Computing system, power supply module and computing device

Abstract

The embodiment of the invention provides a computing system, a power supply module and computing equipment. The computing system comprises a power conversion device, an energy storage device, a plurality of computing devices and a first management unit. Wherein the input end of the power conversion device is used for being connected with a second alternating-current line, the energy storage device is connected to the output end of the power conversion device, each computing device comprises a first input end and a second input end, the first input end of each computing device is used for being connected with a first alternating-current line, and the second input end of each computing device is used for being connected with a second alternating-current line. The second input end of each computing device is connected with the output end of the power conversion device and the output end of the energy storage device. And the first management unit is electrically connected to the plurality of computing devices, and the first management unit is used for controlling the plurality of computing devices to take electricity from the alternating current line and the energy storage device under the condition that the actual power value of the computing system is greater than the first power threshold value. The computing system is used for providing computing services.

Description

Technical Field

[0001] The present application relates to the field of server technology, and in particular to a computing system, a power supply module and a computing device. Background Art

[0002] Data centers usually include multiple computing devices, and computing devices may include multiple servers, which can provide computing services. Among them, the maximum power of a computing device is limited. Therefore, as the computing power of the server increases, its power gradually increases, resulting in a small number of servers in the computing device, resulting in a low server shelf rate. Summary of the invention

[0003] The purpose of the embodiments of the present application is to provide a computing system, a power supply module and a computing device, which are used to increase the shelf rate of servers in the computing device.

[0004] To achieve the above objectives, the embodiments of the present application provide the following technical solutions:

[0005] In a first aspect, a computing system is provided. The computing system includes a power conversion device, an energy storage device, a plurality of computing devices, and a first management unit. The input end of the power conversion device is used to connect to a second AC line, the energy storage device is connected to the output end of the power conversion device, each computing device includes a first input end and a second input end, the first input end of each computing device is used to connect to the first AC line, and the second input end of each computing device is connected to the output end of the power conversion device and the output end of the energy storage device. The first management unit is electrically connected to a plurality of computing devices, and the first management unit is used to: when the actual power value of the computing system is greater than a first power threshold, control the plurality of computing devices to draw power from the AC line and the energy storage device, wherein drawing power from the AC line includes drawing power from the first AC line, and / or drawing power from the second AC line through the power conversion device.

[0006] In the above-mentioned computing system, the computing device can draw power from the first AC line through the first input terminal, and the second input terminal of the computing device can draw power from the second AC line through the power conversion device. Therefore, both the second AC line and the first AC line can supply power to the computing device. In addition, the second input terminal of the computing device can draw power from the energy storage device. Therefore, the energy storage device can also supply power to the computing device. The first management unit can control multiple computing devices to draw power from the AC line and the energy storage device when the actual power value of the computing system is greater than the first power threshold. Therefore, both the energy storage device and the AC line can supply power to multiple computing devices, thereby increasing the maximum power of the entire cabinet of the computing device, so that more servers can be deployed in the computing device, thereby increasing the shelf rate of servers in the computing device and reducing the vacancy rate of the cabinet without the need to renovate the computer room.

[0007] In one possible implementation, the first management unit is also used to: when the actual power value of the computing system is greater than the first power threshold, control the AC power obtained by multiple computing devices from the AC line so that the power loaded by the AC line is the first power threshold, and the power loaded by the energy storage device is the power difference between the actual power value of the computing device and the first power threshold.

[0008] If the actual power is greater than the first power threshold, the power carried by the AC line is the first power threshold, and the power carried by the energy storage device is the power difference between the actual power value and the first power threshold. The energy storage device can carry the part of the actual power value of the data center that exceeds the first power threshold, thereby increasing the maximum power value of the data center, thereby increasing the rack rate of the data center and reducing the vacancy rate of the cabinets.

[0009] In one possible implementation, the first management unit is electrically connected to the energy storage device; the first management unit is also used to: when the actual power value is greater than the first power threshold, control the output voltage of the power conversion device to be lower than the output voltage of the energy storage device, so that the energy storage device is in a discharging state.

[0010] The first management unit may control the output voltage of the power conversion device to be lower than the output voltage of the energy storage device, so that the energy storage device is in a discharging state, and thus the energy storage device can supply power to the server.

[0011] In one possible implementation, the first management unit is electrically connected to the energy storage device; the first management unit is also used to: obtain the remaining capacity of the energy storage device, adjust the capping power of the computing system according to the remaining capacity, and control the actual power value of the computing system to be less than or equal to the capping power.

[0012] Among them, the first management unit can limit the maximum power of the computing system by adjusting the capped power of the computing system. After determining the capped power of the computing system, the actual power value of the computing system will not exceed the capped power of the computing system, that is, the sum of the powers of multiple computing devices will not exceed the capped power of the computing system. Furthermore, it can be avoided that the actual power value of the computing system exceeds the sum of the powers that can be loaded by the AC line and the energy storage device, so that the computing system can be prevented from losing power due to insufficient remaining capacity of the energy storage device.

[0013] In a possible implementation, the first management unit is further configured to: when an actual power value of the computing system is less than a first power threshold, control the power conversion device to charge the energy storage device.

[0014] Among them, when the actual power value of the computing system is less than the first power threshold, the power conversion device can charge the energy storage device, so as to ensure that the energy storage device has sufficient capacity, and thus the maximum power value of the computing system can be ensured. In addition, when the actual power value of the computing system is greater than the first power threshold, the energy storage device can power multiple computing devices together with the AC line. At this time, the energy storage device is discharged and the energy storage device is not charged. Therefore, the energy storage device can be in only one state, that is, in a charging state or a discharging state, so as to avoid the energy storage device being in both charging and discharging states at the same time, and then the energy storage device can be protected and the service life of the energy storage device can be extended.

[0015] In one possible implementation, the first management unit is also used to: when the actual power value of the computing system is greater than the first power threshold, control an overloaded computing device among the multiple computing devices to draw power from the AC line and the energy storage device; wherein the overloaded computing device is a computing device among the multiple computing devices whose actual power value is greater than the first sub-power threshold, and the product of the first sub-power threshold and the number of computing devices is equal to the first power threshold.

[0016] Among them, when the actual power value of the computing system is greater than the first power threshold, the first management unit can adjust the power supply mode of the overloaded computing device, that is, adjust the power supply from the AC line to the AC line and the energy storage device jointly powering the overloaded computing device, thereby reducing the number of computing devices whose power supply mode is changed.

[0017] In one possible implementation, the first management unit is also used to: when the actual power value of the computing system is greater than the first power threshold, control a steady-state computing device among the multiple computing devices to draw power from the AC line; wherein the steady-state computing device is a computing device among the multiple computing devices whose actual power value is less than or equal to the first sub-power threshold.

[0018] Wherein, when the actual power value of the computing system is greater than the first power threshold, the first management unit does not adjust the power supply mode of the steady-state computing device, and the steady-state computing device is still powered by the AC line. With such an arrangement, when the actual power value of the computing system is greater than the first power threshold, the number of computing devices whose power supply mode is changed can be reduced.

[0019] In one possible implementation, the computing device also includes a second management unit and multiple servers, and the second management unit is communicatively connected to the first management unit and the multiple servers; the first management unit is also used to: when the actual power value of the computing system is greater than the first power threshold, control the overloaded server in the overloaded computing device to draw power from the AC line and the energy storage device through the second management unit of the overloaded computing device; wherein the overloaded server is a server among the multiple servers whose actual power value is greater than the second sub-power threshold, and the product of the second sub-power threshold and the number of servers in a computing device is equal to the first sub-power threshold.

[0020] Among them, when the actual power value of the computing system is greater than the first power threshold, the first management unit can adjust the power supply method of the overloaded server in the overloaded computing device, that is, adjust the power supply from the AC line to the AC line and the energy storage device jointly powering the overloaded computing device, thereby reducing the number of servers whose power supply methods are changed.

[0021] In one possible implementation, the first management unit is further used to: when the actual power value of the computing system is greater than the first power threshold, control the steady-state server in the overloaded computing device to draw power from the AC line through the second management unit of the overloaded computing device.

[0022] When the actual power value of the computing system is greater than the first power threshold, the power supply mode of the steady-state servers in the overloaded computing device is not changed, and thus the number of servers whose power supply modes are changed can be reduced.

[0023] In a possible implementation, the server includes a first switch power supply, a second switch power supply and a plurality of loads, the first switch power supply and the second switch power supply are both connected to the plurality of loads, the plurality of loads include a management chip, the management chip is communicatively connected to the first switch power supply, the second switch power supply and the second management unit, the first switch power supply is a first sub-input terminal of the server, and the second switch power supply is a second sub-input terminal of the server. The first input terminal of the computing device is connected to the first sub-input terminals of the plurality of servers, and the second input terminal of the computing device is connected to the second sub-input terminals of the plurality of servers. The first management unit is used to: when the actual power value of the computing system is greater than the first power threshold, control the first switch power supply and the second switch power supply of the overloaded server in the overloaded computing device to jointly supply power to the plurality of loads in the overloaded server in the overloaded computing device through the second management unit of the overloaded computing device and the management chip of the overloaded server in the overloaded computing device; or control the second switch power supply of the overloaded server in the overloaded computing device to supply power to the plurality of loads in the overloaded server through the second management unit of the overloaded computing device and the management chip of the overloaded server in the overloaded computing device.

[0024] Among them, the first switching power supply can draw power from the first AC line, and the second switching power supply can draw power from the energy storage device, or can draw power from the second AC line through the power conversion device. When the actual power value of the computing system is greater than the first power threshold, the first switching power supply and the second switching power supply of the overloaded server in the overloaded computing device draw power from the first AC line and the energy storage device respectively, or control the second switching power supply of the overloaded server to draw power from the first AC line and the energy storage device at the same time, so that the AC line and the energy storage device can jointly power the overloaded server.

[0025] In one possible implementation, the first management unit is used to: when an actual power value of the computing system is greater than a first power threshold, control a first switching power supply or a second switching power supply of a steady-state server in the overloaded computing device to supply power to multiple loads of the steady-state server in the overloaded computing device through a second management unit of the overloaded computing device and a management chip of the steady-state server in the overloaded computing device; and / or, control a first switching power supply or a second switching power supply of each server in the steady-state computing device to supply power to multiple loads of each server in the steady-state computing device through a second management unit of the steady-state computing device and a management chip of each server in the steady-state computing device.

[0026] In which, when the actual power value of the computing system is greater than the first power threshold, in the steady-state server in the overloaded computing device and each server in the steady-state computing device, the first switching power supply can draw power from the first AC line, or the second switching power supply can draw power from the second AC line through a power conversion device, so that the first switching power supply or the second switching power supply can supply power to the steady-state server in the overloaded computing device.

[0027] In a possible implementation, in a computing device, multiple servers include first-class servers and / or second-class servers, wherein servers that can draw power from a first AC line are first-class servers, and servers that cannot draw power from the first AC line are second-class servers. The first management unit is used to: when an actual power value is greater than a first power threshold, control a first switch power supply and a second switch power supply of a first-class server in an overloaded server in the overloaded computing device through a second management unit of the overloaded computing device and a management chip of a first-class server in an overloaded server in the overloaded computing device to jointly supply power to multiple loads of the first-class server in an overloaded server in the overloaded computing device; and / or, control a second switch power supply of a second-class server in an overloaded server in the overloaded computing device through a second management unit of the overloaded computing device and a management chip of a second-class server in an overloaded server in the overloaded computing device to supply power to multiple loads of the second-class server in an overloaded server in the overloaded computing device.

[0028] When there is a second type of server in the overloaded computing device, the second switching power supply of the second type of server can supply power to multiple loads, thereby preventing the second type of server from losing power.

[0029] In one possible implementation, the first management unit is used to: when an actual power value of the computing system is greater than a first power threshold, control the first switching power supply and the second switching power supply of the first type of servers in the steady-state servers in the overloaded computing device through the second management unit of the overloaded computing device and the management chip of the first type of servers in the steady-state servers in the overloaded computing device to jointly power multiple loads of the first type of servers in the steady-state servers in the overloaded computing device; and / or; control the first type of servers in the steady-state servers in the overloaded computing device through the second management unit of the overloaded computing device and the management chip of the second type of servers in the steady-state servers in the overloaded computing device. The second switching power supply of the second type of server supplies power to multiple loads of the second type of server in the steady-state server in the overloaded computing device; and / or, through the second management unit of the steady-state computing device and the management chip of the first type of server in the steady-state computing device, the first switching power supply and the second switching power supply of the first type of server in the steady-state computing device are controlled to jointly supply power to multiple loads of the first type of server in the steady-state computing device; and / or, through the second management unit of the steady-state computing device and the management chip of the second type of server in the steady-state computing device, the second switching power supply of the second type of server in the steady-state computing device is controlled to supply power to multiple loads of the second type of server in the steady-state computing device.

[0030] When there is a second type of server in a steady-state server or a steady-state computing device, the second switching power supply of the second type of server can supply power to multiple loads, thereby preventing the second type of server from losing power.

[0031] In one possible implementation, the first switching power supply includes a first processing chip, the first processing chip is electrically connected to the management chip, and the first processing chip is used to obtain the input power of the first switching power supply and send it to the management chip; the second switching power supply includes a second processing chip, the second processing chip is electrically connected to the management chip, and the second processing chip is used to obtain the input power of the second switching power supply and send it to the management chip; the management chip is used to determine the actual power value of the server according to the input power of the first switching power supply and the input power of the second switching power supply, and send it to the second management unit; the second management unit is used to determine the actual power value of the computing device according to the actual power values ​​of multiple servers, and send it to the first management unit; the first management unit is used to determine the actual power value of the computing system according to the actual power values ​​of multiple computing devices.

