WO2024255191A1 - 一种刀片式服务器功率控制系统、方法和刀片式服务器 - Google Patents

一种刀片式服务器功率控制系统、方法和刀片式服务器 Download PDF

Info

Publication number
WO2024255191A1
WO2024255191A1 PCT/CN2023/141186 CN2023141186W WO2024255191A1 WO 2024255191 A1 WO2024255191 A1 WO 2024255191A1 CN 2023141186 W CN2023141186 W CN 2023141186W WO 2024255191 A1 WO2024255191 A1 WO 2024255191A1
Authority
WO
WIPO (PCT)
Prior art keywords
power consumption
blade
power
liquid cooling
power supply
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2023/141186
Other languages
English (en)
French (fr)
Inventor
周磊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suzhou Metabrain Intelligent Technology Co Ltd
Original Assignee
Suzhou Metabrain Intelligent Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Suzhou Metabrain Intelligent Technology Co Ltd filed Critical Suzhou Metabrain Intelligent Technology Co Ltd
Priority to US19/115,756 priority Critical patent/US20260113901A1/en
Publication of WO2024255191A1 publication Critical patent/WO2024255191A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/266Arrangements to supply power to external peripherals either directly from the computer or under computer control, e.g. supply of power through the communication port, computer controlled power-strips
    • HELECTRICITY
    • H05ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05KPRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K7/00Constructional details common to different types of electric apparatus
    • H05K7/20Modifications to facilitate cooling, ventilating, or heating
    • H05K7/20709Modifications to facilitate cooling, ventilating, or heating for server racks or cabinets; for data centers, e.g. 19-inch computer racks
    • H05K7/20836Thermal management, e.g. server temperature control
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/16Constructional details or arrangements
    • G06F1/20Cooling means
    • G06F1/206Cooling means comprising thermal management
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/263Arrangements for using multiple switchable power supplies, e.g. battery and AC
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/30Means for acting in the event of power-supply failure or interruption, e.g. power-supply fluctuations
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/30Monitoring
    • G06F11/3058Monitoring arrangements for monitoring environmental properties or parameters of the computing system or of the computing system component, e.g. monitoring of power, currents, temperature, humidity, position, vibrations
    • G06F11/3062Monitoring arrangements for monitoring environmental properties or parameters of the computing system or of the computing system component, e.g. monitoring of power, currents, temperature, humidity, position, vibrations where the monitored property is the power consumption
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F15/00Digital computers in general; Data processing equipment in general
    • G06F15/16Combinations of two or more digital computers each having at least an arithmetic unit, a program unit and a register, e.g. for a simultaneous processing of several programs
    • G06F15/161Computing infrastructure, e.g. computer clusters, blade chassis or hardware partitioning
    • HELECTRICITY
    • H05ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05KPRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K7/00Constructional details common to different types of electric apparatus
    • H05K7/20Modifications to facilitate cooling, ventilating, or heating
    • H05K7/20218Modifications to facilitate cooling, ventilating, or heating using a liquid coolant without phase change in electronic enclosures
    • H05K7/20254Cold plates transferring heat from heat source to coolant
    • HELECTRICITY
    • H05ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05KPRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K7/00Constructional details common to different types of electric apparatus
    • H05K7/20Modifications to facilitate cooling, ventilating, or heating
    • H05K7/20218Modifications to facilitate cooling, ventilating, or heating using a liquid coolant without phase change in electronic enclosures
    • H05K7/20272Accessories for moving fluid, for expanding fluid, for connecting fluid conduits, for distributing fluid, for removing gas or for preventing leakage, e.g. pumps, tanks or manifolds
    • HELECTRICITY
    • H05ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05KPRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K7/00Constructional details common to different types of electric apparatus
    • H05K7/20Modifications to facilitate cooling, ventilating, or heating
    • H05K7/20218Modifications to facilitate cooling, ventilating, or heating using a liquid coolant without phase change in electronic enclosures
    • H05K7/20281Thermal management, e.g. liquid flow control
    • HELECTRICITY
    • H05ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05KPRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K7/00Constructional details common to different types of electric apparatus
    • H05K7/20Modifications to facilitate cooling, ventilating, or heating
    • H05K7/20709Modifications to facilitate cooling, ventilating, or heating for server racks or cabinets; for data centers, e.g. 19-inch computer racks
    • H05K7/20763Liquid cooling without phase change
    • H05K7/20781Liquid cooling without phase change within cabinets for removing heat from server blades
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2200/00Indexing scheme relating to G06F1/04 - G06F1/32
    • G06F2200/20Indexing scheme relating to G06F1/20
    • G06F2200/201Cooling arrangements using cooling fluid
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02DCLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • the present application relates to the field of servers, and in particular to a blade server power control system, method and blade server.
  • a blade server also known as a blade server, Blade Server refers to a server unit that can be plugged into a standard height rack chassis to achieve high availability and high density. Its main structure is a large main chassis, and many blade nodes can be plugged into the main chassis. Each blade node is actually a system motherboard. Blade servers save more space than rack servers. At the same time, the heat dissipation problem is more prominent, and large and powerful fans are often installed in the chassis to dissipate heat.
  • the traditional blade server has introduced cold plate liquid cooling technology (that is, using working fluid as the medium for intermediate heat transfer to transfer heat from the hot zone to a distant place for cooling)
  • this method has the following defects: there is no advantageous adaptation for multi-blade node server systems, the power consumption balance of the whole power supply unit (Power Supply Unit, referred to as PSU) is not achieved, and the power redundancy between multi-blade node servers is not achieved.
  • PSU Power Supply Unit
  • the problem of cold plate power consumption is not reasonably considered in the load control of the power supply unit, and therefore improvement is urgently needed.
  • a blade server power control system comprising:
  • a liquid cooling module for circulating a coolant within each cold plate
  • Multiple power supply units for supplying power to multiple blade nodes and liquid cooling modules
  • the whole machine management control module is connected to each blade node, liquid cooling module and each power supply unit for monitoring the power consumption of each blade node and liquid cooling module and distributing the power consumption evenly to multiple power supply units.
  • the whole machine management control module obtains the first power consumption by acquiring the current value and the voltage value of each blade node.
  • a first current sensor and a first voltage sensor are provided at the input end of the liquid cooling module, and the whole machine management control module obtains the second power consumption through data collected by the first current sensor and the first voltage sensor.
  • the whole machine management control module is further used to obtain a third power consumption of the whole machine operation and a fourth power consumption of the peripheral components of the blade server operation.
  • the whole machine management control module is further used to calculate the sum of all the first power consumption, the second power consumption, the third power consumption and the fourth power consumption to obtain the total power consumption, and use the ratio of the total power consumption to the number of power supply units currently in operation as Output power consumption of each power supply unit.
  • the whole machine management control module is further configured to recalculate the total power consumption and the output power consumption of each power supply unit in response to monitoring a first power consumption increase corresponding to a certain blade node.
  • the whole machine management control module is also used to respond to monitoring a first power consumption increase corresponding to a blade node, and then before recalculating the total power consumption and the output power consumption of each power supply unit, increase the second power consumption of the liquid cooling module based on the first power consumption increase year-on-year.
  • the whole machine management control module is further used to increase the cooling liquid flowing to the cold plate corresponding to a blade node through the liquid cooling module based on the increase in the second power consumption of the liquid cooling module.
  • the whole machine management control module is further configured to recalculate the total power consumption and the output power consumption of each power supply unit in response to monitoring that the first power consumption corresponding to a certain blade node is reduced.
  • the whole machine management control module is further used to:
  • the second power consumption of the liquid cooling module is reduced year-on-year based on the first power consumption reduction.
  • the whole machine management control module is further used to:
  • the cooling liquid flowing to the cold plate corresponding to a blade node is reduced through the liquid cooling module.
  • the whole machine management control module is further used to:
  • represents the coolant flow rate
  • n represents the number of nodes
  • Pn represents the first power consumption
  • PL represents the second power consumption
  • PC represents the third power consumption
  • PW represents the fourth power consumption
  • PFn represents the useful power consumption of the nth blade node for actual computing
  • represents the heat loss of the liquid cooling module and the cold plate module
  • represents the coolant density
  • t represents the control monitoring time of the preset liquid cooling module
  • ⁇ K represents the temperature change of the coolant compared to the most recent control
  • c represents the specific heat capacity of the coolant
  • the liquid cooling module is adjusted to output the coolant at the determined coolant flow rate.
  • the whole machine management control module is further used to:
  • the total power consumption is evenly distributed to the remaining power supply units that have not failed.
  • the multiple power supply units adopt a redundant power supply architecture, wherein the redundant power supply architecture includes at least one backup power supply unit.
  • the whole machine management control module is further used to:
  • a backup power supply unit In response to a power supply unit failure, a backup power supply unit is used to replace the failed power supply unit, and the total power consumption is evenly distributed to the remaining power supply units and the running backup power supply unit.
  • each blade node includes a baseboard management controller, and the baseboard management controller is used to obtain the current value and voltage value of the corresponding blade node and send them to the whole machine management control module.
  • each blade node includes a second current sensor and a second voltage sensor, and the whole machine management control module obtains the current value and the voltage value of the corresponding blade node through the second current sensor and the second voltage sensor.
  • the whole machine management control module is further used to:
  • a blade server power control method comprising:
  • the method further comprises:
  • the process In response to monitoring that the first power consumption corresponding to a certain blade node increases, the process returns to the step of obtaining the first power consumption of each blade node.
  • the method before returning to the step of obtaining the first power consumption of each blade node, the method further includes:
  • the second power consumption of the liquid cooling module is increased year-on-year based on the first power consumption increase.
  • the step of increasing the second power consumption of the liquid cooling module based on the first power consumption increase year-on-year further includes:
  • the cooling liquid flowing to the cold plate corresponding to a blade node is increased through the liquid cooling module.
  • the method further comprises:
  • the process In response to monitoring that the first power consumption corresponding to a certain blade node is reduced, the process returns to the step of obtaining the first power consumption of each blade node.
  • the method before returning to the step of obtaining the first power consumption of each blade node, the method further includes:
  • the second power consumption of the corresponding liquid cooling module is reduced year-on-year based on the first power consumption reduction amount.
  • the step of reducing the second power consumption of the corresponding liquid cooling module based on the first power consumption reduction amount year-on-year also includes:
  • the cooling liquid flowing to the cold plate corresponding to a certain blade node is reduced through the liquid cooling module.
  • the present application provides a blade server, and the blade server includes the above blade server power control system.
  • the above-mentioned blade server power control system has at least the following beneficial effects: advantageously adapting to the multi-blade nodes of the blade server, monitoring the power consumption of each blade node and liquid cooling module through the whole machine management control module, and evenly distributing the power consumption to multiple power supply units, thereby achieving balanced power consumption of the whole machine power supply unit, improving the stability and reliability of the blade server, and improving the cold plate liquid cooling and multi-node adaptability and regulation capabilities.
  • the present application also provides a blade server power control method and a blade server, which can also achieve the above-mentioned technical effects and will not be described in detail here.
  • FIG1 is a schematic diagram of the structure of a blade server power control system provided by an embodiment of the present application.
  • FIG2 is a flow chart of a blade server power control method provided by an embodiment of the present application.
  • FIG3 is a schematic diagram of the working principle of a whole machine management control module provided by another embodiment of the present application.
  • FIG. 4 is a schematic diagram of a whole-machine management control module dynamically regulating the power consumption balance of a PSU provided by another embodiment of the present application.
  • the present application provides a blade server power control system 100.
  • the system includes the following components:
  • the blade node 101 refers to a computing board used to provide computing power.
  • the blade node 101 is used to provide network, data computing, image processing, storage and other services.
  • the functions and configurations of different blade nodes may be the same or different.
  • Multiple blade nodes 101 are inserted into a standard height chassis as follows (1U, 2U, etc.).
  • Multiple blade nodes are used to implement a low-cost server platform of HAHD (High Availability High Density) to handle high-density computing tasks.
  • HAHD High Availability High Density
  • the number of blade nodes 101 can be set according to the business scenario or user needs. This application does not limit the specific number of blade nodes and is only used for example.
  • the cold plate 102 is filled with liquid cooling, that is, the working fluid is used as the medium for intermediate heat transmission to transfer heat from the hot zone to a distant place for cooling.
  • the cold plate 102 and the coolant of this application both adopt the existing cold plate liquid cooling technology.
  • a liquid cooling module 103 used to circulate cooling liquid in each cold plate 102;
  • the liquid cooling module 103 can realize the function of cooling the coolant, for example, reducing the coolant from a relatively high temperature to a pre-set relatively low temperature.
  • the liquid cooling module 103 is usually arranged outside the heat dissipation area and can drive the coolant to circulate between the heat dissipation area and the non-heat dissipation area.
  • a plurality of power supply units 104 used to supply power to the plurality of blade nodes 101 and the liquid cooling module 103;
  • the power supply unit 104 namely PSU, is used to provide power for the server. It converts high-voltage AC power into stable low-voltage DC power to supply various power-consuming components of the server, such as the motherboard, blade nodes, air-cooled heat dissipation devices, liquid cooling modules, etc.
  • the whole machine management control module 105 is connected to each blade node 101, the liquid cooling module 103 and each power supply unit for monitoring the power consumption of each blade node 101 and the liquid cooling module 103 and distributing the power consumption evenly to multiple power supply units. 104 up.
  • a blade server power control system of the present embodiment is advantageously adapted to the multiple blade nodes of the blade server, monitors the power consumption of each blade node and liquid cooling module through the whole machine management control module, and evenly distributes the power consumption to multiple power supply units, thereby achieving balanced power consumption of the whole machine power supply units, improving the stability and reliability of the blade server, and improving the cold plate liquid cooling and multi-node adaptability and regulation capabilities.
  • the whole machine management control module 105 obtains the first power consumption by acquiring the current value and the voltage value of each blade node 101 .
  • the input end of the liquid cooling module 103 is provided with a first current sensor and a first voltage sensor, and the whole machine management control module 105 obtains the second power consumption through data collected by the first current sensor and the first voltage sensor.
  • the whole machine management control module 105 is further used to obtain the third power consumption of the whole machine operation and the fourth power consumption of the peripheral components of the blade server operation.
  • peripheral components refer to components shared by blade nodes 101 in the blade server except for the liquid cooling module 103 and the whole machine management control module 105 mentioned above.
  • the peripheral components may be hard disks, hard disk backplanes, and fans shared by multiple blade nodes 101.
  • the corresponding fourth power consumption refers to the total power consumption of the hard disk, hard disk backplane, and fan.
  • the whole machine management control module 105 is further used to:
  • the ratio of the total power consumption to the number of power supply units currently in operation is taken as the output power consumption of each power supply unit.
  • the whole machine management control module 105 is further used to:
  • the whole machine management control module 105 is further used to:
  • the second power consumption of the liquid cooling module 103 is increased year-on-year based on the first power consumption increase.
  • the liquid cooling module 103 can make corresponding adjustments to take into account the increase in power consumption of the blade node, timely improve the cooling demand, solve the problem of untimely heat dissipation of the server, and significantly improve the timeliness and efficiency of heat dissipation.
  • the whole machine management control module 105 is further used to:
  • the cooling liquid flowing to the cold plate 102 corresponding to a certain blade node 101 is increased through the liquid cooling module 103 .
  • the corresponding blade nodes are quickly cooled down in a targeted manner, which can effectively avoid the problem of local high temperature or uneven heat dissipation of the server as a whole.
  • the whole machine management control module 105 is further used to:
  • the total power consumption and the output power consumption of each power supply unit are recalculated.
  • the whole machine management control module 105 is further used to:
  • the second power consumption of the liquid cooling module 103 is reduced year-on-year based on the first power consumption reduction.
  • the liquid cooling module 103 can make corresponding adjustments while taking into account the reduction of power consumption of the blade node, timely reduce the demand for cooling, avoid waste of resources, and improve heat dissipation efficiency.
  • the whole machine management control module 105 is further used to:
  • the cooling liquid flowing to the cold plate 102 corresponding to a certain blade node 101 is reduced through the liquid cooling module 103 .
  • the cooling capacity of the corresponding blade nodes is reduced in a targeted manner, thereby avoiding the problem of uneven heat dissipation.
  • the whole machine management control module 105 is further used to:
  • represents the coolant flow rate
  • n represents the number of nodes
  • Pn represents the first power consumption
  • PL represents the second power consumption
  • PC represents the third power consumption
  • PW represents the fourth power consumption
  • PFn represents the useful power consumption of the nth blade node for actual computing
  • represents the heat loss of the liquid cooling module and the cold plate module
  • represents the coolant density
  • t represents the control monitoring time of the preset liquid cooling module
  • ⁇ K represents the temperature change of the coolant compared to the most recent control
  • c represents the specific heat capacity of the coolant
  • the liquid cooling module is adjusted to output the coolant at the determined coolant flow rate.
  • the whole machine management control module 105 is further used to:
  • the total power consumption is evenly distributed to the remaining power supply units that have not failed.
  • the plurality of power supply units 104 adopt a redundant power supply architecture, wherein the redundant power supply architecture includes at least one backup power supply unit.
  • the whole machine management control module 105 is further used to:
  • a backup power supply unit In response to a power supply unit failure, a backup power supply unit is used to replace the failed power supply unit, and the total power consumption is evenly distributed to the remaining power supply units and the running backup power supply unit.
  • each blade node 101 includes a baseboard management controller, which is used to obtain the current value and voltage value of the corresponding blade node 101 and send them to the whole machine management control module 105 .
  • each blade node usually has a separate baseboard management controller
  • the baseboard management controller itself has the function of monitoring each device on the node. Current, voltage and other data are directly obtained from the baseboard management controller, avoiding the need to individually modify the blade nodes. This can be achieved using existing conventional blade servers, which helps save costs.
  • each blade node 101 includes a second current sensor and a second voltage sensor, and the whole machine management control module 105 obtains the current value and the voltage value of the corresponding blade node 101 through the second current sensor and the second voltage sensor.
  • the whole machine management control module 105 is further used to:
  • the output voltage and output current of each power supply unit are adjusted to the same value, so that the power consumption is evenly distributed to the multiple power supply units 104 .
  • the present application further provides a blade server power control method 200, specifically, the method includes the following steps:
  • Step 201 obtaining a first power consumption of each blade node
  • Step 202 obtaining a second power consumption of a liquid cooling module that provides heat dissipation for all blade nodes
  • Step 203 obtaining a third power consumption of the whole machine management control module and a fourth power consumption of the peripheral components of the blade server;
  • Step 204 calculating the total power consumption based on the first power consumption, the second power consumption, the third power consumption and the fourth power consumption;
  • Step 205 evenly distribute the total power consumption to multiple power supply units.
  • a blade server power control method of the present embodiment is advantageously adapted to the multiple blade nodes of the blade server, monitors the power consumption of each blade node and liquid cooling module through the whole machine management control module, and evenly distributes the power consumption to multiple power supply units, thereby achieving balanced power consumption of the whole machine power supply units, improving the stability and reliability of the blade server, and improving the cold plate liquid cooling and multi-node adaptability and regulation capabilities.
  • the method further comprises:
  • the process In response to monitoring that the first power consumption corresponding to a certain blade node increases, the process returns to the step of obtaining the first power consumption of each blade node.
  • returning to the step of obtaining the first power consumption of each blade node further includes:
  • the second power consumption of the liquid cooling module is increased year-on-year based on the first power consumption increase.
  • the step of increasing the second power consumption of the liquid cooling module based on the first power consumption increase year-on-year further includes:
  • the cooling liquid flowing to the cold plate corresponding to a blade node is increased through the liquid cooling module.
  • the method further comprises:
  • the process In response to monitoring that the first power consumption corresponding to a certain blade node is reduced, the process returns to the step of obtaining the first power consumption of each blade node.
