Server system

Abstract

The invention discloses a server system, and relates to the technical field of servers, the server system comprises a heat dissipation node and a service node, the heat dissipation node is provided with a heat dissipation controller, and the service node is provided with a service controller; wherein the service controller is connected with the heat dissipation controller, and the heat dissipation controller is connected with a power supply interface of the service node; wherein the service controller is also used for receiving a power supply control request and forwarding the power supply control request to the heat dissipation controller, and the power supply control request is used for requesting to carry out power supply control on the service node; and the heat dissipation controller is also used for responding to the power supply control request and carrying out power supply control on the service node through the power supply interface of the service node. Through the server system, the technical problem of relatively low control efficiency of the server system in related technologies is solved, and the technical effect of improving the control efficiency of the server system is achieved.

Description

Technical Field

[0001] This application relates to the field of server technology, and more particularly to a server system. Background Technology

[0002] Due to the rapid development of internet services, the demand for servers is increasing. In related technologies, if it's necessary to control the power-on / off / restart of business nodes in a server system, it can only be achieved through on-site operations such as pulling the business nodes, and remote power supply control and remote maintenance are not possible. This design of the server system results in low control efficiency. Summary of the Invention

[0003] This application provides a server system to at least solve the problem of low control efficiency of server systems in related technologies.

[0004] This application provides a server system, including: a heat dissipation node and a service node. A heat dissipation controller is configured on the heat dissipation node, and a service controller is configured on the service node. The service controller is connected to the heat dissipation controller, and the heat dissipation controller is connected to the power supply interface of the service node. The service controller is used to control the service node; the heat dissipation controller is used to control the heat dissipation node to cool the service node. The service controller is also used to receive power control requests and forward them to the heat dissipation controller, wherein the power control requests request power control for the service node. The heat dissipation controller is also used to respond to the power control requests and control the power supply to the service node through the power supply interface of the service node.

[0005] Through this application, a service controller set on a service node is connected to a heat dissipation controller set on a heat dissipation node. The heat dissipation controller is also connected to the power supply interface of the service node. When the service controller is used to control the service node, it is also used to receive power supply control requests and forward the power supply control requests to the heat dissipation controller. When the heat dissipation controller is used to control the heat dissipation node to cool the service node, it is also used to respond to the power supply control requests and control the power supply of the service node through the power supply interface of the service node. The power supply control requests received by the service controller are transferred to the heat dissipation controller, which is not located on the service node but on the heat dissipation node. This ensures that the execution of the power supply control requests is not affected by the power supply status of the service node where the service controller is located, thus guaranteeing the effective implementation of remote control. Therefore, it can solve the technical problem of low control efficiency of server systems in related technologies and achieve the technical effect of improving the control efficiency of server systems. Attached Figure Description

[0006] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0007] Figure 1 This is a schematic diagram of a server system according to an embodiment of this application;

[0008] Figure 2 This is a schematic diagram of a dual-node power supply architecture according to an embodiment of this application;

[0009] Figure 3 This is a schematic diagram of the power-on process of a dual-node server system according to an embodiment of this application;

[0010] Figure 4 This is a schematic diagram of a power supply control process for a dual-node server system according to an embodiment of this application. Detailed Implementation

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0012] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0013] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0014] The embodiments of this application provide a server system. The server system provided in the embodiments of this application will be described in detail in conjunction with the composition of the server system and the interaction between the various components of the server system.

[0015] This embodiment provides a server system. Figure 1 This is a schematic diagram of a server system according to an embodiment of this application, such as... Figure 1 As shown, the server system includes: a heat dissipation node and a service node. A heat dissipation controller is installed on the heat dissipation node, and a service controller is installed on the service node. The service controller is connected to the heat dissipation controller, and the heat dissipation controller is connected to the power supply interface of the service node.

[0016] The service controller is used to control the service nodes; the heat dissipation controller is used to control the heat dissipation nodes to cool the service nodes.

[0017] The service controller is also used to receive power control requests and forward them to the heat dissipation controller. The power control request is used to request power control for the service node. The heat dissipation controller is also used to respond to the power control request and control the power supply to the service node through the power supply interface of the service node.

[0018] Through the above server system, the business controller set on the business node is connected to the heat dissipation controller set on the heat dissipation node. The heat dissipation controller is also connected to the power supply interface of the business node. When the business controller is used to control the business node, it is also used to receive power control requests and forward the power control requests to the heat dissipation controller. When the heat dissipation controller is used to control the heat dissipation node to cool the business node, it is also used to respond to the power control requests and control the power supply of the business node through the power supply interface of the business node. The power control requests received by the business controller are transferred to the heat dissipation controller located on the heat dissipation node, which is not located on the business node. This makes the execution of the power control requests unaffected by the power supply status of the business node where the business controller is located, ensuring the effective implementation of remote control. Therefore, it can solve the technical problem of low control efficiency of server systems in related technologies and achieve the technical effect of improving the control efficiency of server systems.

[0019] Optionally, in the embodiments of this application, the server system may include, but is not limited to, at least a heat dissipation node and a service node.

