optionally, in step 2, the large model is a decision tree model, a random forest model, or a neural network model.

[0020] By adopting the above technical solution, when the corresponding device parameters cannot be found in the configuration library, the cooling strategy and the corresponding sensor threshold cannot be obtained from the configuration library, and intelligent means are used to automatically generate the corresponding optimal cooling strategy and sensor threshold:

[0021] Data collection: First, collect relevant data from the new device, including but not limited to manufacturer, power, and operating mode, which reflect the current operating status and cooling requirements of the device.

[0022] Feature extraction: Extract heat dissipation-related features that have an important impact on the formulation of heat dissipation strategies from the data, including the device model, specifications, power consumption level, heat dissipation area, air duct design, etc. At the same time, the trend of device temperature values ​​over time will be analyzed to capture the dynamic heat dissipation needs of the device.

[0023] Using large model reasoning or algorithm application:

[0024] ① Rule-based reasoning: If the system adopts rule-based reasoning, it will match the preset rule set with the characteristics of the current device to recommend a suitable cooling strategy. These rules come from the provided rule library, etc.

[0025] ② Machine learning algorithm: If the system uses a machine learning algorithm, the system will use historical data (including the cooling strategies and effects of known devices) to train the model. After receiving the characteristics of the new device, the model will make predictions based on these characteristics and output a preliminary cooling strategy. Algorithms include decision trees, random forests, neural networks, etc., which can handle complex nonlinear relationships and interactions between data.

[0026] Strategy optimization and verification: The initially generated cooling strategy needs to go through an optimization step to ensure its effectiveness and efficiency in actual applications. Optimization may involve fine-tuning of strategy parameters, comparative evaluation of different strategies, etc. In addition, the system will verify the optimized strategy. This is done by setting the generated cooling strategy and sensor thresholds in the actual device.

[0027] Policy output and application: After optimization and verification, the system will output the final cooling strategy and sensor threshold and apply them to the new device.

[0028] Continuous monitoring and feedback: During the application of the cooling strategy, the system will continue to monitor the operating status and cooling effect of the device. If it is found that the strategy does not meet expectations or the device has an abnormal situation, the system will adjust the cooling strategy and sensor threshold in time.

[0029] At the end of the process, the system continues to monitor the operating status of the new device and automatically adjusts the cooling strategy as needed. At the same time, the new device and its corresponding cooling strategy and sensor threshold information are added to the configuration library for quick identification and application in the future.

[0030] Implementation method: Through a large number of experiments and data analysis, the most suitable cooling strategy and sensor threshold are determined for each device and model, and this information is stored in a local data file to form a configuration library. When the corresponding cooling strategy and sensor threshold cannot be found in the configuration library, the system will automatically generate and continuously optimize.

[0031] Optionally, a hardware service module is used to detect parameters of a newly inserted device. The hardware service module detects whether new hardware is inserted. When a new device is inserted, the hardware service module queries the device parameters of the newly inserted hardware through an interface and transmits them to the information acquisition module based on the DBUS protocol.

[0032] By adopting the above technical solution, the hardware service module (such as GPU-Manager, Backplane-Manager, etc.) is used to manage specific hardware and provide detailed information of the hardware to DBUS, including model, specification, manufacturer, etc., to ensure the stable operation of the entire system.

[0033] Optionally, the parameters of the device include the model, specification, power consumption level, heat dissipation area or air duct design of the device.

[0034] A device for generating a dynamically switchable heat dissipation strategy based on OpenBMC, used to implement a method for generating a dynamically switchable heat dissipation strategy based on OpenBMC, the device comprising a dynamic strategy module, a hardware service module, an information acquisition module and an FSC fan speed controller;

[0035] The dynamic strategy module stores predefined parameters for different devices and models and the corresponding cooling strategies and sensor threshold configuration library;

[0036] The hardware service modules include the raid card service module, the backplane service module and the psu service module, which are used to detect newly inserted devices;

[0037] The information acquisition module communicates with the hardware service module to connect and interact with the device parameters;

[0038] The newly inserted device, hardware service module, information acquisition module and dynamic strategy module are interconnected based on DBUS protocol communication.