  • returning to the step of obtaining the first power consumption of each blade node further includes:
  • the second power consumption of the corresponding liquid cooling module is reduced year-on-year based on the first power consumption reduction amount.
  • the step of reducing the second power consumption of the corresponding liquid cooling module based on the first power consumption reduction amount year-on-year also includes:
  • the cooling liquid flowing to the cold plate corresponding to a certain blade node is reduced through the liquid cooling module.
  • the present embodiment takes a blade server including n blade nodes as an example to explain the solution of the present application in detail, a blade server power control system
  • the power input of the blade server is jointly supported by n PSUs, and may also involve modules commonly used between blades, such as hard disks, hard disk backplanes, fans, cold plates and whole machine management control modules.
  • the whole machine management control module realizes the heat dissipation control of the cold plate, including controlling the liquid replenishment and the coolant flow rate management, and the power supply of the cold liquid module is also controlled by the whole machine management control module;
  • a multi-node blade server involves multiple PSU power supplies, and the power supply mode is to power on separately, and the whole machine management control module is used for unified management and power consumption balance control.
  • the power on the PSU can achieve balanced output, so as to ensure the normal operation of the server well, and when a single node has a burst of traffic, timely intervention in power consumption balance is performed to achieve flow balance, and ensure data reliability in the business operation of the server, as well as the reliability and life of the stable operation of the server itself.
  • power redundancy is set, ranging from N+1 to N+N redundancy.
  • N represents the number of PSUs required to support the normal operation of the system
  • the following numbers represent the number of PSUs that can be abnormally operated.
  • the server system and the whole machine management control module are also required to support the redundant design of the server support, and its implementation method cannot affect the power balance of the blade server, and when the redundancy takes effect, the whole machine management control module needs to quickly complete the power rebalancing of multiple PSUs.
  • the whole machine management control module can be used to detect the total power consumption required by the blade nodes and the liquid cooling module.
  • Detecting the power consumption of blade nodes is already a mature solution.
  • the total power consumption of each node is calculated through the sensor on the BMC chip (baseboard controller) of each node, and the relevant data is transmitted to the whole machine management control module through the signal line.
  • This application needs to increase the power consumption detection of the liquid cooling module.
  • the liquid cooling module it only needs to realize the cooling function of the coolant, so the sensor only needs to be equipped with a temperature sensor under standard conditions.
  • a current sensor is added to the input end of the entire liquid cooling module. Since its power supply voltage is constant (usually set at 12V DC), the power consumption of the liquid cooling module can be calculated by the whole machine management control module.
  • the above calculation method will result in a more accurate power consumption, and the whole machine management control module will also more accurately obtain the power consumption of each blade node and liquid cooling module.
  • the whole machine power consumption can be obtained through the total output current and input voltage. Since the power consumption output of the PSU is realized through the whole machine management control module, the whole machine management control module can directly obtain the relevant power consumption data in this case.
  • the BMC works normally, it will switch to the above-mentioned accurate power consumption acquisition state of the blade.
  • the whole machine management control module can also be used to distribute the power consumption to all working PSUs.
  • the output voltage and input voltage of each PSU are usually the same. In this case, to achieve balanced power consumption, you only need to balance the current.
  • the whole machine management control module After obtaining the value of the whole machine power consumption M, the whole machine management control module obtains the number of all PSUs currently in place N and distributes the current evenly to each PSU.
  • the power consumption that each PSU needs to bear is: M/N
  • the input current that needs to be borne is: (M/N)/input voltage of the computer room.
  • the input voltage of the computer room is determined according to the PSU support and the actual situation of the computer room. There are many possibilities such as 110V, 220V, 380V, etc.
  • the whole machine management control module can dynamically adjust the power consumption balance between PSUs.
  • the node's current sensor will immediately detect the current change on the demand loop.
  • the BMC current appears This information about the increase in demand is sent back to the whole machine management control module, which starts to increase the power supply, that is, gradually increase the input current and evenly distribute it to each PSU.
  • the above slow increase may still not meet the actual power demand of the node with increased power demand.
  • the current increase strategy needs to make corresponding changes to the whole machine input current according to the current increase value sent back by the BMC.
  • the current growth rate needs to be slightly higher than the required power consumption growth rate. When the current grows to the required current value, it will be called back.
  • the whole machine management control module can regulate the flow rate and fan speed of the liquid cooling module.
  • the cooling liquid module will synchronously increase the flow rate of the cooling liquid or the speed of the cooling fan.
  • the current of the cooling liquid module needs to change synchronously.
  • the current increased by the above-mentioned whole machine control management module needs to cover both the power consumption required by the node and the power consumption required to drive the heat dissipation work.
  • the increased current margin also needs to cover the current required for air cooling and liquid cooling heat dissipation regulation, and synchronously call back to the adapted current after a round of regulation is completed, so as to avoid power consumption waste and excessive current supply resulting in the need to increase the power consumption again in the future, resulting in insufficient overall power consumption, and avoid abnormalities in subsequent regulation.
  • the whole machine management control module dynamically adjusts the power consumption and current sharing.
  • the present embodiment provides a blade server power control system, which is advantageously adapted to a multi-node server system to achieve balanced power consumption of the entire PSU and power redundancy between multi-node servers to ensure high reliability and security of server products.
  • the present application also provides a blade server, which includes a standard machine chassis, a hard disk, a hard disk backplane, a fan, and a network connection module shared by multiple blade nodes.
  • the blade server also includes the blade server power control system of the above embodiments, and the system includes: multiple blade nodes; a cold plate that heat exchanges with each blade node, and the cold plate contains coolant; a liquid cooling module, which is used to circulate the coolant in each cold plate; multiple power supply units, which are used to power multiple blade nodes and liquid cooling modules; a whole machine management control module, which is communicatively connected to each blade node, the liquid cooling module and each power supply unit, and is used to monitor the power consumption of each blade node and the liquid cooling module, and evenly distribute the power consumption to multiple power supply units.
  • the whole machine management control module obtains the first power consumption by acquiring the current value and the voltage value of each blade node.
  • a first current sensor and a first voltage sensor are provided at the input end of the liquid cooling module, and the whole machine management control module obtains the second power consumption through data collected by the first current sensor and the first voltage sensor.
  • the whole machine management control module is further used to obtain a third power consumption of the whole machine operation and a fourth power consumption of the peripheral components of the blade server operation.
  • the whole machine management control module is further used to calculate the sum of all first power consumption, second power consumption, third power consumption and fourth power consumption to obtain the total power consumption, and use the ratio of the total power consumption to the number of power supply units currently in operation as the output power consumption of each power supply unit.
  • the whole machine management control module is further configured to recalculate the total power consumption and the output power consumption of each power supply unit in response to monitoring a first power consumption increase corresponding to a certain blade node.
  • the whole machine management control module is further configured to, in response to monitoring a first power consumption increase corresponding to a blade node, calculate a first power consumption increase based on a year-on-year increase in the first power consumption before recalculating the total power consumption and the output power consumption of each power supply unit. Add the second power consumption of the liquid cooling module.
  • the whole machine management control module is further used to increase the cooling liquid flowing to the cold plate corresponding to a blade node through the liquid cooling module based on the increase in the second power consumption of the liquid cooling module.
  • the whole machine management control module is further configured to recalculate the total power consumption and the output power consumption of each power supply unit in response to monitoring that the first power consumption corresponding to a certain blade node is reduced.
  • the whole machine management control module is further used to:
  • the second power consumption of the liquid cooling module is reduced year-on-year based on the first power consumption reduction.
  • the whole machine management control module is further used to:
  • the cooling liquid flowing to the cold plate corresponding to a blade node is reduced through the liquid cooling module.
  • the whole machine management control module is further used to:
  • represents the coolant flow rate
  • n represents the number of nodes
  • Pn represents the first power consumption
  • PL represents the second power consumption
  • PC represents the third power consumption
  • PW represents the fourth power consumption
  • PFn represents the useful power consumption of the nth blade node for actual computing
  • represents the heat loss of the liquid cooling module and the cold plate module
  • represents the coolant density
  • t represents the control monitoring time of the preset liquid cooling module
  • ⁇ K represents the temperature change of the coolant compared to the most recent control
  • c represents the specific heat capacity of the coolant
  • the liquid cooling module is adjusted to output the coolant at the determined coolant flow rate.
  • the whole machine management control module is further used to:
  • the total power consumption is evenly distributed to the remaining power supply units that have not failed.
  • the multiple power supply units adopt a redundant power supply architecture, wherein the redundant power supply architecture includes at least one backup power supply unit.
  • the whole machine management control module is further used to:
  • a backup power supply unit In response to a power supply unit failure, a backup power supply unit is used to replace the failed power supply unit, and the total power consumption is evenly distributed to the remaining power supply units and the running backup power supply unit.
  • each blade node includes a baseboard management controller, and the baseboard management controller is used to obtain the current value and voltage value of the corresponding blade node and send them to the whole machine management control module.
  • each blade node includes a second current sensor and a second voltage sensor, and the whole machine management control module obtains the current value and the voltage value of the corresponding blade node through the second current sensor and the second voltage sensor.
  • the whole machine management control module is further used to:
  • the output voltage and output current of each power supply unit are adjusted to the same value, thereby evenly distributing the power consumption to multiple power supply units.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Hardware Design (AREA)
  • Thermal Sciences (AREA)
  • Computing Systems (AREA)
  • Quality & Reliability (AREA)
  • Power Engineering (AREA)
  • Human Computer Interaction (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Power Sources (AREA)

