Heat dissipation equipment management method and device, equipment and readable storage medium

Abstract

The invention provides a heat dissipation equipment management method, device and equipment and a readable storage medium, and the method comprises the steps that heat dissipation parameters of all nodes are obtained respectively, and the heat dissipation parameters comprise operation power, air inlet air temperature and air outlet air temperature; according to the heat dissipation parameter of each node, the air volume parameter of each node is obtained; and according to the air volume requirement of a heat dissipation object corresponding to the target heat dissipation equipment, the output air volume of the target heat dissipation equipment is adjusted, the heat dissipation object comprises a node set composed of one or more adjacent nodes in the server cluster, and the air volume requirement of the heat dissipation object is associated with the air volume parameter of each node of the node set. According to the technical scheme, the air volume of each node is accurately calculated in a non-intrusive mode, the output of the heat dissipation equipment is dynamically adjusted based on the total demand of the node set, accurate matching of the refrigerating capacity and the heat dissipation demand of the equipment is achieved, local hot spots and excessive refrigeration are effectively eliminated, energy consumption is remarkably reduced, and the heat dissipation reliability of the system is improved.

Description

Technical Field

[0001] This specification relates to the field of communication technology, and in particular to a heat dissipation device management method, apparatus, device, and readable storage medium. Background Technology

[0002] Data centers, as the core carriers of digital infrastructure, are undergoing unprecedented expansion and density upgrades. The power density of a single server rack has rapidly increased from several kilowatts to tens or even hundreds of kilowatts. The resulting thermal management challenges have become a key bottleneck restricting the reliability and energy efficiency of data centers. In high-density deployment scenarios, if the large amount of heat generated by servers operating under continuous high load cannot be dissipated in a timely and effective manner, it will not only lead to processor throttling and system performance degradation, but may also cause serious consequences such as hardware overheating damage, data loss, and even business interruption. Accurately grasping the actual cooling air requirements of servers and implementing precise airflow organization and management are core technical requirements for ensuring the stable operation of data center infrastructure and extending equipment lifespan.

[0003] Data center cooling strategies primarily rely on traditional room-level environmental temperature control architectures. This architecture deploys cooling equipment such as Computer Room Air Conditioning (CRAC) or Computer Room Air Handlers (CRAH) to maintain the overall room temperature within a preset safe range. In practice, existing systems typically use temperature sensors at key locations within the room to monitor macroscopic environmental parameters such as air conditioning supply and return air temperatures, and the average temperature within the room. Maintenance personnel then manually or semi-automatically adjust the cooling equipment's outlet air temperature or fan speed based on experience or simple threshold rules. This temperature control model, with the room as the smallest unit of management, is designed based on the assumption of uniform heat distribution within the room and consistent cooling requirements for all equipment. It ensures that even servers in the most unfavorable locations receive sufficient cooling capacity through overall environmental overcooling.

[0004] In the aforementioned solutions, data center-level temperature control lacks precise quantitative assessment methods for the actual airflow requirements of individual servers or racks. It fails to establish a precise mathematical relationship between server power, temperature, and required cooling airflow, forcing maintenance personnel to rely on rough estimates based on experience. This prevents them from knowing whether each device has truly received the minimum necessary airflow to meet its cooling needs. Secondly, due to the coarse monitoring granularity, existing technologies struggle to identify localized hot spots or insufficient cooling in individual racks caused by poor airflow organization. When the monitoring system detects a server triggering a high-temperature alarm, the equipment is often already overheated. This passive response approach has a significant time lag, failing to provide early warning and proactive intervention for potential heat dissipation risks. To avoid equipment failures caused by localized overheating, data center operations typically adopt a conservative overcooling strategy, reducing the overall ambient temperature or increasing air conditioning output to cover the worst heat dissipation scenarios. This directly leads to a significant waste of cooling energy, resulting in persistently high Power Usage Effectiveness (PUE) values ​​in data centers. Furthermore, the above solution cannot dynamically adjust the allocation of cooling resources according to real-time changes in server load. When some racks are at low load while others are at full load, the uniform cooling method is still used, which further exacerbates the low energy efficiency.

[0005] While the problem of quantitative assessment can be solved by directly measuring airflow by installing hardware devices such as wind speed sensors and air volume meters at the air inlets of each server or rack, such hardware measurement solutions not only require large-scale modification of existing infrastructure and increase additional hardware procurement and deployment costs, but the sensors themselves are also affected by various factors such as rack layout, cable obstruction, and airflow interference from nearby devices, making it difficult to guarantee measurement accuracy. At the same time, the introduction of a large number of sensors also increases the complexity of the system and maintenance costs. Summary of the Invention

[0006] In view of this, this specification provides a method, apparatus, device, and readable storage medium for managing heat dissipation equipment, in order to improve the aforementioned problem of difficulty in accurately and cost-effectively quantifying and evaluating heat dissipation configurations.

[0007] The specific technical solution is as follows: This specification provides a method for managing heat dissipation equipment, applied to a management device that manages the heat dissipation equipment of a server cluster, the server cluster comprising several nodes. The method includes: acquiring the heat dissipation parameters of each node, the heat dissipation parameters including operating power, inlet air temperature, and outlet air temperature; acquiring the airflow parameters of each node based on the heat dissipation parameters of each node; and adjusting the output airflow of the target heat dissipation equipment according to the airflow requirements of the target heat dissipation equipment, the heat dissipation equipment comprising a set of nodes consisting of one or more adjacent nodes in the server cluster, the airflow requirements of the heat dissipation equipment being associated with the airflow parameters of each node in the node set.

[0008] As a technical solution, obtaining the airflow parameters of each node based on the heat dissipation parameters of each node includes: selecting a node in the server cluster as the target node; obtaining the inlet and outlet temperature difference based on the inlet and outlet air temperatures of the target node; and using the quotient obtained by dividing the operating power by the product of the air specific heat capacity, air density, and inlet and outlet temperature difference as the airflow parameters of the target node; and iterating through each node in the server cluster, repeating the above steps to obtain the airflow parameters of each node.

[0009] As a technical solution, the heat dissipation device includes a precision air conditioner in the physical space where the server cluster is located, the heat dissipation object includes a target rack corresponding to the precision air conditioner, and the target rack includes multiple nodes; the step of adjusting the output air volume of the target heat dissipation device according to the air volume requirement of the heat dissipation object corresponding to the target heat dissipation device includes: accumulating the air volume parameters of each node included in the target rack to obtain the total air volume parameter corresponding to the target rack, and adjusting the output air volume of the precision air conditioner according to the total air volume parameter.

[0010] As a technical solution, a special node within the heat dissipation object is identified, and the physical position of the special node is adjusted. The ratio of the airflow parameter of the special node to the average airflow parameter of all nodes within the heat dissipation object is greater than a first threshold or less than a second threshold, wherein the second threshold is less than the first threshold.