[0032] The actual power value of the computing device can be equal to the sum of the powers of multiple servers, and the actual power value of the computing system can be equal to the sum of the powers of multiple computing devices. The first processing chip can send the input power of the first switching power supply obtained by it to the management chip, and the second processing chip can send the input power of the second switching power supply obtained by it to the management chip. The management chip can determine the power of the server where the management chip is located based on the sum of the input power of the first switching power supply and the input power of the second switching power supply. The management chip can send the power of the server obtained by it to the second management unit, so the second management unit can determine the power of the computing device where it is located based on the power of multiple servers. The second management unit can send the power of the computing device obtained by it to the first management unit, so the first management unit can obtain the actual power value of the computing system based on the power of multiple computing devices.

[0033] In one possible implementation, the computing system also includes: a switch, the switch includes a first communication interface and multiple second communication interfaces, the first communication interface and the multiple second communication interfaces are connected, the first communication interface is connected to the first management unit, and the second communication interface is connected to the second management unit.

[0034] The first communication interface is connected to the first management unit, and the second communication interface is connected to the second management unit of the computing device. The first communication interface is connected to the second communication interface, so the first management unit can communicate with the second management unit of the computing device through the first communication interface and the second communication interface.

[0035] In one possible implementation, the computing device further includes a cabinet, and the plurality of servers are disposed in the cabinet; the energy storage device is located outside the cabinet.

[0036] Wherein, the cabinet can play a protective role for the server. When the AC line supplies power to the computing device, the AC line can supply power to multiple servers. When the energy storage device supplies power to the computing device, the energy storage device can supply power to at least some of the servers in the computing setting. Wherein, the cabinet can play a protective role for the server. When the AC line supplies power to the computing device, the AC line can supply power to multiple servers. When the energy storage device supplies power to the computing device, the energy storage device can supply power to at least some of the servers in the computing setting.

[0037] In a possible implementation, the first management unit is further configured to: when an actual power value of the computing system is less than or equal to a first power threshold, control the plurality of computing devices to draw power from the AC line.

[0038] The first power threshold may be equal to the maximum power that the AC line can provide to the computing system. At this time, multiple computing devices all draw power from the AC line and will not exceed the AC maximum power, thereby ensuring normal operation of the computing devices.

[0039] In one possible implementation, the first management unit is also used to: when the actual power value is less than or equal to the first power threshold, control the first switching power supply and the second switching power supply of the first type of server in the computing device to jointly power multiple loads of the first type of server in the computing device through the second management unit of the computing device and the management chip of the first type of server in the computing device; and / or, control the second switching power supply of the second type of server in the computing device to power multiple loads of the second type of server in the computing device through the second management unit of the computing device and the management chip of the second type of server in the computing device.

[0040] Among them, when the actual power value is less than or equal to the first power threshold, when there is a second type of server in any computing device, the second switching power supply of the second type of server can power multiple loads, thereby avoiding power failure of the second type of server.

[0041] In one possible implementation, the computing system also includes: a first power distribution device and a second power distribution device, the input end of the first power distribution device is connected to the first AC line, and the output end of the first power distribution device is electrically connected to the first input end of multiple computing devices; the input end of the second power distribution device is connected to the output end of the energy storage device and the output end of the power conversion device, and the output end of the second power distribution device is connected to the second input end of the multiple computing devices.

[0042] For example, the first power distribution device and the second power distribution device can both be terminal cabinets.

[0043] In a second aspect, a power supply module is provided, which includes: an energy storage device, a first management unit and a power conversion device. The input end of the power conversion device is used to connect to the second AC line, and the output end of the power conversion device is used to connect to the second input end of multiple computing devices in the computing system. The energy storage device is connected to the output end of the power conversion device, and the output end of the energy storage device is used to connect to the second input end of multiple computing devices; the first management unit is used to be electrically connected to multiple computing devices, and the first management unit is used to: when the actual power value of the computing system is greater than the first power threshold, control the multiple computing devices to draw power from the AC line and the energy storage device, wherein drawing power from the AC line includes drawing power from the first AC line, and / or drawing power from the second AC line through the power conversion device, and the first AC line is connected to the first input end of multiple computing devices.

[0044] Among them, the first management unit can control multiple computing devices to draw power from the AC line and the energy storage device when the actual power value of the computing system is greater than the first power threshold. Therefore, both the energy storage device and the AC line can supply power to multiple computing devices, thereby increasing the maximum power of the entire cabinet of the computing device, so that more servers can be deployed in the computing device, thereby increasing the shelf rate of servers in the computing device and reducing the vacancy rate of the cabinet without the need to renovate the computer room.

[0045] Based on the power supply module provided by the second aspect, in one possible implementation, the first management unit is also used to: when the actual power value of the computing system is greater than the first power threshold, control the AC power obtained by multiple computing devices from the AC line so that the power carried by the AC line is the first power threshold, and the power carried by the energy storage device is the power difference between the actual power value of the computing device and the first power threshold.

[0046] Based on the power supply module provided by the second aspect, in one possible implementation, the first management unit is electrically connected to the energy storage device; the first management unit is also used to: obtain the remaining capacity of the energy storage device, adjust the capping power of the computing system according to the remaining capacity, and control the actual power value of the computing system to be less than or equal to the capping power.

[0047] Based on the power supply module provided in the second aspect, in a possible implementation, the first management unit is also used to: when the actual power value of the computing system is less than the first power threshold, control the power conversion device to charge the energy storage device.

[0048] Based on the power supply module provided by the second aspect, in one possible implementation, the first management unit is also used to: when the actual power value of the computing system is greater than the first power threshold, control an overloaded computing device among multiple computing devices to draw power from the AC line and the energy storage device; wherein the overloaded computing device is a computing device among the multiple computing devices whose actual power value is greater than the first sub-power threshold, and the product of the first sub-power threshold and the number of computing devices is equal to the first power threshold.

[0049] Based on the power supply module provided by the second aspect, in a possible implementation, the computing device also includes a second management unit and multiple servers, and the second management unit is communicatively connected to the first management unit and the multiple servers; the first management unit is also used to: when the actual power value of the computing system is greater than the first power threshold, control the overloaded server in the overloaded computing device to draw power from the AC line and the energy storage device through the second management unit of the overloaded computing device; wherein the overloaded server is a server among the multiple servers whose actual power value is greater than the second sub-power threshold, and the product of the second sub-power threshold and the number of servers in a computing device is equal to the first sub-power threshold.

[0050] Based on the power supply module provided by the second aspect, in a possible implementation, the server includes a first switching power supply, a second switching power supply and multiple loads, the first switching power supply and the second switching power supply are both connected to the multiple loads, the multiple loads include a management chip, the management chip is communicatively connected to the first switching power supply, the second switching power supply and the second management unit, the first switching power supply is the first sub-input terminal of the server, and the second switching power supply is the second sub-input terminal of the server; the first input terminal of the computing device is connected to the first sub-input terminals of the multiple servers, and the second input terminal of the computing device is connected to the second sub-input terminals of the multiple servers; in a computing device, the multiple servers include first-class servers and / or second-class servers, wherein the server that can draw power from the first AC line is the first-class server, and The server that cannot draw power from the first AC line is a second-category server; the first management unit is used to: when the actual power value is greater than the first power threshold, control the first switching power supply and the second switching power supply of the first-category server in the overloaded servers in the overloaded computing device through the second management unit of the overloaded computing device and the management chip of the first-category server in the overloaded servers in the overloaded computing device to jointly supply power to multiple loads of the first-category server in the overloaded servers in the overloaded computing device; and / or, control the second switching power supply of the second-category server in the overloaded servers in the overloaded computing device through the second management unit of the overloaded computing device and the management chip of the second-category server in the overloaded servers in the overloaded computing device to supply power to multiple loads of the first-category server in the overloaded computing device.

[0051] Based on the power supply module provided in the second aspect, in a possible implementation, the first management unit is also used to: when the actual power value of the computing system is less than or equal to the first power threshold, control multiple computing devices to draw power from the AC line.

[0052] In a third aspect, a computing device is provided, the computing device comprising: a second management unit and a plurality of servers. The first sub-input terminals of the plurality of servers are used to be connected to the first AC line, and the second sub-input terminals of the plurality of servers are used to be connected to the output terminal of the power conversion device of the power supply module and the output terminal of the energy storage device of the power supply module, and the input terminal of the power conversion device is connected to the second AC line. The second management unit is connected to the plurality of servers in communication, and the second management unit is used to be connected to the first management unit of the power supply module in communication. In the case where the actual power value of the computing system is greater than the first power threshold, if the actual power value of the computing device is greater than the first sub-power threshold, the second management unit is used to: under the control of the first management unit, control the overloaded server to draw power from the AC line and the energy storage device, wherein drawing power from the AC line includes drawing power from the first AC line, and / or drawing power from the second AC line through the power conversion device. Among them, the overloaded server is a server among the plurality of servers whose actual power value is greater than the second sub-power threshold, the product of the second sub-power threshold and the number of servers is equal to the first sub-power threshold, and the product of the first sub-power threshold and the number of computing devices in the computing system is equal to the first power threshold.

[0053] Among them, when the actual power value of the computing system is greater than the first power threshold, if the actual power value of the computing device is greater than the first sub-power threshold, the first management unit can adjust the power supply method of the overloaded server in the computing device, that is, adjust the power supply from the AC line to the AC line and the energy storage device to jointly power the overloaded computing device, so as to increase the maximum power of the entire cabinet of the computing device, so that more servers can be deployed in the computing device, thereby increasing the server rack rate in the computing device and reducing the vacancy rate of the cabinet without the need to renovate the computer room. In addition, adjusting the power supply method of the overloaded server can reduce the number of servers whose power supply method needs to be changed.

[0054] Based on the computing device provided in the third aspect, in a possible implementation, the server includes a first switching power supply, a second switching power supply and a plurality of loads, the first switching power supply and the second switching power supply are both connected to the plurality of loads, the plurality of loads include a management chip, the management chip is communicatively connected with the first switching power supply, the second switching power supply and the second management unit, the first switching power supply is a first sub-input terminal of the server, and the second switching power supply is a second sub-input terminal of the server. In a computing device, the plurality of servers include first-class servers and / or second-class servers, wherein the servers that can draw power from the first AC line are first-class servers, and the servers that cannot draw power from the first AC line are second-class servers.

[0055] When the actual power value is greater than the first power threshold, if the actual power value of the computing device is greater than the first sub-power threshold, the second management unit is used to: under the control of the first management unit, control the first switching power supply and the second switching power supply of the first type of server in the overloaded server through the management chip of the first type of server in the overloaded server to jointly power multiple loads of the first type of server in the overloaded server; and / or, control the second switching power supply of the second type of server in the overloaded server through the management chip of the second type of server in the overloaded server to power multiple loads of the first type of server in the overloaded server. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the present application, the following is a brief introduction to the drawings required for use in some embodiments of the present application. Obviously, the drawings described below are only drawings of some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings. In addition, the drawings described below can be regarded as schematic diagrams, and are not intended to limit the actual size of the product involved in the embodiments of the present application, the actual process of the method, the actual timing of the signal, etc.

[0057] Figure 1 A structural diagram of a data center provided for some embodiments of the present application;

[0058] Figure 2 for Figure 1 A block diagram of a computing device in ;

[0059] Figure 3 for Figure 1 The structural block diagram of the computing system in

[0060] Figure 4 for Figure 2 A graph showing how the power P of the computing device in the figure changes with time t;

[0061] Figure 5 is a block diagram of a computing system according to some embodiments;

[0062] Figure 6 is another block diagram of a computing system according to some embodiments;

[0063] Figure 7 for Figure 5 Another structural block diagram of the computing system in;

[0064] Figure 8 is another structural block diagram of a computing system according to some embodiments. DETAILED DESCRIPTION

[0065] The following will be combined with the accompanying drawings to clearly and completely describe the technical solutions in some embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments provided by the present disclosure, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present disclosure.

[0066] Unless the context requires otherwise, throughout the specification and claims, the term "comprise" and other forms thereof, such as the third person singular form "comprises" and the present participle form "comprising", are to be interpreted as open, inclusive, that is, "including, but not limited to". In the description of the specification, the terms "some embodiments", "example" or "some examples" and the like are intended to indicate that specific features, structures, materials or characteristics associated with the embodiment or example are included in at least one embodiment or example of the present disclosure. The schematic representation of the above terms does not necessarily refer to the same embodiment or example. In addition, the specific features, structures, materials or characteristics may be included in any one or more embodiments or examples in any appropriate manner.

[0067] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present disclosure, unless otherwise specified, "plurality" means two or more.

[0068] “A and / or B” includes the following three combinations: A only, B only, and a combination of A and B.

[0069] Figure 1 A structural diagram of a data center provided for some embodiments of the present application.

[0070] See also Figure 1 The embodiment of the present application provides a data center 2000. The data center 2000 is a global collaborative specific device network used to implement data transmission, acceleration, display, calculation and storage functions on the Internet (Chinese name: Internet) network infrastructure.

[0071] The data center 2000 may include a plurality of computing systems 4000 , and the computing systems 4000 may include a display cabinet 3000 and a plurality of computing devices 1000 , and the display cabinet 3000 and the plurality of computing devices 1000 may be displayed in a computer room.

[0072] The input end of a header cabinet 3000 can be connected to an AC line, and the output end of the header cabinet 3000 is connected to the input ends of multiple computing devices 1000, wherein the header cabinet 3000 can draw power from the AC line and distribute the electrical energy to the multiple computing devices 1000 electrically connected to it.

[0073] For example, the AC line may provide 220V AC power.

[0074] By way of example, the plurality of computing devices 1000 of the computing system 4000 may be arranged in a row. By way of example, the computing devices 1000 may be a whole rack server.

[0075] Figure 2 for Figure 1 A structural diagram of the computing device 1000 in FIG.

[0076] See also Figure 2 The computing device 1000 may include a cabinet 100 and a plurality of servers 200 , wherein input ends of the plurality of servers 200 may be electrically connected to output ends of a header cabinet 3000 , and the servers 200 may then draw power from an AC line through the header cabinet 3000 .