[0039] In summary, the present invention includes at least one of the following beneficial technical effects:

[0040] The present invention can provide a method and device for generating a dynamically switchable heat dissipation strategy based on OpenBMC, and construct a configuration library containing optimal heat dissipation strategies and sensor thresholds corresponding to a variety of devices and models. The configuration library predefines a series of tested heat dissipation strategies and sensor thresholds based on factors such as the hardware characteristics, power consumption, and heat dissipation capabilities of the device. According to the device type, model, and other information obtained by the device information acquisition module, the corresponding configuration is searched in the heat dissipation strategy and sensor threshold configuration library. Using technologies such as hash tables, database queries, or configuration file parsing, the corresponding thresholds and strategies are quickly and accurately matched. According to parameters such as temperature provided by the sensor, and according to the heat dissipation strategy adjusted in real time by Dynamic-policy, the speed of the internal fan of the server is adjusted in real time to cope with different heat dissipation requirements generated by different devices and models, thereby solving the technical problems of rigid sensor thresholds and inflexible switching of heat dissipation strategies in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flow chart of a method for generating a dynamically switchable heat dissipation strategy based on OpenBMC implemented in the present invention. DETAILED DESCRIPTION

[0042] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0043] The embodiment of the present invention discloses a method and a device for generating a dynamically switchable heat dissipation strategy based on OpenBMC.

[0044] Reference Figure 1 , a method for generating a dynamically switchable cooling strategy based on OpenBMC includes the following steps:

[0045] Step 1: The dynamic strategy module builds a heat dissipation strategy and sensor threshold configuration library, which contains the optimal heat dissipation strategy configuration data and corresponding sensor thresholds for various devices and models;

[0046] Step 2: When a new device is connected to the system, the sensor threshold and cooling strategy switching interface definition are added, the information acquisition module obtains the parameters of the newly connected device, and interacts with the dynamic strategy module to exchange device parameters;

[0047] Step 3: The dynamic strategy module matches the corresponding optimal cooling strategy configuration data and the corresponding sensor threshold in the cooling strategy and sensor threshold configuration library based on the device parameters;

[0048] Step 4. If the match is successful, the dynamic strategy module exchanges the optimal cooling strategy configuration data and the corresponding sensor threshold with the newly connected device based on the DBUS protocol;

[0049] Step 5: The newly connected device executes the optimal heat dissipation strategy configuration data and the corresponding sensor threshold to dissipate heat.

[0050] Build a configuration library that contains the optimal cooling strategies and sensor thresholds for a variety of devices and models. The configuration library predefines a series of tested cooling strategies and sensor thresholds based on the hardware characteristics, power consumption, cooling capacity and other factors of the device. According to the device type, model and other information obtained by the device information acquisition module, find the corresponding configuration in the cooling strategy and sensor threshold configuration library. Use hash tables, database queries or configuration file parsing and other technologies to quickly and accurately match the corresponding thresholds and strategies, and adjust the speed of the internal fan of the server in real time according to the temperature and other parameters provided by the sensor and the cooling strategy adjusted in real time by Dynamic-policy to cope with the different cooling requirements generated by different devices and models. The FSC module provides an interface for modifying the cooling strategy, including pid, setpoint and linear speed regulation strategies, so as to solve the technical problems of rigid sensor thresholds and inflexible switching of cooling strategies in existing technologies.

[0051] If the corresponding device parameters are not matched in step 4, the following method is used to generate the optimal cooling strategy configuration data and the corresponding sensor threshold;

[0052] Step 41, extracting multiple related features that have an important impact on the formulation of the heat dissipation strategy from the device parameters;

[0053] Step 42, inputting multiple associated features into the trained large model, and the large model outputs a preliminary heat dissipation strategy;

[0054] Step 43, optimize the preliminary heat dissipation strategy to output the final heat dissipation strategy and sensor threshold, and continue to execute steps 4 and 5.