[0011] This specification also provides a heat dissipation equipment management device, applied to a management device that manages the heat dissipation equipment of a server cluster. The server cluster includes several nodes. The device includes: a first module for acquiring the heat dissipation parameters of each node, including operating power, inlet air temperature, and outlet air temperature; a second module for acquiring the airflow parameters of each node based on its heat dissipation parameters; and a third module for adjusting the output airflow of the target heat dissipation equipment according to the airflow requirements of the target heat dissipation equipment, wherein the target heat dissipation equipment includes a set of nodes consisting of one or more adjacent nodes in the server cluster, and the airflow requirements of the target heat dissipation equipment are associated with the airflow parameters of each node in the set of nodes.

[0012] As a technical solution, obtaining the airflow parameters of each node based on the heat dissipation parameters of each node includes: selecting a node in the server cluster as the target node; obtaining the inlet and outlet temperature difference based on the inlet and outlet air temperatures of the target node; and using the quotient obtained by dividing the operating power by the product of the air specific heat capacity, air density, and inlet and outlet temperature difference as the airflow parameters of the target node; and iterating through each node in the server cluster, repeating the above steps to obtain the airflow parameters of each node.

[0013] As a technical solution, the heat dissipation device includes a precision air conditioner in the physical space where the server cluster is located, the heat dissipation object includes a target rack corresponding to the precision air conditioner, and the target rack includes multiple nodes; the step of adjusting the output air volume of the target heat dissipation device according to the air volume requirement of the heat dissipation object corresponding to the target heat dissipation device includes: accumulating the air volume parameters of each node included in the target rack to obtain the total air volume parameter corresponding to the target rack, and adjusting the output air volume of the precision air conditioner according to the total air volume parameter.

[0014] As a technical solution, it also includes a fourth module for identifying special nodes within the heat dissipation object and prompting adjustments to the physical position of the special nodes. The ratio of the airflow parameter of the special node to the average airflow parameter of each node within the heat dissipation object is greater than a first threshold or less than a second threshold, wherein the second threshold is less than the first threshold.

[0015] This specification also provides an electronic device, including a processor and a readable storage medium storing machine-executable instructions that can be executed by the processor, the processor executing the machine-executable instructions to implement the aforementioned heat dissipation device management method.

[0016] This specification also provides a readable storage medium storing machine-executable instructions that, when invoked and executed by a processor, cause the processor to implement the aforementioned heat dissipation device management method.

[0017] The technical solutions provided in this specification offer at least the following beneficial effects: By accurately calculating the airflow of each node in a non-intrusive manner and dynamically adjusting the output of the heat dissipation equipment based on the total demand of the node set, the cooling capacity and the heat dissipation demand of the equipment are precisely matched, effectively eliminating local hot spots and over-cooling, significantly reducing energy consumption and improving the reliability of system heat dissipation. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings used in the description of the embodiments of this specification or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings of the embodiments of this specification.

[0019] Figure 1 This is a flowchart of a heat dissipation device management method in one embodiment of this specification; Figure 2 This is a structural diagram of a heat dissipation equipment management device according to one embodiment of this specification; Figure 3 This is a hardware structure diagram of an electronic device according to one embodiment of this specification.

[0020] Reference numerals: Module 1 21, Module 22, Module 3 23. Detailed Implementation

[0021] The terminology used in the embodiments described herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this specification. The singular forms “a,” “described,” and “the” as used in this specification and claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to any and all possible combinations comprising one or more of the associated listed items.

[0022] It should be understood that although the terms first, second, third, etc., may be used to describe various information in embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this specification, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" may also be interpreted as "when," "when," or "in response to a determination."

[0023] In view of the above, this specification provides a method, apparatus, device and readable storage medium for managing heat dissipation equipment, so as to at least improve one of the above-mentioned technical problems.

[0024] The specific technical solution is described below.

[0025] In one embodiment, this specification provides a heat dissipation device management method applied to a management device that manages the heat dissipation devices of a server cluster, the server cluster comprising several nodes. The method includes: acquiring heat dissipation parameters for each node, the heat dissipation parameters including operating power, inlet air temperature, and outlet air temperature; acquiring airflow parameters for each node based on the heat dissipation parameters of each node; and adjusting the output airflow of the target heat dissipation device according to the airflow requirement of the target heat dissipation device corresponding to the heat dissipation object, the heat dissipation object comprising a node set consisting of one or more adjacent nodes in the server cluster, the airflow requirement of the heat dissipation object being associated with the airflow parameters of each node in the node set.

[0026] This method applies to a centralized management device, which can physically be a standalone server or a high-availability cluster, but logically serves as the intelligent hub of the entire data center thermal management system. This management device manages the thermal devices that work in conjunction with the server cluster via network protocols. These devices are typically precision air conditioning (CRAC) units located within the server room, and may also include auxiliary cooling devices such as fan walls at the rack level. The server cluster consists of a large number of computing nodes, which can be standalone servers or individual blades within blade servers. These nodes are the basic units that generate heat and are the ultimate target of thermal management.

[0027] Specifically, such as Figure 1 This includes the following steps, the order of which can be changed depending on the needs of the actual application scenario: Step S11: Obtain the heat dissipation parameters of each node, including operating power, inlet air temperature, and outlet air temperature.

[0028] The management device establishes a secure communication connection with the baseboard management controller (BMC) of each compute node via standard out-of-band management protocols such as IPMI or Redfish. The acquired thermal parameters are fundamental to the computation and primarily include three key physical quantities: the node's real-time operating power, the node's inlet air temperature, and the node's outlet air temperature. Operating power is typically obtained directly by the node's BMC through monitoring the motherboard power supply or power module; it accurately reflects the node's heat generation rate under the current compute load, and is usually measured in watts. Inlet air temperature is obtained through temperature sensors deployed at the front of the node (usually near the hard drive bays or inside the front panel of the chassis); it represents the initial temperature of the air entering the node to cool the electronic components. Outlet air temperature is obtained through temperature sensors deployed at the rear of the node (usually in the PCIe slot area or inside the rear panel of the chassis, near the fan vents); it represents the final temperature of the air that has absorbed heat from the node and is about to be discharged into the server room's hot aisle. The management device initiates a polling of the BMC of all nodes in the cluster at a preset sampling period, such as every 30 seconds or every minute, to obtain these key telemetry data in batches and store them in a time series database for subsequent calculation and analysis.

[0029] Step S12: Obtain the airflow parameters for each node based on the heat dissipation parameters of each node.

[0030] The airflow parameter here specifically refers to the minimum volumetric flow rate of cooling air required for this node to maintain its current thermal equilibrium. Its calculation is based on the classical thermodynamic formula, which states that heat equals the product of mass, specific heat capacity, and temperature difference.

[0031] Specifically, for a computing node in a stable heat dissipation state, the heat generated per unit time should be equal to the heat carried away by the air flowing through it per unit time. The heat generated by the node per unit time is numerically equal to its operating power. The heat carried away by the air is equal to the product of the air mass flow rate, the air's specific heat capacity at constant pressure, and the temperature difference between the air before and after it flows through the node (i.e., the difference between the outlet air temperature and the inlet air temperature).