[0077] A plurality of servers 200 may be located in the cabinet 100, so that the cabinet 100 can protect the servers 200. The plurality of servers 200 may be stacked in the same cabinet 100 to improve the space utilization of the data center 2000 and reduce the space cost.

[0078] In some examples, in addition to the server 200, the computing device 1000 may also include a workstation, a memory, a switch or a repeater, etc.

[0079] Figure 3 for Figure 1 A structural block diagram of the computing system 4000 in FIG.

[0080] See also Figure 3 , the AC line may include a first AC line and a second AC line. The computing system 4000 includes two terminal cabinets 3000, and the computing device 1000 may include a first input terminal and a second input terminal. The input terminal of one terminal cabinet 3000 is connected to the first AC line, and the output terminal is connected to the first input terminals of the plurality of computing devices 1000, and the input terminal of the other terminal cabinet 3000 is connected to the second AC line, and the output terminal is connected to the second input terminals of the plurality of computing devices 1000.

[0081] The data center 2000 includes a plurality of computing systems 4000, and the plurality of computing systems 4000 can share a first AC line and a second AC line. When one of the first AC line and the second AC line cannot supply power to the computing device 1000, the other of the first AC line and the second AC line can supply power to the computing device 1000, thereby preventing the computing device 1000 from losing power.

[0082] For example, both the first AC line and the second AC line can provide 220V AC power.

[0083] The maximum power that the AC line can provide to the computing system 4000 is the AC maximum power (Pamx1), and the actual power of the AC line can be less than or equal to the AC maximum power.

[0084] The maximum power that the first AC line can provide is equal to the AC maximum power, the maximum power that the second AC line can provide is equal to the AC maximum power, and the maximum value of the sum of the powers that the first AC line and the second AC line can provide is the AC maximum power. For example, the actual power provided by the first AC line is W1, and the actual power provided by the second AC line is W2, and W1+W2=AC maximum power. When W1 is equal to the AC maximum power, W2 is equal to 0; when W1 is equal to 0, W2 is equal to the AC maximum power.

[0085] In the related art, the maximum power (Pmax) of the computing system 4000 may be equal to the AC maximum power, ie, Pmax = Pamx1. In the early design stage of the data center 2000, the AC maximum power has been determined, so the maximum power of the computing system 4000 has been determined in the early design stage of the data center 2000.

[0086] Among them, the power of the computing system 4000 may include the sum of the power of multiple computing devices 1000 and the power of the header cabinet 3000. In the embodiment of the present application, the power of the header cabinet 3000 is negligible. At this time, the power of the computing system 4000 may include the power of multiple computing devices 1000.

[0087] The following describes how to design the maximum power of the entire cabinet of the computing device 1000 based on the maximum AC power.

[0088] Figure 4 for Figure 2 FIG. 1 is a graph showing a change in power P of the computing device 1000 over time t.

[0089] See also Figure 4, the power of the computing device 1000 is related to the business run by the computing device 1000. When the computing device 1000 runs different businesses, the power of the computing device 1000 may be different. The power of the computing device 1000 may include the basic power of the whole cabinet (Ptyp2) and the maximum power of the whole cabinet (Pamx2). Among them, the basic power of the whole cabinet is the power when the computing device is running stably, and the maximum power of the whole cabinet is the power when the computing device 1000 is running at full load. When the computing device 1000 is running stably, the power of the computing device 1000 is less than or equal to the basic power of the whole cabinet. When the business of the computing device 1000 increases, the power of the computing device 1000 is greater than the basic power of the whole cabinet and less than the maximum power of the whole cabinet. When the business of the computing device 1000 reaches a peak, the power of the computing device 1000 can be equal to the maximum power of the whole cabinet.

[0090] The maximum power of the computing system 4000 may be equal to the maximum power of the entire cabinet of the computing device 1000 multiplied by the number of computing devices 1000 N1, that is, Pmax=Pamx2×N1. The maximum power of the computing system 4000 depends on the maximum AC power. Furthermore, when designing the maximum power of the entire cabinet of the computing device 1000, the maximum AC power may be referred to to prevent the sum of the maximum powers of the entire cabinet of multiple computing devices 1000 from exceeding the maximum AC power.

[0091] For example, the maximum power of the entire cabinet of computing devices 1000 may be equal to the maximum AC power divided by the number of computing devices 1000, that is, Pamx2 = Pamx1 ÷ N1.

[0092] For example, the maximum AC power is 50KW, that is, Pamx1 = 50KW, and the number of computing devices 1000 in the computing system 4000 can be 10, that is, N1 = 10, so the maximum power of the entire cabinet of the computing device 1000 = Pamx1 ÷ N1 = 50KW ÷ 10 = 5KW, that is, the maximum power of the entire cabinet of the computing device 1000 can be equal to 5KW, that is, Pamx2 = 5KW. At this time, the basic power of the entire cabinet of the computing device 1000 can be 4KW, Ptyp2 = 4KW.

[0093] Among them, the computing device 1000 includes multiple servers 200. The power of the server 200 is related to the business run by the server 200. When the business run by the server 200 is different, the power of the server 200 may be different. The power of the server 200 may include a single basic power (Ptyp3) and a single maximum power (Pamx3), wherein the single basic power is the power of the server 200 when it is running stably, and the single maximum power is the power of the server 200 when it is running at full load. When the server 200 is running stably, the power of the server 200 is less than or equal to the single basic power. When the business of the server 200 increases, the power of the server 200 is greater than the single basic power and less than the single maximum power. When the business of the server 200 reaches a peak, the power of the server 200 may be equal to the single maximum power.

[0094] For example, the basic power of a single server 200 may be equal to 357W, and the maximum power of a single server may be equal to 500W.

[0095] In the related art, the number of servers 200 in the computing device 1000 can be deployed according to the maximum power of the entire cabinet of the computing device 1000 and the maximum power of a single server 200 Pamx3. For example, the number of servers 200 (N2) can be equal to the maximum power of the entire cabinet of the computing device 1000 divided by the maximum power of a single server 200, that is, N2 = Pamx2 ÷ Pamx3. For example, the maximum power of the entire cabinet of the computing device 1000 is 5KW, and the maximum power of a single server 200 is 500W. At this time, N2 = Pamx2 ÷ Pamx3 = 5KW ÷ 500W = 10, and then, 10 servers 200 can be deployed in the computing device 1000.

[0096] As the computing power of the server 200 gradually increases, the maximum power of a single server 200 gradually increases. Therefore, the number of servers 200 that can be placed in the cabinet 100 of a computing device 1000 decreases, resulting in a high vacancy rate of the cabinet 100. If the number of servers 200 continues to increase in the cabinet 100, the maximum power of the entire cabinet of a single computing device 1000 will exceed the maximum power of the entire cabinet of the computing device 1000, and then the maximum power of the computing system 4000 will exceed the maximum AC power. At this time, the computer room needs to be renovated simultaneously. However, the transformation of the computing system 4000 is difficult and takes a long time. Therefore, the maximum power of the entire cabinet of a single computing device 1000 is difficult to increase, and it is difficult to solve the problem of a high vacancy rate of the cabinet 100.

[0097] In this embodiment, model energy consumption training can be performed on each server 200 in the computing device 1000 to count the historical power changes of each server 200, and the capped power of each server 200 can be corrected according to the actual power of each server 200. After correcting the capped power of each server 200, the power of the entire cabinet of the computing device 1000 has a margin, and then the number of servers 200 deployed in the computing device 1000 can be increased, thereby increasing the density of the servers 200.

[0098] Among them, the capped power of server 200 can be used to limit the actual power of server 200. After the capped power of server 200 is determined, the actual power of server 200 will not exceed the capped power. If the business of server 200 suddenly increases at a certain moment, causing the actual power of server 200 to increase, the power of at least part of the equipment in server 200 can be reduced, for example, by reducing the speed of the fan in server 200 to reduce the power of the fan, thereby reducing the actual power of server 200, and thus the actual power of server 200 is made below the capped power of server 200.

[0099] Table 1 is a comparison table of capped powers of each server 200 in the computing device 1000 before and after the capped powers are corrected.

[0100]

[0101] Table 1

[0102] Please refer to Table 1. For example, the computing device 1000 includes 10 servers 200. In Table 1, the 10 servers 200 are sequentially referred to by the labels U1, U2, U3, ..., U10. Before the capped power of the 10 servers 200 is corrected, the capped power of each server 200 is 1000W. At this time, the total power budget of the computing device 1000 is 10000W. The total power budget of the computing device 1000 can be understood as the maximum total power required by the computing device 1000 to ensure the normal operation of all power-consuming devices in the computing device 1000 (such as servers 200 and switches, etc.).

[0103] After the capped power of the ten servers 200 is corrected, the capped power of server U1 is 400 W, the capped power of server U2 is 800 W, the capped power of server U3 is 600 W, the capped power of server U4 is 1200 W, the capped power of server U5 is 800 W, the capped power of server U6 is 500 W, the capped power of server U7 is 1000 W, the capped power of server U8 is 400 W, the capped power of server U9 is 700 W, and the capped power of server U10 is 1100 W. At this time, the total power budget of the computing device 1000 is 7500 W.

[0104] Compared with before the power capping of server 200 was corrected, the total power budget of computing device 1000 was reduced by 2500W. Therefore, the computing device 1000 can have a power margin of 2500W, and then three more servers 200 can be added to the cabinet 100, thereby increasing the deployment density of server 200, reducing the vacancy rate of cabinet 100, and increasing the computing power of computing device 1000.

[0105] In some of the above embodiments, the computing device reduces the vacancy rate of the cabinet 100 by correcting the capped power of the computing device 1000 .

[0106] Next, some other embodiments of how to reduce the vacancy rate of the cabinet 100 are introduced.

[0107] Figure 5 is a block diagram of a computing system 4000 according to some embodiments.

[0108] See also Figure 5 , an embodiment of the present application provides a computing system 4000. The computing system 4000 may include an energy storage device 300, a first management unit 400, a power conversion device 500, a plurality of computing devices 1000, a first power distribution device 3001 and a second power distribution device 3002. The number of computing devices 1000 may be 2, 3 or even more. In some embodiments of the present application, the number of computing devices 1000 is not limited. Figure 5 In the embodiment of the present application, five computing devices 1000 are taken as an example. It can be understood that the energy storage device 300, the first management unit 400, the power conversion device 500, the first power distribution device 3001 and the second power distribution device 3002 constitute a power supply module to supply power to multiple computing devices 1000. Of course, in other embodiments, the power supply module can also only include the energy storage device 300, the first management unit 400, and the power conversion device 500.

[0109] The input end of the power conversion device 500 is connected to the second AC line.

[0110] An input end of the energy storage device 300 is connected to an output end of the power conversion device 500 .

[0111] Each computing device 1000 includes a first input terminal and a second input terminal. The first input terminal of each computing device 1000 is connected to the first AC line, and the second input terminal of each computing device 1000 is connected to the output terminal of the power conversion device 500 and the output terminal of the energy storage device 300 .

[0112] The first management unit 400 may be connected to the output terminal of the power conversion device 500 , so that the power conversion device 500 may provide power for the first management unit 400 .

[0113] In some examples, the first input terminal of each computing device 1000 may be directly connected to the first AC line. In other examples, the first input terminal of each computing device 1000 may be connected to the first AC line through the first power distribution device 3001. In this case, the input terminal of the first power distribution device 3001 is connected to the first AC line, and the output terminal of the first power distribution device 3001 is connected to the first input terminal of each computing device 1000. For example, the first power distribution device 3001 may be a terminal cabinet 3000.

[0114] The first AC line can supply power to multiple computing devices 1000 through the first power distribution device 3001 .

[0115] In some examples, the second input terminal of each computing device 1000 may be directly connected to the output terminal of the energy storage device 300 and the output terminal of the power conversion device 500, and in other examples, the second input terminal of each computing device 1000 may be connected to the output terminal of the energy storage device 300 through the second power distribution device 3002. In this case, the input terminal of the second power distribution device 3002 may be connected to the output terminal of the power conversion device 500 and the output terminal of the energy storage device 300, and the output terminal of the second power distribution device 3002 may be connected to the second input terminal of each computing device 1000. For example, the second power distribution device 3002 may be a cabinet 3000.

[0116] The computing device 1000 may include multiple servers 200, and the servers 200 include a first sub-input terminal and a second sub-input terminal, wherein the first input terminals of the multiple computing devices 1000 are connected to the first sub-input terminals of the multiple servers 200. The second input terminals of the multiple computing devices 1000 are connected to the second sub-input terminals of the multiple servers 200. The first sub-input terminals of the multiple servers 200 can all be electrically connected to the first AC line through the first power distribution device 3001, so the first AC line can supply power to the multiple servers 200 through the first power distribution device 3001. The second sub-input terminals of the multiple servers 200 are electrically connected to the output terminal of the energy storage device 300 and the output terminal of the power conversion device 500 through the second power distribution device 3002.

[0117] The input end of the power conversion device 500 is connected to the second AC line. The power conversion device 500 can convert the AC signal received at the input end into a DC signal, and the DC signal can be output from the output end.

[0118] For example, the power conversion device 500 may be an alternating current-to-direct current (AC / DC) converter.

[0119] The power conversion device 500 can provide 180V-400V DC power. For example, the voltage of the DC power signal provided by the power conversion device 500 can be 240V. In addition, the voltage of the DC power signal provided by the power conversion device 500 can be 48V.

[0120] The output end of the power conversion device 500 can be electrically connected to the input end of the energy storage device 300 , so that the power conversion device 500 can charge the energy storage device 300 .

[0121] For example, the energy storage device 300 can provide a direct current of 240 V. For example, the energy storage device 300 can include a battery.