[0055] During the execution of the cooling strategy in step 5, the device continues to interact with the running status. If it is determined that the cooling status of the device does not meet expectations or the device is abnormal, the cooling strategy and sensor threshold are adjusted and the cooling effect is verified again until the cooling expectation is met.

[0056] The optimal cooling strategy and corresponding sensor threshold of the new device after optimization and verification are added to the cooling strategy and sensor threshold configuration library.

[0057] In step 2, the large model is a decision tree model, a random forest model or a neural network model.

[0058] When the corresponding device parameters cannot be found in the configuration library, the cooling strategy and the corresponding sensor threshold cannot be obtained from the configuration library. Intelligent means are used to automatically generate the corresponding optimal cooling strategy and sensor threshold:

[0059] Data collection: First, collect relevant data from the new device, including but not limited to manufacturer, power, and operating mode, which reflect the current operating status and cooling requirements of the device.

[0060] Feature extraction: Extract heat dissipation-related features that have an important impact on the formulation of heat dissipation strategies from the data, including the device model, specifications, power consumption level, heat dissipation area, air duct design, etc. At the same time, the trend of device temperature values ​​over time will be analyzed to capture the dynamic heat dissipation needs of the device.

[0061] Using large model reasoning or algorithm application:

[0062] ① Rule-based reasoning: If the system adopts rule-based reasoning, it will match the preset rule set with the characteristics of the current device to recommend a suitable cooling strategy. These rules come from the provided rule library, etc.

[0063] ② Machine learning algorithm: If the system uses a machine learning algorithm, the system will use historical data (including the cooling strategies and effects of known devices) to train the model. After receiving the characteristics of the new device, the model will make predictions based on these characteristics and output a preliminary cooling strategy. Algorithms include decision trees, random forests, neural networks, etc., which can handle complex nonlinear relationships and interactions between data.

[0064] Strategy optimization and verification: The initially generated cooling strategy needs to go through an optimization step to ensure its effectiveness and efficiency in actual applications. Optimization may involve fine-tuning of strategy parameters, comparative evaluation of different strategies, etc. In addition, the system will verify the optimized strategy. This is done by setting the generated cooling strategy and sensor thresholds in the actual device.

[0065] Policy output and application: After optimization and verification, the system will output the final cooling strategy and sensor threshold and apply them to the new device.

[0066] The principle of automatically generating cooling strategies and sensor thresholds:

[0067] 1) Basic data generation

[0068] a. Dataset Construction

[0069] Device feature set: defines a set of device features, such as {manufacturer ID, model ID, cooling system type, device insertion position}.

[0070] Historical performance records: Generate a series of historical performance records for each device feature combination, each record contains {timestamp, workload, fan speed, whether an alarm is triggered}.

[0071] b. Data generation logic

[0072] Generate data using preset rules to ensure that the data fluctuates within a reasonable range and simulate performance changes under different conditions. For example, the temperature will rise under high workload, and the fan speed needs to be increased to prevent overheating; when the ambient temperature is high, the heat dissipation effect may deteriorate, and the fan speed needs to be adjusted earlier

[0073] 2) Example feature selection of decision tree model: Features that have a greater impact on the cooling strategy and alarm threshold are selected as inputs to the decision tree, such as ambient temperature range and workload.

[0074] Node partitioning: Divide the data set into different subsets according to the different values ​​of the features, and each subset corresponds to a node of the decision tree.

[0075] Leaf node: Each leaf node stores a set of recommended cooling strategies and alarm thresholds, such as {setpoint value, pid value, sensor threshold}.

[0076] 3) Learning Algorithms

[0077] Decision tree training: Use information gain and Gini impurity indicators to select the best partitioning features and partitioning points to build a decision tree.