Abstract

本申请涉及服务器领域,尤其涉及一种刀片式服务器功率控制系统、方法和刀片式服务器。所述系统包括:多个刀片节点;与每个刀片节点热交换的冷板,所述冷板内有冷却液;液冷模块,用于使所述冷却液在每个所述冷板内循环;多个电源单元,用于为所述多个刀片节点和所述液冷模块供电;整机管理控制模块,与每个刀片节点、所述液冷模块以及每个电源单元通信连接,用于监控每个刀片节点和所述液冷模块的功耗,并将所述功耗均衡分布到所述多个电源单元上。本申请的方案实现整机供电单元功耗均衡,提升刀片式服务器的稳定性和可靠性,提升冷板液冷及多节点适配性以及调控能力。

Description

一种刀片式服务器功率控制系统、方法和刀片式服务器
相关申请的交叉引用
本申请要求于2023年06月13日提交中国专利局,申请号为202310699361.9,申请名称为“一种刀片式服务器功率控制系统、方法和刀片式服务器”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及服务器领域,尤其涉及一种刀片式服务器功率控制系统、方法和刀片式服务器。
背景技术
刀片服务器(又称刀片式服务器,Blade Server)是指在标准高度的机架式机箱内可插装多个卡式的服务器单元,实现高可用和高密度。其主要结构为一大型主体机箱,主体机箱的内部可插上许多刀片节点,其中每一个刀片节点实际上就是一块系统主板。刀片服务器比机架式服务器更节省空间,同时,散热问题也更突出,往往要在机箱内装上大型强力风扇来散热。
目前,传统的刀片服务器虽然引入了冷板式液冷技术(即利用工作流体作为中间热量传输的媒介,将热量由热区传递到远处再进行冷却),但是此种方式存在以下缺陷:没有针对多刀片节点服务器系统进行优势性的适配,整机供电单元(Power Supply Unit,简称PSU)功耗均衡未实现,多刀片节点服务器之间的电源冗余未实现,此外在供电单元负载控制方面也未有合理考虑冷板功耗的问题,因而亟需改进。
发明内容
有鉴于此,有必要针对以上技术问题,提供一种刀片式服务器功率控制系统、方法和刀片式服务器。
根据本申请的第一方面,提供了一种刀片式服务器功率控制系统,系统包括:
多个刀片节点;
与每个刀片节点热交换的冷板,冷板内有冷却液;
液冷模块,用于使冷却液在每个冷板内循环;
多个电源单元,用于为多个刀片节点和液冷模块供电;
整机管理控制模块,与每个刀片节点、液冷模块以及每个电源单元通信连接,用于监控每个刀片节点和液冷模块的功耗,并将功耗均衡分布到多个电源单元上。
在一些实施例中,整机管理控制模块通过获取每个刀片节点的电流值和电压值以得到第一功耗。
在一些实施例中,液冷模块的输入端设置有第一电流传感器和第一电压传感器,整机管理控制模块通过第一电流传感器和第一电压传感器采集的数据得到第二功耗。
在一些实施例中,整机管理控制模块还用于获取自身工作的第三功耗以及刀片式服务器外围部件工作的第四功耗。
在一些实施例中,整机管理控制模块进一步用于计算所有的第一功耗、第二功耗、第三功耗和第四功耗的总和以得到总功耗,并将总功耗与当前在运行的电源单元数量的比值作为 每个电源单元的输出功耗。
在一些实施例中,整机管理控制模块还用于响应于监测到某一刀片节点对应的第一功耗增加,则重新计算总功耗以及每个电源单元的输出功耗。
在一些实施例中,整机管理控制模块还用于响应于监测到某一刀片节点对应的第一功耗增加,则在重新计算总功耗以及每个电源单元的输出功耗之前,基于第一功耗增加量同比增加液冷模块的第二功耗。
在一些实施例中,整机管理控制模块还用于基于液冷模块的第二功耗的增加量通过液冷模块增加流向某一刀片节点对应的冷板的冷却液。
在一些实施例中,整机管理控制模块还用于响应于监测到某一刀片节点对应的第一功耗减少,则重新计算总功耗以及每个电源单元的输出功耗。
在一些实施例中,整机管理控制模块还用于:
响应于监测到某一刀片节点对应的第一功耗减少,则在重新计算总功耗以及每个电源单元的输出功耗之前,基于第一功耗减少量同比减少液冷模块的第二功耗。
在一些实施例中,整机管理控制模块还用于:
基于液冷模块的第二功耗减少量通过液冷模块减少流向某一刀片节点对应的冷板的冷却液。
在一些实施例中,整机管理控制模块还用于:
根据下述公式确定液冷模块冷却液流速;
其中,μ表示冷却液流速,n表示节点数量,Pn表示第一功耗,PL表示第二功耗,PC表示第三功耗,PW表示第四功耗,PFn表示第n个刀片节点的用于实际运算的有用功功耗,σ表示液冷模块和冷板模块的散热损耗,ρ表示冷却液密度,t表示预设液冷模块的调控监控时间,ΔK表示相比于最近一次调控冷却液的温度变化,c表示冷却液的比热容;
将液冷模块调整为以所确定的冷却液流速输出冷却液。
在一些实施例中,整机管理控制模块还用于:
响应于某一电源单元故障,则将总功耗均衡分布到剩余未发生故障的电源单元上。
在一些实施例中,多个电源单元采用冗余供电架构,其中,冗余供电架构至少包括一个备份电源单元。
在一些实施例中,整机管理控制模块还用于:
响应于某一电源单元故障,则采用备份电源单元替换故障的电源单元,并将总功耗均衡分布到剩余电源单元和在运行的备份电源单元上。
在一些实施例中,每个刀片节点均包含基板管理控制器,基板管理控制器用于获取对应的刀片节点的电流值和电压值并发送给整机管理控制模块。
在一些实施例中,每个刀片节点均包含第二电流传感器和第二电压传感器,整机管理控制模块通过第二电流传感器和第二电压传感器获取对应的刀片节点的电流值和电压值。
在一些实施例中,整机管理控制模块进一步用于:
将每个电源单元的输出电压和输出电流均调整为相同值,从而将功耗均衡分布到多个电 源单元。
根据本申请的第二方面,提供了一种刀片式服务器功率控制方法,方法包括:
获取每个刀片节点的第一功耗;
获取为所有刀片节点提供散热的液冷模块的第二功耗;
获取整机管理控制模块工作的第三功耗以及刀片式服务器外围部件工作的第四功耗;
基于第一功耗、第二功耗、第三功耗和第四功耗计算总功耗;
将总功耗均衡分布到多个电源单元。
在一些实施例中,方法还包括:
响应于监测到某一刀片节点对应的第一功耗增加,则返回获取每个刀片节点的第一功耗的步骤。
在一些实施例中,返回获取每个刀片节点的第一功耗的步骤之前,还包括:
基于第一功耗增加量同比增加液冷模块的第二功耗。
在一些实施例中,基于第一功耗增加量同比增加液冷模块的第二功耗的步骤,还包括:
基于对应液冷模块的第二功耗的增加量通过液冷模块增加流向某一刀片节点对应的冷板的冷却液。
在一些实施例中,方法还包括:
响应于监测到某一刀片节点对应的第一功耗减少,则返回获取每个刀片节点的第一功耗的步骤。
在一些实施例中,返回获取每个刀片节点的第一功耗的步骤之前,还包括:
基于第一功耗减少量同比减少对应液冷模块的第二功耗。
在一些实施例中,基于第一功耗减少量同比减少对应液冷模块的第二功耗的步骤,还包括:
基于对应液冷模块的第二功耗减少量通过液冷模块减少流向某一刀片节点对应的冷板的冷却液。