[0122] The energy storage device 300 may be arranged outside the cabinet 100. The energy storage device 300 may include a plurality of batteries, and the plurality of batteries may be centrally placed in the machine room, which is conducive to centralized management and fire protection of the plurality of batteries.

[0123] In addition, the output end of the power conversion device 500 can also be electrically connected to the first management unit 400, so that the first management unit 400 can supply power to the first management unit 400. In some other examples, the first management unit 400 can be connected to the output end of the energy storage device 300, so that the energy storage device 300 supplies power to the first management unit 400.

[0124] In some examples, the first management unit 400 may be disposed in a housing of the power conversion device 500 , that is, the power conversion device 500 may include the first management unit 400 .

[0125] The output end of the power conversion device 500 may also be electrically connected to the second input end of the computing device 1000 , so that the DC signal generated by the power conversion device 500 may be input into the computing device 1000 to provide power for the computing device 1000 .

[0126] In some examples, the first management unit 400 may also be communicatively connected with the energy storage device 300 , the power conversion device 500 , and a plurality of computing devices 1000 , and the first management unit 400 may control the power conversion device 500 to charge the energy storage device 300 .

[0127] For example, the first management unit 400 and the energy storage device 300 may be communicatively connected via an I2C (Inter-Integrated Circuit) bus or a CAN (Controller Area Network) bus.

[0128] For example, the first management unit 400 and the power conversion device 500 may be communicatively connected via an I2C bus or a CAN bus.

[0129] The first management unit 400 is also electrically connected to the plurality of computing devices 1000, and the first management unit 400 can send management signals to the plurality of computing devices 1000. The first management unit 400 can include a first control chip, and the first control chip can include, for example, any one of control chips having management control functions, such as a BMC and an FPGA (Field-Programmable Gate Array), or a microcontroller unit.

[0130] In this embodiment, the first management unit 400 is used to control the multiple computing devices 1000 to draw power from the AC line when the actual power value (P) of the computing system 4000 is less than or equal to the first power threshold (P1). For example, the first power threshold may be equal to the maximum AC power, that is, P1 = Pamx1. At this time, the multiple computing devices 1000 all draw power from the AC line and will not exceed the maximum AC power, thereby ensuring the normal operation of the computing devices 1000.

[0131] The first management unit 400 can also be used to: when the actual power value of the computing system 4000 is greater than the first power threshold, control multiple computing devices 1000 to draw power from the AC line and the energy storage device 300. At this time, the AC line and the energy storage device 300 jointly power the multiple computing devices 1000.

[0132] Wherein, taking power from the AC line includes taking power from a first AC line and / or taking power from a second AC line through a power conversion device 500 .

[0133] When the computing device 1000 obtains power from the second AC line through the power conversion device 500 , the power conversion device 500 converts the AC power signal obtained from the first AC line into a DC power signal and outputs it to the second input terminal of the computing device 1000 .

[0134] The above-mentioned multiple computing devices 1000 draw power from the AC line and the energy storage device 300, including: multiple computing devices 1000 all draw power from the AC line, and some computing devices 1000 draw power from the energy storage device 300; or, each computing device 1000 draws power from the AC line and the energy storage device 300.

[0135] When multiple computing devices 1000 draw power from the AC line, the following scenarios may be included: for example, multiple computing devices 1000 all draw power from the first AC line. For another example, multiple computing devices 1000 all draw power from the second AC line through the power conversion device 500. For another example, some computing devices 1000 draw power from the first AC line, while the remaining computing devices 1000 draw power from the second AC line through the power conversion device 500. For another example, all computing devices 1000 draw power from the first AC line and also draw power from the second AC line through the power conversion device 500. For another example, some computing devices 1000 draw power from the first AC line; the remaining computing devices 1000 draw power from both the first AC line and the second AC line through the power conversion device 500.

[0136] In which, when the first AC line is normal and the first sub-input terminals of all servers 200 in all computing devices 1000 can draw power from the first AC line, multiple computing devices 1000 draw power from the first AC line, and at least some computing devices 1000 draw power from the energy storage device 300. At this time, the AC signal provided by the first AC line can enter the computing device 1000 from the first input terminal of the computing device 1000, and the DC signal provided by the energy storage device 300 can enter the computing device 1000 from the second input terminal of the computing device 1000.

[0137] When the first AC line is abnormal, but the second AC line is normal, and the second sub-input terminals of all servers 200 in all computing devices 1000 can draw power from the second AC line through the power conversion device 500, multiple computing devices 1000 draw power from the second AC line, and at least some computing devices 1000 draw power from the energy storage device 300. The second input terminals of at least some computing devices 1000 simultaneously receive the DC power output by the power conversion device 500 and the DC power output by the energy storage device 300.

[0138] When the first AC line and the second AC line are both normal, and the first sub-input terminal of at least some servers 200 cannot draw power from the first AC line, the second sub-input terminal of the server 200 with abnormal first sub-input terminal can draw power from the second AC line through the power conversion device 500, and the first sub-input terminal of the server 200 with normal first sub-input terminal can draw power from the first AC line. In this paragraph, "at least some servers 200" can be located in at least one computing device 1000. And at least some of the multiple computing devices 1000 can draw power from the energy storage device 300. Among them, the abnormal first sub-input terminal means that the first sub-input terminal of the server 200 cannot draw power from the first AC line, and the normal first sub-input terminal means that the first sub-input terminal of the server 200 can draw power from the first AC line.

[0139] In some examples, the actual power value of the computing system 4000 may be the sum of the power of all power-consuming devices in the computing system 4000 except the power conversion device 500 and the energy storage device 300. For example, the actual power value of the computing system 4000 may be the sum of the power of multiple computing devices 1000, the power of the first management unit 400, the power of the first power distribution device 3001, and the power of the second power distribution device 3002. In some other examples, the power of the first power distribution device 3001, the second power distribution device 3002, and the first management unit 400 may be ignored. In this case, the actual power value of the computing system 4000 may be the sum of the actual power of multiple computing devices 1000.

[0140] When the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 can control at least some of the computing devices 1000 to draw power from the energy storage device 300 in addition to controlling the multiple computing devices 1000 to draw power from the AC line. At this time, the energy storage device 300 can be in a discharging state and can provide power for the computing device 1000.

[0141] For example, the first management unit 400 controls the energy storage device 300 to supply power to the computing device 1000 in the following manner: the first management unit 400 can control the output voltage of the energy storage device 300 to be the same as or similar to the output voltage of the power conversion device 500, so that the energy storage device 300 and the power conversion device 500 can jointly supply power to the computing device. When the output voltage of the energy storage device 300 and the output voltage of the power conversion device 500 are similar, the voltage difference between the output voltage of the energy storage device 300 and the output voltage of the power conversion device 500 is less than or equal to a preset value, and as long as the voltage difference between the output voltage of the energy storage device 300 and the output voltage of the power conversion device 500 is less than or equal to the preset value, the energy storage device 300 and the power conversion device 500 can jointly supply power to the computing device 1000.

[0142] At this time, the maximum power of the computing system 4000 can be equal to the sum of the AC maximum power and the power that the energy storage device 300 can provide to the computing system 4000. Therefore, the maximum power of the computing system 4000 is greater than the AC maximum power, that is, Pmax>Pamx1. In the related art, the maximum power of the computing system 4000 is equal to the AC maximum power, that is, Pmax=Pamx1. In summary, the embodiment of the present application can increase the maximum power of the computing system 4000 by setting the energy storage device 300 in the computing system 4000. The computing system 4000 includes multiple computing devices 1000. On the basis of increasing the maximum power of the computing system 4000, the maximum power of the entire cabinet of the computing device 1000 can be increased, so that more servers 200 can be deployed in the computing device 1000, thereby increasing the shelf rate of the server 200 in the computing device 1000 and reducing the vacancy rate of the cabinet 100 without the need to renovate the computer room.

[0143] Please continue reading Figure 5 In this embodiment, when the actual power value of the computing system 4000 is less than or equal to the first power threshold, the first management unit 400 can control multiple computing devices 1000 to draw power from the AC line. In this case, the sum of the AC power obtained by multiple computing devices 1000 from the AC line is the actual power value of the computing system 4000. Among them, the actual power value of the computing system 4000 is equal to the sum of the power (Pac1) provided by the first AC line and the power (Pac2) provided by the second AC line, that is, P = Pac1 + Pac2. At this time, Pac1 + Pac2 < Pamx1 or Pac1 + Pac2 = Pamx1, so that when the AC line supplies power to multiple computing devices 1000, the sum of the actual power values ​​of the multiple computing devices 1000 exceeds the maximum AC power, causing the computing devices 1000 to lose power.

[0144] In this embodiment, the first management unit 400 is also used to: when the actual power value of the computing system 4000 is greater than the first power threshold, control the AC power obtained by multiple computing devices 1000 from the AC line so that the power loaded by the AC line is the first power threshold, and the power loaded by the energy storage device 300 is the power difference between the actual power value of the computing system 4000 and the first power threshold.

[0145] Among them, the AC power obtained by multiple computing devices 1000 from the AC line can be the sum of the powers provided by the AC line to the multiple computing devices 1000, and the AC power obtained by multiple computing devices 1000 from the AC line can be equal to the first power threshold, and is equal to the sum of the power (Pac1) provided by the first AC line to the multiple computing devices 1000 and the power (Pac2) provided by the second AC line to the multiple computing devices 1000 through the power conversion device 500, that is, P1=Pac1+Pac2.

[0146] The power carried by the energy storage device 300 may be the sum of the powers provided by the energy storage device 300 to the plurality of computing devices 1000. In this case, the power carried by the energy storage device 300 is equal to P-P1.

[0147] In some examples, the first power threshold may be equal to the AC maximum power. In the embodiment of the present application, when designing the whole cabinet basic power of the computing device 1000, the first power threshold (ie, the AC maximum power) may be referred to.

[0148] For example, the basic power of the entire cabinet of the computing device 1000 may be equal to the first power threshold (ie, the maximum AC power) divided by the number of computing devices 1000 (N1). In this case, Ptyp2 = Pamx1 ÷ N1.

[0149] In the related art, the maximum power of the whole cabinet of the computing device 1000 can be equal to the maximum AC power divided by the number of computing devices 1000, that is, Pamx2 = Pamx1 ÷ N1. Furthermore, the maximum power of the whole cabinet of the computing device 1000 in the related art can be equal to the basic power of the whole cabinet of the computing device 1000 in the embodiment of the present application.

[0150] In the related art and the embodiments of the present application, the maximum power of the entire cabinet of the computing device 1000 is greater than the basic power of the entire cabinet of the computing device 1000. Therefore, the embodiments of the present application can increase the maximum power of the entire cabinet of the computing device 1000.

[0151] For example, the first power threshold (i.e., the maximum AC power) is 50KW, i.e., P1=Pmax1=50KW, and the computing system 4000 may include 10 computing devices 1000, i.e., N1=10. Therefore, the basic power of the entire cabinet of the computing device 1000 = P1÷N1=50KW÷10=5KW, i.e., Ptyp2=5KW. Therefore, the maximum power of the entire cabinet of the computing device 1000 is greater than 5KW. In the related art, the maximum power of the entire cabinet of the computing device 1000 = Pamx1÷N1=50KW÷10=5KW, i.e., Pamx2=5KW. Therefore, the embodiment of the present application can increase the maximum power of the entire cabinet of the computing device 1000.

[0152] The number of servers 200 in the computing device 1000 can refer to the maximum power of the entire cabinet of the computing device 1000 and the maximum power of a single server 200. At this time, the number of servers 200 (N2) can be equal to the maximum power of the entire cabinet of the computing device 1000 divided by the maximum power of a single server 200, that is, N2 = Pamx2 ÷ Pamx3. Therefore, when the maximum power of the entire cabinet of the computing device 1000 is increased, the number of servers 200 deployed in the computing device 1000 can be increased, thereby improving the shelf rate of the servers 200.

[0153] For example, when the basic power of the entire cabinet of the computing device 1000 is equal to 5KW, that is, Ptyp2=5KW, the maximum power of the entire cabinet of the computing device 1000 can be designed to be 7KW, that is, Pamx2=7KW. And the basic power of a single server 200 is 357W, that is, Ptyp3=357W, and the maximum power of a single server is 500W, that is, Pamx3=500W. At this time, the number of servers 200 = the maximum power of the entire cabinet of the computing device 1000 ÷ the maximum power of a single server 200, that is, N2=Pamx2÷Pamx3=7KW÷500W=14. In the related art, the maximum power of the entire cabinet of the computing device 1000 is equal to 5KW. At this time, the number of servers 200 in the computing device 1000 can be equal to the maximum power of the entire cabinet of the computing device 1000 divided by the maximum power of a single server, which is 500W, that is, N2=5KW÷500W=10. In the embodiment of the present application, 14 servers 200 can be set in the computing device 1000. Therefore, the embodiment of the present application can increase the number of servers 200 deployed in the computing device 1000, thereby improving the shelf rate of the servers 200.

[0154] In addition, when the power of all computing devices 1000 reaches the maximum power of the entire cabinet, at this time, the maximum power of the entire cabinet multiplied by the number of computing devices 1000 will be greater than the first power threshold (i.e., the maximum AC power). In an embodiment of the present application, when the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 can control multiple computing devices 1000 to draw power from the energy storage device 300, thereby avoiding power failure of the computing devices 1000 in the computing system 4000.

[0155] For example, the maximum power of the whole cabinet of computing device 1000 is equal to 7KW, that is, Pamx2=7KW, and the computing system 4000 includes 10 computing devices 1000. When the actual power of multiple computing devices 1000 reaches the maximum power of the whole cabinet of computing device 1000, at this time, the actual power value of computing system 4000 is equal to the product of the maximum power of the whole cabinet of computing device 1000 and the number of computing devices 1000, that is, P=Pamx2×N1=7KW×10=70KW, where 70KW is greater than 50W, that is, 70W is greater than the first power threshold (that is, the maximum AC power). At this time, the part exceeding the first power threshold (that is, the maximum AC power) can be provided by the energy storage device 300, that is, the energy storage device 300 can provide 20KW of power.