[0032] By simultaneously solving these two equations, the required mass flow rate can be calculated. To obtain a more intuitive volumetric flow rate, the mass flow rate needs to be further divided by the density of air. Air density is not constant but varies with temperature, making it a variable that needs to be determined based on actual conditions. In a preferred embodiment of this method, the air density value is dynamically determined using a lookup table. The management device has a pre-stored or dynamically updated table of the correspondence between air density and temperature, based on the physical properties under standard atmospheric pressure.

[0033] For example, when the inlet air temperature is 25 degrees Celsius, the air density, according to the table, is approximately 1.185 kg / m³; when the inlet air temperature rises to 35 degrees Celsius, the air density decreases to approximately 1.146 kg / m³. The specific heat capacity of air at constant pressure can be considered a constant within the typical temperature range of a computer room, usually taken as 1005 joules per kilogram of temperature.

[0034] Therefore, for any node, the formula for calculating its required volumetric flow rate (usually in cubic meters per second) can be specifically expressed as: the node operating power (watts) divided by the product of the air constant pressure specific heat capacity (1005 joules per kilogram per degree Celsius), the air density (kilograms per cubic meter, obtained from the table of inlet air temperature), and the temperature difference between the node inlet and outlet (degree Celsius).

[0035] Through this series of calculations, the management equipment can non-invasively and with high accuracy estimate the actual cooling requirements of each node at any given moment without installing any additional airflow sensors. This software- and model-based calculation method is the cornerstone for achieving zero hardware modification costs and refined management.

[0036] Step S13: Adjust the output airflow of the target heat dissipation device according to the airflow requirements of the target heat dissipation device.

[0037] Directly using raw sampled data for calculations and decisions may introduce noise due to instantaneous errors in sensors or minor jitter in network transmission, resulting in short-term spikes or troughs in the calculated airflow parameters, known as "glitch" values. If these glitch values ​​are directly used to control cooling equipment, they may cause system oscillations, leading to frequent and unnecessary acceleration and deceleration of the precision air conditioner, ultimately reducing efficiency and shortening equipment lifespan. To address this issue, this method introduces a data smoothing and verification step after calculating the airflow parameter for a single sampling point. The management device maintains a data sliding window for each node, such as a one-hour window, which stores the sequence of airflow parameters calculated for all sampling periods within that window in chronological order. When a newly calculated airflow parameter arrives, the system compares it with the statistical characteristics (such as the mean and standard deviation) of the existing data in the current window. If the new data deviates from the mean by more than a preset multiple (e.g., three standard deviations), the data point is marked as an outlier. Outliers are not directly used for subsequent decisions. Instead, a robust interpolation strategy is employed to replace them. For example, the arithmetic mean of the preceding and following valid data points (if available) can be used, or the moving average of the data within the window can be used to replace the outlier. This sliding window filtering algorithm effectively smooths out random interference, resulting in a smoother and more reliable airflow parameter curve that reflects the actual cooling demand of the node, laying a solid data foundation for subsequent precise control.

[0038] Establish a logical mapping relationship between heat dissipation devices and computing nodes. The management device maintains a global heat dissipation topology mapping table. This table is pre-configured by the system administrator or through an automatic discovery mechanism, defining the heat dissipation objects that each heat dissipation device (e.g., the precision air conditioner numbered CRAC-01) is responsible for cooling.

[0039] A heat dissipation object is a set of one or more physically adjacent nodes in a server cluster. The most common heat dissipation object is a complete server rack, where all compute nodes, network switches, and storage devices installed from top to bottom constitute a heat dissipation object, cooled by corresponding precision air conditioners. In more sophisticated deployments, the upper and lower halves of a rack may be handled by different air conditioners or different air ducts of the same air conditioner. In such cases, the heat dissipation object can be defined as the set of nodes in the upper half and the set of nodes in the lower half of the rack. By querying this mapping table, the management device can quickly summarize the sum of the airflow parameters of all member nodes within the heat dissipation object corresponding to any target heat dissipation device at the current moment. This sum represents the total airflow requirement of that heat dissipation object.

[0040] The management device first obtains the current operating parameters of the target heat dissipation device (such as a precision air conditioner) from its control system through a standard interface (such as Modbus TCP, BACnet, etc.). The most critical parameter is its rated air volume or current actual output air volume (the unit is usually cubic meters per second or cubic feet per minute).

[0041] Subsequently, the system compares the calculated total airflow requirement of the object to be cooled with the capacity of the cooling equipment, introducing an engineering safety margin or elasticity coefficient, such as 0.9. If the current output airflow of the cooling equipment is less than the product of the total airflow requirement of the object to be cooled and the elasticity coefficient (i.e., equipment airflow < 0.9 * total required airflow), it is determined that the current cooling supply is insufficient and there is a risk of overheating. At this time, the management device generates a control command and sends it to the target cooling equipment through the network, instructing it to increase the fan speed or opening to increase the output airflow, ensuring that enough cool air is delivered to the corresponding node.

[0042] Conversely, if the current output airflow of the cooling device is significantly greater than the total required airflow (e.g., more than 1.2 times the total required airflow), it indicates a clear over-airflow, i.e., over-cooling. In this case, the management device will instruct the target cooling device to appropriately reduce its output airflow until it meets a reasonable upper limit. This dynamic adjustment mechanism achieves real-time matching between cooling capacity and heat load, thereby minimizing energy waste while ensuring heat dissipation safety.

[0043] In one implementation, obtaining the airflow parameters of each node based on the heat dissipation parameters of each node includes: selecting a node in the server cluster as the target node; obtaining the inlet and outlet temperature difference based on the inlet and outlet air temperatures of the target node; and using the quotient obtained by dividing the operating power by the product of the air specific heat capacity, air density, and inlet and outlet temperature difference as the airflow parameters of the target node; and iterating through each node in the server cluster, repeating the above steps to obtain the airflow parameters of each node.

[0044] In one embodiment, the heat dissipation device includes a precision air conditioner in the physical space where the server cluster is located, the heat dissipation object includes a target rack corresponding to the precision air conditioner, and the target rack includes multiple nodes; the step of adjusting the output air volume of the target heat dissipation device according to the air volume requirement of the heat dissipation object corresponding to the target heat dissipation device includes: accumulating the air volume parameters of each node included in the target rack to obtain the total air volume parameter corresponding to the target rack, and adjusting the output air volume of the precision air conditioner according to the total air volume parameter.

[0045] In one implementation, a special node within the heat dissipation object is identified, and a prompt is made to adjust the physical position of the special node. The ratio of the airflow parameter of the special node to the average airflow parameter of all nodes within the heat dissipation object is greater than a first threshold or less than a second threshold, wherein the second threshold is less than the first threshold.

