An edge-computing-based intelligent cabinet heat dissipation control method and system

Abstract

The application relates to the technical field of heat dissipation control, and provides a cabinet intelligent heat dissipation control method and system based on edge computing, which comprises the following steps: acquiring temperature measurement values in a cabinet in real time; acquiring operation load data of equipment in the cabinet; inferring an expected range of the temperature measurement values in the cabinet based on the operation load data; determining a reliability index of the temperature measurement values according to the deviation degree of the temperature measurement values from the expected range and the time variation characteristics of the temperature measurement values; processing the temperature measurement values according to the reliability index to obtain comprehensive temperature feedback for heat dissipation control; and adjusting the operation state of a heat dissipation execution mechanism in the cabinet according to the comprehensive temperature feedback. The application can effectively identify and process abnormal temperature measurement values caused by sensor performance degradation or data drift, thereby avoiding suboptimal heat dissipation decisions made by a system based on inaccurate data, and solving the problems of low efficiency and difficult diagnosis of a heat dissipation control system in the prior art.

Description

Technical Field

[0001] This application relates to the field of rack heat dissipation control technology, and more specifically, to a rack intelligent heat dissipation control method and system based on edge computing. Background Technology

[0002] In modern high-density data centers, edge computing-based intelligent thermal control systems are typically deployed to ensure stable server operation and effective energy consumption management. These systems monitor real-time temperature changes within the server racks and predict heat generation based on server operating status, dynamically adjusting fan speeds accordingly. However, with prolonged high-intensity operation in data centers, temperature sensors within the racks may experience slight performance degradation or data drift. These sensors are not completely ineffective, and their output data is still considered normal under standard diagnostic procedures.

[0003] Specifically, in a rack-mounted intelligent heat dissipation control system that incorporates server internal operating status information for heat prediction and uses temperature sensor feedback for adaptive optimization of the prediction logic, some temperature sensors may experience slight performance degradation or data drift over long-term operation. This sensor drift can cause the prediction logic of the edge computing unit to incorrectly treat the drifted data as real physical feedback during the adaptive learning process, thereby learning and solidifying an inaccurate mapping relationship between heat generation and heat dissipation requirements.

[0004] Ultimately, the system may make suboptimal cooling decisions in localized areas, resulting in unnecessary overcooling, leading to energy waste and noise, or hidden undercooling that accelerates component aging. More challenging is that, because the sensors are not completely faulty, the system's internal diagnostic mechanisms often fail to detect these subtle drifts. All temperature readings and fan speeds appear to be within normal ranges, making it difficult for maintenance personnel to detect and pinpoint the true root cause of the problem. This issue causes the thermal control system to make suboptimal or even incorrect decisions in localized areas, resulting in unnecessary energy waste or accelerated hardware aging, and the problem is difficult to detect and diagnose accurately and in a timely manner. Summary of the Invention

[0005] This application provides a method and system for intelligent heat dissipation control of server racks based on edge computing, in order to solve at least one of the above-mentioned technical problems.

[0006] To achieve the above objectives, this application adopts the following technical solution: In a first aspect, this application discloses a smart heat dissipation control method for server racks based on edge computing, including: Real-time acquisition of temperature measurements inside the server rack; Obtain operating load data for equipment within the cabinet; Based on operational load data, infer the expected range of internal temperature measurements within the server rack; The reliability index of the temperature measurement value is determined based on the degree of deviation between the temperature measurement value and the expected range, as well as the time variation characteristics of the temperature measurement value. Based on reliability indicators, the temperature measurement values ​​are processed to obtain comprehensive temperature feedback for heat dissipation control. Based on comprehensive temperature feedback, adjust the operating status of the heat dissipation actuators inside the cabinet.

[0007] Secondly, this application also discloses an intelligent heat dissipation control system for server racks based on edge computing, comprising: Temperature measurement acquisition module, used to acquire temperature measurement values ​​inside the cabinet; The operating load data acquisition module is used to acquire the operating load data of the equipment in the cabinet; The expected range inference module is used to infer the expected range of the measured temperature values ​​inside the cabinet based on the operating load data; The reliability index determination module is used to determine the reliability index of the temperature measurement value based on the degree of deviation between the temperature measurement value and the expected range, as well as the time variation characteristics of the temperature measurement value. The temperature measurement value processing module is used to process the temperature measurement values ​​according to reliability indicators to obtain comprehensive temperature feedback for heat dissipation control. The heat dissipation actuator adjustment module is used to adjust the operating status of the heat dissipation actuator in the cabinet based on comprehensive temperature feedback.

[0008] Compared with the prior art, this application has at least the following beneficial effects: This application effectively solves the problem of unnecessary energy waste or accelerated hardware aging, and can promptly detect and accurately diagnose the root cause of the problem, thereby achieving more precise and reliable intelligent heat dissipation control, significantly improving the operating efficiency and equipment stability of the data center. Attached Figure Description

[0009] Figure 1 A flowchart illustrating an edge computing-based intelligent heat dissipation control method for server racks provided in this application; Figure 2 This is a schematic diagram of a rack-based intelligent heat dissipation control system provided in this application. Detailed Implementation

[0010] The technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0011] like Figure 1 As shown in the figure, this application proposes a rack intelligent heat dissipation control method based on edge computing, including: Real-time acquisition of temperature measurements inside the server rack; Obtain operating load data for equipment within the cabinet; Based on operational load data, infer the expected range of internal temperature measurements within the server rack; The reliability index of the temperature measurement value is determined based on the degree of deviation between the temperature measurement value and the expected range, as well as the time variation characteristics of the temperature measurement value. Based on reliability indicators, the temperature measurement values ​​are processed to obtain comprehensive temperature feedback for heat dissipation control. Based on comprehensive temperature feedback, adjust the operating status of the heat dissipation actuators inside the cabinet.

[0012] Among them, the temperature measurement value refers to the ambient temperature data collected in real time by temperature sensors deployed inside the cabinet. This data is the basic input for heat dissipation control.

[0013] Operating load data refers to the current working status and resource utilization of IT equipment such as servers and network devices in the rack, such as CPU utilization, memory usage, and I / O throughput. These data are closely related to the heat output of the equipment.

[0014] The expected range refers to the reasonable range in which the measured temperature inside the cabinet should be under the current operating load conditions. This range can be inferred through historical data analysis, physical model prediction, or machine learning model.

[0015] The degree of deviation refers to the difference between the actual temperature measurement and the inferred expected range, and is an important basis for judging temperature data anomalies.

[0016] Time-varying characteristics refer to the pattern of temperature measurement changes over time, such as its rate of change, fluctuation frequency, and trend. These characteristics help identify sensor drift or transient anomalies.

[0017] The reliability index is a quantitative assessment of the trustworthiness of temperature measurements. This index takes into account both the degree of deviation and the characteristics of time variation, and is used to guide subsequent data processing.

[0018] Comprehensive temperature feedback is temperature data that has undergone reliability assessment and processing. It reflects the true thermal state inside the cabinet more accurately than the original temperature measurement and serves as a direct basis for heat dissipation control decisions.

[0019] Cooling actuators refer to the physical devices inside the server rack used to regulate heat dissipation, such as cooling fans, coolant pumps, and air conditioning units. The adjustment of their operating status directly affects the temperature inside the server rack.

[0020] The implementation environment of this application is typically a smart cabinet in a data center or server room, which is equipped with various IT devices, temperature sensors, heat dissipation actuators, and edge computing units for data acquisition, processing, and control.

[0021] The following section provides a more detailed description of the edge computing-based intelligent heat dissipation control method for server racks presented in this application: First, it's necessary to acquire real-time temperature measurements inside the server rack. This can be achieved by deploying multiple temperature sensors within the rack. For example, thermistors, thermocouples, or infrared temperature sensors can be installed at different locations such as the front, middle, rear, top, and bottom of the rack to obtain temperature data at multiple points. These sensors can be standalone or integrated into the server or PDU (Power Distribution Unit). The analog signals collected by the sensors are converted from analog to digital and then transmitted to the edge computing unit via wired (e.g., RS485, Ethernet) or wireless (e.g., Wi-Fi, Zigbee) communication methods.

[0022] Simultaneously, it's necessary to acquire operational load data from the devices within the rack. This can be achieved in several ways. For example, performance counters such as CPU utilization, memory usage, and network traffic can be periodically read from devices like servers and switches via SNMP (Simple Network Management Protocol). Alternatively, more detailed process-level resource usage information can be collected in real-time through an agent program installed in the server's operating system. Furthermore, real-time power consumption data can be obtained from the devices via the Power Management Unit (PDU), as power consumption and heat generation are typically positively correlated. This data is also transmitted to the edge computing unit for processing.

[0023] After obtaining temperature measurements and operating load data, the next step is to infer the expected range of the internal temperature measurements of the server rack based on the operating load data. One approach is to pre-build a machine learning model based on historical data, such as a regression model or a neural network model. This model takes the equipment's operating load data as input and outputs the predicted value of the internal temperature of the server rack under the current load, along with its confidence interval, i.e., the expected range. For example, when CPU utilization increases, the model will predict an increase in temperature and provide the corresponding expected temperature range. Another approach is to use a physical thermal model for inference. This model considers parameters such as the server rack structure, airflow organization, and equipment thermal design power (TDP), and calculates the theoretical heat output based on the real-time operating load, thereby inferring the expected temperature range.

[0024] Subsequently, the reliability index of the temperature measurement is determined based on the degree of deviation between the measured temperature value and the expected range, as well as the time-varying characteristics of the temperature measurement. For assessing the degree of deviation, the actual temperature measurement can be compared with the upper and lower limits of the expected range. For example, if the measured value exceeds the expected range, the deviation is large; if the measured value is within the expected range, the deviation is small. For assessing the time-varying characteristics, the rate of change of the temperature measurement over a short period can be monitored. For example, if the temperature measurement shows a large jump within a few seconds, or exhibits irregular high-frequency fluctuations over a period of time, it indicates abnormal time-varying characteristics. The reliability index can be a value between 0 and 1, where 1 represents complete reliability and 0 represents complete unreliability. For example, when the deviation is small and the time-varying characteristics are stable, the reliability index is high; conversely, when the deviation is large or the time-varying characteristics are abnormal, the reliability index is low.

[0025] Based on reliability metrics, temperature measurements are processed to obtain comprehensive temperature feedback for heat dissipation control. One approach is to directly mark a temperature measurement as abnormal when the reliability metric falls below a preset threshold, reducing its weight in the comprehensive temperature feedback calculation, or even excluding it altogether. For example, a weighted average method can be used, averaging the temperature measurements from multiple sensors, with the weights determined by their respective reliability metrics. Sensors with higher reliability metrics contribute more to the comprehensive temperature feedback. Another approach is not to directly exclude measurements with low reliability metrics, but rather to smooth or filter them, such as using a moving average filter or a Kalman filter, to reduce the impact of abnormal fluctuations.

[0026] Finally, based on the overall temperature feedback, the operating status of the cooling actuators inside the cabinet is adjusted. For example, if the overall temperature feedback indicates that the internal temperature of the cabinet is too high, the system will increase the speed of the cooling fans or activate additional cooling units. If the overall temperature feedback indicates that the temperature is moderate or too low, the system will reduce the fan speed or reduce cooling input to save energy. This adjustment can be linear, i.e., the higher the temperature, the greater the heat dissipation intensity; or it can be based on a PID (proportional-integral-derivative) control algorithm to achieve more precise and stable temperature control.

[0027] The edge computing-based intelligent heat dissipation control method for server racks in this application can effectively identify and correct sensor drift or abnormal data by introducing the evaluation and processing of the reliability index of temperature measurement values. This avoids suboptimal heat dissipation decisions caused by minor sensor degradation in traditional systems, and significantly improves the accuracy and energy efficiency of heat dissipation control.

[0028] Specifically, this application does not rely solely on the original temperature measurements, but introduces a reliability metric to assess the credibility of these measurements.

[0029] First, by acquiring temperature measurements inside the cabinet and operating load data of the equipment within the cabinet, basic data is provided for subsequent temperature inference and reliability assessment.

[0030] Secondly, the expected range of internal temperature measurements within the server rack is inferred based on operational load data. This allows the system to have a reasonable expectation of temperature measurements, thus providing a reference for identifying anomalies. For example, when the server load is low, the system expects the temperature to be low as well; if the temperature measurement is abnormally high at this time, there may be a problem.

[0031] Next, based on the degree of deviation of the temperature measurement from the expected range and the time-varying characteristics of the temperature measurement, the reliability index of the temperature measurement is determined. This step is crucial to this application, as it comprehensively considers both static deviation and dynamic changes, enabling a more holistic assessment of the sensor's health. For example, even if a measurement value is within the expected range, it may be deemed to have low reliability if it fluctuates frequently and drastically within a short period.

[0032] Finally, the temperature measurements are processed according to reliability indicators to obtain comprehensive temperature feedback for heat dissipation control, and the operating status of the heat dissipation actuator is adjusted accordingly. This means that the system no longer blindly trusts all sensor data, but rather weights, corrects, or excludes data based on its reliability, thereby obtaining more realistic and reliable temperature feedback to guide heat dissipation decisions.

[0033] In summary, this application significantly improves the intelligence level and operating efficiency of the edge computing-based intelligent heat dissipation control system for server racks by introducing an evaluation and processing mechanism for the reliability index of temperature measurement values. It solves the problem that traditional systems are unable to effectively cope with the subtle degradation of sensors, and has significant technological progress and practical value.

[0034] However, in practical applications, relying solely on these passive monitoring indicators may not fully reflect the true operating status of the cabinet's cooling system, especially in complex scenarios such as slow sensor drift, decreased cooling system efficiency, or deteriorating equipment component health. This passive reliability assessment may lead to a delay in the discovery of potential problems, thereby affecting the timeliness and accuracy of cooling control, and could even adversely affect the long-term stable operation of equipment within the cabinet.

[0035] In some embodiments, the step of determining the reliability index of the temperature measurement value based on the degree of deviation of the temperature measurement value from the expected range and the time variation characteristics of the temperature measurement value further includes: Obtain operational data on the cooling actuators within the server rack; Obtain component health output data for devices within the rack; Based on operational input data and component health output data, assess the heat dissipation costs and component benefits of the rack area; When the heat dissipation cost continues to exceed the preset threshold or the component benefit continues to be lower than the preset threshold, the strategy exploration mode is triggered. In strategy exploration mode, the heat dissipation target of the rack area is proactively adjusted, and the operating intensity of the heat dissipation actuator is forcibly increased; In this exploration process, the reliability index of temperature measurements was re-evaluated, and the reliability index of the re-evaluated temperature measurements was cross-validated by combining the temperature difference pattern between the internal temperature of the equipment in the cabinet and the ambient temperature of the cabinet.

[0036] Specifically, operational input data refers to the resources consumed by the cooling actuators during operation, such as fan speed, power consumption, and coolant flow rate, reflecting the cost of maintaining the rack temperature. Component health output data refers to the operating status and performance of critical equipment within the rack, such as CPU utilization, hard disk I / O speed, memory error rate, and power module output stability, which are directly related to equipment reliability and service quality.

[0037] Assessing the cooling costs and component benefits of a server rack area based on operational input data and component health output data involves establishing a comprehensive evaluation model by quantifying the relationship between cooling input and equipment performance output. Cooling costs can be understood as the energy or resource consumption required to maintain a specific temperature, while component benefits represent the performance and stability the equipment can provide under current cooling conditions. For example, excessively high cooling costs without a corresponding improvement in component benefits may indicate low cooling efficiency or abnormal sensor readings.

[0038] When the heat dissipation cost consistently exceeds a preset threshold or the component efficiency consistently falls below a preset threshold, the system will trigger a strategy exploration mode. This mode is a proactive diagnostic mechanism designed to verify the reliability of temperature measurements through targeted intervention. In strategy exploration mode, the system will proactively adjust the heat dissipation target for the rack area, such as temporarily setting a lower temperature target and forcibly increasing the operating intensity of the heat dissipation actuators, such as increasing the fan speed to maximum or activating additional cooling units.

[0039] During this exploration, the system re-evaluates the reliability of temperature measurements. To improve the accuracy of the evaluation, it also cross-validates the temperature difference pattern between the internal temperature of the equipment within the rack and the ambient temperature of the rack. The temperature difference pattern refers to the stable temperature difference range and variation pattern that should be maintained between the internal temperature of critical equipment (such as CPUs and GPUs) and the ambient temperature of the rack under normal operating conditions. If, after forcibly increasing the heat dissipation intensity, the temperature measurements fail to change according to the expected temperature difference pattern, or if the temperature difference pattern becomes abnormal, it further confirms the potential unreliability of the temperature measurements.

[0040] This application overcomes the limitations of passive reliability assessment by introducing an evaluation of heat dissipation costs and component benefits, along with a strategy exploration mode and cross-validation mechanism. When a mismatch occurs between the investment in the heat dissipation system and the output of equipment performance—for example, a continuous increase in heat dissipation energy consumption without a corresponding improvement in equipment performance—this provides an early warning signal, indicating potential problems such as sensor drift, decreased heat dissipation efficiency, or equipment overheating. By triggering the strategy exploration mode, the system can proactively intervene in the heat dissipation actuators, forcibly changing heat dissipation conditions and observing the response of temperature measurements and temperature difference patterns. This proactive detection mechanism allows the system to move beyond relying solely on historical data and expected ranges of temperature measurements. Instead, it incorporates external intervention and comprehensive analysis of multi-dimensional data (operational investment, component health, temperature difference patterns) to more comprehensively and accurately verify the authenticity and reliability of temperature measurements. Therefore, even slow sensor drift or subtle changes in heat dissipation efficiency can be detected and corrected promptly, ensuring that heat dissipation control decisions are based on highly reliable temperature feedback.

[0041] The following is a specific example to illustrate this.

[0042] Suppose a temperature sensor inside a server rack, after long-term operation, begins to show a slight upward drift in its reading. However, the magnitude of this change remains within the expected range and time-varying characteristics mentioned above. Therefore, a basic reliability assessment alone might not detect this problem in time. However, as the sensor continues to drift, in order to maintain the set temperature inside the rack, the operating intensity of cooling actuators (such as fans) is gradually increased, leading to a continuous increase in their operational input data (such as fan speed and power consumption). Simultaneously, the component health output data of the equipment within the rack (such as CPU average utilization and response time) does not show a corresponding performance improvement, and may even lead to a decrease in overall energy efficiency due to increased cooling energy consumption.

[0043] The system will detect when the heat dissipation cost (high energy consumption) consistently exceeds a preset threshold, while the component benefit (performance output) consistently falls below a preset threshold, thus triggering a strategy exploration mode. In this mode, the system will proactively adjust the heat dissipation target for the rack area to a lower temperature and forcibly increase the fan speed to maximum. During this process, the system will reassess the reliability index of the temperature measurement values. If the temperature drop reported by the drift sensor is significantly less than that of other normal sensors or does not conform to the expected temperature difference pattern (e.g., the temperature difference between the CPU core temperature and the rack ambient temperature), for example, after forced cooling, the rack ambient temperature drops by 5°C, but the sensor reports a temperature drop of only 1°C, while other normal sensors report a temperature drop of 4°C, then the system will significantly reduce the reliability index of the sensor based on this abnormal temperature difference pattern change. Based on this reduced reliability index, the weight of the sensor's reading in the comprehensive temperature feedback calculation will be significantly reduced, or even temporarily excluded, thereby avoiding errors in heat dissipation control decisions caused by erroneous temperature feedback and ensuring the accuracy and efficiency of the rack cooling system.

[0044] However, in practical applications, temperature sensors may be affected by transient interference or environmental noise, causing brief, drastic jumps or continuous high-frequency fluctuations in temperature measurements. If these abnormal data are not effectively identified and processed, they may mislead the assessment of reliability indicators, thereby affecting the accuracy and stability of heat dissipation control. To address this, this application further proposes improving the accuracy of reliability assessment by conducting transient validity checks and continuously monitoring high-frequency fluctuations when determining the reliability indicators of temperature measurements.

[0045] In some embodiments, the step of determining the reliability index of the temperature measurement value based on the degree of deviation of the temperature measurement value from the expected range and the time variation characteristics of the temperature measurement value further includes: Perform an instantaneous validity check on the temperature measurement value, wherein the instantaneous validity check includes comparing the current temperature measurement value with the previous or most recent temperature measurement values; When the temperature measurement value fluctuates beyond a preset range within a preset time period, the reliability index of the temperature measurement value is reduced. Continuously monitor temperature measurements within a preset time window; When temperature measurements consistently exhibit high-frequency fluctuations, the reliability index of the temperature measurements is reduced.

[0046] Specifically, instantaneous validity checking refers to the immediate assessment of the currently acquired temperature measurement value to identify whether it is abnormal instantaneous data. This check is achieved by comparing the current temperature measurement value with the previous or several recent historical measurements considered stable. Its purpose is to quickly detect and filter out abnormal readings caused by momentary sensor malfunctions, electromagnetic interference, or other transient factors. For example, if the difference between the current measurement value and the previous stable value suddenly exceeds a certain preset threshold, it can be preliminarily determined that the measurement value may have a momentary anomaly.

[0047] Specifically, when the temperature measurement value fluctuates beyond a preset range within a preset time period, it indicates an unexpected and drastic change in temperature data. Such drastic fluctuations typically do not conform to physical laws or the actual operating conditions of the system, thus requiring a reduction in the reliability index of the temperature measurement value. The preset time period and preset range can be adjusted based on factors such as the physical characteristics of the server rack, the response speed of the cooling system, and the accuracy of the sensors to ensure effective identification of abnormal fluctuations rather than normal temperature changes.

[0048] Continuously monitoring temperature measurements within a preset time window is intended to capture long-term trends and fluctuation patterns in temperature data. This preset time window can be a sliding window, such as displaying data from the past few seconds, tens of seconds, or minutes. Continuous monitoring reveals high-frequency fluctuations that are not instantaneous jumps but rather persistent occurrences.

[0049] When temperature measurements consistently exhibit high-frequency fluctuations, it typically indicates potential issues such as sensor malfunction, improper installation, or unstable airflow within the cabinet. These high-frequency fluctuations make it difficult for temperature data to accurately reflect the true thermal state of the cabinet; therefore, it's necessary to reduce the reliability index of this temperature measurement. Lowering the reliability index assigns a lower weight to this measurement in subsequent comprehensive temperature feedback calculations, thereby reducing its negative impact on thermal control decisions.

[0050] This application effectively addresses the problem of insufficient identification and handling of transient anomalies and persistent noise when assessing the reliability of temperature measurements by introducing a mechanism for instantaneous validity checking and continuous monitoring of high-frequency fluctuations. Specifically, the instantaneous validity check can quickly capture drastic jumps in temperature measurements caused by sudden interference. By comparing these jumps with stable historical data, the reliability of such anomalies can be promptly identified and reduced. Simultaneously, continuous monitoring of temperature measurements within a preset time window can identify persistent high-frequency fluctuations caused by sensor or environmental factors. While these fluctuations may not manifest as drastic jumps, they still affect the accuracy of temperature data. Through the identification and handling of these two anomalies, this application can more comprehensively and accurately assess the true reliability of temperature measurements, avoiding misjudgments in heat dissipation control due to erroneous data.