[0156] In summary, if the actual power is greater than the first power threshold, the first management unit 400 controls the AC power obtained by the multiple computing devices 1000 from the AC line so that the power loaded by the AC line is the first power threshold, and the power loaded by the energy storage device 300 is the power difference between the actual power value of the computing device 1000 and the first power threshold. Among them, the energy storage device 300 can load the part of the actual power value of the computing system 4000 that exceeds the first power threshold, and then can increase the maximum power value of the computing system 4000, thereby increasing the shelf rate of the computing system 4000 and reducing the vacancy rate of the cabinet 100.

[0157] Please continue reading Figure 5 In this embodiment, the first management unit 400 is also used to: obtain the actual power value (P2) of each computing device 1000, and can determine the overloaded computing device 1001 and the steady-state computing device 1002 among the multiple computing devices 1000 according to the actual power value of the computing device 1000.

[0158] The computing device 1000 may be divided into an overloaded computing device 1001 and a steady-state computing device 1002 according to the actual power value of the computing device 1000. The overloaded computing device 1001 is a computing device whose power is greater than a first sub-power threshold among the multiple computing devices 1000, and the steady-state computing device 1002 is a computing device whose power is less than or equal to the first sub-power threshold among the multiple computing devices 1000, and the product of the first sub-power threshold and the number of computing devices 1000 is equal to the first power threshold.

[0159] In some examples, the actual power value of the computing device 1000 may be equal to the sum of the powers of all power-consuming devices in the computing device 1000, for example, it may be equal to the sum of the powers of multiple servers 200, the powers of workstations, the powers of memories, the powers of switches, and the powers of repeaters. In other examples, except for the server 200, the powers of switches, memories, and other power-consuming devices may be ignored, in which case the actual power value P2 of the computing device 1000 may be equal to the sum of the powers of multiple servers 200 in the computing device 1000.

[0160] In some examples, the first sub-power threshold may be equal to the entire cabinet base power of the computing device 1000 .

[0161] In the case where the actual power value of the computing system 4000 is less than or equal to the first power threshold, at least some of the multiple computing devices 1000 are steady-state computing devices 1002. At this time, the multiple computing devices 1000 may be steady-state computing devices 1002, or some of the computing devices 1000 are overloaded computing devices 1001, while the remaining computing devices 1000 are steady-state computing devices 1002. At this time, the steady-state computing devices 1002 and the overloaded computing devices 1001 in the multiple computing devices 1000 both draw power from the AC line.

[0162] When the actual power value of the computing system 4000 is greater than the first power threshold, at least some of the computing devices 1000 are overloaded computing devices 1001. At this time, all computing devices 1000 may be overloaded computing devices 1001; or, some of the computing devices 1000 are overloaded computing devices 1001, and the rest of the computing devices 1000 are stable computing devices 1002.

[0163] The first management unit 400 is also used to: when the actual power value of the computing system 4000 is greater than the first power threshold, control the overloaded computing device 1001 among the multiple computing devices 1000 to draw power from the AC line and the energy storage device 300, so that the AC line and the energy storage device 300 jointly power the overloaded computing device 1001.

[0164] Among them, when the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 can adjust the power supply mode of the overloaded computing device 1001, that is, adjust the power supply from the AC line to the AC line and the energy storage device 300 jointly powering the overloaded computing device 1001, thereby reducing the number of computing devices 1000 whose power supply mode is changed.

[0165] When the actual power value of the computing system 4000 is greater than the first power threshold, the power provided by the AC line to the overloaded computing device 1001 may be greater than, less than or equal to the first sub-power threshold. In the embodiment of the present application, the power provided by the AC line to the overloaded computing device 1001 is not limited.

[0166] When the actual power value of the computing system 4000 is greater than the first power threshold, if the computing system 4000 also includes a steady-state computing device 1002, the first management unit 400 can control the steady-state computing device 1002 in the plurality of computing devices 1000 to draw power from the AC line. At this time, all servers 200 in the steady-state computing device 1002 can draw power from the AC line.

[0167] The first management unit 400 does not adjust the power supply mode of the steady-state computing device 1002. At this time, the steady-state computing device 1002 is still powered by the AC line. In this way, when the actual power value of the computing system 4000 is greater than the first power threshold, the number of computing devices 1000 whose power supply mode is changed can be reduced.

[0168] Figure 6 is another structural block diagram of a computing system 4000 according to some embodiments.

[0169] See also Figure 6 In this embodiment, the computing system 4000 further includes a switch 800, and the computing device 1000 may further include a second management unit 700. The second management unit 700 is disposed in the cabinet 100, and the second management unit 700 may be communicatively connected to the first management unit 400, and the second management unit 700 may be communicatively connected to a plurality of servers 200 in the computing device 1000 where the second management unit 700 is located.

[0170] For example, the second management unit 700 may include a second control chip, which may include, for example, any one of a BMC, an FPGA (Field-Programmable Gate Array), or a microcontroller unit, etc., which have management and control functions.

[0171] Among them, the second management unit 700 is used to obtain the actual power value of the computing device 1000 where it is located, and send the actual power value of the computing device 1000 to the first management unit 400, and the first management unit 400 can obtain the actual power value of the computing system 4000 based on the actual power values ​​of multiple computing devices 1000.

[0172] For example, the actual power value of the computing system 4000 may be equal to the sum of the actual power values ​​of the plurality of computing devices 1000 .

[0173] In this embodiment, the second management unit 700 may be electrically connected to a plurality of servers 200 in the computing device 1000 where the second management unit 700 is located. The second management unit 700 is used to obtain the actual power value of each server 200, and determine the actual power value of the computing device 1000 where the second management unit 700 is located according to the actual power value of each server 200.

[0174] For example, the actual power value of the computing device 1000 is equal to the sum of the actual power values ​​of the multiple servers 200 in the computing device 1000.

[0175] Figure 7 for Figure 5 Another structural block diagram of the computing system 4000 in FIG. Figure 7 Only one computing device 1000 is illustrated, and of course more computing devices 2000 may be included in the computing system 4000. Figure 7 The computing device 1000 shown in FIG. 1 overloads the computing device 1001 .

[0176] See also Figure 7 The second management unit 700 can also determine whether the server 200 is an overloaded server 201 or a steady-state server 202 according to the power of the server 200, wherein the overloaded server 201 is a server whose actual power value is greater than the second sub-power threshold, and the steady-state server 202 is a server whose actual power value is less than or equal to the second sub-power threshold, wherein the product of the second sub-power threshold and the number of servers 200 in a computing device 1000 is equal to the first sub-power threshold.

[0177] For example, the second sub-power threshold may be equal to a single basic power Ptyp3 of the server 200 .

[0178] Among them, when the actual power value of the computing system 4000 is less than or equal to the first power threshold, each computing device 1000 may include an overloaded server 201 and / or a steady-state server 202. At this time, the overloaded server 201 and the steady-state server 202 both draw power from the AC line.

[0179] When the actual power value of the computing system 4000 is greater than the first power threshold, at least some of the servers 200 in the overload computing device 1001 are overloaded servers 201. When some of the servers 200 are overloaded servers 201, the remaining servers 200 may be steady-state servers 202.

[0180] In this embodiment, the first management unit 400 is further used to: when the actual power value of the computing system 4000 is greater than the first power threshold, control the overload server 201 in the overload computing device 1001 to draw power from the AC line and the energy storage device 300 through the second management unit 700 of the overload computing device 1001. At this time, the second management unit 700 can control the overload server 201 to draw power from the AC line and the energy storage device 300 under the control of the first management unit 400.

[0181] Among them, when the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 can adjust the power supply mode of the overloaded server 201 in the overloaded computing device 1001, that is, adjust the power supply from the AC line to the AC line and the energy storage device 300 jointly powering the overloaded computing device 1001, thereby reducing the number of servers 200 whose power supply mode is changed.

[0182] In addition to the overloaded server 201, if the overloaded computing device 1001 also includes a steady-state server 202, the first management unit 400 is further configured to: when the actual power value of the computing system 4000 is greater than the first power threshold, control the steady-state server 202 in the overloaded computing device 1001 to draw power from the AC line through the second management unit 700 of the overloaded computing device 1001.

[0183] The power supply mode of the steady-state servers 200 in the overloaded computing device 1001 is not changed, so the number of servers 200 whose power supply modes are changed can be reduced.

[0184] Please continue reading Figure 7 In this embodiment, the computing device 1000 may further include two power distribution units (PDU) 600 . The server 200 may include a first switching power supply 210 , a second switching power supply 220 , and a plurality of loads 230 .

[0185] For example, the multiple loads 230 in the server 200 may include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a hard disk, a fan, a network card, a PCIE (Peripheral Component Interconnect Express) card, a CPLD (Complex Programmable Logic Device), a BMC (Baseboard Manager Controller) and other devices that require power supply, which are not listed here one by one.

[0186] It should be noted that different computing devices have different names for BMC, for example, some computing devices are called BMC, some computing devices are called iLO (Integrated lights-out), and some computing devices are called Integrated Dell Remote Access Controller (iDRAC). Whether it is called BMC, iLO or iDRAC, it can be understood as the BMC in the embodiment of the present invention.

[0187] Among them, the output end of the first power distribution device 3001 is electrically connected to a power distribution unit 600 in the computing device 1000 through the first input end of the computing device 1000, and the power distribution unit 600 is electrically connected to the input end of the first switching power supply 210 of multiple servers 200 in the computing device 1000, and the output end of the first switching power supply 210 is electrically connected to multiple loads 230. In this way, the AC signal provided by the first AC line can be transmitted to the first switching power supplies of multiple servers 200 through the first power distribution device 3001 and the power distribution unit 600. The first switching power supply 210 can convert the AC signal it receives into a DC signal to power multiple loads 230. For example, the voltage of the DC signal generated by the first switching power supply 210 can be 12V.

[0188] The first switch power supply 210 can be used as the first sub-input terminal of the server 200 where the first switch power supply 210 is located. The abnormality of the first sub-input terminal of the server 200 may include an abnormality of the first switch power supply 210 or an abnormality between the first switch power supply 210 and the first AC line.

[0189] The output end of the second power distribution device 3002 is electrically connected to another power distribution unit 600 through the second input end of the computing device 1000, and the power distribution unit 600 is electrically connected to the input end of the second switching power supply 220 of the plurality of servers 200 in the computing device 1000. The output end of the second switching power supply 220 is electrically connected to the plurality of loads 230.

[0190] Among them, the second switching power supply 220 can be electrically connected to the output end of the energy storage device 300 and the output end of the power conversion device 500 through the above-mentioned another power distribution unit 600 and the second power distribution device 3002. Therefore, the DC power signal provided by the energy storage device 300 can be transmitted to the second switching power supply 220 of the server 200 through the second power distribution device 3002 and the above-mentioned another power distribution unit 600. The AC power signal provided by the second AC line can be converted into a DC power signal by the power conversion device 500, and can be transmitted to the second switching power supply 220 of the server 200 through the second power distribution device 3002 and the above-mentioned another power distribution unit 600.

[0191] The second switching power supply 220 can regulate the voltage of the DC power signal provided by the energy storage device 300 and the DC power signal provided by the power conversion device 500, and output the regulated DC power signal to the multiple loads 230, thereby supplying power to the multiple loads 230. For example, the voltage of the DC power signal generated by the second switching power supply 220 can be 12V.

[0192] The second switch power supply 220 can be used as the second sub-input terminal of the server 200 where the second switch power supply 220 is located. The abnormality of the second sub-input terminal of the server 200 may include an abnormality of the second switch power supply 220 or an abnormality between the second switch power supply 220 and the second AC line.

[0193] Each power distribution unit 600 is provided with a plurality of sockets, and the input end of the first switching power supply 210 and the input end of the second switching power supply 220 of the server 200 can be plugged into the corresponding sockets on the power distribution unit 600 through cables.

[0194] The plurality of loads 230 may include a management chip 231. The management chip 231 may be communicatively connected with the first switching power supply 210 and the second switching power supply 220.

[0195] For example, the management chip 231 may be a BMC.

[0196] In some examples, the management chip 231 may obtain the input power of the first switching power supply 210 and the input power of the second switching power supply 220, and determine the actual power value of the server 200 where the management chip 231 is located according to the input power of the first switching power supply 210 and the input power of the second switching power supply 220. For example, the actual power value of the server 200 may be equal to the sum of the input power of the first switching power supply 210 and the input power of the second switching power supply 220.

[0197] Next, the manner in which the management chip 231 obtains the input power of the first switching power supply 210 and the input power of the second switching power supply 220 is introduced.

[0198] Please continue reading Figure 7 The first switching power supply 210 includes a first processing chip 211 , which is electrically connected to the management chip 231 . The first processing chip 211 is used to obtain input power of the first switching power supply 210 and send it to the management chip 231 .

[0199] The second switching power supply 220 includes a second processing chip 221 . The second processing chip 221 is electrically connected to the management chip 231 . The second processing chip 221 is used to obtain input power of the second switching power supply 220 and send it to the management chip 231 .

[0200] For example, the first processing chip 211 and the second processing chip 221 may both be DSPs (Digital Signal Processors).

[0201] The management chip 231 is used to determine the actual power value of the server 200 according to the input power of the first switching power supply 210 and the input power of the second switching power supply 220 , and send the actual power value to the second management unit 700 .

[0202] For example, the actual power value of the server 200 is equal to the sum of the input power of the first switching power supply 210 and the power of the second switching power supply 220 .

[0203] The second management unit 700 is used to determine the actual power value P2 of the computing device 1000 according to the actual power values ​​of the multiple servers 200 , and send it to the first management unit 400 .

[0204] The first management unit 400 is used to determine the actual power value of the computing system 4000 according to the actual power values ​​of the plurality of computing devices 1000 .