根据本申请的第二方面,本申请提供了一种刀片式服务器,刀片式服务器包括以上的刀片式服务器功率控制系统。
上述一种刀片式服务器功率控制系统至少具备以下有益效果:针对刀片式服务器的多刀片节点进行优势性的适配,通过整机管理控制模块监控每个刀片节点和液冷模块的功耗,并将功耗均衡分布到多个供电单元上,实现整机供电单元功耗均衡,提升刀片式服务器的稳定性和可靠性,提升冷板液冷及多节点适配性以及调控能力。
此外,本申请还提供了一种片式服务器功率控制方法,以及一种刀片式服务器,同样能实现上述技术效果,这里不再赘述。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的实施例。
图1为本申请一个实施例提供的一种刀片式服务器功率控制系统的结构示意图;
图2为本申请一个实施例提供的一种刀片式服务器功率控制方法的流程图;
图3为本申请另一个实施例提供的整机管理控制模块工作原理示意图;
图4为本申请另一个实施例提供的整机管理控制模块动态调控PSU的功耗均衡的示意图。
【附图标记说明】
100:刀片式服务器功率控制系统;
101:刀片节点;
102:冷板;
103:液冷模块;
104:电源单元;
105:整机管理控制模块。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚明白,以下结合具体实施例,并参照附图,对本申请实施例进一步详细说明。
需要说明的是,本申请实施例中所有使用“第一”和“第二”的表述均是为了区分两个相同名称非相同的实体或者非相同的参量,可见“第一”“第二”仅为了表述的方便,不应理解为对本申请实施例的限定,后续实施例对此不再一一说明。
在一个实施例中,请参照图1所示,本申请提供了一种刀片式服务器功率控制系统100,具体来说,系统包括以下部件:
多个刀片节点101;
在本实施例中,刀片节点101是指用来提供计算能力的计算板,例如使用刀片节点101提供网络、数据运算、图像处理、存储等服务,在具体实施过程中不同刀片节点的功能及配置可以相同也可以不同。多个刀片节点101插设在标准高度的机箱如下(1U、2U等等),多个刀片节点用于实现HAHD(High Availability High Density,高可用高密度)的低成本服务器平台,以处理高密度计算任务。请继续参照图1,其示出了n个刀片节点101,其中n为大于等于二的正整数,在具体实施过程中,刀片节点101的数量可以依据业务场景或者用户需求设定,本申请不限制刀片节点的具体数量,仅用于举例说明。
与每个刀片节点101热交换的冷板102,冷板102内有冷却液(图中未示出);
在本实施例中,冷板102内填充液冷,即利用工作流体作为中间热量传输的媒介,将热量由热区传递到远处再进行冷却,本申请的冷板102和冷却液均采用现有冷板式液冷技术。
液冷模块103,用于使冷却液在每个冷板102内循环;
在本实施例中,液冷模块103能够实现对冷却液的降温功能,例如将冷却液从一个相对较高的温度降低到预先设定的相对较低的温度,液冷模块103通常设置在散热区外部,且能够驱动冷却液在散热区和非散热区之间循环流动。
多个电源单元104,用于为多个刀片节点101和液冷模块103供电;
在本实施例中,供电单元104即PSU,用于为服务器提供供电,它将高压交流电转换成稳定的低压直流电供给服务器的各个用电部件,例如主板、刀片节点、风冷散热器件、液冷模块等等。
整机管理控制模块105,与每个刀片节点101、液冷模块103以及每个电源单元通信连接,用于监控每个刀片节点101和液冷模块103的功耗,并将功耗均衡分布到多个电源单元 104上。
本实施例的一种刀片式服务器功率控制系统,针对刀片式服务器的多刀片节点进行优势性的适配,通过整机管理控制模块监控每个刀片节点和液冷模块的功耗,并将功耗均衡分布到多个供电单元上,实现整机供电单元功耗均衡,提升刀片式服务器的稳定性和可靠性,提升冷板液冷及多节点适配性以及调控能力。
在一些实施例中,整机管理控制模块105通过获取每个刀片节点101的电流值和电压值以得到第一功耗。
在一些实施例中,液冷模块103的输入端设置有第一电流传感器和第一电压传感器,整机管理控制模块105通过第一电流传感器和第一电压传感器采集的数据得到第二功耗。
在一些实例里中,整机管理控制模块105还用于获取自身工作的第三功耗以及刀片式服务器外围部件工作的第四功耗。
在本实施例中,外围部件指的是刀片式服务器中除以上所指的液冷模块103、整机管理控制模块105以外刀片节点101共用的部件,例如外围部件可以是多个刀片节点101共用的硬盘、硬盘背板、风扇,相应的第四功耗就是指硬盘、硬盘背板、风扇的功耗总和。
在一些实施例中,整机管理控制模块105进一步用于:
计算所有的第一功耗、第二功耗、第三功耗和第四功耗的总和以得到总功耗;
将总功耗与当前在运行的电源单元数量的比值作为每个电源单元的输出功耗。
在一些实施例中,整机管理控制模块105还用于:
响应于监测到某一刀片节点101对应的第一功耗增加,则重新计算总功耗以及每个电源单元的输出功耗。
本实施例中,在单刀片节点突发流量的时刻,及时进行功耗均衡干预,以达到均流,并在服务器的业务运行中保障数据可靠性,以及服务器本身稳定运行地可靠性和寿命。
在一些实施例中,整机管理控制模块105还用于:
响应于监测到某一刀片节点101对应的第一功耗增加,则在重新计算总功耗以及每个电源单元的输出功耗之前,基于第一功耗增加量同比增加液冷模块103的第二功耗。
在本实施例中,液冷模块103能够兼顾刀片节点的功耗增加做出相应的调整,及时的提升降温的需求,解决了服务器出现散热不及时的问题,显著提升散热的时效性和效率。
在一些实施例中,整机管理控制模块105还用于:
基于液冷模块103的第二功耗的增加量通过液冷模块103增加流向某一刀片节点101对应的冷板102的冷却液。
在本实施例中,针对刀片节点功率提升的情况,有针对性的对相应刀片节点进行快速降温,可有效避免局部果然高温或服务器整体散热不均衡的问题。
在一些实施例中,整机管理控制模块105还用于:
响应于监测到某一刀片节点101对应的第一功耗减少,则重新计算总功耗以及每个电源单元的输出功耗。
本实施例,在单刀片节点流量降低的时刻,及时进行功耗均衡干预,以达到均流,可避免浪费资源,并在服务器的业务运行中保障数据可靠性,以及服务器本身稳定运行地可靠性和寿命。
在一些实施例中,整机管理控制模块105还用于:
响应于监测到某一刀片节点101对应的第一功耗减少,则在重新计算总功耗以及每个电源单元的输出功耗之前,基于第一功耗减少量同比减少液冷模块103的第二功耗。
在本实施例中,液冷模块103能够兼顾刀片节点的功耗降低做出相应的调整,及时的减少降温的需求,避免发生资源浪费,提高散热效率。
在一些实施例中,整机管理控制模块105还用于:
基于液冷模块103的第二功耗减少量通过液冷模块103减少流向某一刀片节点101对应的冷板102的冷却液。