[0020] Optionally, in the embodiments of this application, a heat dissipation node may refer to, but is not limited to, a part of a system (such as a server, computer, data center equipment, etc.) specifically responsible for thermal management. The main responsibility of the heat dissipation node may be, but is not limited to, monitoring and regulating the system temperature to ensure that the system operates within the optimal temperature range and prevent performance degradation or hardware damage caused by overheating. The heat dissipation node may be equipped with physical components such as heat sinks, fans, heat fins, liquid cooling systems, and heat dissipation controllers that control the operation of these components.

[0021] Optionally, in the embodiments of this application, the heat dissipation node may include, but is not limited to: an air-cooled system (or fan plate): equipped with multiple fans, heat sinks and temperature sensors, and the heat dissipation controller dynamically adjusts the fan speed according to the data of the temperature sensors to achieve the optimal cooling effect; a liquid-cooled system (or liquid-cooled plate): removes heat through liquid circulation, and the liquid-cooled system may include, but is not limited to, a pump, heat exchanger, temperature controller, etc., and the heat dissipation controller adjusts the liquid flow rate and cooling power according to the temperature of the liquid-cooled system.

[0022] Optionally, in the embodiments of this application, the business node may be, but is not limited to, the part of a server or other IT infrastructure responsible for executing business logic and data processing. A business node may be, but is not limited to, a single server, or a node in a server cluster, primarily running applications, databases, network services, etc.

[0023] Optionally, in this embodiment, the business nodes may include, but are not limited to: Web server nodes: running website services, processing HTTP requests, and providing web page content; database server nodes: storing and managing data, performing data queries and transaction processing; and computing server nodes: high-performance computing nodes used for computationally intensive tasks such as scientific computing and data analysis.

[0024] Optionally, in this embodiment, a service controller may be provided on the service node, but is not limited to. The service controller may be, but is not limited to, a module / component / device in the service node used to manage and control the service processor and other hardware resources. The service processor may be, but is not limited to, the core processing unit of the service node, and may be, but is not limited to, responsible for executing business logic operations and service requests. Optionally, the service processor may include, but is not limited to, CPU (Central Processing Unit), GPU (Graphics Processing Unit), FPGA (FPGA Programmable Gate Array), etc.

[0025] Optionally, in the embodiments of this application, the service controller may be, but is not limited to, a BMC (Baseboard Management Controller) or a custom control chip.

[0026] Optionally, in this embodiment, the service controller may, but is not limited to, support remote monitoring and management via network connections or other connection methods. For example, the BMC, as a service controller, may, but is not limited to, be an embedded controller integrated on a server motherboard, and may, but is not limited to, implement remote monitoring and management through the IPMI standard interface.

[0027] Optionally, in the embodiments of this application, the basic functions of the service controller may include, but are not limited to, monitoring the health status of the server and collecting and reporting hardware information. Through the design of this application, the service controller may be endowed with more advanced functions, such as being able to receive power control requests and forward these requests to the thermal controller, thereby achieving fine-grained control over the power supply status of service nodes.

[0028] Optionally, in this embodiment, the basic function of the thermal controller may include, but is not limited to, dynamically adjusting the operating state of the heat dissipation components based on the hardware temperature, typically the fan speed and the activity of the heat sink. Through the design of this application, the thermal controller integrates the ability to respond to power supply control. The thermal controller can dynamically control the power supply interface of the service node according to the request of the service controller, achieving more refined power supply management.

[0029] Optionally, in this embodiment, by responding to the power supply control request through a thermal controller located on a node of another system, the power supply status of the service node where the service controller receiving the power supply control request is located can be prevented from affecting the execution of the power supply control request.

[0030] Optionally, in the embodiments of this application, the power supply control request may be, but is not limited to, a request received by the service controller from the remote management system or the user. The power supply control request may be, but is not limited to, used to control the power supply status of the service node, and may be either a power-on request or a power-off request.

[0031] Optionally, in this embodiment, the heat dissipation node may also be equipped with a network interface and a local area network (LAN) switch, but not limited to. The network interface may be connected to the service controller on the service node via the LAN switch, and the service controller may receive power control requests from the network via the LAN switch and the network interface, but not limited to. By providing a network interface and a LAN switch on the heat dissipation node, enabling it to communicate with the service controller on the service node via the LAN and receive power control requests, this not only enhances the ability to remotely monitor and precisely control the heat dissipation system, improving the efficiency and flexibility of server management, but also improves the overall stability and scalability of the server system through network redundancy and modular design, facilitating future maintenance and upgrades.

[0032] Optionally, in the embodiments of this application, the power supply control request may, but is not limited to, directly or indirectly indicate the need to control the power supply to the service node. For example, the power supply control request may, but is not limited to, directly request to control the power supply to the service node, or the power supply control request may, but is not limited to, requesting to make the upgrade operation of the service node / components (such as the service processor) set on the service node effective. Since making the upgrade operation effective requires restarting the service node, the power supply control request actually indirectly indicates the need to control the power supply to the service node.