[0205] In addition to obtaining the actual power value of the server 200 , the management chip 231 can also control the first switch power supply 210 to be a main power supply and control the second switch power supply 220 to be a backup power supply.

[0206] For example, the management chip 231 is used to detect the input voltage of the first switching power supply 210 and the input voltage of the second switching power supply 220, and set the switching power supply whose input voltage is an AC voltage as the main power supply, and set the switching power supply whose input voltage is a DC voltage as the backup power supply. In an embodiment of the present application, the first switching power supply 210 is electrically connected to the first AC line through the first power distribution device 3001, and the input voltage of the first switching power supply 210 is an AC voltage. The second switching power supply 220 is electrically connected to the energy storage device 300 and the power conversion device 500 through the first power distribution device 3001, and the voltages input from the energy storage device 300 and the power conversion device 500 to the second switching power supply 220 are both DC voltages, so the input voltage of the second switching power supply 220 is a DC voltage. Then the management chip 231 controls the first switching power supply 210 as the main power supply and controls the second switching power supply 220 as the backup power supply.

[0207] When the actual power value of the computing system 4000 is less than or equal to the first power threshold, if the first switching power supply 210 can draw power from the first AC line, the management chip 231 can control the first switching power supply 210 to supply power to multiple loads 230, and the second switching power supply 220 does not supply power to multiple loads 230.

[0208] For example, the management chip 231 can control the output voltage of the first switching power supply 210 (i.e., the main power supply) to be greater than the output voltage of the second switching power supply 220 (i.e., the backup power supply), so that the first switching power supply 210 supplies power to multiple loads 230. At this time, the power provided by the second switching power supply 220 can be 0. Of course, in other embodiments, the first switching power supply 210 can be controlled to supply power to multiple loads 230, while the second switching power supply 220 does not supply power to multiple loads 230 in other ways. The embodiment of the present application does not limit the method of controlling the first switching power supply 210 to supply power to multiple loads 230, while the second switching power supply 220 does not supply power to multiple loads 230. As long as the first switching power supply 210 can supply power to multiple loads 230, while the second switching power supply 220 does not supply power to multiple loads 230, any method is acceptable.

[0209] If the first switching power supply 210 cannot draw power from the first AC line, that is, the first switching power supply 210 cannot supply power to the multiple loads 230 , the management chip 231 can control the second switching power supply 220 to supply power to the multiple loads 230 .

[0210] When the first AC line is abnormal, or when the first switch power supply 210 is abnormal, the first switch power supply 210 cannot draw power from the second AC line.

[0211] In this embodiment, the first management unit 400 is further used to: when the actual power value is less than or equal to the first power threshold, control the first switching power supply 210 or the second switching power supply 220 to power multiple loads 230 through the second management unit 700 and the management chip 231.

[0212] At this time, the first switching power supply 210 can obtain power from the first AC line, and the second switching power supply 220 can obtain power from the second AC line through the power conversion device 500 .

[0213] Please continue reading Figure 7 , when the actual power value of the computing system 4000 is greater than the first power threshold, if the overloaded server 201 in the overloaded computing device 1001 can draw power from the first AC line.

[0214] The first management unit 400 can be used to: when the actual power value of the computing system 4000 is greater than the first power threshold, control the first switching power supply 210 and the second switching power supply 220 of the overloaded server 201 in the overloaded computing device 1001 through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 of the overloaded server 201 in the overloaded computing device 1001 to jointly power multiple loads 230 in the overloaded server 201 in the overloaded computing device 1001.

[0215] At this time, in the overload server 201 of the overload computing device 1001, the first switching power supply 210 can draw power from the first AC line, and the second switching power supply 220 can draw power from the energy storage device 300, so as to supply power to multiple loads 230 in the overload server 201. At this time, the first AC line and the energy storage device 300 jointly supply power to the overload server 201.

[0216] In addition, if the overloaded server 201 in the overloaded computing device 1001 cannot draw power from the first AC line, the first management unit 400 can also be used to: control the second switch power supply 220 of the overloaded server 201 in the overloaded computing device 1001 to supply power to the multiple loads 230 in the overloaded server 201 through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 of the overloaded server 201 in the overloaded computing device 1001.

[0217] At this time, the first switching power supply 210 of the overloaded server 201 of the overloaded computing device 1001 cannot draw power from the first AC line, while the second switching power supply 220 can draw power from the second AC line through the power conversion device 500, or from the energy storage device 300. Therefore, the energy storage device 300 and the second AC line can jointly supply power to the overloaded server 201.

[0218] When the actual power value of the computing system 4000 is greater than the first power threshold, in addition to the overloaded server 201 , the overloaded computing device 1001 may further include a steady-state server 202 .

[0219] The first management unit 400 can be used to: when the actual power value of the computing system 4000 is greater than the first power threshold, control the first switching power supply 210 or the second switching power supply 220 of the steady-state server 202 in the overloaded computing device 1001 to power multiple loads 230 of the steady-state server 202 in the overloaded computing device 1001 through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 of the steady-state server 202 in the overloaded computing device 1001.

[0220] At this time, the steady-state server 202 in the overloaded computing device 1001 can draw power from the first AC line through the first switching power supply 210 to supply power to the steady-state server 202; or, the second switching power supply 220 of the steady-state server 202 in the overloaded computing device 1001 can draw power from the second AC line through the power conversion device 500. At this time, the AC line supplies power to the steady-state server 202 in the overloaded computing device 1001.

[0221] In addition to the overload computing device 1001, the plurality of computing devices 1000 further include a steady-state computing device 1002. In the case where the plurality of computing devices 1000 further include a steady-state computing device 1002, the first management unit 400 may be used to: control the first switch power supply 210 or the second switch power supply 220 of each server 200 in the steady-state computing device 1002 to supply power to the plurality of loads 230 of each server 200 in the steady-state computing device 1002 through the second management unit 700 of the steady-state computing device 1002 and the management chip 231 of each server in the steady-state computing device 1002. At this time, the AC line supplies power to each server 200 in the steady-state computing device 1002.

[0222] Please continue reading Figure 7 , Figure 7 The computing device 1000 shown in FIG. 1 overloads the computing device 1001 .

[0223] When the actual power value P of the computing system 4000 is greater than the first power threshold value P1, at least some computing devices 1000 are overloaded computing devices 1001; and in the overloaded computing devices 1001, at least some servers 200 are overloaded servers 201. At this time, in an overloaded computing device 1000, if some servers 200 are overloaded servers 201, the remaining servers 200 are stable servers 202. If all servers 200 are overloaded servers 201, the number of stable servers 202 is zero.

[0224] In addition, the overloaded server 201 may include a first-category server 200A and / or a second-category server 200B.

[0225] Of course, the steady-state server 202 may also include the first type server 200A and / or the second type server 200B.

[0226] The server that can obtain power from the first AC line is a first-type server 200A, and the server that cannot obtain power from the first AC line is a second-type server 200B.

[0227] When the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 controls the overloaded server 201 in the overloaded computing device 1001 in at least one of the following manners 1 and 2.

[0228] Method 1: The first management unit 400 is used to: control the first switching power supply 210 and the second switching power supply 220 of the first type of server 200A in the overloaded server 201 in the overloaded computing device 1001 through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 of the first type of server 200A in the overloaded server 201 in the overloaded computing device 1001 to jointly power multiple loads 230 of the first type of server 200A in the overloaded server 201 in the overloaded computing device 1001.

[0229] The first switch power supply 210 of the first type of server 200 in the overloaded server 201 in the overloaded computing device 1001 takes power from the first AC line, and the second switch power supply 220 of the first type of server 200A in the overloaded server 201 in the overloaded computing device 1001 takes power from the energy storage device 300, so that the first switch power supply 210 and the second switch power supply 220 jointly supply power to the multiple loads 230 of the first type of server 200A where the first and second switch power supplies are located. The power provided by the first AC line to the first switch power supply 210 may be greater than, equal to, or less than the second sub-power threshold.

[0230] Method 2: The first management unit 400 is used to: control the second switching power supply 220 of the second type server 200B in the overloaded server 201 in the overloaded computing device 1001 to power multiple loads 230 of the second type server 200B in the overloaded server 201 in the overloaded computing device 1001 through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 of the second type server 200B in the overloaded server 201 in the overloaded computing device 1001.

[0231] At this time, the second switching power supply 220 of the second type server 200B in the overloaded server 201 in the overload computing device 1001 draws power from the second AC line through the power conversion device 500, and the first management unit 400 controls the second switching power supply 220 of the second type server 200B to draw power from the energy storage device 300, so that the second switching power supply 220 draws power from the second AC line and the energy storage device 300 at the same time to supply power to the multiple loads 230 of the second type server 200B where it is located. The power provided by the second AC line to the second switching power supply 220 may be greater than, equal to, or less than the second sub-power threshold.

[0232] In which, when the actual power value of the computing system 4000 is greater than the first power threshold, in the overloaded computing device 1001, when some of the multiple servers 200 may be overloaded servers 201, the remaining servers 200 are steady-state servers 202, wherein the steady-state servers 202 may include first-class servers 200A and / or second-class servers 200B.

[0233] When the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 is used to control the steady-state server 202 in the overloaded computing device 1001 in at least one of the following methods 3 and 4.

[0234] Method 3: The first management unit 400 is used to: control the first switching power supply 210 and the second switching power supply 220 of the first type of server 200A in the steady-state server 202 in the overloaded computing device 1001 through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 of the first type of server 200A in the steady-state server 202 in the overloaded computing device 1001 to jointly power multiple loads 230 of the first type of server 200A in the steady-state server 202 in the overloaded computing device 1001.

[0235] At this time, the first switching power supply 210 of the first type server 200A in the steady-state server 202 in the overload computing device 1001 draws power from the first AC line, so that the first switching power supply 210 supplies power to the multiple loads 230 of the first type server 200A where it is located.

[0236] Method 4: The first management unit 400 is used to: control the second switching power supply 220 of the second type server 200B in the steady-state server 202 in the overloaded computing device 1001 to power multiple loads 230 of the second type server 200B in the steady-state server 202 in the overloaded computing device 1001 through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 of the second type server 200B in the steady-state server 202 in the overloaded computing device 1001.

[0237] At this time, the second switching power supply 220 of the second type server 200B in the steady-state server 202 in the overload computing device 1001 draws power from the second AC line through the power conversion device 500, so that the second switching power supply 220 supplies power to multiple loads 230 of the second type server 200B where it is located.

[0238] For example, when the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 can generate a first management signal and send it to the second management unit 700 that controls the overloaded computing device 1001. The second management unit 700 can then control at least some of the servers 200 in the overloaded computing device 1001 to draw power from the energy storage device 300 based on the first management signal.

[0239] Among them, the second management unit 700 can generate at least one of the first sub-management signal, the second sub-management signal, the third sub-management signal and the fourth sub-management signal according to the first management signal. Specifically, the second management unit 700 can send different sub-management signals to the server 200 in the computing device 1000 according to the type of the server 200. For example, if the server 200 is a first-class server 200A in the overloaded server 201, the second management unit 700 can generate a first sub-management signal for the server; for example, if the server is a second-class server 200B in the overloaded server 201, the second management unit 700 can generate a second sub-management signal for the server; for example, if the server is a first-class server 200A in the steady-state server 202, the second management unit 700 can generate a third sub-management signal for the server; for example, if the server is a second-class server 200B in the steady-state server 202, the second management unit 700 can generate a fourth sub-management signal for the server.

[0240] The second management unit 700 can send the first sub-management signal to the management chip 231 of the first type server 200A in the overloaded server 201 in the overloaded computing device 1001 where it is located, and the management chip 231 can control the first switching power supply 210 and the second switching power supply 220 of the first type server 200A where it is located to jointly power multiple loads 230 according to the first sub-management signal.

[0241] The second management unit 700 can send a second sub-management signal to the management chip 231 of the second type server 200B in the overload server 201 of the overload computing device 1001 where it is located. The management chip can control the second switching power supply 220 of the second type server 200B where it is located according to the second sub-management signal to draw power from the power conversion device 500 and the energy storage device 300 to supply power to multiple loads 230 of the second type server 200B.

[0242] The second management unit 700 can send the third sub-management signal to the management chip 231 of the first type server 200A in the steady-state server 202 of the overloaded computing device 1001 where it is located. The management chip 231 controls the first switching power supply 210 of the first type server 200A where it is located to power multiple loads 230 of the first type server 200A according to the third sub-management signal.

[0243] The second management unit 700 can send the fourth sub-management signal to the management chip 231 of the second type server 200B in the steady-state server 202 of the overloaded computing device 1001 where it is located. The management chip 231 can control the second switching power supply 220 of the second type server 200B where it is located according to the fourth sub-management signal to draw power from the power conversion device 500 to supply power to multiple loads 230 of the second type server 200B.

[0244] In the case that some of the multiple computing devices 1000 are overloaded computing devices 1001 , the remaining computing devices 1000 may be steady-state computing devices 1002 .

[0245] The first management unit 400 is used to control the steady-state computing device 1002 in at least one of the following modes 5 and 6:

[0246] Method 5: The first management unit 400 is used to: control the first switching power supply 210 of the first type server 200A in the steady-state computing device 1002 to power multiple loads 230 of the first type server 200A in the steady-state computing device 1002 through the second management unit 700 of the steady-state computing device 1002 and the management chip 231 of the first type server 200A in the steady-state computing device 1002.

[0247] Mode 6: The first management unit 400 is used to control the second switch power supply 220 of the second type server 200B in the steady-state computing device 1002 to supply power to multiple loads of the second type server 200B.

[0248] For example, the first management unit 400 may generate a second management signal and send it to the second management unit 700 of the steady-state computing device 1002 , and the second management unit 700 of the steady-state computing device 1002 may control each server 200 in the steady-state computing device 1002 to draw power from the AC line.