[0046] In one implementation, the management device is responsible for the centralized control of the heat dissipation equipment of the server cluster, which includes several independently operating nodes (i.e., single servers). Through the full-process management of accurately collecting the heat dissipation parameters of each node, scientifically calculating the air volume requirements, and dynamically adjusting the output air volume of the heat dissipation equipment, the heat dissipation efficiency of the server cluster is optimized and the energy consumption is minimized.

[0047] Taking data center scenarios as an example, management devices are usually deployed on dedicated servers or cloud platforms in the form of data center management software, and have the ability to communicate with the BMC (Baseboard Management Controller) and heat dissipation equipment (mainly precision air conditioning) control systems of each node in the server cluster.

[0048] The nodes of the server cluster are distributed across different racks. Each rack is deployed according to a standardized unit (U) configuration. For example, a data center may have 10 rows of racks, each row containing 20 cabinets. Each cabinet supports 42U deployments, with a total of 800 servers deployed as cluster nodes. Each server is configured with an independent BMC module. Additionally, 15 precision air conditioners are deployed in different areas of the data center for cooling purposes. Each precision air conditioner controls 1-2 rows of racks within a specific area (i.e., the cooling target is the set of all nodes contained in the racks within that area).

[0049] Before implementing thermal management, the management equipment needs to complete basic configuration work. Maintenance personnel use the visual interface of the data center management software to bind each server node to its actual deployment location, clarifying the rack, cabinet, and unit information to which the node belongs. For example, "Node S1-001 is deployed in the 10U position of the 3rd cabinet in the 1st row of racks," and "Node S2-056 is deployed in the 32U position of the 8th cabinet in the 2nd row of racks." Simultaneously, based on the air supply range, airflow organization characteristics, and rack distribution of the data center air conditioner, a correspondence is established between precision air conditioners and the objects to be cooled. For example, "Precision air conditioner AC-01 corresponds to all nodes in racks 1-2," and "Precision air conditioner AC-08 corresponds to all nodes in racks 7-8." This configuration information is stored in the management equipment's database, forming a complete topology mapping relationship, providing a basis for subsequent parameter collection, airflow calculation, and airflow adjustment.

[0050] The heat dissipation parameters of each node are acquired separately, including the operating power, inlet air temperature, and outlet air temperature. For operating power acquisition, the management device establishes a communication connection with the BMC software of each node based on IPMI (Intelligent Platform Management Interface), utilizing the BMC module to monitor the server's power consumption data in real time. Each server's BMC has a built-in power sensor capable of accurately collecting real-time power consumption of core components such as the CPU, memory, hard drive, and motherboard, and summarizing it as the overall system operating power. The sampling frequency can be configured by maintenance personnel according to actual needs; the default setting is once every 5 seconds to ensure that power fluctuations caused by changes in server load are captured.

[0051] For example, when node S1-001 is under low load (such as running only basic services), the BMC collects an operating power of 85W. When the node starts an AI training task, the load increases sharply, and the operating power climbs to 320W within one minute. The management device continuously captures this change. During the collection process, the management device adds a timestamp and node identifier to each power data point, such as "2024-05-20 14:32:10 - Node S1-001 - Operating Power 285W", and stores it in the time-series database for later use in calculating the average power.

[0052] The acquisition of inlet and outlet air temperatures relies on temperature sensors controlled by the server's BMC module. Each server is designed with high-precision NTC temperature sensors deployed at the inlet (usually located on the front panel of the server) and outlet (usually located on the rear panel of the server) to accurately sense changes in inlet and outlet air temperatures.

[0053] The management equipment sends acquisition commands to the temperature sensors via BMC software. The acquisition frequency is consistent with the operating power, i.e., once every 5 seconds, to ensure the time synchronization of temperature and power data. For example, the inlet temperature sensors of node S1-001 are deployed on the left, right, and middle sides of the front panel. During acquisition, the average value of the three sensors is taken as the inlet air temperature of this node to avoid temperature measurement deviations caused by uneven local airflow. The outlet temperature sensors are deployed in the cooling fan outlet area of ​​the rear panel, and the multi-sensor averaging method is also used to improve measurement accuracy.

[0054] Assuming at a certain data acquisition moment, the temperatures measured by the sensors on the left, middle, and right sides of node S1-001 are 23.5℃, 23.7℃, and 23.6℃ respectively, then the inlet air temperature of this node is calculated as (23.5 + 23.7 + 23.6) / 3 = 23.6℃. The temperatures measured by the three sensors at the outlet are 38.2℃, 38.5℃, and 38.3℃ respectively, so the outlet air temperature is (38.2 + 38.5 + 38.3) / 3 = 38.3℃. After acquiring the inlet and outlet air temperatures, the management device immediately calculates the temperature difference Δt, i.e., Δt = outlet air temperature - inlet air temperature. In the example above, Δt = 38.3 - 23.6 = 14.7℃. This temperature difference data is then associated with and stored with the corresponding power data to provide a basis for subsequent airflow calculations.

[0055] During parameter acquisition, the management equipment also needs to establish a data verification mechanism to ensure that the acquired heat dissipation parameters are true and valid. For operating power data, if a acquired value exceeds the rated power range of the node (e.g., the rated power of node S1-001 is 350W, but 420W power data is acquired) or deviates from the power value at an adjacent acquisition time by more than 50% (e.g., the power was 90W at the previous moment, and suddenly 300W is acquired at the current moment, with no load change record), then the data is determined to be abnormal data, discarded, and the average of the valid power data at the previous moment and the next moment is used to complete it; for temperature data, if the acquired inlet or outlet air temperature exceeds the sensor's measurement range (-40℃ to 125℃), or the temperature difference Δt is negative (i.e., the outlet air temperature is lower than the inlet air temperature, which does not conform to the heat dissipation logic), then the data is determined to be abnormal, and the average of the adjacent valid data is used to complete it. For example, if node S2-034 collects inlet air temperature of 18.5℃ and outlet air temperature of 17.2℃ at a certain moment, with Δt = -1.3℃, the management device determines this set of temperature data to be abnormal. Querying the previous valid data, the inlet air temperature is found to be 19.2℃ and outlet air temperature 28.5℃, and the next valid data is found to be 19.3℃ and outlet air temperature 28.7℃. Therefore, the completed inlet air temperature for the current moment is (19.2 + 19.3) / 2 = 19.25℃, and the outlet air temperature is (28.5 + 28.7) / 2 = 28.6℃, with Δt = 28.6 - 19.25 = 9.35℃. This data verification mechanism effectively avoids the impact of abnormal data caused by sensor failure, communication interference, and other factors on subsequent processes, ensuring data quality.