[0033] Optionally, in this embodiment, the service controller may receive power control requests in the following ways, but are not limited to: receiving them through a network interface, such as receiving them from a remote operation and maintenance system via the IPMI protocol or a custom network protocol; receiving them through a serial interface, such as UART or RS-232, from a connected device or console; or receiving them through a management interface, such as a command-line interface or a graphical user interface, directly generated by user operation.

[0034] Optionally, in this embodiment, the way the service controller forwards the power control request to the thermal controller may include, but is not limited to, the following: forwarding via the I2C bus, i.e., using the I2C protocol to send the power control request as a data packet to the thermal controller; forwarding via the SPI bus, i.e., transmitting the power control request via the serial peripheral device interface; and forwarding via GPIO (General Purpose Input / Output) pins, i.e. using direct pin connections to transmit the power control request through changes in high and low levels.

[0035] Optionally, in this embodiment, in addition to receiving and forwarding power control requests, the service controller may also, but is not limited to, classify power control requests upon receipt and apply different preliminary processing methods based on the classification. For example, the service controller may, but is not limited to, detecting whether there are pending power control requests in the request queue upon receiving a power control request; detecting the request type of the received power control request to obtain the target request type, detecting the urgency type of the received power control request to obtain the target urgency type, and detecting the source type of the received power control request to obtain the target source type, wherein the request type indicates the power control purpose requested by the power control request, the urgency type indicates the urgency level of the power control request, and the source type indicates the user account level initiating the power control request; matching the target request value for the target request type from the corresponding request types and request values, matching the target urgency value for the target urgency type from the corresponding urgency types and urgency values, and matching the target urgency value from the corresponding urgency types and urgency values. The source type and source value are matched with the corresponding target source value for the target source type. The weights of the target request value, target urgency value, and target source value are calculated based on the request weight, urgency weight, and source weight. The request weight indicates the degree of influence of the request type on the power control request processing order, the urgency weight indicates the degree of influence of the urgency type on the power control request processing order, and the source weight indicates the degree of influence of the source type on the power control request processing order. If the weight sum is greater than or equal to the processing threshold, the received power control request is placed in the priority processing position of the request queue. The request queue includes priority processing positions and general processing positions. Power control requests in the priority processing position are forwarded to the thermal controller by the business controller before being processed by the thermal controller, compared to those in the general processing position. If the weight sum is less than the processing threshold, the received power control request is placed in the general processing position of the request queue. By assigning higher weights to urgent requests, the business controller can ensure that in the event of emergencies such as server overheating or hardware failure, these requests are processed first, allowing for timely power cut-off or restoration and protecting equipment from damage. Linking power control request processing to the account level of the requester allows for automatic adjustment of request processing priority based on user permission levels. This prevents unauthorized operations from interfering with server system stability while meeting the needs of users at different levels. By dynamically adjusting the processing order of power control requests, the power supply status of server nodes can be rationally arranged without affecting business continuity, optimizing energy use and reducing unnecessary power consumption.High-priority power control requests can be executed quickly, especially in emergencies, enabling timely measures to prevent the spread of faults and improve the overall stability and security of the system.

[0036] For example, the business controller in server system A may receive, but is not limited to, the following power control requests: Request A: Power-on request, used to start a backup server node to cope with upcoming peak traffic. Request type: Power-on, Emergency type: Medium, Source type: Operations administrator. Request B: Emergency power-off request, an overheated business node is detected, requiring immediate power-off to prevent hardware damage. Request type: Emergency power-off, Emergency type: High, Source type: System self-check. Request C: User request, requesting to restart the business node where their personal account resides to resolve a minor application failure. Request type: Restart, Emergency type: Low, Source type: Ordinary user. The corresponding processing flow includes: The business controller first detects power control requests waiting to be processed in the request queue. Next, it extracts and matches attributes for each request to obtain the target request value, target emergency value, and target source value. It calculates the weight sum of each request, where request B, due to its extremely high emergency type and relatively low request and source weights, still has a total weight sum far exceeding the processing threshold, and is therefore considered a high priority. Requests A and C are assigned corresponding comprehensive weights based on their different attributes. They are then assigned to either a priority processing position or a general processing position depending on whether they exceed a processing threshold. Ultimately, request B is immediately placed in the priority processing position and forwarded to the thermal controller, which quickly performs an emergency power-down operation to prevent hardware damage caused by overheating. Requests A and C are then placed in either the priority processing position or the general processing position, respectively, based on their comprehensive weights, awaiting further processing.

[0037] Optionally, in the embodiments of this application, the power supply control request may be responded to by a heat dissipation controller, but is not limited to the following: by directly controlling the electronic fuse, that is, by reading the register bit status of the corresponding service node in the heat dissipation register and directly controlling the on / off state of the electronic fuse to change the power supply state; by providing feedback confirmation, that is, after responding to the request of the service controller, the heat dissipation controller can send a confirmation signal to indicate that the power supply state has been adjusted; or by using delay control, that is, after receiving the power-down request, the heat dissipation controller can set a certain delay before performing the power-down operation to ensure safety or meet special test requirements.