在本实施例中,针对刀片节点功率降低的情况,有针对性的减少相应刀片节点降温能力,可避免散热不均衡的问题。
在一些实施例中,整机管理控制模块105还用于:
根据下述公式确定液冷模块冷却液流速;
其中,μ表示冷却液流速,n表示节点数量,Pn表示第一功耗,PL表示第二功耗,PC表示第三功耗,PW表示第四功耗,PFn表示第n个刀片节点的用于实际运算的有用功功耗,σ表示液冷模块和冷板模块的散热损耗,ρ表示冷却液密度,t表示预设液冷模块的调控监控时间,ΔK表示相比于最近一次调控冷却液的温度变化,c表示冷却液的比热容;
将液冷模块调整为以所确定的冷却液流速输出冷却液。
在一些实施例中,整机管理控制模块105还用于:
响应于某一电源单元故障,则将总功耗均衡分布到剩余未发生故障的电源单元上。
在一些实施例中,多个电源单元104采用冗余供电架构,其中,冗余供电架构至少包括一个备份电源单元。
在一些实施例中,整机管理控制模块105还用于:
响应于某一电源单元故障,则采用备份电源单元替换故障的电源单元,并将总功耗均衡分布到剩余电源单元和在运行的备份电源单元上。
在一些实施例中,每个刀片节点101均包含基板管理控制器,基板管理控制器用于获取对应的刀片节点101的电流值和电压值并发送给整机管理控制模块105。
在本实施例中,由于通常每个刀片节点都具有单独的基板管理控制器,基板管理控制器本身具有监控节点上各个器件的功能,直接从基板管理控制器获取电流、电压等数据,避免了对刀片节点进行单独改造,利用现有常规刀片服务器就能够实现,有助于节省成本。
在一些实施例中,每个刀片节点101均包含第二电流传感器和第二电压传感器,整机管理控制模块105通过第二电流传感器和第二电压传感器获取对应的刀片节点101的电流值和电压值。
在本实施例中,为了保证功率控制的稳定性,单独设置电流、电压传感器提升了安全性,在基板管理控制器不正常工作、或还未正常启动的情况下仍能够保证整机功耗可实现均流,当基板管理控制器正常工作后,再切换到刀片对应的基板管理控制器准确功耗获取状态,具有较好的稳定性。
在一些实施例中,整机管理控制模块105进一步用于:
将每个电源单元的输出电压和输出电流均调整为相同值,从而将功耗均衡分布到多个电源单元104。
在一些实施例中,请结合图2所示,本申请还提供了一种刀片式服务器功率控制方法200,具体来说方法包括以下步骤:
步骤201,获取每个刀片节点的第一功耗;
步骤202,获取为所有刀片节点提供散热的液冷模块的第二功耗;
步骤203,获取整机管理控制模块工作的第三功耗以及刀片式服务器外围部件工作的第四功耗;
步骤204,基于第一功耗、第二功耗、第三功耗和第四功耗计算总功耗;
步骤205,将总功耗均衡分布到多个电源单元。
本实施例的一种刀片式服务器功率控制方法,针对刀片式服务器的多刀片节点进行优势性的适配,通过整机管理控制模块监控每个刀片节点和液冷模块的功耗,并将功耗均衡分布到多个供电单元上,实现整机供电单元功耗均衡,提升刀片式服务器的稳定性和可靠性,提升冷板液冷及多节点适配性以及调控能力。
在一些实施例中,方法还包括:
响应于监测到某一刀片节点对应的第一功耗增加,则返回获取每个刀片节点的第一功耗的步骤。
在一些实施例中,返回获取每个刀片节点的第一功耗的步骤,还包括:
在返回获取每个刀片节点的第一功耗的步骤之前,基于第一功耗增加量同比增加液冷模块的第二功耗。
在一些实施例中,基于第一功耗增加量同比增加液冷模块的第二功耗的步骤,还包括:
基于对应液冷模块的第二功耗的增加量通过液冷模块增加流向某一刀片节点对应的冷板的冷却液。
在一些实施例中,方法还包括:
响应于监测到某一刀片节点对应的第一功耗减少,则返回获取每个刀片节点的第一功耗的步骤。
在一些实施例中,返回获取每个刀片节点的第一功耗的步骤,还包括:
在返回获取每个刀片节点的第一功耗的步骤之前,基于第一功耗减少量同比减少对应液冷模块的第二功耗。
在一些实施例中,基于第一功耗减少量同比减少对应液冷模块的第二功耗的步骤,还包括:
基于对应液冷模块的第二功耗减少量通过液冷模块减少流向某一刀片节点对应的冷板的冷却液。
在又一个实施例中,为了便于理解本申请的方案,本实施例以包括n个刀片节点的刀片式服务器为例详细说明本申请方案,一种刀片式服务器功率控制系统,刀片服务器的电源输入由n个PSU共同支撑,此外可能还涉及刀片间共同使用的模块,如硬盘、硬盘背板、风扇、冷板及整机管理控制模块,在原先风冷多节点的基础上,由整机管理控制模块实现对冷板的散热控制,包括控制补液、冷却液流速管理,冷液模块的供电同样由整机管理控制模块进行管控;
请参照图3所示,整机管理控制模块工作原理示如下:多节点刀片式服务器涉及多个PSU供电,其供电方式为分别上电,统一由整机管理控制模块进行管理和功耗均衡控制。PSU上的功率在服务器正常工作的情况下,达到均衡输出,才能良好地保障服务器正常运行,并且在单个节点突发流量的时刻,及时进行功耗均衡干预,以达到均流,并在服务器的业务运行中保障数据可靠性,以及服务器本身稳定运行地可靠性和寿命。为保障业务数据的稳定性,设定有电源冗余,从N+1到N+N冗余量不等。其中,N是代表支持系统正常运行需要的PSU数量,后面的数字代表可以不正常工作的做多的PSU数量。在保障冷板服务器正常运作的情况下,在保障PSU均流的基础上,也需要服务器系统和整机管理控制模块进行服务器支持的冗余设计支持,其实现方式也不能影响到刀片服务器的功率均衡,并且在冗余发生起效时,需要由整机管理控制模块快速完成多个PSU的功率再平衡。
下面将详细说明整机管理控制模块的具体功能实现:
在实际应用中,整机管理控制模块可以用于侦测刀片节点和液冷模块需要的总功耗。
侦测刀片节点的功耗已经是成熟的方案,通常通过每个节点自身的BMC芯片(基板控制器)上的传感器进行该节点总功耗的计算,并通过信号线将相关数据传送给整机管理控制模块。本申请需要增加对液冷模块的功耗侦测。对于液冷模块来说,其仅需要实现对冷却液的降温功能,因此传感器在标准情况下只需要配备温度传感器。为实现其功耗的准确判定,整个液冷模块的输入端增设电流传感器,由于其供电电压恒定(通常设定在12V直流),液冷模块的功耗可由整机管理控制模块通过计算得出。
需要说明的是,采用如上计算方式将会得到较为准确的功耗,并且整机管理控制模块也将较为准确的获得每个刀片节点及液冷模块的功耗。而在节点的BMC模块不正常工作、或还未正常启动的情况下,整机功耗可以通过总的输出电流及输入电压获得。由于PSU的功耗输出均通过整机管理控制模块实现,整机管理控制模块可在这种情况下直接获得相关功耗数据,当BMC正常工作后,再切换到上述的分刀片准确功耗获取状态。
在实际应用中,整机管理控制模块还可以用于将功耗均流分布到全部工作PSU上。