[0078] Logistic regression training: The model parameters are adjusted through the gradient descent optimization algorithm to minimize the error between the predicted value and the actual value.

[0079] 4) Forecasting and decision making

[0080] Real-time data input: When new equipment data (such as real-time temperature, workload, ambient temperature) is input, it is converted into the feature vector required by the model.

[0081] Model prediction: Use the trained model to make predictions, obtain recommended cooling strategies and alarm thresholds, and update them to the cooling strategy and sensor threshold configuration library.

[0082] Strategy execution: The dynamic strategy module adjusts the cooling system (such as setting new PID and setpoint values) and the alarm system (such as updating the alarm threshold) based on the model prediction results.

[0083] 5) Feedback and adjustment Performance monitoring: Continuously monitor the actual cooling effect and stability of the equipment and record new performance data. Model update: Add new performance data to the training set and retrain the model to optimize its performance.

[0084] Strategy optimization: Based on the results of model updates and feedback from performance monitoring, the cooling strategy and sensor thresholds are automatically adjusted in a timely manner to adapt to changes in device status and environmental conditions, and new devices and their corresponding cooling strategies and sensor thresholds are updated to local data files to form a configuration library.

[0085] 6) Generate policy configuration example

[0086] The automatically generated sensor thresholds and cooling strategies are shown in Table 1, which generate the PID cooling control strategies and temperature sensor thresholds corresponding to CPU, DDR, and NVME SSD of different manufacturers:

[0087] Table 1

[0088]

[0089] Continuous monitoring and feedback: During the application of the cooling strategy, the system will continue to monitor the operating status and cooling effect of the device. If it is found that the strategy does not meet expectations or the device has an abnormal situation, the system will adjust the cooling strategy and sensor threshold in time.

[0090] At the end of the process, the system continues to monitor the operating status of the new device and automatically adjusts the cooling strategy as needed. At the same time, the new device and its corresponding cooling strategy and sensor threshold information are added to the configuration library for quick identification and application in the future.

[0091] Implementation method: Through a large number of experiments and data analysis, the most suitable cooling strategy and sensor threshold are determined for each device and model, and this information is stored in a local data file to form a configuration library. When the corresponding cooling strategy and sensor threshold cannot be found in the configuration library, the system will automatically generate and continuously optimize.

[0092] The hardware service module is used to detect the parameters of the newly inserted device. The hardware service module detects whether new hardware is inserted. When a new device is inserted, the hardware service module queries the device parameters of the newly inserted hardware through the interface and transmits them to the information acquisition module based on the DBUS protocol.

[0093] Hardware service module (such as GPU-Manager, Backplane-Manager, etc.): used to manage specific hardware and provide detailed information of the hardware to DBUS, including model, specification, manufacturer, etc., to ensure the stable operation of the entire system.

[0094] The parameters of the device include the model, specifications, power consumption level, heat dissipation area or air duct design of the device.

[0095] A device for generating a dynamically switchable heat dissipation strategy based on OpenBMC, used to implement a method for generating a dynamically switchable heat dissipation strategy based on OpenBMC, the device comprising a dynamic strategy module, a hardware service module, an information acquisition module and an FSC fan speed controller;

[0096] The dynamic strategy module stores predefined parameters for different devices and models and the corresponding cooling strategies and sensor threshold configuration library;

[0097] The hardware service modules include the raid card service module, the backplane service module and the psu service module, which are used to detect newly inserted devices;

[0098] The information acquisition module communicates with the hardware service module to connect and interact with the device parameters;

[0099] The newly inserted device, hardware service module, information acquisition module and dynamic strategy module are interconnected based on DBUS protocol communication.

[0100] The following is a specific case to illustrate the implementation principle of the method and device for generating a dynamically switchable cooling strategy based on OpenBMC:

[0101] 1. Add sensor threshold and cooling strategy switching interface definition:

[0102] In this embodiment, the sensor threshold and heat dissipation strategy need to be provided to dbus (an inter-process communication method);

[0103] The specific steps are as follows: by loading the configuration file, the data to be modified and displayed is placed on the dbus to provide interface services.