[0038] Optionally, in this embodiment, for the server system involved in this application, the power supply voltage can be output by the server system's PSU, but is not limited to. This power supply voltage can be directly output to the heat dissipation node and the service node via cables, and the power supply link of the heat dissipation node is on. However, since the default control of the heat dissipation node to the power supply interface of the service node is to disable the power supply link at this time, the power supply link of the service node can be, but is not limited to, still in a disconnected state until the heat dissipation controller on the heat dissipation node starts working. Then, the control of the heat dissipation node to the power supply interface of the service node switches to allow the power supply link to be on, and the power supply link of the service node is on. The service node then performs the power-on logic internally, and after the service node is powered on, it starts the normal working mode. Through the above method, a phased power supply control mechanism is adopted, that is, the PSU outputs the power supply voltage to the heat dissipation node first, and then the power supply link is allowed to be on to the service node under the control of the heat dissipation controller. This design can bring several benefits. First, in the initial stage of server power-on, the heat dissipation system starts before the service node, which can provide a good heat dissipation environment for the subsequent power-on and operation of the service node. This avoids instantaneous temperature rise caused by rapid power-on, reduces thermal shock to hardware, extends server lifespan, and improves overall system stability. Secondly, by controlling the power supply interfaces of service nodes through the heat dissipation controllers on the heat dissipation nodes, the power supply can be dynamically adjusted according to actual needs during server operation, achieving refined energy management. For example, under low load or idle conditions, the heat dissipation controllers can reduce or shut down unnecessary power supply to service nodes, thereby saving energy and reducing unnecessary power consumption. Finally, the server system proposed in this application, combined with phased power supply control, enables remote and precise control of service nodes. Maintenance personnel can remotely activate the heat dissipation nodes and then selectively control the power supply status of service nodes, completing management activities such as restarting, powering off, and restoring power to server nodes without on-site operation, greatly improving maintenance efficiency and convenience.

[0039] As an optional implementation, a service node includes multiple service sub-nodes, each of which is equipped with a service sub-controller. The service controller includes multiple service sub-controllers.

[0040] Among them, multiple service sub-controllers are connected to the heat dissipation controller, and the power supply interface on each service sub-node is connected to the heat dissipation controller;

[0041] Among them, the target service sub-controller set on the target sub-node among multiple service sub-nodes is also used to receive power-down requests and forward the power-down requests to the heat dissipation controller. The power-down request is used to request the power-down of the alternative sub-nodes among multiple service sub-nodes, and the power supply control request includes the power-down request.

[0042] The heat dissipation controller is also used to respond to power-down requests and control the power-down of the alternative sub-nodes through the power supply interface of the alternative sub-nodes according to the power supply status of the service nodes. The power supply status is used to indicate the situation of power supply to multiple service sub-nodes.

[0043] Optionally, in this embodiment of the application, the server system may, but is not limited to, set up multiple business sub-nodes, and each business sub-node may, but is not limited to, set up a business sub-controller.

[0044] Optionally, in this embodiment, each service sub-controller may, but is not limited to, be connected to a thermal controller. Optionally, for some large server systems, multiple thermal controllers may, but are not limited to, be configured to control heat dissipation nodes. In this case, multiple service sub-controllers may, but are not limited to, be divided into multiple service sub-controller groups. Service sub-controllers within the same service sub-controller group may, but are not limited to, be connected to the corresponding thermal controller of that group. The thermal controller is responsible for responding to power control requests forwarded by the service sub-controllers within its group. Correspondingly, the power supply interface on each service sub-node may, but is not limited to, also be connected to the thermal controller corresponding to the service sub-controller group to which the service sub-controller belongs. Through the above, by dividing multiple service sub-controllers into multiple service sub-controller groups, and establishing a connection between each group and a specific thermal controller, rapid response and processing of power control requests can be achieved. In emergency situations, such as hardware failure or overheating, the service sub-controller group can immediately send a power control request to its corresponding thermal controller. The thermal controller can react quickly, for example, by cutting off the power or adjusting the cooling strategy, effectively preventing potential system crashes and protecting the hardware from damage. In large-scale server cluster systems, adopting a grouped thermal control architecture can significantly improve management efficiency. Maintenance personnel or automated systems can send power control requests to specific business sub-controller groups without operating the entire system. This not only reduces the complexity of network communication but also avoids unnecessary resource consumption, making maintenance activities more efficient and precise.

[0045] Optionally, in this embodiment, the service sub-controllers set on each service sub-node can, but are not limited to, be connected to the network interface set on the heat dissipation node through the local area network switch set on the heat dissipation node, and each service sub-controller can, but is not limited to, receive power control requests from the network through the local area network switch and the network interface.

[0046] Optionally, in this embodiment, the target sub-node may be, but is not limited to, the same business sub-node as the candidate sub-node, or it may be a different business sub-node. That is, the power supply control requirements of the candidate sub-node can be sent to the business sub-controller on the candidate sub-node, or to the business sub-controller on other power-on business sub-nodes other than the candidate sub-node. This setting simplifies the management complexity of the server system.

[0047] As an optional implementation, a heat dissipation register is also provided on the heat dissipation node; wherein, the heat dissipation register includes multiple register bits corresponding one-to-one with each service sub-node; wherein, the target service sub-controller is also used to receive a power-down request; generate a power-down signal for the alternative sub-node according to the power-down request; and write the power-down signal of the alternative sub-node into the alternative register bit corresponding to the alternative sub-node in the multiple register bits.

[0048] Optionally, in the embodiments of this application, a heat dissipation register may be provided on the heat dissipation node, and the power control requirements expressed by the power control request may be transmitted from the service controller on the service node (or the service sub-controller on the service sub-node) to the heat dissipation controller on the heat dissipation node through the heat dissipation register.

[0049] Optionally, in this embodiment, a heat dissipation register is introduced. This heat dissipation register may include, but is not limited to, multiple register bits, each register bit corresponding to a specific service sub-node in the server system. This approach allows the target service sub-controller to generate a power-down signal for a specified alternative sub-node when it receives a power-down request, and to directly write this power-down signal into the register bit corresponding to the alternative sub-node in the heat dissipation register, thereby remotely controlling the power supply status of the alternative sub-node.

[0050] Optionally, in this embodiment, the target service sub-controller may, but is not limited to, extract the candidate sub-node ID to be powered down from the power-down request upon receiving it. The target service sub-controller then searches for the corresponding register bit in the heat dissipation register, i.e., the candidate register bit, based on the found ID. A binary power-down signal representing the candidate sub-node's power-down instruction is then generated and written into the candidate register bit of the heat dissipation register via a communication protocol such as a data bus or I2C interface.

[0051] In this way, the target service sub-controller can precisely control each register bit in the heat dissipation register, thereby achieving detailed management and remote control of the power supply status of the corresponding service sub-nodes. This overcomes the limitations of traditional centralized control in multi-sub-node environments and improves the operation and maintenance efficiency and energy management capabilities of the server system.

[0052] As an optional implementation, the heat dissipation controller is also used to: detect whether the control signal stored in each register bit changes from a power supply signal to a power-off signal; when the control signal stored in the candidate register bit changes from a power supply signal to a power-off signal, detect the signal status of the control signals stored in multiple register bits to obtain the power supply status of the service node; and control the candidate sub-node to power down in a manner corresponding to the power supply status through the power supply interface of the candidate sub-node.

[0053] Optionally, in the embodiments of this application, different power supply states may correspond to different power-off methods, but are not limited to. For example, the power-off method corresponding to a power supply state in which all control signals stored in multiple register bits are power-off signals is to first control the alternative sub-node to power off, and then control the alternative sub-node to power on after the alternative sub-node has been powered off for a period of time; the order placement method corresponding to a power supply state in which at least one of the control signals stored in multiple register bits is a power supply signal is to control the alternative sub-node to power off.

[0054] Through the above, by continuously monitoring the control signal status stored in each register bit of the heat dissipation register by the heat dissipation controller, the transition from power supply signal to power-off signal of the service sub-node can be identified in real time. This allows for accurate determination of the current power supply status of the service node, and based on this status, the power supply interface of the alternative sub-node can be precisely controlled to power it off. This achieves dynamic monitoring and efficient, automated management of the power supply status of the service nodes in the server system, significantly improving the flexibility and response speed of operation and maintenance, while ensuring the stable operation of the server system and the rational allocation of energy.

[0055] As an optional implementation, the heat dissipation controller is also configured to: control the power supply interface of the alternative sub-node to disconnect the power supply link of the alternative sub-node when the power supply state indicates that the control signals stored in the multiple register bits are all power-off signals; control the power supply interface of the alternative sub-node to connect the power supply link of the alternative sub-node after a target duration of disconnection of the power supply link of the alternative sub-node; and update the control signals stored in the alternative register bits.

[0056] Optionally, in this embodiment of the application, when the power supply status is used to indicate that the control signals stored in multiple register bits are all power-off signals, that is, when performing a power-down operation on the alternative sub-node will cause all business sub-nodes in the entire server system to be in a power-down state, in order to avoid the failure of remote control, the power supply interface of the alternative sub-node can be controlled to connect the power supply link of the alternative sub-node, that is, to power on the alternative sub-node, after disconnecting the power supply link of the alternative sub-node for a target duration.

[0057] The above steps ensure that when all business sub-nodes receive power-down commands and are in a power-off state, the thermal controller can automatically maintain a certain power-off interval before reconnecting the power supply link of the selected sub-node, avoiding the risk of remote control failure. At the same time, by updating the control signals of the alternative register bits in the thermal register, the power supply status of each business sub-node in the server system is monitored, enhancing the reliability of remote management, ensuring the server system's recovery capability and continuous operation under extreme conditions, and improving operation and maintenance efficiency and user experience.