[0056] Based on the specific heat capacity formula and the energy balance relationship of server heat dissipation, non-intrusive airflow calculation is achieved through software modeling. No new hardware is required; it relies entirely on existing sensor data. For air-cooled servers, all heat generated during operation is carried away by the cooling air; therefore, the heat generated by the server is equal to the heat carried away by the air. The basic specific heat capacity formula is Q=cmΔt, where Q represents heat (unit: joules J), c represents the specific heat capacity of the substance (unit: joules per kilogram per degree Celsius J / kg·℃), m represents the mass of the substance (unit: kilograms kg), and Δt represents the temperature change (unit:℃). In server heat dissipation scenarios, Q is also equal to the heat generated by the server over a certain period of time, i.e., Q=P*T (P is the average power of the server, unit: watts W; T is time, unit: seconds S); the mass of the air is m=ρV (ρ is the air density, unit: kilograms per cubic meter kg / m³; V is the air volume, unit: cubic meters m³). Substituting the above relationship into the specific heat capacity formula, we can derive the formula for calculating airflow: V / T = P / (cρΔt), where V / T is the airflow per unit time (unit: m³ / s), which is the required airflow parameter for the node.

[0057] In the specific implementation process, the management equipment needs to first clarify the value standards of each constant in the formula to ensure the consistency and accuracy of the calculation. Among them, the specific heat capacity of air, c, is taken as 1005 J / kg・℃ in the temperature range of 0-60℃ (the normal temperature range of data center computer rooms). This value is an industry-recognized standard constant and has been pre-stored in the calculation model of the management equipment. The air density ρ is related to the air temperature. Since the air temperature at the server inlet has been collected by sensors, the management equipment can look up the preset temperature-density correspondence table based on the air temperature at the inlet to obtain the corresponding air density ρ. The preset temperature-density correspondence table is based on the physical properties of air under standard atmospheric pressure and covers the temperature range that may occur in data centers. The specific correspondences are as follows: 1.205 kg / m³ at 20℃, 1.185 kg / m³ at 25℃, 1.165 kg / m³ at 30℃, 1.146 kg / m³ at 35℃, 1.128 kg / m³ at 40℃, 1.11 kg / m³ at 45℃, 1.093 kg / m³ at 50℃, 1.076 kg / m³ at 55℃, and 1.06 kg / m³ at 60℃. If the collected inlet air temperature is not at the corresponding temperature node in the table (e.g., 22.3℃), the management device uses linear interpolation to calculate the corresponding air density. For example, 22.3℃ is between 20℃ (1.205 kg / m³) and 25℃ (1.185 kg / m³), with a temperature difference of 2.3℃. The total temperature range is 5℃, and the density difference is 0.02 kg / m³. Therefore, the interpolated air density ρ = 1.205 - (2.3 / 5) * 0.02 = 1.205 - 0.0092 = 1.1958 kg / m³, retaining four decimal places to ensure calculation accuracy.

[0058] To calculate the server's average power P, the management device uses a sliding time window method. It calculates the average value over a period of time based on the collected real-time power data to avoid deviations in airflow parameter calculations caused by instantaneous power fluctuations. The length of the sliding time window can be configured by the operations and maintenance personnel, with a default setting of 1 minute. This means that the average power of the node is calculated every 1 minute, and the window contains power data from 12 collection points (collection frequency of 5 seconds / time). For example, node S1-001 collected 12 power data points within the sliding window from 14:30:00 to 14:31:00: 285W, 288W, 290W, 287W, 292W, 295W, 293W, 289W, 291W, 286W, 294W, and 290W. The management device calculated the arithmetic mean of these 12 data points to obtain the average power P within the window: P = (285+288+290+287+292+295+293+289+291+286+294+290) / 12 = 3450 / 12 = 287.5W. If abnormal power data exists within the sliding window (identified and supplemented through the data verification mechanism described above), the supplemented valid data is used for averaging to ensure the accuracy of the average power.

[0059] The temperature difference Δt is the average value of the difference between the inlet and outlet air temperatures within the sliding time window, consistent with the calculation window of the average power P, ensuring their temporal correlation. For example, for node S1-001, within the aforementioned 1-minute sliding window, the temperature differences Δt at each sampling moment are: 14.7℃, 14.8℃, 15.0℃, 14.9℃, 15.1℃, 15.2℃, 15.0℃, 14.8℃, 14.9℃, 14.7℃, 15.3℃, and 15.1℃. Therefore, the average temperature difference Δt = (14.7 + 14.8 + 15.0 + 14.9 + 15.1 + 15.2 + 15.0 + 14.8 + 14.9 + 14.7 + 15.3 + 15.1) / 12 = 179.5 / 12 ≈ 14.96℃, rounded to two decimal places for subsequent calculations.

[0060] After obtaining the average power P, specific heat capacity of air c, air density ρ, and average temperature difference Δt, the management equipment substitutes these parameters into the airflow calculation formula V / T = P / (cρΔt) to calculate the airflow parameters of the node. Taking node S1-001 as an example, given P=287.5W, c=1005 J / kg・℃, ρ is calculated to be 1.194 kg / m³ based on the average inlet air temperature of 23.8℃ in this window, and Δt≈14.96℃, then the airflow parameter V / T=287.5 / (1005×1.194×14.96). First, calculate the denominator: 1005 × 1.194 ≈ 1199.97, 1199.97 × 14.96 ≈ 17951.55. Then calculate the ratio of the numerator to the denominator: 287.5 / 17951.55 ≈ 0.0160 m³ / s. Converting to the commonly used unit m³ / h, it is 0.0160 × 3600 ≈ 57.6 m³ / h. That is, the airflow parameter of node S1-001 is 57.6 m³ / h, which represents the minimum cooling airflow required for this node to maintain normal heat dissipation.

[0061] To further improve the accuracy of airflow parameters, anomaly correction is performed on the calculated airflow data, using a sliding window method to remove glitch data. The sliding window length is set to 1 hour, containing 60 airflow data points for 1-minute calculation cycles. The management device calculates the mean μ and standard deviation σ of all airflow data within the window and sets an anomaly threshold of μ ± 2σ. If any airflow data point exceeds this threshold, it is identified as glitch data (i.e., anomalies caused by sudden load changes, temporary sensor fluctuations, etc.). For example, for node S2-056, the mean μ of the 60 airflow data points within the 1-hour sliding window is 62.3 m³ / h, and the standard deviation σ is 3.5 m³ / h. The threshold range is 62.3 ± 7 = 55.3 m³ / h to 69.3 m³ / h. If the airflow data at any given moment is 75.8 m³ / h, exceeding the upper threshold, it is identified as glitch data. For abnormal spike data, the management equipment will not use it directly, but will correct it by averaging the two adjacent valid air volume data before and after the abnormal data. For example, if the air volume before the abnormal data is 68.5 m³ / h and the air volume after the abnormal data is 67.9 m³ / h, then the corrected air volume data is (68.5+67.9) / 2=68.2 m³ / h, to ensure the continuity and stability of the time series data of air volume parameters.