[0249] For example, the second management unit 700 of the steady-state computing device 1002 may generate a fifth sub-management signal and / or a sixth sub-management signal according to the second management signal. Specifically, the second management unit 700 may send different sub-management signals to the server 200 according to the type of the server 200 in the computing device 1000. For example, if the server 200 is a first-type server 200A, the second management unit 700 may generate a fifth sub-management signal for the server; if the server 200 is a second-type server 200B, the second management unit 700 may generate a sixth sub-management signal for the server.

[0250] The second management unit 700 can send the fifth sub-management signal to the management chip 231 of the first type server 200A in the steady-state computing device 1002 where it is located, and the management chip 231 can control the switching power supply 210 of the first type server 200A where it is located to power multiple loads 230 of the first type server 200A.

[0251] The second management unit 700 can send the sixth sub-management signal to the management chip 231 of the second type server 200B in the steady-state computing device 1002 where it is located. The management chip 231 can control the second switching power supply 220 of the second type server 200B where it is located to draw power from the power conversion device 500 to power multiple loads 230 of the second type server 200B.

[0252] Please refer again Figure 6 In this embodiment, when the actual power value of the computing system 4000 is less than or equal to the first power threshold: multiple computing devices 1000 can all be steady-state computing devices 1002; or, some of the multiple computing devices 1000 are steady-state computing devices 1002, and the rest are overloaded computing devices 1001. At this time, the first management unit 400 can control all computing devices 1001 (including overloaded computing devices 1001 and steady-state computing devices 1002) to draw power from the AC line.

[0253] The computing device 1000 may include a first type server 200A and / or a second type server 200B.

[0254] At this time, in the computing device 1000, some servers 200 are first-category servers 200A, and the remaining servers 200 are second-category servers 200B; or all servers 200 are first-category servers 200A; or all servers 200 are second-category servers 200B.

[0255] The first management unit 400 is also used for: when the actual power value is less than or equal to the first power threshold, controlling the first switching power supply 210 of the first type server 200A in the computing device 1000 to supply power to multiple loads 230 of the first type server 200A in the computing device 1000 through the second management unit 700 of the computing device 1000 and the management chip 231 of the first type server 200A in the computing device 1000; and / or, controlling the second switching power supply 220 of the second type server 200B in the computing device 1000 to supply power to multiple loads 230 of the second type server 200B in the computing device 1000 by taking power from the power conversion device 500 through the second management unit 700 of the computing device 1000 and the management chip 231 of the second type server 200B in the computing device 1000.

[0256] Therefore, when a second type server 200B appears in the computing device 1000, the first management unit 400 controls the second switching power supply 220 of the second type server 200B to supply power to multiple loads 230 through the management chip 231 of the second type server 200, thereby avoiding power failure of the second type server 200B.

[0257] For example, when the actual power value of the computing system 4000 is less than or equal to the first power threshold, the first management unit 400 may generate a third management signal and send it to the second management unit 700 of the computing device 1000. The second management unit 700 of the computing device 1000 may control the server 200 to draw power from the AC line according to the third management signal.

[0258] For example, the second management unit 700 can generate a seventh sub-management signal and / or an eighth sub-management signal based on the third management signal, and send the seventh sub-management signal to the management chip 231 of the first type of server 200A in the computing device 1000, and send the eighth sub-management signal to the management chip 231 of the second type of server 200B in the computing device 1000.

[0259] The management chip 231 of the first type server 200A controls the first switch power supply 210 of the first type server 200A to supply power to the multiple loads 230 of the first type server 200A according to the seventh sub-management signal.

[0260] The management chip 231 of the second type server 200B controls the second switch power supply 220 of the second type server 200B to obtain power from the power conversion device 500 according to the eighth sub-management signal to supply power to the multiple loads 230 of the second type server 200B.

[0261] Among them, when there is a second type of server 200B in the computing system 4000, the first management unit 400 can control the second switch power supply 220 of the second type of server 200B through the second management unit 700 and the management chip 231 to power multiple loads 230 of the second type of server 200B by taking power from the power conversion device 500. In some other examples, if there is a second type of server 200B in the computing system 4000, the management chip 231 of the second type of server 200B can automatically control the second switch power supply 220 to take power from the power conversion device 500 to power multiple loads 230 of the second type of server 200B.

[0262] In this embodiment, when the actual power value of the computing system 4000 is less than or equal to the first power threshold, the first management unit 400 can control the first switch power supply 210 of a certain overloaded server 201 to supply power to the multiple loads 230, and at this time, the second switch power supply 220 does not supply power. For example, the first management unit 400 can control the output voltage of the first switch power supply 210 of the overloaded server 201 to be greater than the output voltage of the second switch power supply 220 through the management chip 231, so that the first switch power supply 210 supplies power to the multiple loads 230, and the second switch power supply 220 does not supply power to the multiple loads 230. Of course, in other embodiments, the first switch power supply 210 can be controlled to supply power to the multiple loads 230, and the second switch power supply 220 does not supply power to the multiple loads 230 in other ways. The embodiment of the present application does not limit the method of controlling the first switching power supply 210 to supply power to multiple loads 230 while the second switching power supply 220 does not supply power to multiple loads 230. Any method is acceptable as long as it is possible to achieve that the first switching power supply 210 supplies power to multiple loads 230 while the second switching power supply 220 does not supply power to multiple loads 230.

[0263] When the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 can control the first switch power supply 210 and the second switch power supply 220 to jointly power multiple loads 230 through the management chip 231. At this time, the second switch power supply 220 can draw power from the energy storage device 300, and the first switch power supply 210 can draw power from the first AC line. For example, the management chip 231 can control the output voltage of the second switch power supply 220 of the overloaded server 201 to be the same as or similar to the output voltage of the first switch power supply 210. When the output voltage of the second switch power supply 220 is similar to the output voltage of the first switch power supply 210, the voltage difference between the output voltage of the second switch power supply 220 and the output voltage of the first switch power supply 210 is less than or equal to the preset threshold. As long as the voltage difference between the output voltage of the second switch power supply 220 and the output voltage of the first switch power supply 210 is less than or equal to the preset threshold, the first switch power supply and the second switch power supply can jointly power multiple loads 230.

[0264] Of course, in other embodiments, other methods can be used to realize that the first switching power supply 210 and the second switching power supply 220 jointly supply power to multiple loads 230. The embodiment of the present application does not limit the implementation method of the first switching power supply 210 and the second switching power supply 220 jointly supply power to multiple loads 230.

[0265] In this embodiment, when the actual power value of the computing system 4000 is greater than the first power threshold, the sum of the actual powers of all steady-state computing devices 1002 is equal to the product of the number of steady-state computing devices 1002 and the first sub-power threshold, and the servers 200 in the overloaded computing device 1001 are all overloaded servers 201. At this time, the first management unit 400 can control the AC power obtained by the overloaded server 201 in the overloaded computing device 1001 from the AC line, so that the power obtained by the overloaded server 201 from the AC line is the second sub-power threshold, and the power obtained by the overloaded server 201 from the energy storage device 300 is the difference between the actual power value of the overloaded server 201 and the second sub-power threshold.

[0266] For example, the overloaded server 201 is a first type server 200A, that is, the first switching power supply 210 of the overloaded server 201 is normal, and the first management unit 400 can control the output power of the first switching power supply 210 of the overloaded server 201 through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 in the overloaded computing device 1001 that controls the overloaded server 201. At this time, the output power of the first switching power supply 210 is equal to the second sub-power threshold, and the output power of the second switching power supply 220 is equal to the difference between the actual power value of the overloaded server 201 and the second sub-power threshold.

[0267] For example, the management chip 231 can control the output voltage of the first switching power supply 210 to be lower than the output voltage of the second switching power supply 220, so as to limit the output power of the first switching power supply 210 to the first sub-power threshold. It can be understood that when the output voltage of the first switching power supply 210 is adjusted to be lower than the output voltage of the second switching power supply 220, it is necessary to ensure that the first switching power supply 210 and the second switching power supply 220 can jointly power multiple loads 230. For example, the voltage difference between the output voltage of the second switching power supply 220 and the first switching power supply 210 is less than or equal to the preset threshold. When the voltage difference between the output voltage of the second switching power supply 220 and the first switching power supply 210 is less than or equal to the preset threshold, the first switching power supply 210 and the second switching power supply 220 can jointly power multiple loads 230, and the output power of the first switching power supply 210 can be limited.

[0268] Of course, the output power of the first type of server 200A can be related to the output voltage and output current of the first type of server 200A. In some other examples, the management chip 231 can also control the output current of the first switching power supply 210 of the first type of server 200A in the overloaded server 201 in the overloaded computing device 1001, thereby controlling the output power of the first switching power supply 210 of the first type of server 200A in the overloaded server 201 in the overloaded computing device 1001, so that the output power of the first switching power supply 210 is the second sub-power threshold.

[0269] In this embodiment, when the actual power value of the computing system 4000 is greater than the first power threshold, the first management unit 400 can determine the third sub-power threshold based on the power difference between the actual power value of the computing system 4000 and the first power threshold and the number of overloaded servers 201 in all overloaded computing devices 1001, wherein the third sub-power threshold is the power obtained by the overloaded computing device 201 from the energy storage device 300. In some examples, the third sub-power thresholds corresponding to each overloaded server 201 in the overloaded computing device 1001 are equal.

[0270] The first management unit 400 can limit the AC power obtained from the AC line by the overloaded server 201 in the overloaded computing device 1001 according to the third sub-power threshold. For example, the AC power obtained from the AC line by the overloaded server 201 in the overloaded computing device 1001 is the difference between the actual power value of the overloaded server 201 and the third sub-power threshold.

[0271] For example, when the actual power value of the computing system is greater than the first power threshold, the actual power value of the computing system 4000 is 70KW, and the first power threshold is 50KW. At this time, the difference between the actual power value of the computing system 4000 and the first power threshold is 20KW. The number of overloaded servers 201 in the overloaded computing device 4000 is 100, and thus, the first management unit 400 can determine that the third sub-power threshold is equal to 200W.

[0272] For example, if the actual power value of each overloaded server 201 is 700W, the first management unit 400 can limit the output power of the first switching power supply 210 of the overloaded server 201 to 500W through the second management unit 700 of the overloaded computing device 1001 and the management chip 231 of the overloaded server 201 in the overloaded computing device 1001.

[0273] If the actual power value of one overloaded server 201 among the 100 overloaded servers 201 is 800W, the first management unit 400 can control the output power of the first switch power supply 210 of the overloaded server 201 to be the difference between the actual power value of the overloaded server 201 and the third sub-power threshold through the management chip 231 of the overloaded server 201, that is, the output power of the first switch power supply 210 is 600W. The third sub-power threshold is 200W.

[0274] The actual power value of another overloaded server 201 is 600W, and the first management unit 400 can control the output power of the first switch power supply 210 of the overloaded server 201 to be the difference between the actual power value of the overloaded server 201 and the third sub-power threshold through the management chip 231 of the overloaded server 201, that is, the output power of the first switch power supply 210 is 400W. The third sub-power threshold is 200W.

[0275] Please continue reading Figure 6 In this embodiment, the first management unit 400 can also be electrically connected to the energy storage device 300. The first management unit 400 is also used to: when the actual power value of the computing system 4000 is greater than the first power threshold, obtain the remaining capacity of the energy storage device 300, adjust the capping power of the computing system 4000 according to the remaining capacity, and control the actual power value of the computing system 4000 to be less than or equal to the capping power.

[0276] Among them, when the remaining capacity of the energy storage device 300 is different, the power that the energy storage device 300 can load is different. The energy storage device 300 can be used together with the AC line to power multiple computing devices 1000 in the computing system 4000. Therefore, when the remaining capacity of the energy storage device 300 is different, the maximum power of the computing system 4000 is different.

[0277] The capped power of the computing system 4000 can limit the maximum power of the computing system 4000. After determining the capped power of the computing system 4000, the actual power value of the computing system 4000 will not exceed the capped power of the computing system 4000, that is, the sum of the actual power values ​​of multiple computing devices 1000 will not exceed the capped power of the computing system 4000.

[0278] If at a certain moment, the business of the computing system 4000 suddenly increases, causing the actual power value P of the computing system 4000 to increase, the actual power value of the computing system 4000 can be reduced by reducing the actual power value of at least part of the computing devices 1000 in the computing system 4000, thereby making the actual power value of the computing system 4000 less than or equal to the capped power of the computing system 4000.

[0279] For example, the first power threshold (i.e., the maximum AC power) is 50KW. Assuming that the power that the energy storage device 300 can load is 10KW, and the actual power value P of the computing system 4000 is 55KW, at this time, the power loaded by the AC line is 50W, and the power loaded by the energy storage device 300 is 5KW. At this time, the energy storage device 300 can load the power difference between the actual power value of the computing system 4000 and the first power threshold. If at a certain moment, the capacity of the energy storage device 300 is reduced, and the power it can load is only 3KW, while the actual power value of the computing system 4000 is still 55KW, at this time, the power that the AC line can load is 50KW, and the power that the energy storage device 300 needs to load is 5KW. However, since the remaining capacity of the energy storage device 300 can only load 3KW of power, which is less than 5KW, it may cause the computing system 4000 to lose power. In the embodiment of the present application, the capping power of the computing system 4000 can be adjusted according to the remaining capacity of the energy storage device 300. Therefore, the capping power of the computing system 4000 can be modified to 53KW. At this time, the power that the energy storage device 300 needs to load is 3KW. Therefore, the energy storage device 300 can load the power difference between the actual power value P of the computing system 4000 and the first power threshold. This can avoid the inability to load the power difference between the actual power value P of the computing system 4000 and the first power threshold due to insufficient remaining capacity of the energy storage device 300, thereby avoiding power failure of the computing system 4000.