[0058] As an optional implementation, an electronic fuse is installed between the heat dissipation controller and the power supply interface of each service sub-node.

[0059] The heat dissipation controller is also used to send a power-off enable signal to the alternative electronic fuse corresponding to the alternative sub-node when the power supply state indicates that the control signals stored in multiple register bits are all power-off signals. The power-off enable signal is used to control the alternative electronic fuse to blow. It also detects the power-off time length from the current time to the time when the power-off enable signal was sent. When the power-off time length reaches the target duration, it sends a power-on enable signal to the alternative electronic fuse. The power-on enable signal is used to control the alternative electronic fuse to turn on.

[0060] Optionally, in this embodiment of the application, the power supply control of the service sub-node by the heat dissipation controller can be, but is not limited to, controlled by the melting or turning on of the electronic fuse corresponding to the service sub-node.

[0061] Optionally, in this embodiment, the electronic fuse is an intelligent circuit protection device that combines the power-off protection function of a traditional fuse with the intelligent control capabilities of modern electronic technology. Unlike traditional fuses that irreversibly blow under overload or short-circuit conditions, electronic fuses can dynamically cut off or restore power to the circuit after receiving a control signal.

[0062] Optionally, in the embodiments of this application, the basic function of the electronic fuse set in the server system may be, but is not limited to, protecting the electrical safety of the corresponding power supply link. In this application, the characteristics of the electronic fuse are further combined to realize the power-on and power-off control of the service sub-nodes by controlling the electronic fuse.

[0063] Based on the above, by setting electronic fuses between the power supply interfaces of the heat dissipation controller and each business sub-node, and by using the power-off enable signal and power-on enable signal sent by the heat dissipation controller to intelligently control the state of the electronic fuses, precise management of the power supply status of the business sub-nodes in the server system is achieved.

[0064] As an optional implementation, the heat dissipation controller is also configured to: control the power supply interface of the alternative sub-node to disconnect the power supply link of the alternative sub-node when the power supply state indicates that at least one of the control signals stored in the plurality of register bits is a power supply signal.

[0065] Optionally, in the embodiments of this application, when the power supply state is used to indicate that at least one of the control signals stored in the multiple register bits is a power supply signal, that is, when performing the power-down operation on the alternative sub-node will not cause all business sub-nodes in the entire server system to be in a power-down state, the risk of remote control failure caused by the power-down operation on the alternative sub-node is small. Therefore, the power-down operation on the alternative sub-node can be performed directly, without having to power on the alternative sub-node directly at intervals, thus saving the power consumption of the server system.

[0066] As an optional implementation, the corresponding control signal is a reference service sub-controller set on the reference sub-node of the power supply signal, which is also used to receive a power-on request and forward the power-on request to the heat dissipation controller. The power-on request is used to request power to the alternative sub-node after the heat dissipation controller controls the power supply interface of the alternative sub-node to disconnect the power supply link of the alternative sub-node. The power supply control request also includes a power-on request. The heat dissipation controller is also used to respond to the power-on request and control the power supply interface of the alternative sub-node to connect the power supply link of the alternative sub-node.

[0067] Optionally, in this embodiment of the application, the power supply control request may include, but is not limited to, a power-on request in addition to a power-off request.

[0068] Optionally, in this embodiment of the application, after the alternative sub-node is powered down, the alternative sub-node can be powered on by receiving a power-on request through a service sub-controller on another service sub-node that is still powered on (i.e., a reference service sub-controller set on the reference sub-node).

[0069] The above features significantly enhance the internal communication and collaboration capabilities of the server system. By establishing a request forwarding mechanism between the thermal controller and the business sub-controller, fine-grained control over the power supply status of alternative sub-nodes is achieved. When the server system needs to power off other nodes while maintaining power to some nodes, and then power them back on, a smooth transition is ensured without manual intervention, improving the automation level and flexibility of the server system's operation and maintenance.

[0070] As an optional implementation, the reference service sub-controller is also configured to: receive a power-on request; generate a power supply signal for a candidate sub-node based on the power-on request; and write the power supply signal of the candidate sub-node into a candidate register.

[0071] Optionally, in this embodiment, similar to the previous one, the power supply control requirement expressed by the power supply control request can be, but is not limited to, transmitted from the service sub-controller on the service sub-node to the heat dissipation controller on the heat dissipation node through the heat dissipation register.

[0072] As an optional implementation, an electronic fuse is installed between the heat dissipation controller and the power supply interface of each service sub-node; wherein, the heat dissipation controller is also used to detect whether the control signal stored in each register has changed from a power-off signal to a power-on signal; when the control signal stored in the alternative register changes from a power-off signal to a power-on signal, a power supply enable signal is sent to the alternative electronic fuse corresponding to the alternative sub-node, wherein the power supply enable signal is used to control the alternative electronic fuse to conduct.

[0073] Through the above, by setting up electronic fuses between the heat dissipation controller and the power supply interface of the business sub-nodes, dynamic monitoring and intelligent control of the power supply status of the server system are realized. When the heat dissipation controller detects that the business sub-node has changed from a power-off state to a power-required state, it immediately responds and sends a power enable signal to activate the corresponding electronic fuse, thereby restoring power in a timely manner. This ensures that each business sub-node in the server system receives flexible and precise energy supply according to the real-time changes in current business needs, thereby improving the overall response speed, automation management capabilities and energy utilization efficiency of the system.

[0074] As an optional implementation, this application also provides a design method for remotely powering down child nodes in a server system.

[0075] In related technologies, the current power supply scheme for dual-node server systems is as follows: the server system is powered by a PSU, usually by two PSUs. The PSUs supply the P12V voltage output by the PSUs to all nodes and boards of the server system through the PDB power board. Figure 2 This is a schematic diagram of a dual-node power supply architecture according to an embodiment of this application. The specific topology of the power supply architecture of a dual-node server system in the related art is as follows: Figure 2 As shown. However, this power supply design leads to difficulties in remote operation and maintenance (if you want to power off a certain node server, you can only do so by pulling out a single node or powering down the entire machine; other methods cannot be used for single-node remote power-off operation and maintenance), and it also makes it impossible to perform corresponding node power-off cycle tests in the whole machine power-off AC cycle test, which poses a potential threat to the stability of the server system.

[0076] To address the problems in the aforementioned related technologies, this application provides a design method for remote power-off of child nodes in a server system, combining logic chip and circuit design methods with software processing. It should be noted that this application uses a two-node server system topology for illustration, but the solution is not limited to two-node server systems; server systems with more or fewer child nodes can also apply the solution provided in this application.

[0077] Figure 3 This is a schematic diagram illustrating the power-on process of a dual-node server system according to an embodiment of this application. Figure 3 As shown, the power supply provides power to the service sub-nodes and fan boards (i.e., the aforementioned heat dissipation nodes) via cable. The power enters the motherboard or fan board of the sub-nodes, and each service sub-node is equipped with an Efuse (i.e., the aforementioned electronic fuse) for power input protection. The power-on process of the dual-node server system may include, but is not limited to, the following steps: Step ①, after inserting the PSU, the PSU outputs voltage P12V_STBY, which is directly output to the FCB (fan board), node 1, and node 2 via cable; Step ②, after P12V_STBY enters the fan board, the fan board CPLD (i.e., the aforementioned heat dissipation controller) starts working; Step ③, after the fan board CPLD starts working, it sends an enable signal Power_EN (i.e., the aforementioned power supply enable signal) to the Efuse of node 1 and node 2, controlling the node Efuse to enable and open; Step ④, after the node Efuse outputs P12V, it begins to supply power to the internal logic of the node, and the node starts normal operation mode after power-on.

[0078] Based on the above topology, it is possible, but not limited to, to remotely control the power-on and power-off of child nodes using a combination of hardware and software. Figure 4 This is a schematic diagram of a power supply control process for a dual-node server system according to an embodiment of this application. Figure 4As shown, in step (1), the server system is working normally, and the external network can control the BMC (i.e., the aforementioned service controller) of the two nodes through the network switch (LAN SW). When node 1 needs to be powered down, the external network connects to the BMC of node 2 through the network switch. In step (2), after the external network connects to the BMC of node 2, the BMC of node 2 writes the CPLD register (i.e., the aforementioned heat dissipation register) on the fan board (a CPLD register may be set on the fan board, but is not limited to one, and each CPLD on the fan board can access the CPLD register). In step (3), the CPLD on the fan board reads the value stored in the CPLD register and sends Power_DISABLE (i.e., the aforementioned power-off enable signal) to control the Efuse of node 1 to be turned off. This will realize the power-off function of node 1, and the node can be remotely controlled. The remote control implementation scheme of node 2 is similar to that of node 1.

[0079] Optionally, in this embodiment, since the fan board and node 2 are always working normally, the power_EN command of node 1 can also be sent to control the Efuse to open, thus realizing the power-on function of node 1. When a single node is in place, since there are no other nodes controlling it, the CPLD on the fan board can selectively change the power-down command to a command that delays power-on for a period of time after power-down, according to the in-place status of each node in the server system. This enables remote control of the server system with a single node in place. In summary, the power-on and power-off control of a single node in the complete server system can be realized without affecting the operation of other nodes.

[0080] The above features enable real-time monitoring and one-click operation and maintenance of the remote power failure status of multi-node server systems. Maintenance personnel do not need to be physically present on-site; they can simply use the remote management platform to perform operations such as restarting and status detection of the power failure nodes, greatly improving maintenance efficiency and significantly shortening fault recovery time.

[0081] Furthermore, in terms of AC cycle testing for the stability of multi-node server systems, the solution provided in this application can simulate the operating status of multi-node servers under different loads and network environments through systematic test design, and conduct comprehensive testing on scenarios such as AC power switching and voltage fluctuations, accurately capturing potential risks that may affect stability, thereby ensuring the stability of the server more reliably and reducing the risk of business interruption due to hardware stability issues.