[0062] After calculating the airflow parameters for a single node, the airflow data for all nodes is categorized and stored according to their respective racks, cabinets, and precision air conditioning control areas. A correlation is established between airflow parameters and node locations and the objects being cooled, for example, "Node S1-001 - Rack 1-3 - Precision Air Conditioner AC-01 - Airflow parameter 57.6 m³ / h" and "Node S2-056 - Rack 2-8 - Precision Air Conditioner AC-01 - Airflow parameter 62.3 m³ / h". This provides data support for subsequent statistics on the airflow requirements of the objects being cooled and for adjusting the airflow of the cooling equipment. Simultaneously, the airflow parameters for each node are updated in real time, with an update frequency consistent with the 1-minute calculation cycle, ensuring that the airflow data reflects changes in the node's cooling requirements in real time.

[0063] The heat dissipation object is a set of nodes consisting of one or more adjacent nodes in the server cluster. Specifically, it corresponds to all nodes within the control area of ​​the target heat dissipation device (precision air conditioner). Its air volume requirement is the sum of the air volume parameters of all nodes in the set. The management device dynamically adjusts the operating parameters of the heat dissipation device by comparing the air volume requirement of the heat dissipation object with the current output air volume of the target heat dissipation device, so as to achieve a precise match between the cooling capacity and the actual demand.

[0064] In practical implementation, the management equipment first needs to calculate the airflow demand of the target heat dissipation device corresponding to the heat dissipation object. Since the heat dissipation object consists of multiple adjacent nodes, and the airflow parameters of each node are calculated and stored in real time, the management equipment will periodically summarize the current airflow parameters of all nodes within the heat dissipation object according to the preset correspondence between precision air conditioners and heat dissipation objects. The summarization cycle is consistent with the airflow parameter update cycle, that is, it is summarized once every 1 minute. The summarization method is arithmetic summation, that is, the total airflow demand Q of the heat dissipation object = Σ (airflow parameter of node 1 + airflow parameter of node 2 + ... + airflow parameter of node n), where n is the total number of nodes within the heat dissipation object.

[0065] For example, the heat dissipation target of precision air conditioner AC-01 is all nodes in the first and second rows of racks, which includes a total of 160 servers. At 14:31:00, the management equipment summarizes the current air volume parameters of these 160 nodes, where node S1-001 is 57.6 m³ / h, node S1-002 is 61.2 m³ / h, ..., node S2-080 is 59.8 m³ / h. By summing, we get Q_total = 57.6 + 61.2 + ... + 59.8 = 9480 m³ / h, that is, the current air volume requirement of this heat dissipation target is 9480 m³ / h.

[0066] To ensure the accuracy of the aggregated airflow demand, the management device verifies the online status of nodes during the aggregation process. If a node is unable to collect heat dissipation parameters and calculate airflow parameters due to faults, offline status, or other reasons, the management device will use the average airflow parameter of that node for the same historical period as a substitute value for aggregation. For example, if node S1-045 is offline due to a hardware fault and cannot obtain real-time airflow data, the management device will find that the average airflow parameter of this node for the same time period (14:30-14:31) over the past 7 days is 58.5 m³ / h. In this aggregation, 58.5 m³ / h will be used as the airflow parameter for this node to avoid significant deviations in the aggregated airflow demand value due to the offline status of individual nodes. Meanwhile, the management equipment will verify the rationality of the aggregated air volume demand, calculate the mean μ_total and standard deviation σ_total of the air volume parameters of all nodes within the heat dissipation object. If the aggregated Q_total exceeds the range of n×(μ_total ± σ_total), the aggregated result is determined to be abnormal. The management equipment will re-check the air volume parameters of each node, eliminate abnormal node data, and re-aggregate to ensure the reliability of the air volume demand data.

[0067] After obtaining the airflow requirements of the target cooling device, the current output airflow and rated airflow parameters of the target cooling equipment are acquired. The control system of the target cooling equipment (precision air conditioner) has a communication interface (such as RS485 or Ethernet interface). The management device establishes a connection with the control system through this interface to read the operating parameters of the precision air conditioner in real time, including the current output airflow, rated airflow, and operating status. For example, the management device communicates with the control system of precision air conditioner AC-01 and obtains that the current output airflow of the air conditioner is 8800 m³ / h, the rated airflow range is 6000-12000 m³ / h, and the operating status is normal. At the same time, the management device queries the preset elasticity coefficient parameter. The elasticity coefficient is used to adjust the matching accuracy of airflow supply and demand, avoiding frequent adjustments of the air conditioner due to small fluctuations in airflow demand. The value range of the elasticity coefficient is 0.9-1.1, which can be configured by maintenance personnel according to the cooling requirements of the computer room. The default value is 0.95. The role of the elasticity coefficient is to set a reasonable range for air volume supply and demand. When the current output air volume of the target heat dissipation equipment is within the range of (air volume demand × lower limit of elasticity coefficient) to (air volume demand × upper limit of elasticity coefficient), it is determined that the cooling capacity matches the demand and no adjustment is required; if it exceeds this range, air volume adjustment is required.

[0068] The management equipment determines whether adjustment is needed and the direction and magnitude of adjustment by comparing the current output airflow of the target heat dissipation equipment with the airflow demand and elasticity coefficient of the heat dissipation object. The specific judgment logic is as follows: First, calculate the lower limit threshold Q_lower = Q_total × (1 - elasticity coefficient) and the upper limit threshold Q_upper = Q_total × (1 + elasticity coefficient); if the current output airflow Q_current...<q 下，则判定供冷不足，需要增加输出风量；若当前输出风量 q 当前>If Q_up, it is determined that there is excess cooling and the output airflow needs to be reduced; if Q_down ≤ Q_current ≤ Q_up, it is determined that supply and demand are matched, and the current output airflow remains unchanged. For example, the total airflow requirement for the cooling object corresponding to the precision air conditioner AC-01 is Q_total = 9480 m³ / h, and the elasticity coefficient is 0.95. Then Q_down = 9480 × (1 - 0.05) = 9480 × 0.95 = 9006 m³ / h, Q_up = 9480 × (1 + 0.05) = 9480 × 1.05 = 9954 m³ / h, while the current output airflow Q_current = 8800 m³ / h.

[0069] The calculation of the airflow adjustment range adopts a stepped adjustment strategy to avoid sudden changes in airflow organization in the computer room caused by a large one-time adjustment, which would affect the server's heat dissipation stability. The stepped adjustment strategy determines the adjustment range based on the difference between the current output airflow and the threshold, specifically set as follows: if the difference (Q_down - Q_current) or (Q_current - Q_up) ≤ 5% × Q_total, the adjustment range is 10% × rated airflow; if 5% × Q_total < difference ≤ 10% × Q_total, the adjustment range is 15% × rated airflow; if the difference > 10% × Q_total, the adjustment range is 20% × rated airflow. Taking the precision air conditioner AC-01 as an example, Qtotal = 9480 m³ / h, Qcurrent = 8800 m³ / h, the difference = 9006 - 8800 = 206 m³ / h, 5% × Qtotal = 474 m³ / h, 206 m³ / h ≤ 474 m³ / h, therefore the adjustment range is 10% × rated air volume. Given that the rated air volume of this air conditioner is 12000 m³ / h, the adjustment range = 12000 × 10% = 1200 m³ / h. The management equipment calculates the target output air volume = current output air volume + adjustment range = 8800 + 1200 = 10000 m³ / h.