每个PSU的输出电压及输入电压通常情况下都是相同的,在这种情况下,要想实现功耗的均衡,只需要做到电流均衡即可。获取到整机功耗M的数值后,整机管理控制模块获取当前在位的全部PSU数量N,将电流平均的分布到每个PSU上。此时,每个PSU需要负担的功耗为:M/N,而需要承受的输入电流即为:(M/N)/机房的输入电压。需要说明的是机房的输入电压是按照PSU支持及实际机房的情况来决定的,存在110V、220V、380V等多种可能。
在实际应用中,在某刀片的功耗突然发生变化,整机管理控制模块可以动态调控PSU间的功耗均流。
由于刀片间运行的业务流存在差异,可能存在某个刀片节点的功耗突然变化的情况。功耗突然降低,其对于整机的影响较小,整机系统损坏的可能性也不大;但整机系统的功耗突然增加的情况下,若没有整机管理模块的介入控制,可能造成单PSU的电流过大,从而导致该PSU损坏或征集系统崩溃宕机,极大影响业务的部署和使用。为避免出现上述情况,整机的功耗是由整机管理控制模块来控制,不再由单节点自身来使用其对应的PSU。即:将PSU的功耗资源作为整体处理,可视作一个功耗池,需求的功耗值均从该功耗池中获取。
请参照图4所示,当某节点的部件需求的功耗突然增加时,该节点的电流传感器将在第一时间侦测到需求回路上的电流变化。当该信息被基板控制器BMC获取到后,BMC电流出现 需求增加的这一信息回传给整机管理控制模块,由整机管理控制模块开始增加功耗供给,即逐渐提高输入电流并均分到各PSU上。但是,如上的缓慢增加有可能仍无法满足功耗需求增加节点的实际功耗需求,此时,电流增加的策略需要根据BMC回传的电流增加数值进行相应的整机输入电流变化。为保障整机的业务不因为功耗不足造成的停机、宕机,需要电流的增长速度略高于需求的功耗增长速度,当电流增长到需求的电流值后,再进行回调。
在实际应用中,整机管理控制模块可以调控液冷模块的流速和风扇转速。
当刀片节点功耗发生变化时,其对应的散热需求也会发生相应变化,此时冷液模块同步增加冷却液的流速或散热风扇转速。在这一过程中,为相应的增加或减少功耗供给,冷液模块的电流需要同步变化。为实现这一目的,并且保证节点需求功耗增加时液冷模块不成为功耗限定的瓶颈,上述的整机控制管理模块上调的电流需要同时涵盖节点所需功耗及为驱动散热工作需求的功耗。另外,多增的电流余量同样需要覆盖风冷和液冷散热调控需要的电流,并在一轮调控完成后同步回调到适应的电流,以免出现功耗浪费和过高的电流供给导致的后续再次需要调高功耗所造成的总体功耗不足,避免后续的调控出现异常。
在实际应用中,某PSU挂掉整机管理控制模块动态调控功耗均流。
电源冗余的实现采用现有的技术,只是除去服务器节点需要的电源供给外,还需要考量液冷模块所需要的电源供给。
本实施例一种刀片式服务器功率控制系统,针对多节点服务器系统进行优势性的适配,实现整机PSU功耗均衡,并且在多节点服务器之间,实现电源冗余,以保障服务器产品的高可靠性和安全性。
在一些实施例中,本申请还提供了一种刀片式服务器,除包括标准机机箱、多个刀片节点间共同使用的硬盘、硬盘背板、风扇、网络连接模块以外,刀片式服务器还包括以上实施例的刀片式服务器功率控制系统,系统包括:多个刀片节点;与每个刀片节点热交换的冷板,冷板内有冷却液;液冷模块,用于使冷却液在每个冷板内循环;多个电源单元,用于为多个刀片节点和液冷模块供电;整机管理控制模块,与每个刀片节点、液冷模块以及每个电源单元通信连接,用于监控每个刀片节点和液冷模块的功耗,并将功耗均衡分布到多个电源单元上。
在一些实施例中,整机管理控制模块通过获取每个刀片节点的电流值和电压值以得到第一功耗。
在一些实施例中,液冷模块的输入端设置有第一电流传感器和第一电压传感器,整机管理控制模块通过第一电流传感器和第一电压传感器采集的数据得到第二功耗。
在一些实施例中,整机管理控制模块还用于获取自身工作的第三功耗以及刀片式服务器外围部件工作的第四功耗。
在一些实施例中,整机管理控制模块进一步用于计算所有的第一功耗、第二功耗、第三功耗和第四功耗的总和以得到总功耗,并将总功耗与当前在运行的电源单元数量的比值作为每个电源单元的输出功耗。
在一些实施例中,整机管理控制模块还用于响应于监测到某一刀片节点对应的第一功耗增加,则重新计算总功耗以及每个电源单元的输出功耗。
在一些实施例中,整机管理控制模块还用于响应于监测到某一刀片节点对应的第一功耗增加,则在重新计算总功耗以及每个电源单元的输出功耗之前,基于第一功耗增加量同比增 加液冷模块的第二功耗。
在一些实施例中,整机管理控制模块还用于基于液冷模块的第二功耗的增加量通过液冷模块增加流向某一刀片节点对应的冷板的冷却液。
在一些实施例中,整机管理控制模块还用于响应于监测到某一刀片节点对应的第一功耗减少,则重新计算总功耗以及每个电源单元的输出功耗。
在一些实施例中,整机管理控制模块还用于:
响应于监测到某一刀片节点对应的第一功耗减少,则在重新计算总功耗以及每个电源单元的输出功耗之前,基于第一功耗减少量同比减少液冷模块的第二功耗。
在一些实施例中,整机管理控制模块还用于:
基于液冷模块的第二功耗减少量通过液冷模块减少流向某一刀片节点对应的冷板的冷却液。
在一些实施例中,整机管理控制模块还用于:
根据下述公式确定液冷模块冷却液流速;
其中,μ表示冷却液流速,n表示节点数量,Pn表示第一功耗,PL表示第二功耗,PC表示第三功耗,PW表示第四功耗,PFn表示第n个刀片节点的用于实际运算的有用功功耗,σ表示液冷模块和冷板模块的散热损耗,ρ表示冷却液密度,t表示预设液冷模块的调控监控时间,ΔK表示相比于最近一次调控冷却液的温度变化,c表示冷却液的比热容;
将液冷模块调整为以所确定的冷却液流速输出冷却液。
在一些实施例中,整机管理控制模块还用于:
响应于某一电源单元故障,则将总功耗均衡分布到剩余未发生故障的电源单元上。
在一些实施例中,多个电源单元采用冗余供电架构,其中,冗余供电架构至少包括一个备份电源单元。
在一些实施例中,整机管理控制模块还用于:
响应于某一电源单元故障,则采用备份电源单元替换故障的电源单元,并将总功耗均衡分布到剩余电源单元和在运行的备份电源单元上。
在一些实施例中,每个刀片节点均包含基板管理控制器,基板管理控制器用于获取对应的刀片节点的电流值和电压值并发送给整机管理控制模块。
在一些实施例中,每个刀片节点均包含第二电流传感器和第二电压传感器,整机管理控制模块通过第二电流传感器和第二电压传感器获取对应的刀片节点的电流值和电压值。
在一些实施例中,整机管理控制模块进一步用于:
将每个电源单元的输出电压和输出电流均调整为相同值,从而将功耗均衡分布到多个电源单元。
以上实施例的各技术特征可以进行任意的组合,为使描述简洁,未对上述实施例中的各个技术特征所有可能的组合都进行描述,然而,只要这些技术特征的组合不存在矛盾,都应当认为是本说明书记载的范围。
以上实施例仅表达了本申请的几种实施方式,其描述较为具体和详细,但并不能因此而 理解为对发明专利范围的限制。应当指出的是,对于本领域的普通技术人员来说,在不脱离本申请构思的前提下,还可以做出若干变形和改进,这些都属于本申请的保护范围。因此,本申请专利的保护范围应以所附权利要求为准。