[0104] For example: sensor threshold interface: used to add, modify and query the sensor threshold of the device. This interface receives the device identification and threshold parameters. Cooling strategy switching interface: performs the policy switching operation according to the device identification and the specified cooling strategy. The interface needs to verify the validity of the strategy and update the system configuration.

[0105] 2. Dynamic hardware identification and monitoring:

[0106] Function description: This function is responsible for obtaining basic data such as device type, model, configuration information, etc. from other modules (for example, obtaining the currently inserted hard disk information from the backplane management module). When the user device is installed or connected to the system, it will automatically obtain the device data of the inserted hardware.

[0107] Implementation method: Encapsulate various hardware service modules (such as raid card service module, backplane service module, psu service module, etc.) to detect whether new hardware is inserted. When a new device is inserted, the system immediately queries the detailed information of the hardware through the interface, including model, specification, manufacturer, etc., and provides it to the dbus. Then the device information acquisition module will search for hardware information on other hardware service modules in a polling (or interrupt) manner, and store it in the corresponding data structure for reference.

[0108] 3. Device identification and matching cooling strategy and sensor threshold configuration library:

[0109] Functional description: Build a configuration library containing the optimal cooling strategies and sensor thresholds for various devices and models. The configuration library predefines a series of tested cooling strategies and sensor thresholds based on the hardware characteristics, power consumption, cooling capacity and other factors of the device. According to the device type, model and other information obtained by the device information acquisition module, find the corresponding configuration in the cooling strategy and sensor threshold configuration library. Use hash tables, database queries or configuration file parsing and other technologies to quickly and accurately match the corresponding thresholds and strategies.

[0110] As shown in Table 2, there are 4 cooling strategies. When it is detected that the hard disk inserted on the rear backplane is a PM9A3 hard disk, the sensor threshold and the setpoint in the cooling strategy provided in the table below are compared and dynamically adjusted and switched through the sensor threshold and cooling strategy switching interface.

[0111] Table 2

[0112]

[0113] The following is an explanation of the English expressions involved:

[0114] BMC (Baseboard Manager Controller) represents a series of monitoring and control functions, and the object of operation is the system hardware. For example, by monitoring the system's temperature, voltage, fan, power supply, etc., and making corresponding adjustments, the system is kept in a healthy state. At the same time, BMC is also responsible for recording various hardware information and log records to prompt users and locate subsequent problems.

[0115] DBUS (an inter-process communication protocol): provides the basis for efficient communication between components in the system. DBUS is responsible for quickly transmitting data information and control instructions within the system, ensuring that modules such as FSC and Sensor can obtain the required information in real time and make corresponding adjustments.

[0116] Hardware service module (such as GPU-Manager, Backplane-Manager, etc.): used to manage specific hardware and provide detailed information of the hardware to DBUS, including model, specification, manufacturer, etc., to ensure the stable operation of the entire system.

[0117] FSC (Fan Speed ​​Control): Fan speed control module. It adjusts the speed of the internal fan of the server in real time according to the temperature and other parameters provided by the sensor and the cooling strategy adjusted in real time by Dynamic-policy to meet the different cooling requirements of different devices and models. The FSC module provides an interface for modifying the cooling strategy, including pid, setpoint and linear speed control strategy.

[0118] Sensor: Sensors throughout the system monitor key parameters of hardware devices such as temperature and power consumption in real time and generate alarms based on the provided thresholds. These sensor data are transmitted to the BMC and FSC as an important basis for adjusting the fan speed.

[0119] Dynamic-policy: The dynamic policy module stores predefined parameters for different devices and models, including heat dissipation strategies and sensor thresholds. First, when a device is connected or the device model changes, the module will dynamically monitor the type and model of the device currently inserted in the hardware service module, and intelligently adjust the sensor threshold and heat dissipation strategy based on the predefined parameters.

[0120] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating a dynamically switchable heat dissipation strategy based on OpenBMC, characterized in that: The following steps are involved: Step 1: The dynamic strategy module builds a heat dissipation strategy and sensor threshold configuration library, which contains the optimal heat dissipation strategy configuration data and corresponding sensor thresholds for various devices and models; Step 2: When a new device is connected to the system, the sensor threshold and cooling strategy switching interface definition are added, the information acquisition module obtains the parameters of the newly connected device, and interacts with the dynamic strategy module to exchange device parameters; Step 3: The dynamic strategy module matches the corresponding optimal cooling strategy configuration data and the corresponding sensor threshold in the cooling strategy and sensor threshold configuration library based on the device parameters; Step 4. If the match is successful, the dynamic strategy module exchanges the optimal cooling strategy configuration data and the corresponding sensor threshold with the newly connected device based on the DBUS protocol; Step 5: The newly connected device executes the optimal heat dissipation strategy configuration data and the corresponding sensor threshold to dissipate heat.

2. The method for generating a dynamically switchable heat dissipation strategy based on OpenBMC according to claim 1, characterized in that: If the corresponding device parameters are not matched in step 4, the following method is used to generate the optimal cooling strategy configuration data and the corresponding sensor threshold; Step 41, extracting multiple related features that have an important impact on the formulation of the heat dissipation strategy from the device parameters; Step 42, inputting multiple associated features into the trained large model, and the large model outputs a preliminary heat dissipation strategy; Step 43, optimize the preliminary heat dissipation strategy to output the final heat dissipation strategy and sensor threshold, and continue to execute steps 4 and 5.

3. The method for generating a dynamically switchable heat dissipation strategy based on OpenBMC according to claim 2, characterized in that: During the execution of the cooling strategy in step 5, the device continues to interact with the running status. If it is determined that the cooling status of the device does not meet expectations or the device is abnormal, the cooling strategy and sensor threshold are adjusted and the cooling effect is verified again until the cooling expectation is met.

4. The method for generating a dynamically switchable heat dissipation strategy based on OpenBMC according to claim 3, characterized in that: The optimal cooling strategy and corresponding sensor threshold of the new device after optimization and verification are added to the cooling strategy and sensor threshold configuration library.

5. The method for generating a dynamically switchable heat dissipation strategy based on OpenBMC according to claim 4, characterized in that: In step 2, the large model is a decision tree model, a random forest model or a neural network model.

6. The method for generating a dynamically switchable heat dissipation strategy based on OpenBMC according to claim 1, characterized in that: The hardware service module is used to detect the parameters of the newly inserted device. The hardware service module detects whether new hardware is inserted. When a new device is inserted, the hardware service module queries the device parameters of the newly inserted hardware through the interface and transmits them to the information acquisition module based on the DBUS protocol.

7. The method for generating a dynamically switchable heat dissipation strategy based on OpenBMC according to claim 2, characterized in that: The parameters of the device include the model, specifications, power consumption level, heat dissipation area or air duct design of the device.

8. A device for generating a dynamically switchable heat dissipation strategy based on OpenBMC, characterized in that: Used to implement the method for generating a dynamically switchable cooling strategy based on OpenBMC as described in any one of claims 1 to 7, the device includes a dynamic strategy module, a hardware service module, an information acquisition module and an FSC fan speed controller; The dynamic strategy module stores predefined parameters for different devices and models and the corresponding cooling strategies and sensor threshold configuration library; The hardware service modules include the raid card service module, the backplane service module and the psu service module, which are used to detect newly inserted devices; The information acquisition module communicates with the hardware service module to connect and interact with the device parameters; The newly inserted device, hardware service module, information acquisition module and dynamic strategy module are interconnected based on DBUS protocol communication.

Images