[0280] In summary, in the embodiment of the present application, the remaining capacity of the energy storage device 300 is obtained by the first management unit 400, and the capping power of the computing system 4000 is adjusted according to the remaining capacity, so that the actual power value P of the computing system 4000 can be prevented from exceeding the sum of the powers that can be loaded by the AC line and the energy storage device 300, thereby preventing the computing system 4000 from losing power due to insufficient remaining capacity of the energy storage device 300.

[0281] In some examples, the capped power of computing system 4000 may be adjusted by modifying the capped power of some or all computing devices 1000 .

[0282] Illustratively, at least some of computing devices 1000 have different capped powers.

[0283] For example, the capped power of each computing device 1000 is the same.

[0284] In some examples, the capped power of computing device 1000 may be adjusted by modifying the capped power of some or all of servers 200 in computing device 1000 .

[0285] Please continue reading Figure 6 In this embodiment, the first management unit 400 is also used to: when the actual power value of the computing system 4000 is less than the first power threshold, control the power conversion device 500 to charge the energy storage device 300.

[0286] When the actual power value of the computing system 4000 is less than the first power threshold, the energy storage device 300 does not supply power to the multiple computing devices 1000. At this time, the power conversion device 500 charges the energy storage device 300, which can ensure that the energy storage device 300 has sufficient capacity, thereby increasing the power value that the energy storage device 300 can load, thereby increasing the maximum power of the computing system 4000. In addition, when the actual power value of the computing system 4000 is less than the first power threshold, that is, when the energy storage device 300 does not supply power to the multiple computing devices 1000, the power conversion device 500 charges the energy storage device 300, which can also avoid the energy storage device 300 being in both charging and discharging states at the same time, thereby extending the service life of the energy storage device 300.

[0287] In some examples, the first management unit 400 is used to: when the actual power value of the computing system 4000 is less than the first power threshold, obtain the remaining capacity of the energy storage device 300, and when the remaining capacity of the energy storage device 300 is less than the full-charge capacity of the energy storage device 300, control the power conversion device 500 to charge the energy storage device 300. The full-charge capacity of the energy storage device 300 refers to the electrical energy contained in the energy storage device 300 when it is fully charged. In other words, the maximum value of the remaining capacity of the energy storage device 300 is equal to the full-charge capacity of the energy storage device 300.

[0288] In some embodiments, the first management unit 400 is further used to: when the actual power value is greater than the first power threshold, control the energy storage device 300 and the power conversion device 500 to jointly power the computing device 1000. For example, the first management unit 400 can control the output voltage of the energy storage device 300 to be the same as or similar to the output voltage of the power conversion device 500. When the output voltage of the energy storage device 300 is similar to the output voltage of the power conversion device 500, the voltage difference between the output voltage of the energy storage device 300 and the output voltage of the power conversion device 500 is less than or equal to a preset value. As long as the voltage difference between the output voltage of the energy storage device 300 and the output voltage of the power conversion device 500 is less than or equal to the preset value, the energy storage device 300 and the power conversion device 500 can jointly power the computing device 1000.

[0289] At the same time, the first management unit 400 can also limit the output power of the power conversion device 500 to the difference between the first power threshold and the power provided by the first AC line, that is, so that the power obtained by multiple computing devices 1000 from the AC line is the first power threshold, and the power obtained from the energy storage device 300 is the difference between the actual power value and the first power threshold.

[0290] For example, the first management unit 400 may limit the output power of the power conversion device 500 by controlling the output voltage of the power conversion device 500 to be lower than the output voltage of the energy storage device 300, so as to limit the output power of the power conversion device 500. It is understandable that when the output voltage of the power conversion device 500 is adjusted to be lower than the output voltage of the energy storage device 300, both the energy storage device 300 and the power conversion device 500 need to be in a discharge state to ensure that the energy storage device 300 and the energy storage device 300 can jointly power the computing device 1000. For example, the voltage difference between the output voltage of the energy storage device 300 and the power conversion device 500 is less than or equal to a preset value. When the voltage difference between the output voltage of the energy storage device 300 and the power conversion device 500 is less than or equal to the preset value, the power conversion device 500 and the energy storage device 300 can jointly power the computing device, and the output power of the power conversion device 500 can be limited.

[0291] Of course, the output power of the power conversion device 500 may be related to the output voltage and the output current of the power conversion device 500. In some other examples, the first management unit 400 may also control the output power of the power conversion device 500 by controlling the output current of the power conversion device 500, so as to limit the output power of the power conversion device 500.

[0292] The first management unit 400 is also used to: when the actual power value of the computing system 4000 is less than the first power threshold, control the power conversion device 500 to supply power to the server 200, and the energy storage device 300 does not supply power to the server 200. For example, the first management unit 400 can control the output voltage of the power conversion device 500 to be greater than the output voltage of the energy storage device 300, so that the power conversion device 500 is in a discharging state and the energy storage device 300 is in a non-discharging state. The first management unit 400 can also control the power conversion device 500 to charge the energy storage device 300.

[0293] Of course, when the actual power value of the computing system 4000 is equal to the first power threshold, the first management unit 400 can control the energy storage device 300 to be in a state of neither charging nor discharging.

[0294] Please continue reading Figure 6 In some embodiments, the computing system 4000 further includes: a switch 800. The switch 800 includes a first communication interface and a plurality of second communication interfaces, the first communication interface and the plurality of second communication interfaces are connected, the first communication interface is connected to the first management unit 400, and the second communication interface is connected to the corresponding second management unit 700.

[0295] In some examples, the switch 800 may be disposed in any cabinet 100 .

[0296] For example, the switch 800 may include two switching power supplies, which may be electrically connected to the AC line and the energy storage device 300 respectively, so that the two switching power supplies can respectively draw power from the AC line and the energy storage device 300 to power the switch 800.

[0297] For example, the two switching power supplies of the switch 800 can be electrically connected to the first power distribution device 3001 and the second power distribution device 3002 respectively, so that one switching power supply of the switch 800 can be connected to the first AC line through the first power distribution device 3001, and the other switching power supply can be electrically connected to the power conversion device 500 and the energy storage device 300 through the second power distribution device 3002.

[0298] The first communication interface is connected to the first management unit 400, and the second communication interface is connected to the second management unit 700 of the corresponding computing device 1000. The first communication interface is connected to the second communication interface, so the first management unit 400 can be connected to the second management unit 700 of the computing device 1000 through the first communication interface and the second communication interface.

[0299] For example, the first communication interface and the second communication interface can both be network ports, wherein the first communication interface can be communicatively connected to the first management unit 400 via a network cable, and the second communication interface can be communicatively connected to the second management unit 700 of the computing device 1000 via a network cable.

[0300] Figure 8 is another structural block diagram of a computing system 4000 according to some embodiments.

[0301] See also Figure 8 In this embodiment, the first management unit 400 may include a power acquisition module 410 and a processing module 420. The functions of the power acquisition module 410 and the processing module 420 may be implemented by the first control chip. The power acquisition module 410 is connected to the second management unit 700 of the computing device 1000, and the processing module 420 is connected to the power acquisition module 410 and the second management unit 700 of the computing device 1000.

[0302] For example, the power acquisition module 410 may be connected to the second management unit 700 of the computing device 1000 via the switch 800 .

[0303] The power acquisition module 410 is used to acquire actual power values ​​of the multiple computing devices 1000 , and determine the actual power value of the computing system 4000 according to the actual power values ​​of the multiple computing devices 1000 .

[0304] For example, the second management unit 700 may send the actual power value of the computing device 1000 where it is located to the power acquisition module 410 , and the power acquisition module 410 may determine the actual power value of the computing system 4000 based on the actual power values ​​of the multiple computing devices 1000 .

[0305] The processing module 420 can control multiple computing devices 1000 to draw power from the AC line when the actual power value of the computing system 4000 is less than or equal to the first power threshold; when the actual power value of the computing system 4000 is greater than the first power threshold, the processing module 420 can control multiple computing devices 1000 to draw power from the AC line and the energy storage device 300.

[0306] In addition, the processing module 420 may also control the power conversion device 500 to charge the energy storage device 300 when the actual power value of the computing system 4000 is less than the first power threshold.

[0307] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that a person skilled in the art can think of within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A computing system, characterized in that: include: A power conversion device, wherein an input end of the power conversion device is used to be connected to a second AC line; An energy storage device connected to the output end of the power conversion device; A plurality of computing devices, each computing device comprising a first input terminal and a second input terminal, the first input terminal of each computing device being used to connect to a first AC line, and the second input terminal of each computing device being connected to an output terminal of the power conversion device and an output terminal of the energy storage device; A first management unit is electrically connected to the multiple computing devices, and the first management unit is used to: when the actual power value of the computing system is greater than the first power threshold, control the multiple computing devices to draw power from the AC line and the energy storage device, wherein drawing power from the AC line includes drawing power from the first AC line and / or drawing power from the second AC line through the power conversion device.

2. The computing system according to claim 1, characterized in that The first management unit is also used to: when the actual power value of the computing system is greater than the first power threshold, control the AC power obtained by the multiple computing devices from the AC line so that the power carried by the AC line is the first power threshold, and the power carried by the energy storage device is the power difference between the actual power value of the computing device and the first power threshold.

3. The computing system according to claim 1 or 2, characterized in that: The first management unit is electrically connected to the energy storage device; The first management unit is further used to: obtain the remaining capacity of the energy storage device, adjust the capping power of the computing system according to the remaining capacity, and control the actual power value of the computing system to be less than or equal to the capping power.

4. The computing system according to claim 3, characterized in that: The first management unit is further used to: when the actual power value of the computing system is less than the first power threshold, control the power conversion device to charge the energy storage device.

5. The computing system according to any one of claims 1 to 4, characterized in that: The first management unit is further used to: when the actual power value of the computing system is greater than the first power threshold, control the overloaded computing device among the multiple computing devices to draw power from the AC line and the energy storage device; The overloaded computing device is a computing device among the multiple computing devices whose actual power value is greater than a first sub-power threshold, and the product of the first sub-power threshold and the number of the computing devices is equal to the first power threshold.

6. The computing system according to claim 5, characterized in that: The computing device further includes a second management unit and a plurality of servers, wherein the second management unit is communicatively connected with the first management unit and the plurality of servers; The first management unit is further used to: when the actual power value of the computing system is greater than the first power threshold, control the overload server in the overload computing device to draw power from the AC line and the energy storage device through the second management unit of the overload computing device; The overloaded server is a server among the multiple servers whose actual power value is greater than a second sub-power threshold, and the product of the second sub-power threshold and the number of servers in one of the computing devices is equal to the first sub-power threshold.

7. The computing system according to claim 6, characterized in that: The server comprises a first switching power supply, a second switching power supply and a plurality of loads, wherein the first switching power supply and the second switching power supply are both connected to the plurality of loads, wherein the plurality of loads comprises a management chip, wherein the management chip is communicatively connected to the first switching power supply, the second switching power supply and the second management unit, wherein the first switching power supply is a first sub-input terminal of the server, and the second switching power supply is a second sub-input terminal of the server; the first input terminal of the computing device is connected to the first sub-input terminals of the plurality of servers, and the second input terminal of the computing device is connected to the second sub-input terminals of the plurality of servers; In one of the computing devices, the plurality of servers include first-category servers and / or second-category servers, wherein the servers that can draw power from the first AC line are the first-category servers, and the servers that cannot draw power from the first AC line are the second-category servers; The first management unit is used to: when the actual power value is greater than the first power threshold, control the first switch power supply and the second switch power supply of the first type of servers in the overloaded servers in the overloaded computing device to jointly supply power to the multiple loads of the first type of servers in the overloaded servers in the overloaded computing device through the second management unit of the overloaded computing device and the management chip of the first type of servers in the overloaded servers in the overloaded computing device; and / or, Through the second management unit of the overload computing device and the management chip of the second type of servers in the overload computing device, the second switching power supply of the second type of servers in the overload servers in the overload computing device is controlled to supply power to the multiple loads of the first type of servers in the overload servers in the overload computing device.

8. The computing system according to any one of claims 1 to 7, characterized in that: The first management unit is further used to: when the actual power value of the computing system is less than or equal to the first power threshold, control the multiple computing devices to draw power from the AC line.

9. A power supply module, characterized in that: include: An energy storage device, a first management unit and a power conversion device; The input end of the power conversion device is used to connect to the second AC line, and the output end of the power conversion device is used to connect to the second input ends of multiple computing devices in the computing system; The energy storage device is connected to the output end of the power conversion device, and the output end of the energy storage device is used to connect to the second input ends of multiple computing devices; The first management unit is used to be electrically connected to multiple computing devices, and the first management unit is used to: when the actual power value of the computing system is greater than the first power threshold, control the multiple computing devices to draw power from the AC line and the energy storage device, wherein drawing power from the AC line includes drawing power from the first AC line and / or drawing power from the second AC line through the power conversion device, and the first AC line is connected to the first input terminals of the multiple computing devices.

10. A computing device, characterized in that: include: a second management unit and a plurality of servers; The first sub-input terminals of the plurality of servers are used to be connected to a first AC line, the second sub-input terminals of the plurality of servers are used to be connected to an output terminal of a power conversion device of a power supply module and an output terminal of an energy storage device of the power supply module, and the input terminal of the power conversion device is connected to a second AC line; The second management unit is connected to the plurality of servers for communication, and the second management unit is used to be connected to the first management unit of the power supply module for communication; In the case where the actual power value of the computing system is greater than the first power threshold, if the actual power value of the computing device is greater than the first sub-power threshold, the second management unit is used to: under the control of the first management unit, control the overloaded server to draw power from the AC line and the energy storage device, wherein drawing power from the AC line includes drawing power from the first AC line, and / or drawing power from the second AC line through the power conversion device; Among them, the overloaded server is a server among the multiple servers whose actual power value is greater than the second sub-power threshold, the product of the second sub-power threshold and the number of servers is equal to the first sub-power threshold, and the product of the first sub-power threshold and the number of computing devices in the computing system is equal to the first power threshold.

Images