[0082] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0083] The server system provided in this application has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only intended to help understand the method and core ideas of this application. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A server system, characterized in that, include: A heat dissipation node and a service node, wherein a heat dissipation controller is installed on the heat dissipation node and a service controller is installed on the service node; The service controller is connected to the heat dissipation controller, and the heat dissipation controller is connected to the power supply interface of the service node. The service controller is used to control the service node; the heat dissipation controller is used to control the heat dissipation node to dissipate heat for the service node. The service controller is further configured to receive a power supply control request and forward the power supply control request to the heat dissipation controller, wherein the power supply control request is used to request power supply control for the service node; the heat dissipation controller is further configured to respond to the power supply control request and perform power supply control on the service node through the power supply interface of the service node.

2. The server system according to claim 1, characterized in that, The service node includes multiple service sub-nodes, and each service sub-node is equipped with a service sub-controller. The service controller includes multiple service sub-controllers. Among them, multiple service sub-controllers are connected to the heat dissipation controller, and the power supply interface on each service sub-node is connected to the heat dissipation controller; The target service sub-controller set on the target sub-node among the plurality of service sub-nodes is further configured to receive a power-down request and forward the power-down request to the heat dissipation controller. The power-down request is used to request the power-down of the alternative sub-node among the plurality of service sub-nodes, and the power supply control request includes the power-down request. The heat dissipation controller is also configured to respond to the power-down request and control the alternative sub-node to power down through the power supply interface of the alternative sub-node according to the power supply status of the service node, wherein the power supply status is used to indicate the situation of supplying power to multiple service sub-nodes.

3. The server system according to claim 2, characterized in that, A heat dissipation register is also provided on the heat dissipation node; The heat dissipation register includes multiple register bits that correspond one-to-one with each of the service sub-nodes; The target service sub-controller is further configured to receive the power-down request; generate a power-down signal for the candidate sub-node based on the power-down request; and write the power-down signal of the candidate sub-node into the candidate register corresponding to the candidate sub-node in the plurality of registers.

4. The server system according to claim 3, characterized in that, The heat dissipation controller is also used for: Detect whether the control signal stored in each of the register bits has changed from a power supply signal to a power off signal; When the control signal stored in the alternative register changes from a power supply signal to a power failure signal, the signal status of the control signal stored in the plurality of registers is detected to obtain the power supply status of the service node. The alternative sub-node is controlled to power down in the manner corresponding to the power supply state via its power supply interface.

5. The server system according to claim 4, characterized in that, The heat dissipation controller is also used for: When the power supply status indicates that the control signals stored in the plurality of register bits are all power-off signals, the power supply interface of the alternative sub-node is controlled to disconnect the power supply link of the alternative sub-node. If the power supply link of the alternative sub-node is disconnected for a target duration, control the power supply interface of the alternative sub-node to connect the power supply link of the alternative sub-node. Update the control signals stored in the alternative register bits.

6. The server system according to claim 5, characterized in that, An electronic fuse is installed between the heat dissipation controller and the power supply interface of each of the service sub-nodes. The heat dissipation controller is further configured to: send a power-off enable signal to the alternative electronic fuse corresponding to the alternative sub-node when the power supply state indicates that the control signals stored in the plurality of register bits are all power-off signals; control the alternative electronic fuse to blow; detect the power-off time length from the current time to the time when the power-off enable signal was sent; and send a power-on enable signal to the alternative electronic fuse when the power-off time length reaches the target duration, control the alternative electronic fuse to turn on.

7. The server system according to claim 4, characterized in that, The heat dissipation controller is also used for: When the power supply state indicates that at least one of the control signals stored in the plurality of register bits is a power supply signal, the power supply interface of the alternative sub-node is controlled to disconnect the power supply link of the alternative sub-node.

8. The server system according to claim 7, characterized in that, The corresponding control signal is a reference service sub-controller set on the reference sub-node of the power supply signal, which is also used to receive a power-on request and forward the power-on request to the heat dissipation controller. The power-on request is used to request power to the alternative sub-node after the heat dissipation controller controls the power supply interface of the alternative sub-node to disconnect the power supply link of the alternative sub-node. The power supply control request also includes the power-on request. The heat dissipation controller is also used to respond to the power-on request and control the power supply interface of the alternative sub-node to connect the power supply link of the alternative sub-node.

9. The server system according to claim 8, characterized in that, The reference service sub-controller is further configured to: Receive the power-on request; Generate the power supply signal for the alternative sub-nodes based on the power-on request; Write the power supply signal of the candidate sub-node into the candidate register.

10. The server system according to claim 9, characterized in that, An electronic fuse is installed between the heat dissipation controller and the power supply interface of each of the service sub-nodes. The heat dissipation controller is further configured to detect whether the control signal stored in each of the register bits changes from a power-off signal to a power-on signal; when the control signal stored in the alternative register bit changes from a power-off signal to a power-on signal, a power-on enable signal is sent to the alternative electronic fuse corresponding to the alternative sub-node, wherein the power-on enable signal is used to control the alternative electronic fuse to conduct.

Images