[0070] ​The management device sends an airflow adjustment command to the control system of the precision air conditioner AC-01 via a communication interface. The command specifies a target output airflow of 10,000 m³ / h. Upon receiving the command, the control system adjusts the output airflow by regulating the speed of the fan inside the air conditioner. During the adjustment process, the control system provides real-time feedback on the changes in the current output airflow, and the management device continuously monitors the adjustment progress. The adjustment is considered complete when the actual output airflow of the air conditioner reaches the target output airflow and operates stably for 3 minutes. If any abnormal operation of the air conditioner occurs during the adjustment process (such as fan failure or insufficient air pressure), the control system sends an alarm message to the management device. The management device immediately stops the adjustment operation, sends an alarm notification to the maintenance personnel, and records the abnormality for subsequent troubleshooting.

[0071] In addition to dynamically adjusting the airflow demand based on the cooling target's requirements, the management device also supports optimized adjustments based on the evenness of node airflow distribution, further improving the overall cooling efficiency of the server cluster. Due to differences in load and hardware configuration among different nodes, their airflow demands will vary. This may result in concentrated airflow demands in some areas within the same cooling target, while other areas have lower airflow demands, leading to uneven cooling air distribution. The management device analyzes the airflow parameter distribution of each node within the cooling target, identifies nodes with abnormal airflow demands (i.e., nodes with airflow parameters significantly higher or lower than the average airflow parameter within the cooling target), and provides optimization suggestions for node locations to maintenance personnel. By adjusting the deployment positions of these nodes, a more even distribution of airflow demand within the cooling target is achieved, thereby improving the cooling efficiency of the cooling equipment.

[0072] For example, the heat dissipation target of precision air conditioner AC-08 is 160 nodes in rows 7-8 of the rack. The management equipment calculates the average air volume parameter μ = 63.5 m³ / h and the standard deviation σ = 4.2 m³ / h by statistically analyzing the air volume parameters of all nodes within this heat dissipation target. The criteria for judging abnormal nodes are set as air volume parameter > μ + 2σ (i.e., > 63.5 + 8.4 = 71.9 m³ / h) or < μ - 2σ (i.e., < 63.5 - 8.4 = 55.1 m³ / h). After analysis, it was found that the air volume parameter of node S7-023 is 75.6 m³ / h (above the threshold), and the air volume parameter of node S8-076 is 53.2 m³ / h (below the threshold). Moreover, these two nodes are located at opposite ends of the heat dissipation target. The cooling air supply in the area where node S7-023 is located is relatively tight, while the cooling air supply in the area where node S8-076 is located is relatively sufficient. The management equipment sent optimization suggestions to the operations and maintenance personnel through the visual interface of the data center management software: "It is recommended to swap the deployment locations of node S7-023 (airflow 75.6 m³ / h) and node S8-076 (airflow 53.2 m³ / h) to balance the distribution of airflow demand in the area." After the operations and maintenance personnel completed the node location adjustment according to the suggestion, the management equipment re-collected and calculated the airflow parameters of the two nodes and the airflow distribution of the heat dissipation objects. The adjusted airflow parameter of node S7-023 (formerly S8-076) was 54.1 m³ / h, and the airflow parameter of node S8-076 (formerly S7-023) was 74.8 m³ / h. The airflow demand distribution in the area was more balanced, the heat dissipation efficiency of the precision air conditioner AC-08 improved by 8%, and the average inlet air temperature of the servers in the area decreased by 1.2℃, effectively improving the local hotspot problem.

[0073] Throughout the entire heat dissipation equipment management process, the management equipment also possesses comprehensive monitoring and logging functions. It monitors the heat dissipation parameters and airflow parameters of each node in real time, as well as the operating status and output airflow of each heat dissipation device. All operation records, parameter changes, and adjustment results are stored in the log database, with configurable log data retention time (default 90 days). Maintenance personnel can view the system's operating status in real time through the data center management software's monitoring interface, including temperature, power, and airflow parameter curves for each node, the output airflow of each precision air conditioner, and adjustment records. It also supports data statistics and analysis by node, rack, and precision air conditioner, generating daily, weekly, and monthly reports to provide data support for optimizing data center thermal management. For example, maintenance personnel might find through a monthly report that the average airflow parameter of the nodes in the third rack is consistently higher than other racks, and the corresponding precision air conditioner AC-03 is being adjusted more frequently. This indicates that the server load in that rack is too high, and optimization can be achieved by adjusting service deployment and adding heat dissipation equipment to further improve the overall heat dissipation performance of the data center.

[0074] Furthermore, this heat dissipation equipment management method boasts excellent scalability and compatibility, adapting to server clusters of varying sizes and different types of heat dissipation equipment. For newly added server nodes, simply configuring the node location and BMC communication within the management device automatically brings them into the management scope, collecting heat dissipation parameters and calculating airflow requirements. For newly added precision air conditioners, configuring their correspondence with the heat dissipation target enables coordinated airflow adjustment. Simultaneously, this method supports integration with existing data center operation and maintenance management platforms and energy management systems, achieving data sharing and coordinated control. For instance, when the energy management system detects excessive energy consumption in the computer room, this method can adjust the operating parameters of the precision air conditioners, prioritizing the heat dissipation needs of core business nodes while reducing the cooling load on non-core nodes, thus achieving a balance between energy consumption and heat dissipation reliability.

[0075] In one implementation, such as Figure 2 This specification also provides a heat dissipation equipment management device, applied to a management device that manages the heat dissipation equipment of a server cluster. The server cluster includes several nodes. The device includes: a first module for acquiring the heat dissipation parameters of each node, including operating power, inlet air temperature, and outlet air temperature; a second module for acquiring the airflow parameters of each node based on the heat dissipation parameters of each node; and a third module for adjusting the output airflow of the target heat dissipation equipment according to the airflow requirements of the target heat dissipation equipment, wherein the heat dissipation equipment includes a set of nodes consisting of one or more adjacent nodes in the server cluster, and the airflow requirements of the heat dissipation equipment are associated with the airflow parameters of each node in the node set.

[0076] In one implementation, obtaining the airflow parameters of each node based on the heat dissipation parameters of each node includes: selecting a node in the server cluster as the target node; obtaining the inlet and outlet temperature difference based on the inlet and outlet air temperatures of the target node; and using the quotient obtained by dividing the operating power by the product of the air specific heat capacity, air density, and inlet and outlet temperature difference as the airflow parameters of the target node; and iterating through each node in the server cluster, repeating the above steps to obtain the airflow parameters of each node.

[0077] In one embodiment, the heat dissipation device includes a precision air conditioner in the physical space where the server cluster is located, the heat dissipation object includes a target rack corresponding to the precision air conditioner, and the target rack includes multiple nodes; the step of adjusting the output air volume of the target heat dissipation device according to the air volume requirement of the heat dissipation object corresponding to the target heat dissipation device includes: accumulating the air volume parameters of each node included in the target rack to obtain the total air volume parameter corresponding to the target rack, and adjusting the output air volume of the precision air conditioner according to the total air volume parameter.

[0078] In one embodiment, a fourth module is further included, used to determine a special node within the heat dissipation object and prompt adjustment of the physical position of the special node, wherein the ratio of the airflow parameter of the special node to the average airflow parameter of each node within the heat dissipation object is greater than a first threshold or less than a second threshold, wherein the second threshold is less than the first threshold.

[0079] The implementation methods of the apparatus are the same as or similar to the corresponding implementation methods, and will not be described again here.

[0080] In one embodiment, this specification provides an electronic device including a processor and a readable storage medium storing machine-executable instructions executable by the processor. The processor executes the machine-executable instructions to implement the aforementioned heat dissipation device management method. From a hardware perspective, a hardware architecture diagram can be found... Figure 3 As shown.

[0081] In one embodiment, this specification provides a readable storage medium storing machine-executable instructions that, when invoked and executed by a processor, cause the processor to implement the aforementioned heat dissipation device management method.

[0082] Here, a readable storage medium can be any electronic, magnetic, optical, or other physical storage device that can contain or store information, such as executable instructions, data, etc. For example, a readable storage medium can be: RAM (Random Access Memory), volatile memory, non-volatile memory, flash memory, storage drives (such as hard disk drives), solid-state drives, any type of storage disk (such as optical discs, DVDs, etc.), or similar storage media, or combinations thereof.

[0083] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, which can take the form of a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email sending and receiving device, game console, tablet computer, wearable device, or any combination of these devices.

[0084] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware.

[0085] Those skilled in the art will understand that embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, this specification can take the form of a completely hardware implementation, a completely software implementation, or an implementation combining software and hardware aspects. Furthermore, embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0086] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments thereof. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0087] Furthermore, these computer program instructions can also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in the process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0088] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0089] Those skilled in the art will understand that embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, this specification can take the form of a completely hardware implementation, a completely software implementation, or an implementation combining software and hardware aspects. Furthermore, this specification can take the form of a computer program product implemented on one or more computer-usable storage media (which may include, but are not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] The above description is merely an embodiment of this specification and is not intended to limit this specification. Various modifications and variations can be made to this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of the claims of this specification.< / q>

Claims

1. A method for managing heat dissipation equipment, characterized in that, The method, applied to a management device that manages the cooling equipment of a server cluster comprising several nodes, includes: The heat dissipation parameters of each node are obtained separately, including operating power, inlet air temperature, and outlet air temperature. Based on the heat dissipation parameters of each node, obtain the airflow parameters of each node; Based on the airflow requirements of the target heat dissipation device, the output airflow of the target heat dissipation device is adjusted. The target heat dissipation device includes a set of nodes consisting of one or more adjacent nodes in a server cluster. The airflow requirements of the target heat dissipation device are related to the airflow parameters of each node in the set of nodes.

2. The method according to claim 1, characterized in that, The process of obtaining the airflow parameters for each node based on its heat dissipation parameters includes: Select a node in a server cluster as the target node. Obtain the inlet and outlet temperature difference based on the inlet and outlet air temperatures of the target node. Divide the operating power by the product of the air specific heat capacity, air density, and inlet and outlet temperature difference, and use the quotient as the air volume parameter of the target node. Iterate through each node of the server cluster, repeating the above steps to obtain the airflow parameters for each node.

3. The method according to claim 1, characterized in that, The heat dissipation equipment includes a precision air conditioner in the physical space where the server cluster is located, and the heat dissipation object includes the target rack corresponding to the precision air conditioner, and the target rack includes multiple nodes; The step of adjusting the output airflow of the target heat dissipation device according to the airflow requirement of the object to which the target heat dissipation device is located includes: The total air volume parameter of the target rack is obtained by summing the air volume parameters of each node included in the target rack, and the output air volume of the precision air conditioner is adjusted according to the total air volume parameter.

4. The method according to claim 1, characterized in that, Also includes: Identify a special node within the heat dissipation object and prompt for adjustment of the physical position of the special node. The ratio of the airflow parameter of the special node to the average airflow parameter of all nodes within the heat dissipation object is greater than a first threshold or less than a second threshold, wherein the second threshold is less than the first threshold.

5. A heat dissipation equipment management device, characterized in that, An apparatus for managing the cooling systems of a server cluster, the server cluster comprising several nodes, the apparatus comprising: The first module is used to obtain the heat dissipation parameters of each node, including operating power, inlet air temperature, and outlet air temperature. The second module is used to obtain the airflow parameters of each node based on the heat dissipation parameters of each node. The third module is used to adjust the output airflow of the target heat dissipation device according to the airflow requirements of the heat dissipation object corresponding to the target heat dissipation device. The heat dissipation object includes a set of nodes consisting of one or more adjacent nodes in the server cluster. The airflow requirements of the heat dissipation object are related to the airflow parameters of each node in the set of nodes.

6. The apparatus according to claim 5, characterized in that, The process of obtaining the airflow parameters for each node based on its heat dissipation parameters includes: Select a node in a server cluster as the target node. Obtain the inlet and outlet temperature difference based on the inlet and outlet air temperatures of the target node. Divide the operating power by the product of the air specific heat capacity, air density, and inlet and outlet temperature difference, and use the quotient as the air volume parameter of the target node. Iterate through each node of the server cluster, repeating the above steps to obtain the airflow parameters for each node.

7. The apparatus according to claim 5, characterized in that, The heat dissipation equipment includes a precision air conditioner in the physical space where the server cluster is located, and the heat dissipation object includes the target rack corresponding to the precision air conditioner, and the target rack includes multiple nodes; The step of adjusting the output airflow of the target heat dissipation device according to the airflow requirement of the object to which the target heat dissipation device is located includes: The total air volume parameter of the target rack is obtained by summing the air volume parameters of each node included in the target rack, and the output air volume of the precision air conditioner is adjusted according to the total air volume parameter.

8. The apparatus according to claim 5, characterized in that, Also includes: The fourth module is used to identify special nodes within the heat dissipation object and prompt adjustments to the physical location of the special nodes. The ratio of the airflow parameter of the special node to the average airflow parameter of all nodes within the heat dissipation object is greater than a first threshold or less than a second threshold, wherein the second threshold is less than the first threshold.

9. An electronic device, characterized in that, include: A processor and a readable storage medium storing machine-executable instructions that can be executed by the processor to implement the method of any one of claims 1-4.

10. A readable storage medium, characterized in that, The readable storage medium stores machine-executable instructions that, when invoked and executed by a processor, cause the processor to implement the method of any one of claims 1-4.

Images