[0051] Suppose a temperature sensor inside the server rack reports a temperature measurement at a certain moment. First, the system performs an instantaneous validity check, comparing the measurement to the average of three temperature measurements from the previous second. If the difference between the current measurement and this average exceeds 5 degrees Celsius (a preset range), the system immediately reduces the reliability index of the current temperature measurement, for example, adjusting its weight from 1.0 to 0.2. Simultaneously, the system continuously monitors temperature measurements over the past 10 seconds (a preset time window). If, within these 10 seconds, the temperature measurement fluctuates within ±1 degree Celsius at a frequency exceeding 3 times per second (high-frequency fluctuation), the system further reduces the reliability index of this fluctuating data, for example, adjusting its weight to 0.1. In this way, even if the sensor experiences transient failures or persistent noise, the system can promptly identify and mitigate its impact on the overall temperature feedback, ensuring the accuracy of thermal control decisions.

[0052] When processing temperature measurements based on reliability metrics to obtain comprehensive temperature feedback, the dynamic changes in the operating load of equipment in different areas within the cabinet may not be fully considered. If processing is based solely on uniform reliability metrics, when the load on equipment in certain areas rises or falls rapidly, the generated comprehensive temperature feedback may not reflect the actual situation of local hot or cold spots in a timely and accurate manner. This can lead to delayed or excessive response in heat dissipation control, affecting heat dissipation efficiency and energy consumption.

[0053] In some embodiments, the step of processing the temperature measurement value according to the reliability index to obtain comprehensive temperature feedback for heat dissipation control includes: Obtain the rate of change of operating load for equipment in each area of ​​the cabinet; Adjust the weight of temperature measurements in the overall temperature feedback calculation based on the rate of change of operating load of equipment in each area; When the rate of change of equipment operating load in a certain area exceeds a preset threshold, the weight of the temperature measurement value in that area in the comprehensive temperature feedback calculation is increased. When the rate of change of equipment operating load in a certain area is lower than a preset threshold, the weight of the temperature measurement value in that area in the comprehensive temperature feedback calculation is reduced. Based on the adjusted weights, the temperature measurements are weighted to obtain a comprehensive temperature feedback for heat dissipation control.

[0054] Specifically, acquiring the rate of change of operating load for equipment in each area of ​​the cabinet refers to the system periodically monitoring the operating load data of equipment deployed in different physical or logical areas within the cabinet, and calculating the magnitude or trend of these load data changes within a specific time window. For example, the rate of change can be determined by comparing the current operating load with the operating load of the previous moment or several moments ago, with the aim of capturing the dynamics of equipment operating status.

[0055] The method of adjusting the weight of temperature measurements in the overall temperature feedback calculation based on the rate of change of equipment load in each area can be understood as dynamically assigning importance to the temperature measurements of different areas in the final overall temperature feedback calculation based on the speed and direction of equipment load changes. When the rate of change of equipment load in a certain area is higher than a preset threshold, it indicates that the equipment in that area may be experiencing or about to experience significant temperature changes. In this case, the weight of the temperature measurements in that area is increased to make the overall feedback more sensitive to the temperature status of that area, so that the heat dissipation system can respond more quickly. Conversely, when the rate of change of equipment load in a certain area is lower than the preset threshold, it indicates that the equipment in that area is operating relatively stably. In this case, the weight of the temperature measurements in that area is decreased to avoid over-focusing on stable areas, thereby optimizing the allocation of overall heat dissipation resources.

[0056] Based on the adjusted weights, the temperature measurements are weighted to obtain a comprehensive temperature feedback for heat dissipation control. Specifically, this involves taking the temperature measurements from each area, after reliability index processing, and calculating a weighted average or other weighted fusion based on their assigned dynamic weights. This generates a comprehensive temperature feedback signal that fully and dynamically reflects the internal thermal environment of the server rack. This comprehensive temperature feedback signal will serve as the basis for subsequent adjustments to the operating status of the heat dissipation actuators within the server rack.

[0057] This application addresses the response lag issue when dealing with rapid changes in localized heat load by incorporating consideration of the rate of change of equipment operating load in each area of ​​the server rack and dynamically adjusting the weight of temperature measurements in the overall temperature feedback calculation. When the rate of change of equipment operating load in a certain area is high, it indicates that the area may heat up rapidly. In this case, by increasing the weight of the temperature measurement in that area, the temperature information of that area takes a more dominant position in the overall feedback, thereby prompting the cooling system to prioritize and respond to this potential hotspot. Conversely, when the rate of change of load is low, reducing the weight avoids excessive heat dissipation to stable areas and optimizes energy use. This dynamic weighting mechanism enables the overall temperature feedback to reflect the actual heat distribution and changing trends inside the server rack more accurately and timely, providing more instructive control basis for the cooling actuators.

[0058] The following is a specific example to illustrate this.

[0059] Suppose a server rack is divided into three zones: Zone A, Zone B, and Zone C. At a certain moment, the servers in Zone A are processing a large amount of data, and their load change rate suddenly increases by 20% from a steady state, exceeding a preset threshold of 15%. Meanwhile, the storage devices in Zone B maintain a stable load with a change rate close to 0%, while the network devices in Zone C experience a slight decrease in load, with a change rate of -5%, below the preset threshold of -3%.

[0060] According to the scheme of this application, the system will identify that the rate of change of operating load in region A is higher than a preset threshold, and therefore will increase the weight of the temperature measurement value of region A in the comprehensive temperature feedback calculation, for example, from 0.3 to 0.5. For region B, since its rate of change of load is lower than the preset threshold (or within a stable range), its weight may remain unchanged or slightly reduced, for example, still 0.3. For region C, since its rate of change of load is lower than the preset threshold, its weight will be reduced, for example, from 0.4 to 0.2.

[0061] Subsequently, the system will weight the temperature measurements from regions A, B, and C according to these adjusted weights to obtain the final comprehensive temperature feedback. For example, if the temperature measurement in region A is 35℃, in region B it is 30℃, and in region C it is 28℃, then the adjusted comprehensive temperature feedback will more likely reflect the higher temperature in region A. This will prompt the cooling actuators to prioritize or increase cooling for region A to quickly address its potential overheating risk. This dynamic adjustment mechanism ensures that cooling resources are intelligently allocated to the regions that need them most, improving the efficiency and response speed of cooling control.

[0062] When processing temperature measurements inside a server rack to obtain comprehensive temperature feedback for heat dissipation control, the focus may be primarily on the reliability assessment of the overall or average temperature, failing to adequately consider the dynamic changes in the operating load of equipment within the rack and the potential for high heat density in localized areas. For example, assuming multiple zones exist within the rack, the operating load and heat generation of equipment in each zone may differ significantly. If heat dissipation control is based solely on uniform reliability indicators or average temperature, some high-load zones may experience insufficient heat dissipation, while other low-load zones may overheat, affecting stable equipment operation and energy efficiency. If these issues are not addressed, hot spots within the rack may experience performance degradation or even malfunction due to inadequate heat dissipation. Furthermore, the inefficiency of the overall heat dissipation strategy can lead to unnecessary energy waste. To address this, this application proposes a more refined method for processing temperature measurements. By dynamically adjusting the weights of temperature measurements in each zone and independently assessing and locally enhancing heat dissipation in high-heat-density areas, more precise and efficient rack heat dissipation control can be achieved.

[0063] In some embodiments, the step of processing temperature measurements according to reliability indicators to obtain comprehensive temperature feedback for heat dissipation control further includes: Identify high heat density areas inside the server rack; Based on reliability indicators, the temperature measurements of high heat density areas are independently evaluated and differentiated from the temperature measurements of other areas in the cabinet to obtain comprehensive temperature feedback for heat dissipation control. When the temperature measurement value of a high heat density area shows a continuous upward trend, it triggers localized enhanced heat dissipation in that area; The results of the localized enhanced heat dissipation are fed back into the calculation of the overall temperature feedback.

[0064] Specifically, identifying high-heat-density areas inside the server rack involves analyzing the configuration, power consumption characteristics, and actual operating data of the equipment within the rack to determine which areas are prone to localized high temperatures due to dense equipment, concentrated power consumption, or limited heat dissipation. The aim is to focus on and manage these critical areas.

[0065] Furthermore, based on reliability metrics, temperature measurements in high heat density areas are independently evaluated and differentiated from those in other areas of the cabinet to obtain comprehensive temperature feedback for heat dissipation control. This means that temperature measurements in identified high heat density areas are no longer simply averaged or uniformly weighted with measurements from other areas, but are analyzed separately based on their specific characteristics, potentially employing stricter thresholds or faster response mechanisms. For example, independent temperature limits can be set for high heat density areas based on reliability metrics, or they can be given higher priority in the comprehensive feedback to ensure effective temperature control in these areas.

[0066] When the temperature measurement in a high heat density area shows a continuous upward trend, local enhanced heat dissipation for that area is triggered. This means that once the system detects that the temperature in a high heat density area is continuously rising and may exceed the safe range, it will immediately activate local heat dissipation enhancement measures for that area. For example, the speed of fans near the area can be increased, or local cooling units can be turned on to quickly reduce the local temperature.

[0067] Therefore, feeding back the results of localized enhanced heat dissipation into the calculation of the overall temperature feedback means that the impact of the localized enhanced heat dissipation measures on the regional temperature will be incorporated back into the calculation process of the overall temperature feedback. This forms a closed-loop control, enabling the heat dissipation strategy to be dynamically adjusted and optimized based on the actual effect, ensuring the effectiveness and sustainability of the heat dissipation measures.

[0068] This application effectively addresses the limitations in handling heterogeneous heat distribution within server racks by incorporating consideration of the rate of change of operating load for equipment in each area and combining this with the identification and differentiated processing of high heat density areas. Specifically, by acquiring the rate of change of operating load for equipment in each area, the system can predict potential heat generation trends and dynamically adjust the weight of temperature measurements in each area in the overall temperature feedback calculation. This dynamic weighting mechanism enables the heat dissipation control to respond more sensitively to rapid changes in local heat load, avoiding lag or inaccuracy caused by averaging. Simultaneously, by identifying and independently evaluating the temperature measurements of high heat density areas and differentiating them, this solution ensures that the temperatures of these critical areas receive priority and more stringent control. When the temperature in a high heat density area shows a continuous upward trend, the system can promptly trigger enhanced local cooling, directly intervening in hotspots to quickly and effectively reduce local temperatures. Furthermore, the results of enhanced local cooling are fed back into the calculation of the overall temperature feedback, forming an adaptive closed-loop control. This allows the heat dissipation strategy to be continuously optimized based on actual cooling effects, ensuring the accuracy and efficiency of heat dissipation control.

[0069] The following is a specific example to illustrate this.

[0070] Suppose a data center rack is equipped with various devices such as high-performance computing servers, storage arrays, and network switches, and the workload of these devices varies significantly at different times.

[0071] First, the system continuously acquires temperature measurements from various areas within the server rack, such as deploying temperature sensors in the server area, storage area, and network equipment area. Simultaneously, the system acquires operational load data from the devices within these areas and calculates the rate of change of operational load for each area. For example, if the CPU utilization in the server area spikes from 20% to 90% within a short period, its rate of change of operational load is considered high.

[0072] Based on this, the system will increase the weight of temperature measurements in the overall temperature feedback calculation as the rate of change of server load in that area increases. This means that temperature changes in that area will have a greater impact on the final overall temperature feedback.

[0073] Simultaneously, the system identifies server areas as high-heat-density zones based on preset equipment configuration information and heat generation characteristics. When the monitored temperature measurement value of the server area shows a continuous upward trend, such as a temperature increase of more than 2 degrees Celsius within 5 consecutive minutes, and approaching the preset local temperature limit, the system will immediately trigger enhanced local cooling for that area. Specifically, it can instruct the corresponding rack fan array for that server area to increase its speed to maximum, or activate local liquid cooling units for auxiliary cooling.

[0074] After implementing localized enhanced cooling measures, the system continuously monitors temperature changes in that area and feeds back the cooling effect (e.g., the rate and magnitude of temperature decrease) into the overall temperature feedback calculation. If the temperature decrease is not significant, the system may further adjust the cooling strategy, such as notifying maintenance personnel to conduct an inspection, or maintaining a high weight for that area in the overall feedback to sustain high-intensity cooling. In this way, even if there are dynamically changing high-heat-load areas inside the cabinet, it ensures that their temperatures are always maintained within a safe range, while avoiding unnecessary excessive cooling of other low-load areas.

[0075] In some embodiments, the step of identifying high heat density areas inside the cabinet includes: Obtain configuration information for devices within the rack; Based on the configuration information and the preset equipment heat output characteristic data, the heat output density of each area inside the computer cabinet is determined. The heat output density is compared with a preset regional heat flux density threshold to identify high heat density areas inside the current cabinet. When the configuration information changes, the calculation and identification process is re-executed to update the definition of high heat density areas.

[0076] Specifically, the configuration information of devices within a server rack can be understood as a set of data describing detailed parameters such as the hardware specifications, models, quantities, and locations of the devices installed inside the rack. For example, it may include CPU type, number of GPUs, memory capacity, power module wattage, and hard drive type and quantity. This information forms the basis for evaluating device heat output. The preset device heat output characteristic data refers to a pre-established data model or database used to describe the heat generated by different types or models of devices under different operating loads. This data can be constructed based on manufacturer-provided specifications, experimental test results, or historical operating data, with the aim of providing accurate input for subsequent heat output density calculations. In practical applications, the heat output density of each area within a computer rack refers to the heat generated per unit volume or unit area in each area after dividing the rack into several logical or physical areas based on the device configuration information and its corresponding heat output characteristic data. For example, this can be calculated by mapping the rated thermal power consumption of a specific device to its physical location within the rack and combining this with the volume or area of ​​that area. Furthermore, the preset regional heat flux density threshold is a critical value pre-set based on factors such as the rack's designed heat dissipation capacity, the equipment's safe operating temperature range, and the system's overall energy efficiency targets. When the heat output density of a certain area exceeds this threshold, it is identified as a high heat density area. In addition, when configuration information changes, such as the addition, removal, replacement, or upgrading of equipment within the rack, the system will re-execute the aforementioned heat output density calculation and high heat density area identification process. This ensures that the definition of high heat density areas always remains consistent with the actual physical configuration and thermal load status of the rack, avoiding inaccurate identification due to configuration changes.

[0077] This application, by acquiring the configuration information of the equipment within the server rack and combining it with preset equipment heat generation characteristic data, can accurately calculate the actual heat generation density of each area inside the rack. This calculation method, based on physical configuration and thermal characteristic data, ensures that the identification of high heat density areas no longer relies solely on a single temperature measurement value, but is based on a deep understanding of the distribution and intensity of heat sources. By comparing the calculated heat generation density with preset regional heat flux density thresholds, it can objectively and accurately identify which areas inside the server rack face a high risk of heat load. Furthermore, when the configuration information of the equipment within the rack changes, this solution can promptly re-execute the calculation and identification process, dynamically updating the definition of high heat density areas. This ensures that the heat dissipation control strategy is always optimized for the latest heat load distribution of the rack, avoiding the waste of heat dissipation resources or insufficient heat dissipation caused by statically defining high heat density areas.

[0078] However, in practical applications, server racks often contain heterogeneous devices of various types and from different manufacturers. These devices may use different communication protocols and data formats to report their configuration information. Without a unified mechanism to effectively acquire, parse, and standardize the configuration information of these heterogeneous devices, it may lead to inefficient configuration information acquisition, inconsistent data, or even errors, thereby affecting the accuracy of identifying high-heat-density areas and the effectiveness of heat dissipation control strategies. To address this, this application further proposes a method to optimize the acquisition of configuration information from devices within the server rack. By establishing a unified standard and adapter mechanism, it ensures the accurate and efficient acquisition and management of configuration information.

[0079] In some embodiments, the step of obtaining the configuration information of the equipment in the rack includes: A device information description specification is established in advance, which defines the configuration parameters and data structure required for the device type to be reported; Obtain the set of device interface adapters, where each adapter in the adapter set corresponds to a communication protocol and data format of one or a class of heterogeneous devices; Select the corresponding device interface adapter from the adapter set based on the device type; Communicate with the device through the selected adapter to obtain the device's original configuration information; The raw configuration information is converted into a data structure defined by the device information description specification using the selected adapter; Store and manage the converted configuration information; When the device's configuration information changes, the configuration information is retrieved again through the corresponding adapter.

[0080] Specifically, the device information description specification refers to a set of predefined rules and templates used to uniformly describe the key configuration parameters of various devices within a rack, such as device model, power consumption characteristics, port layout, firmware version, etc., and specifies the data type, value range, and data structure of these parameters. This specification aims to eliminate the differences in information expression between heterogeneous devices, providing a unified benchmark for subsequent data processing and analysis. The device interface adapter set can be understood as a group of software or hardware modules, each specifically designed to communicate with a specific type or manufacturer of devices. For example, one adapter might communicate with a network switch via the SNMP protocol, another with a power management unit via the Modbus protocol, and still another with a server via a RESTful API. These adapters encapsulate the underlying communication details, freeing upper-layer applications from concern themselves with specific communication protocols and data formats. In practical applications, the appropriate adapter is selected based on the device type, for example, by matching the device's vendor ID or model. The selected adapter is responsible for establishing a connection with the device and obtaining its raw configuration information according to the device's own communication protocol. Raw configuration information refers to unprocessed data directly reported by the device, which may have inconsistent formats. Furthermore, after acquiring the raw configuration information, the selected adapter will parse, transform, and standardize this raw data according to pre-established device information description specifications, making it conform to a unified data structure. For example, if the power consumption in the raw information is in watts, but the specification requires it to be in milliwatts, the adapter will perform the corresponding unit conversion. Thus, the transformed configuration information will be stored in a database or configuration management system and effectively managed to facilitate subsequent high-heat-density area calculations and heat dissipation strategy formulation. When the device's configuration information changes, such as during device upgrades, replacements, or reconfigurations, the system will automatically or on-demand re-acquire and update this configuration information through the corresponding adapter, ensuring data real-time performance and accuracy.

[0081] This application effectively addresses the challenges of obtaining configuration information from heterogeneous devices by introducing a device information description specification and a set of device interface adapters. Specifically, the device information description specification provides a unified data model, ensuring that configuration information is represented in a standardized format regardless of the device's origin. This avoids parsing difficulties and errors caused by inconsistent data formats. Meanwhile, the set of device interface adapters acts as a translator between different devices and the system, with each adapter specializing in the communication protocols and data formats of one or a class of devices. When configuration information for a particular device is needed, the system intelligently selects the appropriate adapter based on the device type. This adapter then communicates with the device at the low level, obtains the raw data, and converts it into standardized data conforming to the specification. This mechanism ensures efficient and accurate acquisition of device configuration information even in complex heterogeneous environments. Furthermore, when configuration information changes, it is re-acquired and updated through the corresponding adapter, guaranteeing the real-time nature of the configuration data and providing a reliable data foundation for the accurate identification of subsequent high-heat-density areas.

[0082] Imagine a server rack containing servers, network switches, and storage devices from different vendors. To obtain configuration information for these devices, a device information description specification is first established. This specification defines common parameters that all devices should report (such as device ID, model, and maximum power consumption) as well as parameters required for specific device types (such as CPU model and memory capacity for servers; number of ports and bandwidth for switches; hard drive capacity and RAID configuration for storage devices). Next, the system maintains a set of device interface adapters, including an adapter for the SNMP protocol (for network switches), an adapter for the IPMI protocol (for servers), and an adapter for the RESTful API (for storage devices). When the system needs to obtain configuration information for a server, it identifies its type and selects the IPMI adapter. This adapter communicates with the server via the IPMI protocol to obtain its raw configuration data, such as CPU model and number of memory modules. Subsequently, the IPMI adapter transforms this raw data according to the preset device information description specification, for example, uniformly representing memory capacity in GB units and encapsulating all parameters into a data structure defined by the specification. This standardized configuration information is stored for subsequent calculations of the server's heat output density. When the server's memory modules are replaced or added, the system will re-acquire and update its configuration information through the IPMI adapter to ensure that the identification of high-heat-density areas is always based on the latest device status.

[0083] However, in practical applications, data transmission can be affected by various factors such as network fluctuations and equipment failures, leading to interruptions, data corruption, or incompleteness in the re-acquired configuration information. If these issues are not addressed, the system may rely on inaccurate or incomplete configuration information for thermal control, thereby affecting the effectiveness of thermal management strategies and the stability of equipment operation within the rack.

[0084] In some embodiments, the step of re-obtaining the configuration information through the corresponding adapter when the device's configuration information changes further includes: Initiate the data transmission integrity check process, which includes: During the data transmission of configuration information, data packets are periodically sent and acknowledgment responses are received from the device to verify the continuity of configuration information data transmission. After the configuration information is transmitted, the received configuration information is verified by comparing the data length of the configuration information or calculating the data fingerprint of the configuration information to ensure the data integrity of the configuration information. When a data transmission interruption of configuration information is detected, an acknowledgment response is missing, or the data verification of configuration information fails, a retransmission mechanism is triggered, and the fault type and occurrence time are recorded. If multiple retransmissions still fail, an alternative data acquisition path is activated to obtain historical configuration information from the local cache or adjacent devices, and the device's configuration information is marked as pending verification.

[0085] Specifically, the data transmission integrity check process refers to a mechanism designed to ensure that configuration information is not lost, damaged, or tampered with during transmission. This includes periodically sending data packets and receiving acknowledgment responses from the device to monitor the activity and stability of the data link in real time, ensuring the continuity of the data flow and preventing incomplete data transmission due to momentary interruptions. The comparison of configuration information data length or data fingerprint calculation, such as generating a checksum through a hash algorithm, aims to perform a comprehensive verification of the received data after transmission is complete, determining whether it is completely consistent with the sender's data, thereby ensuring data integrity. When a transmission interruption, missing acknowledgment response, or data verification failure is detected, a retransmission mechanism is triggered to overcome momentary errors during transmission by repeatedly sending data, improving the success rate of data acquisition. Furthermore, recording the fault type and occurrence time provides a basis for subsequent system diagnosis and maintenance. If multiple retransmissions fail to retrieve data, an alternative data retrieval path is initiated, such as from a locally stored copy of historical data or from other reliable adjacent devices. The purpose is to provide a fault tolerance mechanism to ensure that usable configuration information can still be obtained when the primary path fails, and to mark it as pending verification to prompt the system to further evaluate the reliability of the information.

[0086] This application effectively addresses potential transmission interruptions, data corruption, or incompleteness during configuration information reacquisition by introducing a data transmission integrity check process. Specifically, during configuration information data transmission, periodically sending data packets and receiving confirmation responses from devices allows for real-time monitoring of the data link's health. If a confirmation response is missing, it indicates a potential transmission problem, which the system can promptly detect and address. After data transmission is complete, verification is performed by comparing data lengths or calculating data fingerprints to confirm the completeness and integrity of the received configuration information. Upon detecting any transmission problems or verification failures, the system immediately triggers a retransmission mechanism to attempt to restore correct data transmission. This proactive error detection and correction mechanism significantly improves the reliability of configuration information acquisition. Furthermore, in the event of multiple retransmission failures, the system can activate backup data acquisition paths, retrieving historical configuration information from local caches or adjacent devices. This prevents complete loss of configuration information due to a single transmission path failure, ensuring system resilience under extreme conditions. By marking backup data as pending verification, the system alerts subsequent processing modules to conduct additional reliability assessments of this information, further enhancing system robustness.

[0087] Suppose that the configuration information (e.g., CPU model, memory capacity, number of hard drives, etc.) of a server in a rack changes and needs to be retrieved again through the corresponding device interface adapter. During data transmission, the system periodically sends heartbeat packets or data requests to the server and waits for the server's confirmation response. If no confirmation response is received for several consecutive cycles, the system determines that data transmission may have been interrupted. At this time, the system triggers a retransmission mechanism to re-request the transmission of the server's configuration information. After all configuration information data packets have been received, the system calculates the hash value (data fingerprint) of the received data packets and compares it with the original hash value attached to the data packets sent by the server. If the comparison results are inconsistent, it indicates that the data may have been corrupted or tampered with during transmission, and the system will trigger retransmission again. If, after a preset number of retransmissions (e.g., 3), the system still cannot successfully obtain complete and verified configuration information, it will activate an alternative data acquisition path. For example, the system will first try to find the latest available configuration of the server from the locally stored historical configuration information cache, or request shared configuration information from adjacent devices on the same network segment as the server. Meanwhile, the backup configuration information will be marked as pending verification, prompting the system to perform additional logical checks or manual confirmation before application to ensure its availability. Through this series of mechanisms, even in unstable network environments or with abnormal device communication, the system can obtain reliable device configuration information to the greatest extent possible, thus providing an accurate data foundation for intelligent heat dissipation control of the cabinet.

[0088] However, this verification method primarily focuses on detecting whether data is corrupted or lost during transmission. It struggles to effectively verify the semantic correctness, numerical rationality, or logical consistency of the configuration information itself. For example, even if the data length and fingerprint are correct, the value of a parameter in the configuration information might exceed the actual operating range of the device, or there might be logical conflicts between different parameters. If these problems are not detected, they could lead to deviations in the heat dissipation control strategy or even affect the normal operation of the device. To address this, this application proposes a more refined and intelligent configuration information verification method. By introducing device firmware version information and vendor identification, and combining it with a pre-defined verification rule base, it performs multi-dimensional verification of the configuration information, thereby significantly improving the reliability and accuracy of the configuration information.

[0089] In some embodiments, the steps of verifying the received configuration information after the data transmission of the configuration information is completed, comparing the data length of the configuration information or calculating the data fingerprint of the configuration information to ensure the data integrity of the configuration information include: Obtain the device's firmware version information and manufacturer identifier; Based on the device's firmware version information and the device's manufacturer identifier, select a set of verification rules that matches the device from the preset verification rule library; Based on the selected set of verification rules, the received configuration information is verified in multiple dimensions, including data type checks, numerical range checks, field format checks, and logical consistency checks of key fields, to ensure the data integrity of the configuration information.

[0090] Specifically, obtaining the device's firmware version information and manufacturer identifier means that after receiving the device's configuration information, the system first extracts the current firmware version and manufacturer identifier from the device or its associated metadata. This information is the basis for subsequent accurate verification, because devices from different manufacturers and with different firmware versions may have significant differences in the definition, value range, and logical relationships of their configuration parameters. The pre-set verification rule base can be understood as a database that centrally stores a large number of verification rules for configuration information of different device types, manufacturers, and firmware versions. This rule base can be pre-maintained and updated by the system administrator or device manufacturer. After obtaining the device's firmware version information and manufacturer identifier, the system will search and select the most matching set of verification rules from this rule base based on this identification information. For example, for a specific model of switch, its rule set may include the allowed range of port speeds, the valid range of VLAN IDs, etc. In practical applications, multi-dimensional verification of the received configuration information based on the selected verification rule set refers to using the selected rule set to conduct a thorough and detailed examination of all aspects of the configuration information. Specifically, data type checks ensure that the value of each configuration parameter conforms to its predefined data type; for example, a field representing a port number should not be a string but an integer. Numerical range checks verify that the numerical value of a configuration parameter is between its allowed minimum and maximum values; for example, fan speed cannot be negative, and temperature thresholds cannot exceed physical limits. Field format checks verify that the format of configuration parameters conforms to specifications; for example, IP addresses are in valid dotted decimal format. Logical consistency checks of critical fields assess whether the logical relationships between different configuration parameters are reasonable. For example, if a port is configured to be disabled, its associated bandwidth setting should not be the maximum value of the active state; or, if a power module is configured in redundant mode, at least two power modules must be operational.

[0091] This application first obtains the device's firmware version information and manufacturer identifier, thereby accurately identifying the device's specific model and operating environment. Based on this identification information, the system can dynamically select the most suitable verification rule set from a preset verification rule base. It is precisely because of the use of a rule set highly matched to the device's characteristics that subsequent multi-dimensional verification can specifically and thoroughly check the configuration information for data type, numerical range, field format, and logical consistency of key fields. This mechanism effectively compensates for the shortcomings of relying solely on data length or data fingerprint verification, and can discover and correct configuration data that was not corrupted during transmission but is inherently unreasonable or does not conform to the device's operating specifications, thus fundamentally ensuring the accuracy and validity of the configuration information.

[0092] The following is a specific example to illustrate this.

[0093] Suppose there is a network switch in the rack, and its configuration information needs to be acquired for adjusting the thermal control strategy. First, the system communicates with the switch through the device interface adapter to obtain its configuration information. During this acquisition, the system also obtains the switch's firmware version (Firmware_v3.2) and vendor identifier (Vendor_X). After the configuration information is transmitted, the system no longer simply compares data lengths or calculates data fingerprints. Instead, it uses the Firmware_v3.2 and Vendor_X information to search for and load a set of verification rules that precisely matches this device from a pre-defined verification rule base. This set of verification rules may include the following rules: 1. The data type of the port rate field must be an integer. 2. The port rate value range must be selected from 100Mbps, 1Gbps, and 10Gbps. 3. The IP address field must conform to the IPv4 or IPv6 format specification. 4. If a port is configured to be disabled, its corresponding bandwidth setting must be 0. At this point, if the received configuration information contains an incorrect setting of a port's rate to 200Mbps (not within the preset range), or a disabled port's bandwidth to 10Gbps (not logically consistent), this solution can accurately identify these errors through multi-dimensional verification, even if the data length and fingerprint verification pass. Once an error is detected, the system can trigger a retransmission mechanism, mark the configuration information as untrusted, and activate a backup data acquisition path, thereby ensuring that the configuration information used for thermal control is accurate and effective.

[0094] like Figure 2 As shown in the embodiments, this application also discloses an intelligent heat dissipation control system for server racks based on edge computing, including: Temperature measurement value acquisition module 1 is used to acquire the temperature measurement value inside the cabinet in real time; Operating load data acquisition module 2 is used to acquire operating load data of equipment in the cabinet; Expected range inference module 3 is used to infer the expected range of the measured temperature values ​​inside the cabinet based on the operating load data; The reliability index determination module 4 is used to determine the reliability index of the temperature measurement value based on the degree of deviation between the temperature measurement value and the expected range, as well as the time variation characteristics of the temperature measurement value. Temperature measurement value processing module 5 is used to process temperature measurement values ​​according to reliability indicators to obtain comprehensive temperature feedback for heat dissipation control. The heat dissipation actuator adjustment module 6 is used to adjust the operating status of the heat dissipation actuator in the cabinet based on the comprehensive temperature feedback.

[0095] The edge computing-based intelligent heat dissipation control system for server racks in this application provides a more intelligent, efficient, and robust solution for data center heat dissipation management through its unique modular design and reliability assessment mechanism, demonstrating significant technological advancements and practical value.

[0096] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for intelligent heat dissipation control of server racks based on edge computing, characterized in that, include: Real-time acquisition of temperature measurements inside the server rack; Obtain operating load data for equipment within the cabinet; Based on operational load data, infer the expected range of internal temperature measurements within the server rack; The reliability index of the temperature measurement value is determined based on the degree of deviation between the temperature measurement value and the expected range, as well as the time variation characteristics of the temperature measurement value. Based on reliability indicators, the temperature measurement values ​​are processed to obtain comprehensive temperature feedback for heat dissipation control. Based on comprehensive temperature feedback, adjust the operating status of the heat dissipation actuators inside the cabinet.

2. The intelligent heat dissipation control method for server racks based on edge computing according to claim 1, characterized in that, The step of determining the reliability index of the temperature measurement value based on the degree of deviation between the temperature measurement value and the expected range, and the time variation characteristics of the temperature measurement value, further includes: Obtain operational data on the cooling actuators within the server rack; Obtain component health output data for devices within the rack; Based on operational input data and component health output data, assess the heat dissipation costs and component benefits of the rack area; When the heat dissipation cost continues to exceed the preset threshold or the component benefit continues to be lower than the preset threshold, the strategy exploration mode is triggered. In strategy exploration mode, the heat dissipation target of the rack area is proactively adjusted, and the operating intensity of the heat dissipation actuator is forcibly increased; In this exploration process, the reliability index of temperature measurements was re-evaluated, and the reliability index of the re-evaluated temperature measurements was cross-validated by combining the temperature difference pattern between the internal temperature of the equipment in the cabinet and the ambient temperature of the cabinet.

3. The intelligent heat dissipation control method for server racks based on edge computing according to claim 1, characterized in that, The step of determining the reliability index of the temperature measurement value based on the degree of deviation between the temperature measurement value and the expected range, and the time variation characteristics of the temperature measurement value, further includes: Perform an instantaneous validity check on the temperature measurement value, wherein the instantaneous validity check includes comparing the current temperature measurement value with the previous or most recent temperature measurement values; When the temperature measurement value fluctuates beyond a preset range within a preset time period, the reliability index of the temperature measurement value is reduced. Continuously monitor temperature measurements within a preset time window; When temperature measurements consistently exhibit high-frequency fluctuations, the reliability index of the temperature measurements is reduced.

4. The intelligent heat dissipation control method for server racks based on edge computing according to claim 1, characterized in that, The step of processing the temperature measurement values ​​according to reliability indicators to obtain comprehensive temperature feedback for heat dissipation control includes: Obtain the rate of change of operating load for equipment in each area of ​​the cabinet; Adjust the weight of temperature measurements in the overall temperature feedback calculation based on the rate of change of operating load of equipment in each area; When the rate of change of equipment operating load in a certain area exceeds a preset threshold, the weight of the temperature measurement value in that area in the comprehensive temperature feedback calculation is increased. When the rate of change of equipment operating load in a certain area is lower than a preset threshold, the weight of the temperature measurement value in that area in the comprehensive temperature feedback calculation is reduced. Based on the adjusted weights, the temperature measurements are weighted to obtain a comprehensive temperature feedback for heat dissipation control.

5. The intelligent heat dissipation control method for server racks based on edge computing according to claim 1, characterized in that, The step of processing the temperature measurement values ​​according to reliability indicators to obtain comprehensive temperature feedback for heat dissipation control further includes: Identify high heat density areas inside the server rack; Based on reliability indicators, the temperature measurements of high heat density areas are independently evaluated and differentiated from the temperature measurements of other areas in the cabinet to obtain comprehensive temperature feedback for heat dissipation control. When the temperature measurement value of a high heat density area shows a continuous upward trend, it triggers localized enhanced heat dissipation in that area; The results of the localized enhanced heat dissipation are fed back into the calculation of the overall temperature feedback.

6. The intelligent heat dissipation control method for server racks based on edge computing according to claim 5, characterized in that, The step of identifying high heat density areas inside the cabinet includes: Obtain configuration information for devices within the rack; Based on the configuration information and the preset equipment heat output characteristic data, the heat output density of each area inside the computer cabinet is determined. The heat output density is compared with a preset regional heat flux density threshold to identify high heat density areas inside the current cabinet.

7. The intelligent heat dissipation control method for server racks based on edge computing according to claim 6, characterized in that, The steps for obtaining the configuration information of the devices within the rack include: Establish equipment information description specifications in advance, and define the configuration parameters and data structures required for reporting equipment types. Get the set of device interface adapters; Select the corresponding device interface adapter from the adapter set based on the device type; Communicate with the device through the selected adapter to obtain the device's original configuration information; The raw configuration information is converted into a data structure defined by the device information description specification using the selected adapter; Store and manage the converted configuration information; When the device's configuration information changes, the configuration information is retrieved again through the corresponding adapter.

8. The intelligent heat dissipation control method for server racks based on edge computing according to claim 7, characterized in that, The step of re-obtaining the configuration information through the corresponding adapter when the device's configuration information changes further includes: Initiate the data transmission integrity check process, which includes: During the data transmission of configuration information, data packets are periodically sent and acknowledgment responses are received from the device to verify the continuity of configuration information data transmission. After the configuration information is transmitted, the received configuration information is verified by comparing the data length of the configuration information or calculating the data fingerprint of the configuration information to ensure the data integrity of the configuration information. When a data transmission interruption of configuration information is detected, an acknowledgment response is missing, or the data verification of configuration information fails, a retransmission mechanism is triggered, and the fault type and occurrence time are recorded. If multiple retransmissions still fail, an alternative data acquisition path is activated to obtain historical configuration information from the local cache or adjacent devices, and the device's configuration information is marked as pending verification.

9. The intelligent heat dissipation control method for server racks based on edge computing according to claim 8, characterized in that, The step of verifying the received configuration information after the data transmission of the configuration information is completed, comparing the data length of the configuration information or calculating the data fingerprint of the configuration information to ensure the data integrity of the configuration information includes: Obtain the device's firmware version information and manufacturer identifier; Based on the device's firmware version information and the device's manufacturer identifier, select a set of verification rules that matches the device from the preset verification rule library; Based on the selected set of verification rules, the received configuration information is verified in multiple dimensions, including data type checks, numerical range checks, field format checks, and logical consistency checks of key fields, to ensure the data integrity of the configuration information.

10. A rack-based intelligent heat dissipation control system based on edge computing, characterized in that, include: Temperature measurement acquisition module, used to acquire temperature measurement values ​​inside the cabinet in real time; The operating load data acquisition module is used to acquire the operating load data of the equipment in the cabinet; The expected range inference module is used to infer the expected range of the measured temperature values ​​inside the cabinet based on the operating load data; The reliability index determination module is used to determine the reliability index of the temperature measurement value based on the degree of deviation between the temperature measurement value and the expected range, as well as the time variation characteristics of the temperature measurement value. The temperature measurement value processing module is used to process the temperature measurement values ​​according to reliability indicators to obtain comprehensive temperature feedback for heat dissipation control. The heat dissipation actuator adjustment module is used to adjust the operating status of the heat dissipation actuator in the cabinet based on comprehensive temperature feedback.

Images