Claims (26)

  1. 一种刀片式服务器功率控制系统,其特征在于,所述系统包括:
    多个刀片节点;
    与每个刀片节点热交换的冷板,所述冷板内有冷却液;
    液冷模块,用于使所述冷却液在每个所述冷板内循环;
    多个电源单元,用于为所述多个刀片节点和所述液冷模块供电;
    整机管理控制模块,与每个刀片节点、所述液冷模块以及每个电源单元通信连接,用于监控每个刀片节点和所述液冷模块的功耗,并将所述功耗均衡分布到所述多个电源单元上。
  2. 根据权利要求1所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块通过获取每个刀片节点的电流值和电压值以得到第一功耗。
  3. 根据权利要求2所述的刀片式服务器功率控制系统,其特征在于,所述液冷模块的输入端设置有第一电流传感器和第一电压传感器,所述整机管理控制模块通过所述第一电流传感器和第一电压传感器采集的数据得到第二功耗。
  4. 根据权利要求3所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于获取自身工作的第三功耗以及刀片式服务器外围部件工作的第四功耗。
  5. 根据权利要求4所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块进一步用于:
    计算所有的所述第一功耗、所述第二功耗、所述第三功耗和第四功耗的总和以得到总功耗;
    将所述总功耗与当前在运行的电源单元数量的比值作为每个电源单元的输出功耗。
  6. 根据权利要求5所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    响应于监测到某一刀片节点对应的第一功耗增加,则重新计算所述总功耗以及每个电源单元的输出功耗。
  7. 根据权利要求6所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    响应于监测到某一刀片节点对应的第一功耗增加,则在重新计算所述总功耗以及每个电源单元的输出功耗之前,基于所述第一功耗增加量同比增加所述液冷模块的第二功耗。
  8. 根据权利要求7所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    基于所述液冷模块的第二功耗的增加量通过所述液冷模块增加流向所述某一刀片节点对应的冷板的冷却液。
  9. 根据权利要求5所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    响应于监测到某一刀片节点对应的第一功耗减少,则重新计算所述总功耗以及每个电源单元的输出功耗。
  10. 根据权利要求9所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    响应于监测到某一刀片节点对应的第一功耗减少,则在重新计算所述总功耗以及每个电源单元的输出功耗之前,基于所述第一功耗减少量同比减少所述液冷模块的第二功耗。
  11. 根据权利要求10所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    基于所述液冷模块的第二功耗减少量通过所述液冷模块减少流向所述某一刀片节点对应的冷板的冷却液。
  12. 根据权利要求6或9所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    根据下述公式确定液冷模块冷却液流速;
    其中,μ表示冷却液流速,n表示节点数量,Pn表示所述第一功耗,PL表示所述第二功耗,PC表示所述第三功耗,PW表示所述第四功耗,PFn表示第n个刀片节点的用于实际运算的有用功功耗,σ表示液冷模块和冷板模块的散热损耗,ρ表示冷却液密度,t表示预设液冷模块的调控监控时间,ΔK表示相比于最近一次调控冷却液的温度变化,c表示冷却液的比热容;
    将所述液冷模块调整为以所确定的冷却液流速输出冷却液。
  13. 根据权利要求5所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    响应于某一电源单元故障,则将所述总功耗均衡分布到剩余未发生故障的电源单元上。
  14. 根据权利要求5所述的刀片式服务器功率控制系统,其特征在于,所述多个电源单元采用冗余供电架构,其中,所述冗余供电架构至少包括一个备份电源单元。
  15. 根据权利要求14所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块还用于:
    响应于某一电源单元故障,则采用所述备份电源单元替换故障的电源单元,并将所述总功耗均衡分布到剩余电源单元和在运行的所述备份电源单元上。
  16. 根据权利要求2所述的刀片式服务器功率控制系统,其特征在于,每个刀片节点均包含基板管理控制器,所述基板管理控制器用于获取对应的刀片节点的电流值和电压值并发送给所述整机管理控制模块。
  17. 根据权利要求2所述的刀片式服务器功率控制系统,其特征在于,每个刀片节点均包含第二电流传感器和第二电压传感器,所述整机管理控制模块通过所述第二电流传感器和第二电压传感器获取对应的刀片节点的电流值和电压值。
  18. 根据权利要求1所述的刀片式服务器功率控制系统,其特征在于,所述整机管理控制模块进一步用于:
    将每个电源单元的输出电压和输出电流均调整为相同值,从而将功耗均衡分布到所述多个电源单元。
  19. 一种刀片式服务器功率控制方法,其特征在于,所述方法包括:
    获取每个刀片节点的第一功耗;
    获取为所有刀片节点提供散热的液冷模块的第二功耗;
    获取整机管理控制模块工作的第三功耗以及刀片式服务器外围部件工作的第四功耗;
    基于所述第一功耗、第二功耗、第三功耗和第四功耗计算总功耗;
    将所述总功耗均衡分布到多个电源单元。
  20. 根据权利要求19所述的刀片式服务器功率控制方法,其特征在于,所述方法还包括:
    响应于监测到某一刀片节点对应的第一功耗增加,则返回获取每个刀片节点的第一功耗的步骤。
  21. 根据权利要求20所述的刀片式服务器功率控制方法,其特征在于,在所述返回获取每个刀片节点的第一功耗的步骤之前,还包括:
    基于所述第一功耗增加量同比增加液冷模块的第二功耗。
  22. 根据权利要求21所述的刀片式服务器功率控制方法,其特征在于,所述基于所述第一功耗增加量同比增加液冷模块的第二功耗的步骤,还包括:
    基于对应液冷模块的第二功耗的增加量通过所述液冷模块增加流向所述某一刀片节点对应的冷板的冷却液。
  23. 根据权利要求19所述的刀片式服务器功率控制方法,其特征在于,所述方法还包括:
    响应于监测到某一刀片节点对应的第一功耗减少,则返回获取每个刀片节点的第一功耗的步骤。
  24. 根据权利要求23所述的刀片式服务器功率控制方法,其特征在于,在所述返回获取每个刀片节点的第一功耗的步骤之前,还包括:
    基于所述第一功耗减少量同比减少对应液冷模块的第二功耗。
  25. 根据权利要求24所述的刀片式服务器功率控制方法,其特征在于,所述基于所述第一功耗减少量同比减少对应液冷模块的第二功耗的步骤,还包括:
    基于对应液冷模块的第二功耗减少量通过所述液冷模块减少流向所述某一刀片节点对应的冷板的冷却液。
  26. 一种刀片式服务器,其特征在于,所述刀片式服务器包括权利要求1-18任意一项所述的刀片式服务器功率控制系统。
PCT/CN2023/141186 2023-06-13 2023-12-22 一种刀片式服务器功率控制系统、方法和刀片式服务器 Ceased WO2024255191A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US19/115,756 US20260113901A1 (en) 2023-06-13 2023-12-22 Blade server power control system and method, and blade server

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202310699361.9A CN116430975B (zh) 2023-06-13 2023-06-13 一种刀片式服务器功率控制系统、方法和刀片式服务器
CN202310699361.9 2023-06-13

Publications (1)

Publication Number Publication Date
WO2024255191A1 true WO2024255191A1 (zh) 2024-12-19

Family

ID=87080110

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/141186 Ceased WO2024255191A1 (zh) 2023-06-13 2023-12-22 一种刀片式服务器功率控制系统、方法和刀片式服务器

Country Status (3)

Country Link
US (1) US20260113901A1 (zh)
CN (1) CN116430975B (zh)
WO (1) WO2024255191A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116430975B (zh) * 2023-06-13 2023-09-08 苏州浪潮智能科技有限公司 一种刀片式服务器功率控制系统、方法和刀片式服务器

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120136498A1 (en) * 2010-11-29 2012-05-31 Inventec Corporation Rack server system
CN105302182A (zh) * 2015-11-02 2016-02-03 张存勇 一种服务器液冷散热节能控制系统和方法
CN109917894A (zh) * 2019-03-06 2019-06-21 苏州浪潮智能科技有限公司 一种服务器系统及其供电均衡方法以及电源供应器
CN114461041A (zh) * 2022-01-07 2022-05-10 苏州浪潮智能科技有限公司 一种多节点服务器psu访问及能耗散热管理结构和方法
CN115686117A (zh) * 2023-01-03 2023-02-03 苏州浪潮智能科技有限公司 液冷控制方法、系统、液冷机柜、电子设备及存储介质
CN116430975A (zh) * 2023-06-13 2023-07-14 苏州浪潮智能科技有限公司 一种刀片式服务器功率控制系统、方法和刀片式服务器

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120136498A1 (en) * 2010-11-29 2012-05-31 Inventec Corporation Rack server system
CN105302182A (zh) * 2015-11-02 2016-02-03 张存勇 一种服务器液冷散热节能控制系统和方法
CN109917894A (zh) * 2019-03-06 2019-06-21 苏州浪潮智能科技有限公司 一种服务器系统及其供电均衡方法以及电源供应器
CN114461041A (zh) * 2022-01-07 2022-05-10 苏州浪潮智能科技有限公司 一种多节点服务器psu访问及能耗散热管理结构和方法
CN115686117A (zh) * 2023-01-03 2023-02-03 苏州浪潮智能科技有限公司 液冷控制方法、系统、液冷机柜、电子设备及存储介质
CN116430975A (zh) * 2023-06-13 2023-07-14 苏州浪潮智能科技有限公司 一种刀片式服务器功率控制系统、方法和刀片式服务器

Also Published As

Publication number Publication date
CN116430975B (zh) 2023-09-08
US20260113901A1 (en) 2026-04-23
CN116430975A (zh) 2023-07-14

Similar Documents

Publication Publication Date Title
EP3242185B1 (en) Server rack power management
CN110221946B (zh) 用于对存储系统进行功率分析的方法和装置
JP5254734B2 (ja) 電子システムの電力を管理する方法、コンピュータ・プログラム、及び電子システム
US8250382B2 (en) Power control of servers using advanced configuration and power interface (ACPI) states
US8156358B2 (en) System and method for dynamic modular information handling system power distribution
CN101441506B (zh) 一种刀片服务器系统及其功耗管理方法
US7707443B2 (en) Rack-level power management of computer systems
US8065541B2 (en) Server power consumption controller, and method and computer program for controlling server power consumption
US7272732B2 (en) Controlling power consumption of at least one computer system
US10727553B1 (en) Thermal management system design for battery pack
CN106896884A (zh) Atca架构设备散热管理方法及装置
CN115686117A (zh) 液冷控制方法、系统、液冷机柜、电子设备及存储介质
US11733762B2 (en) Method to allow for higher usable power capacity in a redundant power configuration
WO2004074983A2 (en) Enterprise power and thermal management
US11653479B2 (en) Modular cooling units and advanced distribution hardware for racks
CN113632043B (zh) 可变冗余数据中心电力拓扑
US11457547B1 (en) Phase change thermal management system for electronic racks
US20240080359A1 (en) Multi-server system and method of reducing total power consumption of multiple servers
CN102520783B (zh) 一种可以实现smartrack节能的方法及机柜系统
WO2024255191A1 (zh) 一种刀片式服务器功率控制系统、方法和刀片式服务器
CN101994717A (zh) Atca风扇控制方法及atca机框管理器
TW202418039A (zh) 電腦系統之節電系統
CN115793815A (zh) 服务器的散热控制方法及系统
CN119781600A (zh) 一种电源功率与系统功耗自动控制方法、装置、计算机设备及存储介质
CN109976478B (zh) 多cpu散热控制系统和散热控制方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23941411

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE