Server

Abstract

The embodiment of the invention provides a server, and relates to the technical field of electronic equipment. The server comprises a case, a management board, a node board group and at least one functional module. The chassis includes a first zone and a second zone. In the first direction, the first region is closer to the first side of the chassis than the second region. The management board is arranged in the second area. And the management board is provided with a management module for managing the work of the server. The gusset plate set comprises a plurality of gusset plates which are arranged in a stacked mode. The node plate is provided with a business processing module used for processing business, and the function module is used for cooperating with the business processing module to realize data transmission or processing. Wherein the node plate group and / or the at least one functional module are / is arranged in the first area and are / is in pluggable connection with the management plate, so that the node plate group and / or the at least one functional module can be pulled out or pushed in from the first side of the case. Therefore, maintenance and replacement of the node plate group and / or the function module can be realized without opening the cover plate of the case, the maintenance convenience of the server is improved, and the operation and maintenance convenience is improved.

Description

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office on July 15, 2024, with application number 202410950481.6 and application name “Server”, the entire contents of which are incorporated by reference in this application. Technical Field

[0002] The present application relates to the technical field of computing devices, and in particular to a server. Background Art

[0003] The server includes a chassis, which may include a cover plate, a bottom plate and side walls, wherein the side walls are arranged between the bottom plate and the cover plate and are respectively connected to the bottom plate and the cover plate, so that the cover plate, the bottom plate and the side walls can enclose a receiving cavity. Hardware units of the server, such as node boards and functional modules, are arranged in the receiving cavity.

[0004] Usually, when the hardware unit of the server is inspected, the cover of the chassis needs to be opened so that the hardware unit inside the chassis can be exposed, which affects the convenience of server inspection and causes inconvenience in operation and maintenance. In addition, rack-mounted servers are usually placed in a cabinet. When the server is a rack-mounted server, the server needs to be taken out of the cabinet first, and then the cover needs to be removed, which affects the convenience of server inspection and causes inconvenience in operation and maintenance. Summary of the invention

[0005] The purpose of the embodiments of the present application is to provide a server for improving the convenience of server maintenance, thereby improving the convenience of operation and maintenance.

[0006] The embodiments of the present application provide the following technical solutions:

[0007] On the one hand, an embodiment of the present application provides a server. The server includes a chassis, a management board node board group and at least one functional module. The chassis includes a first area and a second area. Along the first direction, the first area is closer to the first side of the chassis relative to the second area. The management board is arranged in the second area. A management module for managing the work of the server is arranged on the management board. The node board group includes a plurality of node boards arranged in a stacked manner. A business processing module for processing business is arranged on the node board, and the functional module is used to cooperate with the business processing module to realize data transmission or processing. Among them, the node board group and / or at least one functional module are arranged in the first area and are pluggable and connected to the management board. In this way, the node board group and / or at least one functional module can be pulled out or pushed in from the first side of the chassis.

[0008] In the embodiment of the present application, the node board group and / or the functional module are pluggable and connected to the management board, so that the management board, the node board group and the functional module can be physically independent of each other, and they are electrically connected through connectors or cable connections to achieve communication with each other. In this way, the node board group and / or the functional module can be decoupled from the management board separately.

[0009] The first area is close to the first side of the chassis, so that the node board group and / or at least one functional module can be pulled out or pushed in from the first side of the chassis. For example, by pulling out, the node board group and / or at least one functional module arranged in the first area can be decoupled from the management board, and at this time, the node board group and / or at least one functional module arranged in the first area can be separated from the chassis and exposed; by pushing in, the node board group and / or at least one functional module can be arranged in the first area and plugged into the management board.

[0010] When it is necessary to inspect and repair the node board group and / or at least one functional module arranged in the first zone, there is no need to open the upper cover of the chassis. Instead, the node board group and / or at least one functional module can be inspected and repaired by pulling out and pushing in, thereby improving the convenience of operation and maintenance, thereby improving the operation and maintenance efficiency.

[0011] In addition, for rack-mounted servers, the node board group and / or at least one functional module can be overhauled by pulling out and pushing in, without taking the server out of the cabinet, thereby improving the convenience of operation and maintenance and thus improving the efficiency of operation and maintenance.

[0012] In some possible implementations, the chassis also includes a third zone, and along the first direction, the second zone is arranged between the first zone and the third zone, and the third zone is closer to the second side of the chassis relative to the second zone. The node board group is arranged in the first zone and / or the third zone; and / or, at least one functional module is arranged in the first zone and / or the third zone. When the node board group is arranged in the first zone, each node board in the node board group can be respectively pulled out or pushed in from the first side of the chassis by being pluggable connected to the management board respectively. When the node board group is arranged in the third zone, each node board in the node board group can be respectively pulled out or pushed in from the second side of the chassis by being pluggable connected to the management board respectively. When the functional module is arranged in the first zone, it can be pulled out or pushed in from the first side of the chassis by being pluggable connected to the management board. When the functional module is arranged in the third zone, it can be pulled out or pushed in from the second side of the chassis by being pluggable connected to the management board. Such arrangement enables any node board and / or at least one functional module of the node board group to be pulled out or pushed in from the first side of the chassis, and enables any node board and / or at least one functional module of the node board group to be pulled out or pushed in from the second side of the chassis. In this way, when it is necessary to overhaul the node board group and / or at least one functional module arranged in the first zone and the third zone, it is not necessary to open the upper cover of the chassis, but it is possible to overhaul any node board and / or at least one functional module of the node board group by pulling out and pushing in, thereby overhauling the server on both sides (front and rear sides) of the server, improving the convenience of operation and maintenance, thereby improving the efficiency of operation and maintenance. In addition, for a rack-mounted server, the overhaul of the node board group and / or at least one functional module is completed by pulling out and pushing in, without taking the server out of the cabinet, improving the convenience of operation and maintenance, thereby improving the efficiency of operation and maintenance. In addition, the multiple node boards in the node board group are pluggably connected to the management board respectively, so that each node board can be decoupled from the management board separately, thereby improving the convenience of operation and maintenance.

[0013] In some possible implementations, the length of the first zone along the first direction is equal to the length of the third zone along the first direction. This arrangement reduces the restrictions on the node board group and the functional module caused by the physical space of the first zone and the third zone, so that the node board group and / or the functional module can be flexibly set in the first zone and the third zone. In this way, by adjusting the specifications, types and quantities of the node board group and the functional modules set in the first zone and / or the third zone, the server can achieve different configurations without replacing the chassis, thereby improving the adaptability of the chassis, realizing converged production, meeting different needs, and reducing production costs. It can be understood that the outgoing side of the server is usually close to the side where the functional module is located. The functional module is flexibly arranged in the first zone and / or the third zone, so that the server can achieve the first side outgoing line, the second side outgoing line and the first side and the second side outgoing line, meet different needs, and improve the applicability and flexibility of the chassis.

[0014] In some possible implementations, a first opening is formed on the first side of the chassis, and the node plate group and / or functional module disposed in the first zone is configured to be pulled out or pushed in from the first opening. A second opening is formed on the second side of the chassis, and the node plate group and / or functional module disposed in the third zone is configured to be pulled out or pushed in from the second opening. It can be understood that the first zone is in communication with the first opening, so that the node plate group and / or functional module disposed in the first zone can be pulled out or pushed in from the first opening. The third zone is in communication with the second opening, so that the node plate group and / or functional module disposed in the third zone can be pulled out or pushed in from the second opening.

[0015] In some possible implementations, the chassis further includes a first protective plate, which is close to the opening of the chassis relative to the node plate group. The number of the first protective plates is multiple, and the multiple first protective plates are connected to the multiple node plates in the node plate group in a one-to-one correspondence. A first handle bar is provided on any first protective plate, and the first handle bar is used to pull out or push in the node plate connected to the first protective plate. It can be understood that the first protective plate is connected to the node plate, and the first handle bar is provided on the first protective plate. By pulling the first handle bar, the node plate can be pulled out; by pushing the first handle bar, the node plate can be pushed in, which improves the convenience of pulling out and pushing in the node plate, thereby improving the convenience of operation and maintenance. In addition, the multiple first protective plates are connected to the multiple node plates of the node plate group in a one-to-one correspondence. By pulling the first handle bar, any node plate in the node plate group can be pulled out; by pushing the first handle bar, any node plate in the node plate group can be pushed in, which reduces the mutual influence between the multiple node plates in the node plate group and improves the convenience of operation and maintenance. In addition, the first protective plate can be used as a false panel of the chassis to protect the node board group, and can also improve the aesthetics of the chassis.

[0016] In some possible implementations, an indicator board is provided on the first protection plate, and an indicator light and / or a control button corresponding to the working state of the node board are provided on the indicator board. It can be understood that the indicator light can display the working state of the node board connected to the first protection plate, and the control button can control the node board connected to the first protection plate, thereby improving the convenience of operation and maintenance.

[0017] In some possible implementations, the chassis further includes a plurality of partitions extending along a first direction. The plurality of partitions include a first partition and a second partition. The first partition is disposed in the first zone and divides the first zone into a plurality of first sub-zones. The second partition is disposed in the third zone and divides the third zone into a plurality of third sub-zones. The position of the partition along the second direction is adjustable. The second direction intersects with the first direction. With such a configuration, the space within the chassis can be divided, the rational use of the entire internal space of the chassis can be achieved, and the space utilization rate within the chassis can be improved. Furthermore, the position of the partitions (including the first partition and the second partition) along the second direction is adjustable, which improves the flexibility of the space division within the chassis, meets different needs, and enables the chassis to support different configurations of the server.

[0018] In some possible implementations, the management board includes a management board body and a connector module, the management board body and the connector module are electrically connected, and the management module is arranged on the management board body. The connector module is arranged on the management board body; or, the connector module is arranged on one side of the management board body along the first direction and is adjacent to the management board body. The connector module includes a fixed plate and a connector mounted on the surface of the fixed plate. There are multiple connectors, and the node board / or the functional module is configured to be plugged with at least one connector. There are two connector modules, and the two connector modules are arranged at intervals along the first direction. Among the two connector modules, the connector module close to the first side of the chassis is the first connector module, and the connector module close to the second side of the chassis is the second connector module. The connector opening of the first connector module faces the first side of the chassis, and the connector opening of the second connector module faces the second side of the chassis. It can be understood that the connector module is provided to include a connector so that the node board and / or the functional module can support hot plugging with the connector module, thereby enabling hot plugging maintenance of the node board and / or the functional module, improving maintenance convenience, and thus improving operation and maintenance efficiency. In addition, the use of a plug-in method can reduce the number of cables and improve the convenience of operation and maintenance. The connector opening of the first connector module faces the first side of the chassis, which improves the convenience of plugging and separating the node board and / or functional module arranged in the first zone with the connector of the first connector module, thereby improving the convenience of operation and maintenance. The connector opening of the second connector module faces the second side of the chassis, which improves the convenience of plugging and separating the node board and / or functional module arranged in the third zone with the connector of the second connector module, thereby improving the convenience of operation and maintenance. Providing a connector module including a first connector module and a second connector module can reduce the mutual influence between the node board group and / or functional module arranged in the first zone and the node board group and / or functional module arranged in the third zone.

[0019] In some possible implementations, the connector includes a high-speed connector, and at least one of the node board group and the functional module is electrically connected to the high-speed connector for transmitting high-speed signals. This configuration enables high-speed signals to be transmitted between the node board and the connector module, and enables high-speed signals to be transmitted between the functional module and the connector module.

[0020] In some possible implementations, the high-speed connector of the first connector module and the high-speed connector of the second connector module are electrically connected via a high-speed cable, and the high-speed cable is routed along the surface of the management board body. It can be understood that the high-speed connector of the first connector module and the high-speed connector of the second connector module are electrically connected via a high-speed cable, so that high-speed signals can be transmitted between the first connector module and the second connector module. Arranging the high-speed cable to route along the surface of the management board body can reduce the interference of the management board body to the high-speed signal, reduce the loss of the high-speed signal, and improve the quality of the high-speed signal. In addition, the high-speed cable can be routed obliquely along the surface of the management board body without having to be routed along the edge of the bottom plate of the chassis, thereby shortening the length of the high-speed cable, thereby shortening the transmission path of the high-speed signal, and facilitating shortening the transmission time of the high-speed signal.

[0021] In some possible implementations, the server further includes a fan, and the fan discharges air along a first direction. The fan is fixed on the management board body, and there is a gap between the bottom of the fan and the surface of the management board body, and the gap is for high-speed cable routing. It can be understood that the management board body is arranged in the second zone, and the fan is fixed on the management board body, so that the fan can be arranged in the second zone, that is, the fan can be arranged between the first zone and the third zone. The fan discharges air along the first direction, so that the fan can dissipate heat for the node board group and / or functional module arranged in the first zone, and the node board group and / or functional module arranged in the third zone, thereby reducing the risk of the server not being able to work normally due to excessive temperature. There is a gap between the bottom of the fan and the surface of the management board body, and the gap is for high-speed cable routing, which can reduce the mutual influence between the fan and the high-speed cable, and improve the space utilization rate in the chassis.

[0022] In some possible implementations, the connector includes a low-speed connector, the low-speed connector is electrically connected to the management board body, and at least one of the node board group and the functional module is pluggably connected to the low-speed connector for transmitting low-speed signals. This arrangement enables low-speed signals to be transmitted between the management module (BMC) and the node board group, and enables low-speed signals to be transmitted between the BMC and the functional modules, thereby achieving communication between the BMC and the node boards and the functional modules, and enabling the BMC to monitor and manage the node boards and the functional modules.

[0023] In some possible implementations, the node plate group includes a first node plate and a second node plate that are stacked. The first node plate is closer to a cover plate of the chassis relative to the second node plate. The connector includes a first connector and a second connector, and the first connector and the second connector are spaced apart along the height direction of the chassis. The first connector is opposite to the cover plate of the second connector chassis. The first node plate is configured to be electrically connected to the first connector, and the second node plate is configured to be electrically connected to the second connector. This arrangement allows the first node plate to be close to the first connector, and the second node plate to be close to the second connector. In this way, the convenience of electrically connecting the first node plate to the first connector can be improved, and the convenience of electrically connecting the second node plate to the second connector can be improved. In addition, the space utilization of the chassis can be improved.

[0024] In some possible implementations, the management board includes a management board body and a connector module, the management board body and the connector module are electrically connected, and the management module is arranged on the management board body. The connector module is arranged on the management board body; or, the connector module is arranged on one side of the management board body along the first direction and is adjacent to the management board body. The connector module includes a fixed plate and a connector mounted on the surface of the fixed plate. There are multiple connectors, and the node board / or the functional module is configured to be plugged with at least one connector. Among the multiple connectors, the openings of some connectors face the first side of the chassis, and the openings of other connectors face the second side of the chassis. It can be understood that the connector module is provided to include a connector so that the node board and / or the functional module can support hot plugging with the connector module, thereby enabling hot plugging maintenance of the node board and / or the functional module, improving the convenience of maintenance, and thus improving the operation and maintenance efficiency. In addition, the use of a plug-in method can reduce the number of cables and improve the convenience of operation and maintenance. The openings of a portion (one, two or more) of the connectors are oriented toward the first side of the chassis, which improves the convenience of plugging and detaching the connectors of the node board and / or functional module disposed in the first zone with the connector module, thereby improving the convenience of operation and maintenance. The openings of another portion of the connectors are oriented toward the second side of the chassis, which improves the convenience of plugging and detaching the connectors of the node board and / or functional module disposed in the third zone with the connector module, thereby improving the convenience of operation and maintenance. In addition, setting the number of connector modules to one can simplify the structure of the connector module and reduce the cost of the server.

[0025] In some possible implementations, the server also includes a power module, and the power module includes a power output interface. The node board includes a first power input interface, and the functional module includes a second power input interface. The connector includes a first power connector, a second power connector, and a third power connector. The third power connector is electrically connected to the power output interface, the first power connector is electrically connected to one of the first power input interface and the second power input interface, and the second power connector is electrically connected to the other of the first power input interface and the second power input interface. In this way, the power module can supply power to the node board group and the functional module through the management board. In addition, the power module can be decoupled from the management board, which improves the convenience of maintenance of the power module. In addition, different power modules can be replaced according to different configurations of the server, which improves the flexibility of use of the server.

[0026] In some possible implementations, the node board group includes a first node board group and a second node board group, one of the first node board group and the second node board group is arranged in the first zone, and the other is arranged in the third zone. One of the power module and the function module is arranged in the first zone, and is spaced apart from the node board group arranged in the first zone along the second direction. The other of the power module and the function module is arranged in the third zone, and is spaced apart from the node board group arranged in the third zone along the second direction. The second direction intersects with the first direction. It can be understood that setting the number of node board groups to two can improve the computing performance of the server.

[0027] In some possible implementations, the first node plate group and the second node plate group are staggered in the first direction. Such an arrangement can reduce the risk of heat accumulation in the chassis, so that the heat can be evenly distributed in the chassis, which is beneficial to the heat dissipation of the chassis.

[0028] In some possible implementations, a plurality of MCIO connectors are further provided on the node board, and along the first direction, at least one MCIO connector is adjacent to the edge of the node board. The MCIO connectors of each node board in the first node board group are electrically connected, and the MCIO connectors of each node board in the second node board group are electrically connected; or, the MCIO connectors of each node board in the first node board group are electrically connected to the functional module, and the MCIO connectors of each node board in the second node board group are electrically connected to the functional module. In this way, by changing the electrical connection relationship (series or parallel) between the node boards in the node board group and the electrical connection relationship between the high-speed connectors, the server can have different working modes (2P mode or 4P mode), meet different needs, and improve the flexibility of the server. In addition, the operation is simple, and there is no need to set up a structure such as a central backplane, and there is no need to open the cover of the chassis, which improves the convenience of operation and maintenance. In addition, at least one MCIO connector is adjacent to the edge of the node board, which improves the convenience of plugging and unplugging the high-speed cable and the MCIO connector, that is, it improves the convenience of changing the electrical connection relationship between the node boards in the node board group.

[0029] In some possible implementations, the chassis further includes a mounting ear module. The mounting ear module includes a first mounting ear and a second mounting ear, and the first mounting ear and the second mounting ear are arranged on both sides of the chassis along the second direction and are respectively connected to the chassis. The second direction intersects with the first direction. The position of the mounting ear module along the first direction is adjustable. With such an arrangement, the installation position of the mounting ear module can be adjusted according to the application scenario, so that the chassis can meet different configurations of the server, thereby improving the adaptability and use flexibility of the chassis.

[0030] In some possible implementations, the management module includes a baseboard management controller. This configuration allows the management module to be decoupled from the node board and the functional module. In this way, multiple node boards of the node board group are respectively plugged into the management board, and multiple functional modules are respectively plugged into the management board, so that each node board can be maintained separately, and each functional module (such as an IO card, a GPU card, or a DPU card, etc.) can be maintained separately, which improves the convenience and flexibility of operation and maintenance.

[0031] In some possible implementations, the management module includes a first management module and a second management module. The first management module and the second management module are electrically connected. The server includes a first node, a second node, a third node and a fourth node. The first node and the second node are electrically connected to the first management module respectively, and the third node and the fourth node are electrically connected to the second management module respectively. The functional module is used to coordinate with at least one of the first node, the second node, the third node and the fourth node for data transmission or processing. One of the first management module and the second management module is used to manage the functional module. When the server is in the first working state, the first node, the second node, the third node and the fourth node are configured to work separately. The first management module is used to manage at least one of the first node and the second node; and / or, the second management module is used to manage at least one of the third node and the fourth node. It can be understood that the first management module and the second management module are electrically connected so that the first management module and the second management module can communicate without signal convergence on the management board body, and without other logic devices (such as complex programmable logic devices, complex programmable logic device, CPLD) participating in the communication between the first management module and the second management module. In this way, the communication path can be shortened and the efficiency of information interaction between the first management module and the second management module can be improved. The first management module is electrically connected to the first node and the second node, so that the first management module can monitor and manage at least one of the first node and the second node. The second management module is electrically connected to the third node and the fourth node, so that the second management module can monitor and manage at least one of the third node and the fourth node. In this way, two management modules can manage four nodes (one management module manages two nodes), and there is no need to set the number of management modules to be the same as the number of nodes, thereby reducing the number of management modules, simplifying the structure of the server, and reducing the cost of the server. In addition, the management scheme of the server for multiple nodes is simplified.

[0032] In some possible implementations, when the server is in the second working state, the first node and the second node are connected in series to form a first node group, and the third node and the fourth node are connected in series to form a second node group. The first node group and the second node group are configured to work separately. One of the first management module and the second management module is used to manage the first node group and / or the second node group. In this way, one management module can manage two node groups, and there is no need to set the number of management modules to be the same as the number of node groups, which reduces the number of management modules, simplifies the structure of the server, and reduces the cost of the server. And it simplifies the management scheme of the server for multiple nodes. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] 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.

[0034] Figure 1 A top view of a server provided for some embodiments of the present application;

[0035] Figure 2 A top view of a server provided for some other embodiments of the present application;

[0036] Figure 3 A schematic diagram of the structure of a server provided in some embodiments of the present application;

[0037] Figure 4 A schematic diagram of the structure of a connector module provided in some embodiments of the present application;

[0038] Figure 5 A schematic block diagram of the structure of a power supply module provided in some embodiments of the present application;

[0039] Figure 6 A top view of a server provided for some other embodiments of the present application;

[0040] Figure 7 A schematic diagram of the structure of a node plate provided in some embodiments of the present application;

[0041] Figure 8 A schematic block diagram of the structure of a server provided in some embodiments of the present application;

[0042] Fig. 9 A schematic block diagram of the structure of a server provided in some other embodiments of the present application;

[0043] Fig.10 A schematic block diagram of the structure of a server provided in some embodiments of the present application;

[0044] Fig.11 A schematic block diagram of the structure of a server provided in some embodiments of the present application;

[0045] Fig.12 A schematic diagram of the connection relationship between the management module and other components provided in some embodiments of the present application. DETAILED DESCRIPTION

[0046] 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.

[0047] Unless the context requires otherwise, throughout the specification and claims, the term "comprise" and other forms 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 "one embodiment", "some embodiments", "exemplary embodiments", "example", "specific 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.

[0048] 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.

[0049] When describing some embodiments, the expressions "coupled" and "connected" and their derivatives may be used. The term "connected" should be understood in a broad sense. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium. The term "coupled" indicates, for example, that two or more components are in direct physical or electrical contact. The term "coupled" or "communicatively coupled" may also refer to two or more components that are not in direct contact with each other, but still cooperate or interact with each other. The embodiments disclosed herein are not necessarily limited to the contents of this document.

[0050] “A and / or B” includes the following three combinations: A only, B only, and a combination of A and B.

[0051] It should be understood that when a layer or an element is referred to as being on another layer or main board, it may be directly on the other layer or main board, or there may be an intervening layer between the layer or element and the other layer or main board.

[0052] Exemplary embodiments are described herein with reference to cross-sectional views and / or plan views that are idealized exemplary drawings. In the drawings, the thickness of the layers and the area of ​​the regions are exaggerated for clarity. Therefore, variations in the shapes relative to the drawings due to, for example, manufacturing techniques and / or tolerances are conceivable. Therefore, the exemplary embodiments should not be interpreted as being limited to the shapes of the regions shown herein, but include shape deviations due to, for example, manufacturing. For example, an etched region shown as a rectangle will typically have curved features. Therefore, the regions shown in the drawings are schematic in nature, and their shapes are not intended to illustrate the actual shapes of the regions of the device, and are not intended to limit the scope of the exemplary embodiments.

[0053] Figure 1 A top view of a server 1000 provided for some embodiments of the present application.

[0054] See also Figure 1 , an embodiment of the present application provides a server 1000. For example, the type of the server 1000 may be a cabinet server, a rack server, a high-density server, etc., and the embodiment of the present application does not specifically limit the type of the server 1000.

[0055] The server 1000 may include a chassis 1001, and the chassis 1001 is used to accommodate the hardware units of the server 1000. For example, Figure 1 As shown, the server 1000 may include hardware units such as a node board 1002, a function module 1003, a fan 1004, a power module 1005, an input / output interface (IO interface) 1007, and a management module 1008. It is understandable that the server 1000 may also include other hardware units in addition to the above hardware units.

[0056] Depending on the configuration of the server 1000, the height of the chassis 1001 can be divided into 1U, 2U, 4U, etc. Among them, the U digit is used to indicate the size of the chassis 1001 and the number of hardware units that can be accommodated. A lower U digit means that the chassis 1001 is smaller, while a higher U digit means that the chassis 1001 is larger. For example, the height of one U is approximately equal to 44.45 mm (1.75 inches). That is, the height of a 1U chassis is approximately 44.45 mm, the height of a 2U chassis is approximately 88.9 mm, and the height of a 3U chassis is approximately 133.5 mm.

[0057] For example, the chassis 1001 may include a cover plate, a bottom plate and a side wall, wherein the side wall is disposed between the bottom plate and the cover plate and is respectively connected to the bottom plate and the cover plate, so that the cover plate, the bottom plate and the side wall can enclose a receiving cavity. The hardware unit of the server 1000 is disposed in the receiving cavity, so that the chassis 1001 can protect the hardware unit.

[0058] Generally, when the hardware units of the server 1000 need to be inspected, the cover of the chassis 1001 needs to be opened to expose the hardware units in the chassis 1001, which causes inconvenience in operation and maintenance. A rack-mounted server is usually placed in a cabinet. When the server 1000 is a rack-mounted server, the server 1000 needs to be taken out of the cabinet first, and then the cover needs to be removed, which further causes inconvenience in operation and maintenance.

[0059] Figure 2 A top view of a server 1000 provided in some other embodiments of the present application. Based on this, an embodiment of the present application provides a server 1000.

[0060] In some examples, such as Figure 1 and Figure 2 As shown, the server 1000 includes a chassis 1001, a management board 1100, a node board group 1150 and at least one functional module 1003. The chassis 1001 includes a first area A and a second area B. Along the first direction X, the first area A is closer to the first side 1001a of the chassis 1001 than the second area B.

[0061] For example, the chassis 1001 may include a first side 1001a and a second side 1001b that are arranged opposite to each other along a first direction X. The first direction X may be a length direction of the chassis 1001. One of the first side 1001a and the second side 1001b may be understood as the front side of the chassis 1001 (server 1000), and the other may be understood as the rear side of the chassis 1001 (server 1000).

[0062] It can be understood that the first area A and the second area B are virtual areas of the chassis 1001. Figure 1 and Figure 2 As shown, the first area A and the second area B may be arranged along the first direction X, so that the first area A can be closer to the first side 1001 a of the chassis 1001 than the second area B.

[0063] For example, the chassis 1001 may include a partition structure to separate the first area A and the second area B, or the chassis 1001 may not include a structure to separate the first area A and the second area B. The first area A and the second area B may be regular shapes, such as a rectangle or a square. Alternatively, the first area A and the second area B may also be irregular structures. The shapes and areas of the first area A and the second area B may be the same or different.

[0064] like Figure 1 and Figure 2 As shown, the management board 1100 is disposed in the second zone B. The management board 1100 is provided with a management module 1008 for managing the server 1000 .

[0065] For example, the management board 1100 may be fixedly connected to the bottom plate of the chassis 1001 , or the management board 1100 may be detachably connected to the bottom plate of the chassis 1001 .

[0066] In some examples, such as Figure 1 and Figure 2 As shown, the management board 1100 may include a management board body 1110, the management board body 1110 may be a printed circuit board (PCB), and the management module 1008 is disposed on the management board body 1110. For example, the management module 1008 may be a baseboard management controller (BMC). Alternatively, the management module 1008 may also be other managers. The embodiment of the present application takes the management module 1008 as a BMC as an example, and continues to illustrate.

[0067] The management module 1008 is used to manage the hardware units of the server 1000. Taking the management module 1008 as a BMC as an example, the BMC can monitor the temperature and voltage of important hardware of the server 1000, and can provide a remote management interface for the hardware of the server 1000.

[0068] For example, a plurality of management modules 1008 may be provided on the management board body 1110. Figure 1 and Figure 2 As shown, the management module 1008 may include a first management module 1008a and a second management module 1008b. The first management module 1008a and the second management module 1008b are spaced apart from each other on the management board body 1110 and are electrically connected to the management board body 1110 respectively.

[0069] The management module 1008 can be snap-fitted to the management board body 1110. Alternatively, the management module 1008 can also be set on the management board body 1110 by welding or plugging. The connection methods of multiple management modules 1008 and the management board body 1110 can be the same or different. The embodiment of the present application does not further limit the connection method of the management module 1008 and the management board body 1110.

[0070] The server 1000 may include one node board group 1150 or multiple node board groups 1150. The node board group 1150 includes multiple node boards 1002 that are stacked. It can be understood that the multiple node boards 1002 can be stacked along the height direction Z of the chassis 1001. The stacking of multiple node boards 1002 can improve the space utilization of the chassis 1001.

[0071] When there are multiple node plate groups 1150, the number of node plates 1002 included in each node plate group 1150 may be the same or different. In some examples, one node plate group 1150 may include two node plates 1002 stacked together.

[0072] The server 1000 may include one functional module 1003 or multiple functional modules 1003. The number of node board groups 1150 and functional modules 1003 may be the same or different. The embodiment of the present application does not further limit the number of node board groups 1150 and functional modules 1003.

[0073] like Figure 1 and Figure 2 As shown, a business processing module 1201 for processing business is provided on the node board 1002, and the function module 1003 is used to cooperate with the business processing module 1201 to realize data transmission or processing.

[0074] The node board 1002 may be a PCB, and the number of the business processing modules 1201 may be multiple, and the multiple business processing modules 1201 are arranged on the PCB at intervals. It can be understood that the business processing module 1201 is hardware for processing business, and the hardware includes at least a central processing unit (CPU), a memory (such as a dual inline memory module, DIMM), etc., and may also include some chips, such as a platform controller hub (PCH) chip or other chips.

[0075] like Figure 1 and Figure 2 As shown, the service processing module 1201 may include a CPU and a registered dual inline memory module (registered DIMM, RDIMM), and a plurality of RDIMMs may be arranged along the second direction Y on both sides of the CPU.

[0076] The second direction Y intersects with the first direction X. For example, the second direction Y is perpendicular to the first direction Y. The second direction Y may be a width direction of the chassis 1001. The height direction Z of the chassis 1001 intersects with the plane where the first direction X and the second direction Y are located. For example, the height direction Z of the chassis 1001 is perpendicular to the plane where the first direction X and the second direction Y are located.

[0077] For example, the service processing module 1201 may include 8 RDIMMs, or may include 16 or 24 RDIMMs. The embodiment of the present application does not further limit the number of RDIMMs.

[0078] Functional module 1003 is hardware for cooperating with business processing module 1201 to realize data transmission or processing. In some examples, functional module 1003 includes at least one of a hard disk, a graphics processing unit (GPU), and a data processing unit (DPU) card.

[0079] The number of function modules 1003 may be multiple, and the types of the multiple function modules 1003 may be the same or different. For example, the function module 1003 may also include an IO expansion card, a disk array card (raid card), and other network cards (such as OCP network cards) other than the DPU card.

[0080] In some examples, the functional module 1003 may include an adapter card (riser card) and a hard disk backplane, and the riser card and the hard disk backplane are pluggable and connected to the management board 1100. Graphics cards, DPU cards, IO expansion cards, disk array cards (raid cards), other network cards (such as OCP network cards), etc. can be plugged into the riser card. The hard disk is plugged into the hard disk backplane.

[0081] It can be understood that the node board 1002 is electrically connected to the function module 1003, so that the function module 1003 can complete data transmission or processing with the business processing module 1201 of the node board 1002. By matching different types of function modules 1003, the node board 1002 can realize different server 1000 configurations to meet different needs.

[0082] In some examples, the node board group 1150 and / or at least one functional module 1003 are disposed in the first zone A and are pluggably connected to the management board 1100. In this way, the node board group 1150 and / or at least one functional module 1003 can be pulled out or pushed into the first side 1001a of the chassis 1001.

[0083] For example, the first zone A may be provided with a node plate group 1150 , or the first zone A may be provided with at least one functional module 1003 , or the first zone A may be provided with a node plate group 1150 and at least one functional module 1003 .

[0084] When there are multiple node plate groups 1150, the multiple node plate groups 1150 may all be arranged in the first zone A, or a portion (one, two or more node plate groups 1150) of the multiple node plate groups 1150 may be arranged in the first zone A, and another portion may be arranged in other zones (e.g., the third zone C). When there are multiple function modules 1003, the multiple function modules 1003 may all be arranged in the first zone A, or a portion (one, two or more) of the multiple function modules 1003 may be arranged in the first zone A, and another portion may be arranged in other zones (e.g., the third zone C).

[0085] like Figure 1 and Figure 2 As shown, when the node plate group 1150 and the functional module 1003 are arranged in the first zone A, the node plate group 1150 and the functional module 1003 can be arranged at intervals along the second direction Y.

[0086] In the embodiment of the present application, the node board group 1150 and / or the functional module 1003 are pluggable and connected to the management board 1100, so that the management board 1100, the node board group 1150 and the functional module 1003 can be physically independent of each other, and they are electrically connected through connectors or cable connections to achieve communication with each other. In this way, the node board group 1150 and / or the functional module 1003 can be decoupled from the management board 1100 separately.

[0087] The first zone A is close to the first side 1001a of the chassis 1001, so that the node board group 1150 and / or at least one functional module 1003 can be pulled out or pushed in from the first side 1001a of the chassis 1001. For example, by pulling out, the node board group 1150 and / or at least one functional module 1003 arranged in the first zone A can be decoupled from the management board 1100, and at this time, the node board group 1150 and / or at least one functional module 1003 arranged in the first zone A can be separated from the chassis 1001, so as to be exposed; by pushing in, the node board group 1150 and / or at least one functional module 1003 can be arranged in the first zone A and plugged into the management board 1100.

[0088] When it is necessary to inspect and repair the node board group 1150 and / or at least one functional module 1003 arranged in the first zone A, there is no need to open the upper cover of the chassis 1001. Instead, the node board group 1150 and / or at least one functional module 1003 can be inspected and repaired by pulling out and pushing in, thereby improving the convenience of operation and maintenance, thereby improving the operation and maintenance efficiency.

[0089] In addition, for rack-mounted servers, the node board group 1150 and / or at least one functional module 1003 can be inspected and repaired by pulling out and pushing in, without removing the server 1000 from the cabinet, thereby improving the convenience of operation and maintenance and thus improving the efficiency of operation and maintenance.

[0090] In some examples, such as Figure 1 and Figure 2 As shown, the chassis 1001 further includes a third area C. Along the first direction X, the second area B is disposed between the first area A and the third area C, and the third area C is closer to the second side 1001 b of the chassis 1001 than the second area B.

[0091] It can be understood that the third zone C is a virtual zone of the chassis 1001. Figure 1 and Figure 2 As shown, along the first direction X, the first zone A, the second zone B and the third zone C are arranged in sequence, and the first zone A is closer to the first side 1001a of the chassis 1001, the third zone C is closer to the second side 1001b of the chassis 1001, and the second zone B is arranged between the first zone A and the third zone C.

[0092] For example, the chassis 1001 may include a partition structure to separate the second area B and the third area C, or the chassis 1001 may not include a structure to separate the second area B and the third area C. The third area C may be a regular shape, such as a rectangle or a square. Alternatively, the third area C may also be an irregular structure. The shapes and areas of the first area A, the second area B, and the third area C may be the same or different.

[0093] like Figure 1 and Figure 2 As shown, the node plate group 1150 is arranged in the first zone A and / or the third zone C; and / or, at least one functional module 1003 is arranged in the first zone A and / or the third zone C.

[0094] It can be understood that when there are multiple node plate groups 1150, the multiple node plate groups 1150 can all be arranged in the first zone A, or the multiple node plate groups 1150 can also be arranged in the third zone C, or a part of the multiple node plate groups 1150 (one, two or more node plate groups 1150) can be arranged in the first zone A, and the other part can be arranged in the third zone C. When the multiple node plate groups 1150 are respectively arranged in the first zone A and the third zone C, the number of the node plate groups 1150 in the first zone A and the number of the node plate groups 1150 in the third zone C can be the same or different.

[0095] When there are multiple functional modules 1003, the multiple functional modules 1003 may all be arranged in the first area A, or the multiple functional modules 1003 may all be arranged in the third area C, or a part (one, two or more) of the multiple functional modules 1003 may be arranged in the first area A, and the other part may be arranged in the third area C. When the multiple functional modules 1003 are arranged in the first area A and the third area C, respectively, the number of the functional modules 1003 in the first area A and the number of the functional modules 1003 in the third area C may be the same or different.

[0096] That is, the first zone A may be provided with one or more node plate groups 1150, may be provided with one or more functional modules 1003, or may be provided with one or more node plate groups 1150 and one or more functional modules 1003. Similarly, the third zone C may be provided with one or more node plate groups 1150, may be provided with one or more functional modules 1003, or may be provided with one or more node plate groups 1150 and one or more functional modules 1003.

[0097] When the node board group 1150 is arranged in the first zone A, each node board 1002 in the node board group 1150 can be respectively pulled out or pushed in from the first side 1001a of the chassis 1001 by being pluggable connected to the management board 1100. When the node board group 1150 is arranged in the third zone C, each node board 1002 in the node board group 1150 can be respectively pulled out or pushed in from the second side 1001b of the chassis 1001 by being pluggable connected to the management board 1100.

[0098] When the function module 1003 is arranged in the first zone A, it can be pulled out from or pushed into the first side 1001a of the chassis 1001 by being pluggable connected to the management board 1100. When the function module 1003 is arranged in the third zone C, it can be pulled out from or pushed into the second side 1001b of the chassis 1001 by being pluggable connected to the management board 1100.

[0099] It can be understood that the first zone A is close to the first side 1001a of the chassis 1001, and each node board 1002 and / or at least one functional module 1003 of the node board group 1150 arranged in the first zone A is connected to the management board 1100 in a pluggable manner, so that each node board 1002 and / or at least one functional module 1003 of the node board group 1150 can be pulled out or pushed in from the first side 1001a of the chassis 1001. The third zone C is close to the second side 1001b of the chassis 1001, and each node board 1002 and / or at least one functional module 1003 of the node board group 1150 arranged in the third zone C is connected to the management board 1100 in a pluggable manner, so that each node board 1002 and / or at least one functional module 1003 of the node board group 1150 can be pulled out or pushed in from the second side 1001b of the chassis 1001.

[0100] For example, by pulling out, any node plate 1002 and / or at least one functional module 1003 of the node plate group 1150 arranged in the first zone A can be decoupled from the management board 1100, and at this time, any node plate 1002 and / or at least one functional module 1003 of the node plate group 1150 arranged in the first zone A can be separated from the chassis 1001, thereby being exposed; by pushing in, any node plate 1002 and / or at least one functional module 1003 of the node plate group 1150 can be arranged in the first zone A and plugged into the management board 1100.

[0101] Similarly, by pulling out, any node plate 1002 and / or at least one functional module 1003 of the node plate group 1150 arranged in the third zone C can be decoupled from the management board 1100, and at this time, any node plate 1002 and / or at least one functional module 1003 of the node plate group 1150 arranged in the third zone C can be separated from the chassis 1001, thereby being exposed; by pushing in, any node plate 1002 and / or at least one functional module 1003 of the node plate group 1150 can be arranged in the third zone C and plugged into the management board 1100.

[0102] When it is necessary to inspect and repair the node board group 1150 and / or at least one functional module 1003 arranged in the first zone A and the third zone C, there is no need to open the upper cover of the chassis 1001. Instead, any node board 1002 and / or at least one functional module 1003 of the node board group 1150 can be inspected and repaired by pulling out and pushing in, thereby realizing the inspection and repair of the server 1000 on both sides (front and rear sides) of the server 1000, thereby improving the convenience of operation and maintenance, thereby improving the operation and maintenance efficiency.

[0103] In addition, for rack-mounted servers, the node board group 1150 and / or at least one functional module 1003 can be inspected and repaired by pulling out and pushing in, without removing the server 1000 from the cabinet, thereby improving the convenience of operation and maintenance and thus improving the efficiency of operation and maintenance.

[0104] Furthermore, the multiple node boards 1002 in the node board group 1150 are respectively pluggably connected to the management board 1100 , so that each node board 1002 can be individually decoupled from the management board 1100 , thereby improving the convenience of operation and maintenance.

[0105] It can be understood that the management module 1008 (BMC) is set on the management board body 1110 of the management board 1100. In this way, a part of the node boards 1002 (one, two or more) of the node board group 1150 is decoupled from the management board 1100, which will not affect the communication between another part of the node boards 1002 of the node board group 1150 and the multiple functional modules 1003 and the management module 1008. Similarly, a part of the management modules 1008 (one, two or more) of the multiple functional modules 1003 is decoupled from the management board 1100, which will not affect the communication between another part of the functional modules 1003 and the node board group 1150 and the management module 1008.

[0106] That is, the management module 1008 is set on the management board body 1110 of the management board 1100, so that the management module 1008 can be decoupled from the node board 1002 and the functional module 1003. In this way, the multiple node boards 1002 of the node board group 1150 are respectively plugged into the management board 1100, and the multiple functional modules 1003 are respectively plugged into the management board 1100, so that each node board 1002 can be maintained separately, and each functional module 1003 (such as an IO card, a GPU card, or a DPU card, etc.) can be maintained separately, which improves the convenience and flexibility of operation and maintenance.

[0107] Taking the case where the functional module 1003 includes a DPU card as an example, when a part (one, two or more) of the plurality of node boards 1002 is decoupled from the management board 1100 , the multi-host function of the DPU card will not be affected.

[0108] In some examples, a first opening (not shown in the figure) is formed on the first side 1001a of the chassis 1001, and the node plate group 1150 and / or the functional module 1003 disposed in the first zone A are configured to be pulled out from or pushed in through the first opening. A second opening (not shown in the figure) is formed on the second side 1001b of the chassis 1001, and the node plate group 1150 and / or the functional module 1003 disposed in the third zone C are configured to be pulled out from or pushed in through the second opening.

[0109] It can be understood that the first zone A is in communication with the first opening, so that the node plate group 1150 and / or the functional module 1003 disposed in the first zone A can be pulled out from the first opening or pushed in. The third zone C is in communication with the second opening, so that the node plate group 1150 and / or the functional module 1003 disposed in the third zone C can be pulled out from the second opening or pushed in.

[0110] In some examples, a detachable first panel is provided at the first opening, and the first panel closes the first opening. By removing the first panel, the node plate group 1150 and / or the functional module 1003 provided in the first zone A can be exposed. A detachable second panel may be provided at the second opening, and the second panel closes the second opening. By removing the second panel, the node plate group 1150 and / or the functional module 1003 provided in the third zone C can be exposed.

[0111] Figure 3 The schematic diagram of the structure of the server 1000 provided in some embodiments of the present application. In other examples, such as Figure 3 As shown, the chassis 1001 further includes a first protective plate 1301, and the first protective plate 1301 is close to an opening (one of the first opening and the second opening) of the chassis 1001 relative to the node plate group 1150. There are multiple first protective plates 1301, and the multiple first protective plates 1301 are connected to the multiple node plates 1002 in the node plate group 1150 in a one-to-one correspondence.

[0112] The first protective plate 1301 is erected on one side of the node plate group 1150. For example, the first protective plate 1301 is perpendicular or approximately perpendicular to the node plate group 1150, so that the first protective plate 1301 can close the opening (one of the first opening and the second opening) of the chassis 1001. The first protective plate 1301 can completely close the opening of the chassis 1001, or the first protective plate 1301 can also close a part of the opening of the chassis 1001.

[0113] It can be understood that when the node plate group 1150 is arranged in the first zone A, the first protection plate 1301 is close to the first opening of the chassis 1001 relative to the node plate group 1150. When the node plate group 1150 is arranged in the third zone C, the first protection plate 1301 is close to the second opening of the chassis 1001 relative to the node plate group 1150. When multiple node plate groups 1150 are respectively arranged in the first zone A and the third zone C, as shown in FIG. Figure 3 As shown, there are two first protective plates 1301, one of which is close to the first opening of the chassis 1001 relative to the node plate group 1150 set in the first zone A, and the other first protective plate 1301 is close to the second opening of the chassis 1001 relative to the node plate group 1150 set in the third zone C.

[0114] For example, the first protection plate 1301 and the gusset plate 1002 may be connected by bolts or by clamping, etc. The connection methods between different first protection plates 1301 and gusset plates 1002 may be the same or different.

[0115] In some examples, a first handle bar is disposed on any first protective plate 1301 , and the first handle bar is used to pull out or push in the node plate 1002 connected to the first protective plate 1301 .

[0116] It can be understood that the first protective plate 1301 is connected to the node plate 1002, and the first handle bar is set on the first protective plate 1301. By pulling the first handle bar, the node plate 1002 can be pulled out; by pushing the first handle bar, the node plate 1002 can be pushed in, thereby improving the convenience of pulling out and pushing in the node plate 1002, thereby improving the convenience of operation and maintenance.

[0117] In addition, multiple first protective plates 1301 and multiple node plates 1002 of the node plate group 1150 are connected one by one. By pulling the first handle bar, any node plate 1002 in the node plate group 1150 can be pulled out; by pushing the first handle bar, any node plate 1002 in the node plate group 1150 can be pushed in, thereby reducing the mutual influence between the multiple node plates 1002 in the node plate group 1150 and improving the convenience of operation and maintenance.

[0118] In addition, the first protective plate 1301 can be used as a false panel of the chassis 1001 to protect the node board group 1150 , and can also improve the aesthetics of the chassis 1001 .

[0119] For example, when the node plate group 1150 and the functional module 1003 are arranged in the same zone (e.g., the first zone A or the third zone C), the first protective plate 1301 can be connected to the node plate group 1150 and the functional module 1003 arranged in the same zone. At this time, the node plate 1002 and the functional module 1003 can be pulled out by pulling the first handle bar; the node plate 1002 and the functional module 1003 can be pushed in by pushing the first handle bar.

[0120] When the node plate group 1150 and the functional module 1003 are arranged in the same zone (for example, the first zone A or the third zone C), the first protective plate 1301 is only connected to the node plate 1002 but not to the functional module 1003, or when the node plate group 1150 and the functional module 1003 are arranged in different zones, the chassis 1001 may include a second protective plate (not shown in the figure).

[0121] The second protective plate is erected on one side of the functional module 1003 close to the opening (one of the first opening and the second opening), and for example, the second protective plate is perpendicular or approximately perpendicular to the functional module 1003. The second protective plate is connected to the functional module 1003, and for example, the second protective plate and the functional module 1003 can be connected by bolts, or by clamping or the like.

[0122] When the node plate group 1150 and the functional module 1003 are arranged in the same zone (e.g., the first zone A or the third zone C), the first protective plate 1301 and the second protective plate jointly close the opening of the chassis 1001. When the node plate group 1150 and the functional module 1003 are arranged in different zones, the first protective plate 1301 closes one of the first opening and the second opening, and the second protective plate closes the other of the first opening and the second opening.

[0123] The second protective plate may be provided with a second handle bar. By pulling the second handle bar, the functional module 1003 can be pulled out; by pushing the second handle bar, the functional module 1003 can be pushed in.

[0124] In some examples, continue to refer to Figure 3 The first protection plate 1301 is provided with an indicator plate 1302 , and the indicator plate 1302 is provided with indicator lights and / or control buttons corresponding to the working status of the node plate 1002 .

[0125] It can be understood that the indicator light and / or control button are electrically connected to the node plate 1002. The indicator light can display the working state of the node plate 1002 connected to the first protection plate 1301, and the control button can control the node plate 1002 connected to the first protection plate 1301, thereby improving the convenience of operation and maintenance.

[0126] When there are multiple first protective plates 1301, an indicator board 1302 can be set on each first protective plate 1301, so that the working status of each node board 1002 in the node board group 1150 can be obtained according to the indicator lights on different indicator boards 1302, and each node board 1002 in the node board group 1150 can be controlled through the control buttons on different indicator boards 1302, thereby improving the convenience of operation and maintenance.

[0127] Depending on the usage scenario, the server 1000 can have different configurations, such as AI reasoning servers, storage servers, etc. Here, "different configurations" are reflected in at least one dimension, such as the number of CPUs, the number of memories, the number of hard disks, the number of power modules 1005, the number of DPUs, the number of node boards 1002, and the connection relationship between multiple node boards 1002. Due to the different number of hardware caused by different configurations, it is usually necessary to match chassis 1001 of different sizes for assembly. As a result, the chassis 1001 of servers 1000 with different configurations are basically unable to be interchanged, which increases the cost of the server 1000.

[0128] Based on this, in some examples of the present application, the length of the first region A along the first direction X is equal to the length of the third region C along the first direction X.

[0129] It can be understood that setting the first zone A and the third zone C to have the same size in the first direction X reduces the restrictions on the node plate group 1150 and the functional module 1003 caused by the physical space of the first zone A and the third zone C, so that the node plate group 1150 and / or the functional module 1003 can be flexibly set in the first zone A and the third zone C.

[0130] For example, in one configuration of the server 1000, the node board group 1150 may be placed in the first zone A, and at least one functional module 1003 may be placed in the third zone C. In another configuration of the server 1000, the node board group 1150 may be placed in both the first zone A and the third zone C, and at least one functional module 1003 may be placed in one of the first zone A and the third zone C.

[0131] By adjusting the specifications, types and quantity of the node board group 1150 and the functional module 1003 arranged in the first zone A and / or the third zone C, the server 1000 can achieve different configurations without replacing the chassis 1001, thereby improving the adaptability of the chassis 1001, realizing production convergence, meeting different needs and reducing production costs.

[0132] It can be understood that the outgoing line side of the server 1000 is usually close to the side where the functional module 1003 is located. The functional module 1003 is flexibly arranged in the first area A and / or the third area C, so that the server 1000 can realize the outgoing line of the first side 1001a, the outgoing line of the second side 1001b, and the outgoing line of the first side 1001a and the second side 1001b, so as to meet different needs and improve the applicability and flexibility of the chassis 1001.

[0133] In some examples, a length of the first region A along the second direction Y and a length of the third region C along the second direction Y are equal.

[0134] In some examples, such as Figure 3As shown, the chassis 1001 further includes a plurality of partitions 1303, and the partitions 1303 extend along the first direction X. The plurality of partitions 1303 include a first partition 1303a and a second partition 1303b. The first partition 1303a is disposed in the first area A, and divides the first area A into a plurality of first sub-areas. The second partition 1303b is disposed in the third area C, and divides the third area C into a plurality of third sub-areas. The position of the partition 1303 along the second direction Y is adjustable.

[0135] For example, the number of the first partitions 1303a may be one or more, and when the number of the first partitions 1303a is multiple, the multiple first partitions 1303a are spaced apart in the first zone A along the second direction Y. The number of the second partitions 1303b may be one or more, and when the number of the second partitions 1303b is multiple, the multiple second partitions 1303b are spaced apart in the third zone C along the second direction Y. The number of the first partitions 1303a and the number of the second partitions 1303b may be the same or different.

[0136] It can be understood that the first partition 1303a divides the first area A into multiple first sub-areas, and the second partition 1303b divides the third area C into multiple third sub-areas, thereby dividing the space inside the chassis 1001, realizing the rational use of the internal space of the entire chassis 1001, and improving the space utilization rate inside the chassis 1001.

[0137] In addition, the position of the partition 1303 (including the first partition 1303a and the second partition 1303b) along the second direction Y is adjustable, which improves the flexibility of space division within the chassis 1001 and meets different needs, so that the chassis 1001 can support different configurations of the server 1000.

[0138] It can be understood that the node plate group 1150 of the first zone A can be set in a first sub-zone, and the functional module 1003 of the first zone A can be set in a first sub-zone. The node plate group 1150 of the third zone C can be set in a third sub-zone, and the functional module 1003 of the third zone C can be set in a third sub-zone. By adjusting the partition 1303, the setting position of the node plate group 1150 and / or the functional module 1003 in the first zone A and the third zone C can be adjusted, so that the chassis 1001 can support a higher opening rate and meet the heat dissipation requirements.

[0139] For example, a plurality of slide grooves may be provided on the side wall of the chassis 1001, and the plurality of slide grooves are arranged at intervals along the second direction Y. The partition 1303 (including the first partition 1303a and the second partition 1303b) is arranged in different slide grooves, so that the position of the partition 1303 along the second direction Y is adjustable. Alternatively, a plurality of screw holes may be provided on the side wall of the chassis 1001, and the plurality of screw holes are arranged at intervals along the second direction Y. The partition 1303 is connected to different screw holes by bolts, so that the position of the partition 1303 along the second direction Y is adjustable.

[0140] In some examples, such as Figure 2 and Figure 3 As shown, the chassis 1001 also includes a mounting ear module 1400. It can be understood that the mounting ear module 1400 is used to fix the chassis 1001 to the cabinet. The mounting ear module 1400 includes a first mounting ear 1401 and a second mounting ear 1402, which are arranged on both sides of the chassis 1001 along the second direction Y and are respectively connected to the chassis 1001.

[0141] For example, the first hanging ear 1401 and the second hanging ear 1402 can be connected to the chassis 1001 by bolt connection, riveting or clamping. The connection mode between the first hanging ear 1401 and the chassis 1001 and the connection mode between the second hanging ear 1402 and the chassis 1001 can be the same or different.

[0142] For example, Figure 3 As shown, the chassis 1001 may include two mounting ear modules 1400, and the two mounting ear modules 1400 may be spaced apart along the first direction X. For example, the two mounting ear modules 1400 may be close to the first side 1001a and the second side 1001b of the chassis 1001, respectively. Figure 1 and Figure 2 As shown, the chassis 1001 may also include only one mounting ear module 1400. In this case, the mounting ear module 1400 may be close to the first side 1001a of the chassis 1001 or the second side 1001b of the chassis 1001 along the first direction X.

[0143] When the chassis 1001 includes a mounting ear module 1400, the side (the first side 1001a or the second side 1001b) to which the mounting ear module 1400 is close is the front side of the chassis 1001. That is, along the first direction X, when the mounting ear module 1400 is installed close to the first side 1001a of the chassis 1001 (see Figure 1 ), the first side 1001a of the chassis 1001 is the front side; when the ear mounting module 1400 is installed close to the second side 1001b of the chassis 1001, the second side 1001b of the chassis 1001 is the front side (see Figure 2 ).

[0144] In some examples, the position of the ear module 1400 along the first direction X is adjustable. In this way, the installation position of the ear module 1400 can be adjusted according to the application scenario, so that the chassis 1001 can meet different configurations of the server 1000, thereby improving the adaptability and use flexibility of the chassis 1001.

[0145] In some examples, the ear mounting module 1400 is disposed close to the functional module 1003. Figure 1 As shown, when the functional module 1003 is arranged in the first zone A, the ear module 1400 is installed close to the first side 1001a of the chassis 1001. At this time, the first side 1001a of the chassis 1001 is the front side of the server 1000. Figure 2 As shown, when the functional module 1003 is arranged in the third zone C, the ear module 1400 is installed close to the second side 1001 b of the chassis 1001 . At this time, the second side 1001 b of the chassis 1001 is the front side of the server 1000 .

[0146] It can be understood that the front side of the server 1000 is closer to the operation and maintenance personnel than the rear side of the server 1000, that is, the operation and maintenance personnel usually maintain the server 1000 at the front side of the server 1000. The outgoing line side of the server 1000 is usually close to the side where the functional module 1003 is located. In this way, the ear mounting module 1400 is arranged close to the functional module 1003, so that the server 1000 can achieve the outgoing line from the front side, which improves the convenience of operation and maintenance.

[0147] In some examples, at least one of the first hanging ear 1401 and the second hanging ear 1402 may also be provided with a switch for controlling the entire server 1000, an indicator light indicating the entire server 1000, a network interface, a universal serial bus (USB) interface, a display interface, a debugging interface, etc.

[0148] In some examples, such as Figure 1 and Figure 2 As shown, the server 1000 further includes a fan 1004 , and the fan 1004 blows air along a first direction X. The fan 1004 is fixed on the management board body 1110 .

[0149] For example, the fan 1004 can adopt an integrated fan to support the evolution of higher power heat dissipation, thereby improving the overall heat dissipation efficiency of the server 1000. Among them, the integrated fan refers to the integration of multiple separate fans, which has good integrity and compactness, and has lower losses and better heat dissipation performance. For example, the fan 1004 can be an integrated maintenance-free 6056 fan, which has a heat dissipation efficiency of about 10% compared to the separate 6056 fan.

[0150] The fan 1004 can be fixed to the management board body 1110 by plugging or welding. The management module 1008 can control the rotation speed and air volume of the fan 1004, and the management module 1008 can monitor and manage the fan 1004.

[0151] It can be understood that the management board body 1110 is arranged in the second zone B, and the fan 1004 is fixed on the management board body 1110, so that the fan 1004 can be arranged in the second zone B, that is, the fan 1004 can be arranged between the first zone A and the third zone C. The fan 1004 discharges air along the first direction X, so that the fan 1004 can dissipate heat for the node board group 1150 and / or the functional module 1003 arranged in the first zone A, and the node board group 1150 and / or the functional module 1003 arranged in the third zone C, thereby reducing the risk of the server 1000 not being able to work normally due to excessive temperature.

[0152] For example, the fan 1004 may discharge air in a direction from the first side 1001a of the chassis 1001 to the second side 1001b of the chassis 1001 , or the fan 1004 may discharge air in a direction from the second side 1001b of the chassis 1001 to the first side 1001a of the chassis 1001 .

[0153] In some examples, when the mounting ear module 1400 is disposed at different positions, the air outlet direction of the fan 1004 is different. For example, when the mounting ear module 1400 is disposed close to the first side 1001a of the chassis 1001 along the first direction X, the fan 1004 discharges air in a direction from the first side 1001a of the chassis 1001 to the second side 1001b of the chassis 1001. On the contrary, when the mounting ear module 1400 is disposed close to the second side 1001b of the chassis 1001 along the first direction X, the fan 1004 discharges air in a direction from the second side 1001b of the chassis 1001 to the first side 1001a of the chassis 1001. That is, the side close to the mounting ear module 1400 along the first direction X is the air inlet side of the fan 1004.

[0154] It can be understood that the side close to the ear module 1400 is the front side of the server 1000, and the operation and maintenance personnel usually stand in front of the server 1000 for maintenance. Therefore, setting the side close to the ear module 1400 (i.e., the front side of the server 1000) as the air inlet side of the fan 1004 can prevent the wind from the fan 1004 from blowing towards the operation and maintenance personnel, reduce the impact of the wind blown by the fan 1004 on the operation and maintenance personnel, and improve the user experience of the server 1000.

[0155] In some examples, the fan 1004 can be rotated 180° to adjust the air inlet and outlet directions of the fan 1004, or the air inlet and outlet directions of the fan 1004 can be adjusted by changing the rotation direction, tilt angle, etc. of the fan blades.

[0156] In some examples, such as Figure 2 and Figure 3 As shown, the management board 1100 further includes a connector module (cable tray) 1120. The management board body 1110 and the connector module 1120 are electrically connected. It can be understood that the connector module 1120 and the management board body 1110 are both arranged in the second zone B. The management module 1008 is arranged on the management board body 1110, and the management board body 1110 and the connector module 1120 are electrically connected, so that the management module 1008 can be electrically connected to the connector module 1120.

[0157] It is understandable that the connector module 1120 and the management board body 1110 can be electrically connected via a cable or in a pluggable manner. The embodiment of the present application does not further limit the manner in which the connector module 1120 and the management board body 1110 are electrically connected.

[0158] The connector module 1120 is disposed on the management board body 1110; or, the connector module 1120 is disposed on one side of the management board body 1110 along the first direction X and is adjacent to the management board body 1110. In this way, the flexibility of the location of the connector module 1120 can be improved to meet different needs.

[0159] For example, the connector module 1120 and the management board body 1110 can be decoupled. By replacing different connector modules 1120, different configurations of the server 1000 can be met, supporting the evolution of the server 1000 to a higher rate.

[0160] Figure 4 Schematic diagram of the structure of the connector module 1120 provided in some embodiments of the present application. In some examples, such as Figure 4 As shown, the connector module 1120 includes a fixing plate 1121 and a connector 1122 mounted on the surface of the fixing plate 1121. For example, the fixing plate 1121 may be a PCB. When the connector module 1120 is arranged on the management board body 1110, the fixing plate 1121 may be connected to the management board body 1110; when the connector module 1120 is arranged on one side of the management board body 1110 along the first direction X, the fixing plate 1121 may be connected to the cover plate or the bottom plate of the chassis 1001. In this way, the risk of the fixing plate 1121 shaking or offsetting relative to the chassis 1001 can be reduced.

[0161] There may be a plurality of connectors 1122, and the plurality of connectors 1122 are disposed at intervals on the fixing plate 1121. The plurality of connectors 1122 may be used to transmit different types of signals, such as high-speed signals, low-speed signals, and power signals.

[0162] The node board 1002 and / or the functional module 1003 are configured to be plugged into at least one connector 1122. It can be understood that each node board 1002 in the node board group 1150 is respectively plugged into the connector 1122. For example, the node board 1002 can be plugged into one or more connectors 1122, and the functional module 1003 can also be plugged into one or more connectors 1122.

[0163] It can be understood that the node board 1002 and / or the functional module 1003 and at least one connector 1122 can support hot plugging, so that the node board 1002 and / or the functional module 1003 can be hot plugged and maintained, which improves the convenience of maintenance and thus improves the operation and maintenance efficiency. In addition, the use of plug-in can reduce the number of cables and improve the convenience of operation and maintenance.

[0164] In some examples, such as Figure 3 As shown, the number of the connector modules 1120 is two, and the two connector modules 1120 are arranged at intervals along the first direction X. For example, the two connector modules 1120 can be arranged at intervals along the first direction X on the management board body 1110, or the two connector modules 1120 can also be arranged on both sides of the management board body 1110 along the first direction X. The two connector modules 1120 are electrically connected to the management board body 1110 respectively.

[0165] Continue to refer to Figure 3 Among the two connector modules 1120, the connector module 1120 close to the first side 1001a of the chassis 1001 is the first connector module 1120a, and the connector module 1120 close to the second side 1001b of the chassis 1001 is the second connector module 1120b.

[0166] It can be understood that the first connector module 1120a and the second connector module 1120b are only used to distinguish two different connector modules 1120, and do not further limit the structures of the two connector modules 1120. The number and arrangement positions of the connectors 1122 of the first connector module 1120a can be the same as or different from the number and arrangement positions of the connectors 1122 of the second connector module 1120b.

[0167] The connector 1122 of the first connector module 1120 a opens toward the first side 1001 a of the chassis 1001 , and the connector 1122 of the second connector module 1120 b opens toward the second side 1001 b of the chassis 1001 .

[0168] The connector 1122 of the first connector module 1120a opens toward the first side 1001a of the chassis 1001, and when the node plate 1002 and / or the functional module 1003 arranged in the first zone A are pulled out, the node plate 1002 and / or the functional module 1003 arranged in the first zone A can be driven to separate from the connector 1122 of the first connector module 1120a; when the node plate 1002 and / or the functional module 1003 are pushed into the first zone A, the node plate 1002 and / or the functional module 1003 can be plugged into the connector 1122 of the first connector module 1120a, which improves the convenience of plugging and separating the node plate group 1150 and / or the functional module 1003 arranged in the first zone A and the connector 1122 of the first connector module 1120a, thereby improving the convenience of operation and maintenance.

[0169] The connector 1122 of the second connector module 1120b opens toward the second side 1001b of the chassis 1001. When the node plate 1002 and / or the functional module 1003 arranged in the third zone C are pulled out, the node plate 1002 and / or the functional module 1003 arranged in the third zone C can be driven to separate from the connector 1122 of the second connector module 1120b; when the node plate 1002 and / or the functional module 1003 are pushed into the third zone C, the node plate 1002 and / or the functional module 1003 can be plugged into the connector 1122 of the second connector module 1120b, thereby improving the convenience of plugging and separating the node plate 1002 and / or the functional module 1003 arranged in the third zone C and the connector 1122 of the second connector module 1120b, thereby improving the convenience of operation and maintenance.

[0170] It can be understood that the connector module 1120 is provided to include a first connector module 1120a and a second connector module 1120b, and the node plate 1002 and / or the functional module 1003 arranged in the first zone A, and the node plate 1002 and / or the functional module 1003 arranged in the third zone C are respectively plugged into or separated from the first connector module 1120a and the second connector module 1120b, which can reduce the mutual influence between the node plate group 1150 and / or the functional module 1003 arranged in the first zone A and the node plate group 1150 and / or the functional module 1003 arranged in the third zone C.

[0171] In some other examples, the number of connector modules 1120 is one. In this case, among the multiple connectors 1122, the openings of a portion (one, two or more) of the connectors 1122 face the first side 1001a of the chassis 1001, and the openings of another portion of the connectors 1122 face the second side 1001b of the chassis 1001.

[0172] The openings of a portion (one, two or more) of the connectors 1122 are facing the first side 1001a of the chassis 1001. When the node plate 1002 and / or the functional module 1003 arranged in the first zone A are pulled out, the node plate 1002 and / or the functional module 1003 arranged in the first zone A can be driven to separate from the connector 1122 of the connector module 1120; when the node plate 1002 and / or the functional module 1003 are pushed into the first zone A, the node plate 1002 and / or the functional module 1003 can be plugged into the connector 1122 of the connector module 1120, which improves the convenience of plugging and separating the node plate 1002 and / or the functional module 1003 arranged in the first zone A and the connector 1122 of the connector module 1120, thereby improving the convenience of operation and maintenance.

[0173] The opening of another part of the connector 1122 faces the second side 1001b of the chassis 1001. When the node plate 1002 and / or the functional module 1003 arranged in the third zone C are pulled out, the node plate 1002 and / or the functional module 1003 arranged in the third zone C can be driven to separate from the connector 1122 of the connector module 1120; when the node plate 1002 and / or the functional module 1003 are pushed into the third zone C, the node plate 1002 and / or the functional module 1003 can be plugged into the connector 1122 of the connector module 1120, which improves the convenience of plugging and separating the node plate 1002 and / or the functional module 1003 arranged in the third zone C and the connector 1122 of the connector module 1120, thereby improving the convenience of operation and maintenance.

[0174] It can be understood that setting the number of connector modules 1120 to one can simplify the structure of the connector module 1120 and reduce the cost of the server 1000.

[0175] The embodiment of the present application takes the example that the connector module 1120 includes a first connector module 1120a and a second connector module 1120b, and continues to illustrate.

[0176] In some examples, such as Figure 4 As shown, the connector 1122 includes a high-speed connector 1122a, and at least one of the node board group 1150 and the functional module 1003 is electrically connected to the high-speed connector 1122a for transmitting high-speed signals.

[0177] The number of high-speed connectors 1122a can be one or more. The number of high-speed connectors 1122a of the first connector module 1120a and the number of high-speed connectors 1122a of the second connector module 1120b can be the same or different. The node plate group 1150 is electrically connected to the high-speed connector 1122a, so that high-speed signals can be transmitted between the node plate 1002 and the connector module 1120. For example, multiple node plates 1002 in the node plate group 1150 can be electrically connected to the high-speed connector 1122a respectively. The functional module 1003 is electrically connected to the high-speed connector 1122a, so that high-speed signals can be transmitted between the functional module 1003 and the connector module 1120.

[0178] For example, the node board group 1150 and the functional module 1003 can be electrically connected to the high-speed connector 1122a through a high-speed cable (such as a PCIE high-speed cable). In this way, the loss caused by the PCB board during high-speed signal transmission can be reduced, and the connection relationship between the node board group 1150 and / or the functional module 1003 and the connector module 1120 can be changed by plugging and unplugging the cable, meeting different configurations of the server 1000, and enabling the server 1000 to have different working modes.

[0179] It can be understood that a node board 1002 can be electrically connected to one or more high-speed connectors 1122a, and a functional module 1003 can be electrically connected to one or more high-speed connectors 1122a. Different high-speed connectors 1122a can transmit the same high-speed signal or different high-speed signals. For example, the high-speed connector 1122a can be a surelink high-speed connector, which supports the sixth generation of the peripheral component interconnect express standard (Peripheral Component Interconnect Express Generation 6, PCIE Gen6), and can support evolution to 112G.

[0180] When the node board group 1150 and the functional module 1003 are both electrically connected to the connector module 1120, high-speed signals can be transmitted between the node board 1002 and the functional module 1003 through the connector module 1120, so that the functional module 1003 can cooperate with the business processing module 1201 of the node board 1002 to complete data transmission or processing.

[0181] In some examples, the high-speed connector 1122a of the first connector module 1120a and the high-speed connector 1122a of the second connector module 1120b are electrically connected via a high-speed cable, and the high-speed cable is routed along the surface of the management board body 1110 .

[0182] It can be understood that the high-speed cable electrically connects the high-speed connector 1122a of the first connector module 1120a and the high-speed connector 1122a of the second connector module 1120b, so that high-speed signals can be transmitted between the first connector module 1120a and the second connector module 1120b.

[0183] In this way, the node board group 1150 and / or the functional module 1003 arranged in the first zone A can be electrically connected to the node board group 1150 and / or the functional module 1003 arranged in the third zone C through the connector module 1120.

[0184] The high-speed cable is routed along the surface of the management board body 1110, which can reduce the interference of the management board body 1110 on the high-speed signal, reduce the loss of the high-speed signal, and improve the quality of the signal. In addition, the high-speed cable can be routed obliquely along the surface of the management board body 1110, without having to be routed along the edge of the bottom plate of the chassis 1001, which shortens the length of the high-speed cable, thereby shortening the transmission path of the high-speed signal, which is conducive to shortening the transmission time of the high-speed signal.

[0185] In some examples, there is a gap between the bottom of the fan 1004 and the surface of the management board body 1110, and the gap is used for high-speed cable routing. This arrangement can reduce the mutual influence between the fan 1004 and the high-speed cable, and improve the space utilization rate in the chassis 1001.

[0186] In some examples, such as Figure 4 As shown, the connector 1122 includes a low-speed connector 1122b, which is electrically connected to the management board body 1110, and at least one of the node board group 1150 and the functional module 1003 is pluggably connected to the low-speed connector 1122b for transmitting low-speed signals.

[0187] For example, the number of low-speed connectors 1122b can be one or more. The number of low-speed connectors 1122b of the first connector module 1120a and the number of low-speed connectors 1122b of the second connector module 1120b can be the same or different. The node board group 1150 is plugged into the low-speed connector 1122b, so that the low-speed signal can be transmitted between the node board 1002 and the connector module 1120. For example, the multiple node boards 1002 in the node board group 1150 can be electrically connected to the low-speed connector 1122b respectively. The functional module 1003 is plugged into the low-speed connector 1122b, so that the low-speed signal can be transmitted between the functional module 1003 and the connector module 1120.

[0188] A node board 1002 can be pluggably connected to one or more low-speed connectors 1122b, and a functional module 1003 can be pluggably connected to one or more low-speed connectors 1122b. Different low-speed connectors 1122b can transmit the same low-speed signal or different low-speed signals. The low-speed signal includes one or more of I2C, SPI, I3C, and GPIO. It can be understood that the embodiments of the present application do not further limit the types of low-speed signals transmitted between the node board 1002 and the connector module 1120, and between the functional module 1003 and the connector module 1120. For example, the low-speed connector 1122b can be an exmax low-speed connector.

[0189] It can be understood that the management module 1008 can be a BMC, which is arranged on the management board body 1110, and the low-speed connector 1122b is electrically connected to the management board body 1110, so that the low-speed signal can be transmitted between the BMC and the low-speed connector 1122b. When the node board group 1150 is pluggable connected to the low-speed connector 1122b, the node board 1002 and the BMC can communicate through the low-speed signal, so that the BMC can monitor and manage the node board 1002. When the functional module 1003 is pluggable connected to the low-speed connector 1122b, the functional module 1003 and the BMC can communicate through the low-speed signal, so that the BMC can monitor and manage the functional module 1003. In this way, the BMC can monitor and manage the node board 1002 and the functional module 1003.

[0190] In some examples, such as Figure 4 As shown, the connector 1122 includes a first connector 11221 and a second connector 11222, which are spaced apart along the height direction Z of the chassis 1001. The first connector 11221 is closer to the cover of the chassis 1001 than the second connector 11222.

[0191] The node plate group 1150 includes a first node plate 1002a and a second node plate 1002b stacked. The first node plate 1002a is closer to the cover plate of the chassis 1001 than the second node plate 1002b. The first node plate 1002a is configured to be electrically connected to the first connector 11221, and the second node plate 1002b is configured to be electrically connected to the second connector 11222.

[0192] For example, the height of the chassis 1001 can be 2U, so that the first node plate 1002a and the second node plate 1002b can be stacked along the height direction Z of the chassis 1001. The first connector 11221 and the second connector 11222 can be high-speed connectors 1122a or low-speed connectors 1122b. Taking the first connector 11221 and the second connector 11222 as high-speed connectors 1122a as an example, the two high-speed connectors 1122a are spaced apart along the height direction Z of the chassis 1001. Taking the first connector 11221 and the second connector 11222 as low-speed connectors 1122b as an example, the two low-speed connectors 1122b are spaced apart along the height direction Z of the chassis 1001.

[0193] It can be understood that the first connector 11221 is closer to the cover plate of the chassis 1001 relative to the second connector 11222, and the first node plate 1002a is closer to the cover plate of the chassis 1001 relative to the second node plate 1002b, so that the first node plate 1002a can be close to the first connector 11221, and the second node plate 1002b can be close to the second connector 11222. In this way, the convenience of electrically connecting the first node plate 1002a with the first connector 11221 can be improved, and the convenience of electrically connecting the second node plate 1002b with the second connector 11222 can be improved. In addition, the space utilization rate of the chassis 1001 can be improved.

[0194] For example, when the first connector 11221 and the second connector 11222 are low-speed connectors 1122b, the first node plate 1002a and the second node plate 1002b can be electrically connected to the first connector 11221 and the second connector 11222 respectively by plugging. When the first connector 11221 and the second connector 11222 are high-speed connectors 1122a, the first node plate 1002a and the second node plate 1002b can be electrically connected to the first connector 11221 and the second connector 11222 respectively by cables.

[0195] It can be understood that the first connector 11221 and the second connector 11222 can be integrated or separated.

[0196] Figure 5 This is a schematic block diagram of the structure of the power module 1005 provided in some embodiments of the present application. Figure 6 FIG. 1 is a top view of a server 1000 provided in some embodiments of the present application. In some examples, such as Figure 5 and Figure 6 As shown, the server 1000 further includes a power module 1005 , and the power module 1005 includes a power output interface 1005 c .

[0197] For example, the power module 1005 can be arranged in the first area A and / or the third area C. Figure 5 As shown, the power module 1005 may include a power body 1005a and a power distribution board (PDB) 1005b. The power body 1005a is used to provide electrical energy. For example, the power body 1005a may be a power supply unit (PSU). The power distribution board 1005b may be a PCB, and the power output interface 1005c is provided on the power distribution board 1005b.

[0198] The node board 1002 includes a first power supply input interface, and the functional module 1003 includes a second power supply input interface. Figure 6 As shown, the connector 1122 includes a first power connector 1122c, a second power connector 1122d, and a third power connector 1122e. The third power connector 1122e is electrically connected to the power output interface 1005c, the first power connector 1122c is electrically connected to one of the first power input interface and the second power input interface, and the second power connector 1122d is electrically connected to the other of the first power input interface and the second power input interface. In this way, the power module 1005 can supply power to the node board 1002 and the function module 1003 through the management board 1100.

[0199] For example, Figure 6 As shown, the first power connector 1122c can be electrically connected to the first power input interface, so that the power module 1005 can supply power to the node board 1002 through the management board 1100. The second power connector 1122d can be electrically connected to the second power input interface, so that the power module 1005 can supply power to the functional module 1003 through the management board 1100. Alternatively, the first power connector 1122c can be electrically connected to the second power input interface, and the second power connector 1122d can be electrically connected to the first power input interface.

[0200] The embodiment of the present application takes the example that the first power connector 1122c is electrically connected to the first power input interface, and the second power connector 1122d is electrically connected to the second power input interface as an example, and continues to illustrate.

[0201] In some examples, the power output interface 1005c can be electrically connected to the third power connector 1122e through a power cable. In some examples, the third power connector 1122e can be set on the management board body 1110, and a notch is opened on the fixing plate 1121, and the power cable passes through the notch opened on the fixing plate 1121 to electrically connect the power output interface 1005c and the third power connector 1122e.

[0202] The first power supply input interface can be electrically connected to the first power connector 1122c by plugging. For example, the node board 1002 may include multiple first power supply input interfaces, and the multiple first power supply input interfaces can be plugged in one-to-one with the multiple first power connectors 1122c. The second power supply input interface can be electrically connected to the second power connector 1122d by plugging. For example, the functional module 1003 may include multiple second power supply input interfaces, and the multiple second power supply input interfaces can be plugged in one-to-one with the multiple second power connectors 1122d.

[0203] For example, the first power connector 1122c and the second power connector 1122d may be PWR (power) power connectors, and the third power connector 1122e may be a CST power connector.

[0204] In some examples, such as Figure 5 As shown, the power module 1005 may further include a power low-speed interface 1005d, which may be provided on the power distribution board 1005b. The power low-speed interface 1005d is electrically connected to the low-speed connector 1122b via a low-speed cable, and the low-speed connector 1122b is electrically connected to the BMC, so that the BMC and the power module 1005 can communicate with each other, thereby enabling the BMC to monitor and manage the power module 1005. For example, the power low-speed interface 1005d may be a PDB low-speed connector.

[0205] For example, in the 2U chassis 1001 space, the power module 1005 supports the modular hardware system universal redundant power supply (MCRPS). For example, the power module 1005 supports 1+1 redundancy (one for use and one for backup), 2+2 redundancy (two for use and two for backup), and can support up to 6kW (kilowatt) power supply.

[0206] It can be understood that the power module 1005 and the management board 1100 can be decoupled, which improves the convenience of maintenance of the power module 1005. In addition, different power modules 1005 can be replaced according to different configurations of the server 1000, which improves the flexibility of use of the server 1000.

[0207] The fan 1004 is disposed on the management board body 1110 , and the power module 1005 can supply power to the fan 1004 through the management board body 1110 , so that the fan 1004 can work normally.

[0208] In some examples, such as Figure 6 As shown, the node plate group 1150 includes a first node plate group 1150a and a second node plate group 1150b, one of the first node plate group 1150a and the second node plate group 1150b is disposed in the first region A, and the other is disposed in the third region C.

[0209] For example, the first node plate group 1150a may be disposed in the first zone A, and the second node plate group 1150b may be disposed in the third zone C. Alternatively, the first node plate group 1150a may be disposed in the third zone C, and the second node plate group 1150b may be disposed in the first zone A.

[0210] like Figure 6 As shown, one of the power module 1005 and the functional module 1003 is arranged in the first zone A, and is spaced apart from the node plate group 1150 arranged in the first zone A along the second direction Y. The other of the power module 1005 and the functional module 1003 is arranged in the third zone C, and is spaced apart from the node plate group 1150 arranged in the third zone C along the second direction Y.

[0211] For example, the function module 1003 may be arranged in the first zone A, and the power module 1005 may be arranged in the third zone C. Alternatively, the function module 1003 may be arranged in the third zone C, and the power module 1005 may be arranged in the first zone A.

[0212] Setting the number of node board groups 1150 to two can improve the computing performance of the server 1000. For example, this configuration of the server 1000 can be called a high computing performance configuration.

[0213] In some examples, such as Figure 6 As shown, the first node plate group 1150a and the second node plate group 1150b are staggered in the first direction X. It can be understood that the node plate 1002 generates a large amount of heat when working. The first node plate group 1150a and the second node plate group 1150b are staggered in the first direction X, which can reduce the risk of heat accumulation in the chassis 1001, so that the heat can be evenly distributed in the chassis 1001, which is beneficial to the heat dissipation of the chassis 1001.

[0214] The embodiment of the present application takes the first node plate group 1150a being arranged in the first zone A, the second node plate group 1150b being arranged in the third zone C, the functional module 1003 being arranged in the first zone A, and the power supply module 1005 being arranged in the third zone C as an example, and the explanation will be continued by way of example.

[0215] In some examples, such as Figure 6As shown, the functional module 1003 and the first node board group 1150a are arranged at intervals along the second direction Y, and are arranged in the first zone A. For example, the first zone A can be called a computing storage zone. The management board 1100 and the fan 1004 are arranged in the second zone B. For example, the second zone B can be called a fan management zone. The power module 1005 and the second node board group 1150b are arranged at intervals along the second direction Y, and are arranged in the third zone C. For example, the third zone C can be called a computing power zone. The ear module 1400 is arranged along the first direction X close to the first side 1001a of the chassis 1001, that is, the first side 1001a of the chassis 1001 is the front side of the server 1000, and the second side 1001b of the chassis 1001 is the back side of the server 1000.

[0216] like Figure 6 As shown, the functional module 1003 may include a riser card 1003a and a board card 1003c. The riser card 1003a is provided with a slot 1003b. For example, the riser card 1003a may be provided with four slots 1003b. The board card 1003c may be a GPU card or a DPU card, etc., and the board card 1003c can be plugged into the slot 1003b.

[0217] It can be understood that in the drawings of the specification of this application, Figure 6 For example, in order to clearly illustrate the slot 1003b and the board 1003c, the plug-in relationship between the slot 1003b and the board 1003c is not shown, which does not mean that the board 1003c cannot be inserted into the slot 1003b.

[0218] Along the second direction Y, the first zone A and the third zone C are divided into three equal parts, that is, the first zone A and the third zone C include three cages respectively. The first node board group 1150a occupies 2 cage spaces of the first zone A, and the riser card 1003a and the board card 1003c occupy 1 cage space of the first zone A. For example, the riser card 1003a supports double-width full-height full-length cards. The second node board group 1150b occupies 2 cage spaces of the third zone C, and the power module 1005 occupies 1 cage space of the third zone C. For example, the power module 1005 supports 4PSUs and supports 2+2 redundancy.

[0219] It can be understood that in the first node plate group 1150a, the first node plate 1002a and the second node plate 1002b are stacked, and each node plate 1002 occupies 1U of space. In the second node plate group 1150b, the first node plate 1002a and the second node plate 1002b are stacked, and each node plate 1002 occupies 1U of space. The first node plate 1002a is closer to the cover plate than the second node plate 1002b, that is, the first node plate 1002a is arranged in the upper 1U, and the second node plate 1002b is arranged in the lower 1U.

[0220] like Figure 6 As shown, the node board group 1150 (the first node board group 1150a and the second node board group 1150b) occupies 2 cage spaces, and each node board 1002 can support the expansion of 24 RDIMMs. Compared with setting 8 RDIMMs, 16 more RDIMMs can be expanded. The two node board groups 1150 can expand 64 more RDIMMs in total.

[0221] There are two management modules 1008 (BMCs), which are arranged at intervals on the management board body 1110. One of the BMCs is used to monitor and manage the two node boards 1002 in the first node board group 1150a, and the other BMC is used to monitor and manage the two node boards 1002 in the second node board group 1150b. The function module 1003, the power module 1005, the fan 1004, etc. are monitored and managed by at least one of the two BMCs.

[0222] For low-speed signals, the second node board 1002b disposed in the lower 1U may be connected to the low-speed connector 1122b by hard connection (i.e., plugging). For example, the second node board 1002b disposed in the lower 1U may include a bent male for transmitting low-speed signals, and the low-speed connector 1122b may include a bent female for transmitting low-speed signals, and the bent male for transmitting low-speed signals and the bent female for transmitting low-speed signals are plugged.

[0223] The first node board 1002a arranged on the upper 1U can be connected to the low-speed connector 1122b by means of a soft connection (low-speed cable). For example, the first node board 1002a arranged on the upper 1U may include a bent male for transmitting a low-speed signal, the low-speed connector 1122b may include a straight female for transmitting a low-speed signal, and the low-speed cable electrically connects the bent male for transmitting the low-speed signal and the straight female for transmitting the low-speed signal. For example, the low-speed cable electrically connecting the bent male for transmitting the low-speed signal and the straight female for transmitting the low-speed signal may include a bent female-straight male connector. Alternatively, the first node board 1002a arranged on the upper 1U may also be connected to the low-speed connector 1122b by means of a hard connection (i.e., plugging).

[0224] Riser card 1003a may also be connected to low-speed connector 1122b by hard connection (i.e. plug-in). For example, riser card 1003a may include an orthogonal male bend for transmitting low-speed signals, and low-speed connector 1122b may include a female bend for transmitting low-speed signals, and the orthogonal male bend for transmitting low-speed signals is plugged with the female bend for transmitting low-speed signals.

[0225] For high-speed signals, the first node board 1002a, the second node board 1002b and the riser card 1003a can be connected by a soft connection (high-speed cable 1006, see Figure 6 ) is connected to the high-speed connector 1122a.

[0226] For power supply signals, Figure 6 As shown, the power output interface 1005c of the power module 1005 is electrically connected to the third power connector 1122e through a power cable, and supplies power to the node board 1002 and the functional module 1003 through the management board 1100. For example, the number of the third power connectors 1122e can be two.

[0227] There are two first power connectors 1122c, and the two first power connectors 1122c are stacked along the height direction Z of the chassis 1001. The second node board 1002b arranged at the lower 1U includes a bent male for transmitting a power signal, and the first power connector 1122c can be a bent female for transmitting a power signal, and the bent male for transmitting a power signal and the bent female for transmitting a power signal are plugged in. The first node board 1002a arranged at the upper 1U can include a bent male for transmitting a power signal, and the first power connector 1122c can be a bent female for transmitting a power signal, and the bent male for transmitting a power signal and the bent female for transmitting a power signal are plugged in.

[0228] Functional module 1003 (including riser card 1003a and board card 1003c) may include a bent male for transmitting a power signal, and third power connector 1122e may include a bent female for transmitting a power signal, and the bent male for transmitting a power signal and the bent female for transmitting a power signal are plugged in. For example, the bent male for transmitting a power signal of functional module 1003 may be an orthogonal bent male.

[0229] The power signal transmission male connector of the functional module 1003 can be electrically connected to the riser card 1003a and the board card 1003c through a power cable, so that the power module 1005 can supply power to the riser card 1003a and the board card 1003c. For example, the riser card 1003a can use a 2*2 power connector, and the board card 1003c can use a 2*6 power connector to meet the requirements of a 650W (watt) GPU card.

[0230] For example, chassis 1001 supports double-width, full-height, full-length DPU cards. In addition, node board 1002 and function module 1003 both support cable-free hot maintenance. Node board 1002 and function module 1003 achieve low-speed signal and power signal transmission by plugging with connector 1122, reducing the number of low-speed cables and power cables. High-speed signals are transmitted through high-speed cables 1006 (see Figure 6 ) realizes soft connection. By plugging and unplugging the high-speed cable 1006, the electrical connection relationship between different node boards 1002 can be changed, so that the server 1000 can have different working modes.

[0231] Furthermore, the connector module 1120 can be decoupled from the node board 1002 and the functional module 1003, so that the node board 1002 and the functional module 1003 can be plugged and unplugged independently. Taking the board 1003c including the DPU card as an example, when one of the multiple node boards 1002 works abnormally, it does not affect the transmission of signals between other node boards 1002 and the DPU card, thus realizing the multi-host function and increasing maintainability.

[0232] Figure 7 The schematic diagram of the structure of the node plate 1002 provided in some embodiments of the present application. In some examples, such as Figure 7 As shown, a plurality of MCIO connectors 1202 are further disposed on the node board 1002 . Along the first direction X, at least one MCIO connector 1202 is adjacent to an edge of the node board 1002 .

[0233] For example, the MCIO connector 1202 is connected to the CPU signal. There may be multiple MCIO connectors 1202 , and the multiple MCIO connectors 1202 are arranged along the first direction X on both sides of the service processing module 1201 .

[0234] For example, the number of MCIO connectors 1202 may be 16, of which 8 MCIO connectors 1202 are arranged at intervals along the second direction Y and are arranged at one side of the service processing module 1201 along the first direction X. Another 8 service processing modules 1201 are arranged at intervals along the second direction Y and are arranged at the other side of the service processing module 1201 along the first direction X. Each MCIO connector 1202 may be allocated a bandwidth of X8.

[0235] For ease of description, Figure 6 and Figure 7As shown, the eight MCIO connectors 1202 arranged on one side of the service processing module 1201 in the first direction X can be referred to as the first MCIO connector group 1202a, and the eight MCIO connectors 1202 arranged on the other side of the service processing module 1201 (CPU) in the first direction X can be referred to as the first MCIO connector group 1202b. For example, along the first direction X, the second MCIO connector group 1202b is closer to the connector module 1120 relative to the first MCIO connector group 1202a. In some examples, at least one MCIO connector 1202 in the first MCIO connector group 1202a is adjacent to an edge of the node board 1002, and at least one MCIO connector 1202 in the second MCIO connector group 1202b is adjacent to an edge of the node board 1002.

[0236] In some examples, the MCIO connectors 1202 of each node board 1002 in the first node board group 1150a are electrically connected, and the MCIO connectors 1202 of each node board 1002 in the second node board group 1150b are electrically connected. At this time, the first node board 1002a and the second node board 1002b of the first node board group 1150a are connected in series to form a node, and the first node board 1002a and the second node board 1002b of the second node board group 1150b are connected in series to form a node, and the server 1000 works in 2P mode (dual node mode).

[0237] For example, the first node board 1002a and the second node board 1002b of the first node board group 1150a may be interconnected via a unified parallel interface (UPI) bus, and the first node board 1002a and the second node board 1002b of the second node board group 1150b may be interconnected via a UPI bus.

[0238] In one node board group 1150 (first node board group 1150a or second node board group 1150b), three MCIO connectors 1202 of the first node board 1002a and three MCIO connectors 1202 of the second node board 1002b may be electrically connected in a one-to-one correspondence. Alternatively, two, four, five, six, seven or eight MCIO connectors 1202 of the first node board 1002a and two, four, five, six, seven or eight MCIO connectors 1202 of the second node board 1002b may be electrically connected in a one-to-one correspondence.

[0239] The embodiment of the present application takes the example that the three MCIO connectors 1202 of the first node board 1002a and the three MCIO connectors 1202 of the second node board 1002b are electrically connected one by one as an example, and continues to illustrate.

[0240] Taking the first node plate group 1150a as an example, Figure 6 As shown, three MCIO connectors 1202 in a first MCIO connector group 1202a on a second node board 1002b disposed in the lower 1U are electrically connected to three MCIO connectors in a first MCIO connector group 1202a on a first node board 1002a disposed in the upper 1U through high-speed cables 1006 in a one-to-one correspondence.

[0241] On the second node board 1002b of the lower 1U, three MCIO connectors 1202 in the second MCIO connector group 1202b are electrically connected to a high-speed connector 1122a of the first connector module 1120a via a high-speed cable 1006. On the first node board 1002a of the upper 1U, three MCIO connectors 1202 in the second MCIO connector group 1202b are electrically connected to another high-speed connector 1122a of the first connector module 1120a via a high-speed cable 1006.

[0242] The two high-speed connectors 1122a electrically connected to the two node boards 1002 of the first node board group 1150a are arranged along the height direction Z of the chassis 1001 to reduce the space occupied by the connector module 1120 in the first direction X. In addition, the two high-speed connectors 1122a electrically connected to the two node boards 1002 of the first node board group 1150a are electrically connected through the high-speed cable 1006, so that the two node boards 1002 in the first node board group 1150a can be connected in series, so that the server 1000 has a stronger data processing capability.

[0243] Similarly, taking the second node plate group 1150b as an example, Figure 6 As shown, three MCIO connectors 1202 in a first MCIO connector group 1202a on a second node board 1002b disposed in the lower 1U are electrically connected to three MCIO connectors in a first MCIO connector group 1202a on a first node board 1002a disposed in the upper 1U through high-speed cables 1006 in a one-to-one correspondence.

[0244] On the second node board 1002b of the lower 1U, three MCIO connectors 1202 in the second MCIO connector group 1202b are electrically connected to a high-speed connector 1122a of the second connector module 1120b via a high-speed cable 1006. On the first node board 1002a of the upper 1U, three MCIO connectors 1202 in the second MCIO connector group 1202b are electrically connected to another high-speed connector 1122a of the second connector module 1120b via a high-speed cable 1006.

[0245] The two high-speed connectors 1122a electrically connected to the two node boards 1002 of the second node board group 1150b are arranged along the height direction Z of the chassis 1001 to reduce the space occupied by the connector module 1120 in the first direction X. In addition, the two high-speed connectors 1122a electrically connected to the two node boards 1002 of the second node board group 1150b are electrically connected through the high-speed cable 1006, so that the two node boards 1002 stacked in the second node board group 1150b can be connected in series, so that the server 1000 has stronger data processing capabilities.

[0246] It can be understood that the functional module 1003 is electrically connected to the high-speed connector 1122a, so that the first node board group 1150a can access the functional module 1003 after being connected in series into a node, and the second node board group 1150b can access the functional module 1003 after being connected in series into a node.

[0247] In other examples, the MCIO connectors 1202 of each node board 1002 in the first node board group 1150a are respectively electrically connected to the functional module 1003, and the MCIO connectors 1202 of each node board 1002 in the second node board group 1150b are respectively electrically connected to the functional module 1003. At this time, the first node board 1002a and the second node board 1002b of the first node board group 1150a, and the first node board 1002a and the second node board 1002b of the second node board group 1150b can work independently, and the server 1000 works in 4P mode (four-node mode).

[0248] In one node board group 1150 (the first node board group 1150a or the second node board group 1150b), one MCIO connector 1202 of the first node board 1002a and one MCIO connector 1202 of the second node board 1002b may be electrically connected to the function module 1003, respectively. Alternatively, two, three, four or more MCIO connectors 1202 of the first node board 1002a may be electrically connected to the function module 1003; and two, three, four or more MCIO connectors 1202 of the second node board 1002b may be electrically connected to the function module 1003.

[0249] The embodiment of the present application takes an example in which one MCIO connector 1202 of the first node board 1002a and one MCIO connector 1202 of the second node board 1002b are electrically connected to the function module 1003 respectively, and continues to illustrate.

[0250] Taking the first node board group 1150a as an example, one MCIO connector 1202 in the second MCIO connector group 1202b of the first node board 1002a is electrically connected to one high-speed connector 1122a of the first connector module 1120a through the high-speed cable 1006. One MCIO connector 1202 in the second MCIO connector group 1202b of the second node board 1002b is electrically connected to one high-speed connector 1122a of the first connector module 1120a through the high-speed cable 1006.

[0251] It can be understood that the two node plates 1002 of the first node plate group 1150a are electrically connected to the two high-speed connectors 1122a of the first connector module 1120a respectively, and the two high-speed connectors 1122a electrically connected to the two node plates 1002 of the first node plate group 1150a can be spaced apart along the height direction Z of the chassis 1001.

[0252] Taking the second node board group 1150b as an example, one MCIO connector 1202 in the second MCIO connector group 1202b of the first node board 1002a is electrically connected to one high-speed connector 1122a of the second connector module 1120b via the high-speed cable 1006. One MCIO connector 1202 in the second MCIO connector group 1202b of the second node board 1002b is electrically connected to one high-speed connector 1122a of the second connector module 1120b via the high-speed cable 1006.

[0253] It can be understood that the two node plates 1002 of the second node plate group 1150b are electrically connected to the two high-speed connectors 1122a of the second connector module 1120b respectively, and the two high-speed connectors 1122a electrically connected to the two node plates 1002 of the second node plate group 1150b can be arranged at intervals along the height direction Z of the chassis 1001.

[0254] Two high-speed connectors 1122a electrically connected to the two first node board groups 1150a and two high-speed connectors 1122a electrically connected to the second node board group 1150a are electrically connected to the x32 bandwidth high-speed connectors 1122a through four x8 bandwidth high-speed cables.

[0255] The x32 bandwidth high-speed connector 1122a is electrically connected to the riser card 1003a, so that the two node boards 1002 of the two first node board groups 1150a and the two node boards 1002 of the second node board group 1150b can respectively provide PCIE x8 resources to the riser card 1003a, thereby improving the uniformity of resource allocation of the server 1000. For example, four x8 bandwidth high-speed cables are routed between the fan 1004 and the management board body 1110, shortening the length of the high-speed cables and improving the transmission speed of high-speed signals.

[0256] For example, the riser card 1003a includes four slots 1003b, and each slot 1003b is in the form of a PCIE CEM (Card Electromechanical, PCIE card electromechanical specification) direct cable to a high-speed connector 1122a with a bandwidth of x32.

[0257] It can be understood that by changing the electrical connection relationship (series or parallel) between the node boards 1002 in the node board group 1150 and the electrical connection relationship between the high-speed connectors 1122a, the server 1000 can have different working modes (2P mode or 4P mode) to meet different needs, thereby improving the use flexibility of the server 1000. In addition, the operation is simple, and there is no need to set up a structure such as a central backplane, and there is no need to open the cover of the chassis 1001, thereby improving the convenience of operation and maintenance.

[0258] In addition, at least one MCIO connector 1202 is adjacent to the edge of the node board 1002, which improves the convenience of plugging and unplugging the high-speed cable 1006 and the MCIO connector 1202, that is, improves the convenience of changing the electrical connection relationship between the node boards 1002 in the node board group 1150.

[0259] In the embodiment of the present application, the server 1000 is configured differently by adjusting the node board group 1150 and the functional module 1003 arranged in the first zone A and / or the third zone C. For example, in a configuration with high computing performance, the server 1000 is suitable for scenes such as scientific simulation and big data analysis. In addition, the server 1000 has a large memory, a strong data processing capability, and a strong heat dissipation performance, and is suitable for a cloud computing data center, which helps to improve the overall operating efficiency of the cloud computing data center.

[0260] Moreover, the server 1000 has flexible memory capacity and hardware configuration, and can also meet application scenarios such as artificial intelligence and machine learning. In addition, the server 1000 also supports running a large number of concurrent tasks, has strong heat dissipation performance, and meets virtualization and containerization application scenarios. In addition, through more memory and optimized heat dissipation, the server 1000 can handle a large amount of network traffic, support multiple network functions, and meet network function virtualization (NFV) scenarios.

[0261] In some possible implementations, each node board 1002 is monitored and managed by a management module 1008. In this case, four management modules 1008 need to be set up to manage each node board 1002, which is redundant and complex in design and is not conducive to reducing the cost of the server 1000.

[0262] In addition, the collaborative management between the various management modules 1008 completes signal convergence on the management board body 1110, which requires the cooperation of other components (such as logic devices). The management method is complicated and is not conducive to reducing the cost of the server 1000.

[0263] Based on this, continue to refer to Figure 6 In some examples of the present application, the management module 1008 includes a first management module 1008a and a second management module 1008b, and the first management module 1008a and the second management module 1008b are electrically connected.

[0264] For example, the first management module 1008a and the second management module 1008b may be arranged at intervals on the management board body 1110. The first management module 1008a and the second management module 1008b may be electrically connected to the management board body 1110 respectively, so that the first management module 1008a and the second management module 1008b can be electrically connected. The first management module 1008a and the second management module 1008b may both be BMCs.

[0265] In some examples, the first management module 1008a and the second management module 1008b can communicate with each other via an intelligent platform management bus (IPMB).

[0266] It can be understood that the first management module 1008a and the second management module 1008b are electrically connected so that data interaction and arbitration management can be supported between the two.

[0267] Figure 8 The structure of the server 1000 provided in some embodiments of the present application is shown in FIG. Figure 8As shown, the server 1000 includes a first node 1161 , a second node 1162 , a third node 1163 and a fourth node 1164 .

[0268] For example, the first node 1161, the second node 1162, the third node 1163 and the fourth node 1164 may correspond to the first node plate 1002a and the second node plate 1002b in the first node plate group 1150a, and the first node plate 1002a and the second node plate 1002b in the second node plate group 1150b, respectively. Alternatively, at least one of the first node 1161, the second node 1162, the third node 1163 and the fourth node 1164 may also correspond to a node formed by two or more node plates 1002 connected in series.

[0269] In the embodiment of the present application, the first node 1161 corresponds to the first node plate 1002a of the first node plate group 1150a, the second node 1162 corresponds to the second node plate 1002b of the first node plate group 1150a, the third node 1163 corresponds to the first node plate 1002a of the second node plate group 1150b, and the fourth node 1164 corresponds to the second node plate 1002b of the second node plate group 1150b. The illustration is continued by way of example.

[0270] like Figure 8 As shown, the first node 1161 and the second node 1162 are electrically connected to the first management module 1008a respectively, and the third node 1163 and the fourth node 1164 are electrically connected to the second management module 1008b respectively.

[0271] In this way, low-speed signals can be transmitted between the first management module 1008a and the first node 1161, and between the first management module 1008a and the second node 1162, so that the first management module 1008a can monitor and manage the first node 1161 and the second node 1162. In addition, low-speed signals can be transmitted between the second management module 1008b and the third node 1163, and between the second management module 1008b and the fourth node 1164, so that the second management module 1008b can monitor and manage the third node 1163 and the fourth node 1164.

[0272] The function module 1003 is used to coordinate data transmission or processing with at least one of the first node 1161 , the second node 1162 , the third node 1163 and the fourth node 1164 . One of the first management module 1008 a and the second management module 1008 b is used to manage the function module 1003 .

[0273] The functional module 1003 is electrically connected to at least one of the first node 1161, the second node 1162, the third node 1163 and the fourth node 1164, so that the functional module 1003 can cooperate with the business processing module 1201 of at least one of the first node 1161, the second node 1162, the third node 1163 and the fourth node 1164 to transmit or process data.

[0274] The function module 1003 is electrically connected to the management module 1008 through the management board body 1110. For example, the first management module 1008a can manage the function module 1003, or the second management module 1008b can manage the function module 1003. For example, the management module 1008 and the function module 1003 can be managed through out-of-band communication.

[0275] In some examples, when the server 1000 is in the first working state, the first node 1161, the second node 1162, the third node 1163, and the fourth node 1164 are configured to work independently.

[0276] It can be understood that when the server 1000 is in the first working state, the first node 1161, the second node 1162, the third node 1163 and the fourth node 1164 can all be in working state, and the server 1000 is in a four-node working mode. Alternatively, one, two or three of the first node 1161, the second node 1162, the third node 1163 and the fourth node 1164 can also be in working state.

[0277] The first management module 1008a is used to manage at least one of the first node 1161 and the second node 1162; and / or, the second management module 1008a is used to manage at least one of the third node 1163 and the fourth node 1164. For example, the management module 1008 and the nodes can achieve management through out-of-band communication.

[0278] It can be understood that when at least one of the first node 1161 and the second node 1162 is in a working state,

[0279] The first management module 1008a is used to manage nodes in working state. For example, when the first node 1161 is in working state, the first management module 1008a is used to manage the first node 1161; when the second node 1162 is in working state, the first management module 1008a is used to manage the second node 1162; when both the first node 1161 and the second node 1162 are in working state, the first management module 1008a is used to manage the first node 1161 and the second node 1162.

[0280] Similarly, when at least one of the third node 1163 and the fourth node 1164 is in a working state, the second management module 1008b is used to manage the node in a working state. For example, when the third node 1163 is in a working state, the second management module 1008b is used to manage the third node 1163; when the fourth node 1164 is in a working state,

[0281] The second management module 1008b is used to manage the fourth node 1164; when the third node 1163 and the fourth node 1164 are both in working state, the second management module 1008b is used to manage the third node 1163 and the fourth node 116.

[0282] For example, the management module 1008 and the node may transmit signals via the management component transport protocol (MCTP), so that the management module 1008 can manage and monitor the node.

[0283] In an embodiment of the present application, a first management module 1008a and a second management module 1008b are electrically connected so that the first management module 1008a and the second management module 1008b can communicate with each other without signal convergence on the management board body 1110 and without other logic devices (such as complex programmable logic devices, CPLD) participating in the communication between the first management module 1008a and the second management module 1008b.

[0284] In this way, the communication path can be shortened and the efficiency of information interaction between the first management module 1008a and the second management module 1008b can be improved.

[0285] The first management module 1008a is electrically connected to the first node 1161 and the second node 1162, so that the first management module 1008a can monitor and manage at least one of the first node 1161 and the second node 1162. The second management module 1008b is electrically connected to the third node 1163 and the fourth node 1164, so that the second management module 1008b can monitor and manage at least one of the third node 1163 and the fourth node 1164.

[0286] In this way, two management modules 1008 can manage four nodes (one management module 1008 manages two nodes), and there is no need to set the number of management modules 1008 to be the same as the number of nodes, which reduces the number of management modules 1008, simplifies the structure of the server 1000, and reduces the cost of the server 1000. In addition, the management scheme of the server 1000 for multiple nodes is simplified.

[0287] Fig. 9 This is a schematic block diagram of the structure of the server 1000 provided in some other embodiments of the present application. In some examples, such as Fig. 9 As shown, when the server 1000 is in the second working state, the first node 1161 and the second node 1162 (the first node board 1002a and the second node board 1002b in the first node board group 1150a) are connected in series to form the first node group 1160a, and the third node 1163 and the fourth node 1164 (the first node board 1002a and the second node board 1002b in the second node board group 1150b) are connected in series to form the second node group 1160b. The first node group 1160a and the second node group 1160b are configured to work separately.

[0288] For example, when the server 1000 is in the second working state, the first node group 1160a can be regarded as one node, and the second node group 1160b can be regarded as one node. When the server 1000 is in the second working state, the first node group 1160a and the second node group 1160b can both be in working state, and the server 1000 is in a dual-node working mode. Alternatively, one of the first node group 1160a and the second node group 1160b can be in working state, and the server 1000 is in a dual-path single-node working mode.

[0289] One of the first management module 1008a and the second management module 1008b is used to manage the first node group 1160a and / or the second node group 1160b.

[0290] It can be understood that the first management module 1008a and the second management module 1008b are used to manage the node group 1160 in the working state. For example, when the first node group 1160a is in the working state, one of the first management module 1008a and the second management module 1008b is used to manage the first node group 1160a. When the second node group 1160b is in the working state, one of the first management module 1008a and the second management module 1008b is used to manage the second node group 1160b. When both the first node group 1160a and the second node group 1160b are in the working state, one of the first management module 1008a and the second management module 1008b is used to manage the first node group 1160a and the second node group 1160b.

[0291] It can be understood that when one of the first management module 1008a and the second management module 1008b manages the working node group 1160, the other one of the first management module 1008a and the second management module 1008b can be in a standby state.

[0292] One of the first management module 1008a and the second management module 1008b is set to manage the first node group 1160a and / or the second node group 1160b, so that the management module 1008 can monitor and manage the first node group 1160a and the second node group 1160b. In this way, one management module 1008 can manage two node groups, and there is no need to set the same number of management modules 1008 as the number of node groups, which reduces the number of management modules 1008, simplifies the structure of the server 1000, and reduces the cost of the server 1000. In addition, the management scheme of the server 1000 for multiple nodes is simplified.

[0293] In some examples, one of the first management module 1008a and the second management module 1008b is a master management module and the other is a slave management module. When the server 1000 is in the second working state, the master management module is used to manage the first node group 1160a and / or the second node group 1160b.

[0294] Fig.10 This is a schematic block diagram of the structure of the server 1000 provided in some embodiments of the present application. Fig.10 As shown, taking the first management module 1008a as the main management module and the second management module 1008b as the slave management module as an example, the first management module 1008a is used to manage the first node group 1160a and / or the second node group 1160b, and the second management module 1008b is in a standby state. It can be understood that Fig.10 The second management module 1008b in a standby state is shown by a dotted line.

[0295] Setting one of the first management module 1008a and the second management module 1008b as a master management module and the other as a slave management module can avoid management conflicts between the first management module 1008a and the second management module 1008b, so that the master management module in the first management module 1008a and the second management module 1008b can monitor and manage at least one of the first node group 1160a and the second node group 1160b.

[0296] Fig.11 This is a schematic block diagram of the structure of the server 1000 provided in some embodiments of the present application. In some examples, such as Fig.11 As shown, the first management module 1008a includes a first register 1009a. When the first management module 1008a is a master management module, the flag state of the first register 1009a is a first state. When the first management module 1008a is a slave management module, the flag state of the first register 1009a is a second state.

[0297] The second management module 1008b includes a second register 1009b. When the second management module 1008b is a master management module, the flag bit state of the second register 1009b is a first state. When the second management module 1008b is a slave management module, the flag bit state of the second register 1009b is a second state.

[0298] For example, according to the flag bit state of the register (the first register 1009a or the second register 1009b), it can be determined whether the management module (the first management module 1008a or the second management module 1008b) is the master management module or the slave management module. Taking the first management module 1008a as an example, when the first management module 1008a is the master management module, the flag bit state of the first register 1009a is the first state, for example, the flag bit of the first register 1009a can be 1. When the first management module 1008a is the slave management module, the flag bit state of the first register 1009a is the second state, for example, the flag bit of the first register 1009a can be 0.

[0299] Similarly, when the second management module 1008b is the master management module, the flag bit state of the second register 1009b is the first state, for example, the flag bit of the second register 1009b may be 1. When the second management module 1008b is the slave management module, the flag bit state of the second register 1009b is the second state, for example, the flag bit of the second register 1009b may be 0.

[0300] In some examples, the master management module is configured to obtain the flag bit status of the slave management module register, and obtain the master-slave status of the slave management module according to the flag bit status of the slave management module register. When the slave management module switches to the master management module, the master management module switches to the slave management module.

[0301] It can be understood that when the first management module 1008a is the master management module, the first management module 1008a is configured to obtain the flag bit status of the second register 1009b of the second management module 1008b, and obtain the master-slave status of the second management module 1008b according to the flag bit status of the second register 1009b of the second management module 1008b.

[0302] When the second management module 1008b becomes the master management module, the flag of the second register 1009b becomes 1. The first management module 1008a determines that the second management module 1008b is the master management module according to the flag of the second register 1009b, and switches itself (the first management module 1008a) to the slave management module.

[0303] When the second management module 1008b is the master management module, the second management module 1008b is configured to obtain the flag bit status of the first register 1009a of the first management module 1008a, and obtain the master-slave status of the first management module 1008a according to the flag bit status of the first register 1009a of the first management module 1008a.

[0304] When the first management module 1008a becomes the master management module, the flag of the first register 1009a becomes 1. The second management module 1008b determines that the first management module 1008a is the master management module according to the flag of the first register 1009a, and switches itself (the second management module 1008b) to the slave management module.

[0305] For example, the first management module 1008a and the second management module 1008b can communicate with each other through the IPMB protocol to obtain the flag bit status of the other party's register.

[0306] In some examples, the main management module may poll the flag bit status of the register of the slave management module at set intervals, or the main management module may obtain the flag bit status of the register of the slave management module in real time.

[0307] It can be understood that setting the master management module to obtain the flag bit status of the slave management module register, and obtaining the master-slave status of the slave management module based on the flag bit status of the slave management module register can avoid management conflicts between the first management module 1008a and the second management module 1008b, so that the server 1000 can operate normally.

[0308] In some examples, the first management module 1008a is the master management module by default, and the first management module 1008a can communicate with the second management module 1008b via the IPMB bus to inform the second management module 1008b that it is the slave management module by default.

[0309] For example, when the server 1000 is powered on for the first time, the first management module 1008a may be set as the main management module by default. Alternatively, each time the server 1000 is powered on, the first management module 1008a may be set as the main management module by default.

[0310] In some examples, such as Fig.11 As shown, the first management module 1008a is electrically connected to the functional module 1003. When the first management module 1008a is the main management module, the first management module 1008a is used to manage the functional module 1003. When the second management module 1008b is the main management module, the second management module 1008b manages the functional module 1003 through the first management module 1008a.

[0311] For example, the first management module 1008a is electrically connected to the functional module 1003 (e.g., a riser card) so that the first management module 1008a can communicate with the functional module 1003. For example, the first management module 1008a and the functional module 1003 can communicate via an Intelligent Platform Management Interface (IPMI) bus.

[0312] When the first management module 1008 a is a main management module, the first management module 1008 a can monitor and manage the functional module 1003 .

[0313] When the second management module 1008b is the main management module, the first management module 1008a accesses the functional module 1003 and saves the access information. The second management module 1008b accesses the first management module 1008a through the IPMB bus to obtain the information of the functional module 1003, so that the second management module 1008b can monitor and manage the functional module 1003 through the first management module 1008a.

[0314] Alternatively, when the second management module 1008b is the main management module, the second management module 1008b can also access the functional module 1003 via the first management module 1008a, so that the second management module 1008b can monitor and manage the functional module 1003 through the first management module 1008a.

[0315] By adopting the above configuration, when either the first management module 1008a or the second management module 1008b is used as the main management module, both can access the function module 1003 to monitor and manage the function module 1003.

[0316] It can be understood that, in the case where the first management module 1008a is the master management module and the second management module 1008b is the slave management module, if the second management module 1008b needs to access the functional module 1003, the flag bit of the second register 1009b becomes 1. The first management module 1008a determines that the second management module 1008b is the master management module according to the flag bit state of the second register 1009b, and thus switches itself (the first management module 1008a) to the slave management module, so that the second management module 1008b can access the functional module 1003 as the master management module to achieve monitoring and management of the functional module 1003.

[0317] In the case where the second management module 1008b is the master management module and the first management module 1008a is the slave management module, if the first management module 1008a needs to access the function module 1003, the flag bit of the first register 1009a becomes 1. The second management module 1008b determines that the first management module 1008a is the master management module according to the flag bit state of the first register 1009a, and thus switches itself (the second management module 1008b) to the slave management module, so that the first management module 1008a can access the function module 1003 as the master management module to achieve monitoring and management of the function module 1003.

[0318] In some examples, such as Fig.11 As shown, the server 1000 further includes a power module 1005 , and the power module 1005 is electrically connected to one of the first management module 1008 a and the second management module 1008 b .

[0319] For example, the power module 1005 and the management module 1008 (one of the first management module 1008a and the second management module 1008b) can communicate via the IPMI bus. The power module 1005 and the functional module 1003 can be electrically connected to the same management module 1008 or to different management modules 1008.

[0320] The power module 1005 is electrically connected to one of the first management module 1008a and the second management module 1008b, so that one of the first management module 1008a and the second management module 1008b can monitor and manage the power module 1005 to implement power query, abnormality reporting, in-place detection, etc. for the power module 1005.

[0321] In some examples, the first management module 1008a is electrically connected to the power module 1005. When the first management module 1008a is a primary management module, the first management module 1008a is used to manage the power module 1005. When the second management module 1008b is a primary management module, the second management module 1008b manages the power module 1005 through the first management module 1008a.

[0322] For example, the first management module 1008a is electrically connected to the power module 1005, so that the first management module 1008a and the power module 1005 can communicate with each other. For example, the first management module 1008a and the power module 1005 can communicate with each other through an Intelligent Platform Management Interface (IPMI) bus.

[0323] In the case that the first management module 1008 a is a main management module, the first management module 1008 a can monitor and manage the power module 1005 .

[0324] When the second management module 1008b is the main management module, the first management module 1008a accesses the power module 1005 and saves the access information. The second management module 1008b accesses the first management module 1008a through the IPMB bus to obtain the information of the power module 1005, so that the second management module 1008b can monitor and manage the power module 1005 through the first management module 1008a.

[0325] Alternatively, when the second management module 1008b is the main management module, the second management module 1008b can also access the power module 1005 via the first management module 1008a, so that the second management module 1008b can monitor and manage the power module 1005 through the first management module 1008a.

[0326] By adopting the above configuration, when either the first management module 1008a or the second management module 1008b is used as the main management module, both can access the power module 1005 to monitor and manage the power module 1005 .

[0327] It can be understood that, in the case where the first management module 1008a is the master management module and the second management module 1008b is the slave management module, if the second management module 1008b needs to access the power module 1005, the flag bit of the second register 1009b becomes 1. The first management module 1008a determines that the second management module 1008b is the master management module according to the flag bit state of the second register 1009b, and thus switches itself (the first management module 1008a) to the slave management module, so that the second management module 1008b can access the power module 1005 as the master management module to achieve monitoring and management of the power module 1005.

[0328] In the case where the second management module 1008b is the master management module and the first management module 1008a is the slave management module, if the first management module 1008a needs to access the power module 1005, the flag bit of the first register 1009a becomes 1. The second management module 1008b determines that the first management module 1008a is the master management module according to the flag bit state of the first register 1009a, and thus switches itself (the second management module 1008b) to the slave management module, so that the first management module 1008a can access the power module 1005 as the master management module to achieve monitoring and management of the power module 1005.

[0329] In some examples, the first management module 1008a is electrically connected to the fan 1004. The main management module can determine the appropriate speed through the speed adjustment algorithm and control the fan speed through the management board body 1110. In addition, the main management module can detect the fan operation status through the speed feedback module to realize the monitoring and management of the fan.

[0330] Fig.12 A schematic diagram of the connection relationship between the management module and other components provided in some embodiments of the present application. In some examples, such as Fig.12 As shown, the server 1000 further includes a first logic device 1181 and a second logic device 1182. The first management module 1008a is electrically connected to at least one of the first node 1161 and the second node 1162 via an expansion port of the first logic device 1181. The second management module 1008b is electrically connected to at least one of the third node 1164 and the fourth node 1164 via an expansion port of the second logic device 1182.

[0331] It is understandable that in the embodiment of the present application, the first logic device 1181 and the second logic device 1182 may not participate in node control, but serve as expansion ports of the first management module 1008a and the second management module 1008b. For example, the first logic device 1181 and the second logic device 1182 may be CPLDs.

[0332] For example, Fig.12 As shown, the first management module 1008a and the second management module 1008b can use chips that support LVDS tunneling protocol (LTPI). The first management module 1008a can be connected to the first logic device 1181 using the LTPI port, and the second management module 1008b can be electrically connected to the second logic device 1182 using the LTPI port.

[0333] For example, the first management module 1008a can be connected to the first logic device 1181 through the LTPI0 port and the LTPI1 port respectively, and be electrically connected to the first node 1161 through the SGPIO0 port and the UART0 port of the first logic device 1181, and be electrically connected to the second node 1162 through the SGPIO1 port and the UART1 port of the first logic device 1181, so that the first management module 1008a can realize the expansion of low-speed signal ports such as SGPIO and UART through the first logic device 1181.

[0334] The second management module 1008b can be connected to the second logic device 1182 through the LTPI0 port and the LTPI1 port respectively, and is electrically connected to the third node 1163 through the SGPIO0 port and the UART0 port of the second logic device 1182, and is electrically connected to the fourth node 1164 through the SGPIO1 port and the UART1 port of the second logic device 1182, so that the second management module 1008b can realize the expansion of low-speed signal ports such as SGPIO and UART through the second logic device 1182.

[0335] The first logic device 1181 is electrically connected to the second logic device 1182. For example, the first logic device 1181 and the second logic device 1182 can be electrically connected via an SGPIO bus so that the two can exchange information. The first management module 1008a is electrically connected to the first logic device 1181, and the second management module 1008b is electrically connected to the second logic device 1182, so that the first management module 1008a and the second management module 1008b can be electrically connected.

[0336] The management module 1008 is electrically connected to the node through the logic device, so that the logic device can expand the port between the management module 1008 and the node, such as the low-speed signal port such as SGPIO and UART. In addition, the logic device does not need to participate in the control of multiple nodes, which simplifies the control logic of multiple nodes and helps reduce the cost of the server 1000.

[0337] In some examples, such as Fig.12 As shown, the first management module 1008 a is electrically connected to the functional module 1003 and the power module 1005 through the first logic device 1181 .

[0338] For example, Fig.12 As shown, the first management module 1008a can be connected to the first logic device 1181 through the LTPI0 interface and the LTPI1 interface respectively, and is electrically connected to the function module 1003 through the I2C2 port, SGPIO2 port and UART2 port of the first logic device 1181, and the first management module 1008a can be electrically connected to the power module 1005 through the I2C3 port and SGPIO3 port of the first logic device 1181. In this way, the first management module 1008a can achieve expansion of low-speed signal ports such as I2C, SGPIO and UART through the first logic device 1181.

[0339] For example, the management module 1008 (the first management module 1008a and the second management module 1008b) can select a chip that supports Multi Node, so that the management module 1008 can support two PCIE buses, two eSPI buses, two NCSI buses and two I2C buses.

[0340] For example, Fig.12 As shown, the PCIE0 port, the eSPI0 port, and the I2C0 port of the first management module 1008 a can be electrically connected to the first node 1161 without being extended via the first logic device 1181 .

[0341] For example, when the functional module 1003 includes a riser card and a DPU card, the NCSI port of the first management module 1008a needs to be electrically connected to the riser card so that the riser card can support the DPU card. Fig.12 As shown, the server 1000 may include a fifth selector 1195 , and the first interface of the fifth selector 1195 is configured to be connected to the second interface of the fifth selector 1195 or the third interface of the fifth selector 1195 .

[0342] The NCSI0 port of the first management module 1008a is electrically connected to the first interface of the fifth selector 1195, one of the function module 1003 and the power module 1005 is electrically connected to the second interface of the fifth selector 1195, and the other is electrically connected to the third interface of the fifth selector 1195. In this way, the NCSI0 port of the first management module 1008a can be optionally electrically connected to the first node 1161 and the function module 1003 (riser card).

[0343] Continue to refer to Fig.12 The PCIE1 port, eSPI1 port, I2C1 port and NCSI1 port of the first management module 1008a can be electrically connected to the second node 1162 without being extended through the first logic device 1181.

[0344] The PCIE0 port, eSPI0 port, I2C0 port, and NCSI0 port of the second management module 1008b can be electrically connected to the third node 1163 without being extended via the second logic device 1182. The PCIE1 port, eSPI1 port, I2C1 port, and NCSI1 port of the second management module 1008b can be electrically connected to the fourth node 1164 without being extended via the second logic device 1182.

[0345] In some examples, such as Fig.12As shown, the server 1000 further includes a first selector 1191, and the first interface of the first selector 1191 is configured to be connected to the second interface of the first selector 1191 or the third interface of the first selector 1191. The first management module 1008a includes at least one first-class port. One first-class port is electrically connected to the first interface of the first selector 1191. The second interface of the first selector 1191 is electrically connected to the first node 1161, and the third interface of the first selector 1191 is electrically connected to the second node 1162.

[0346] For example, the number of the first class ports may be multiple or more. That is, the first management module 1008a may include one first class port or multiple first class ports. For example, the first class ports may include an I3C port (I3C_MNG port, I3C_SPD port, and I3C_SPD port), a PECI port, and a JTAG port.

[0347] It can be understood that the first type of port is a port that needs to be electrically connected to the first node 1161 and the second node 1162, and the first management module 1008a only includes one. Taking the I3C_MNG port as an example, the first management module 1008a may include one I3C_MNG port, and one I3C_MNG port needs to be electrically connected to the first node 1161 and the second node 1162.

[0348] like Fig.12 As shown, there may be multiple first selectors 1191. The first interface of one first selector 1191 is electrically connected to the I3C_MNG port, and the second interface and the third interface of the first selector 1191 are electrically connected to the first node 1161 and the second node 1162 respectively.

[0349] The first interface of another first selector 1191 is electrically connected to the PECI port, and the second interface and the third interface of the first selector 1191 are electrically connected to the first node 1161 and the second node 1162, respectively. The third interface of another first selector 1191 is electrically connected to the JTAG port, and the second interface and the third interface of the first selector 1191 are electrically connected to the first node 1161 and the second node 1162, respectively. The third interface of another first selector 1191 is electrically connected to the I3C_SPD port, and the second interface and the third interface of the first selector 1191 are electrically connected to the first node 1161 and the second node 1162, respectively. The third interface of another first selector 1191 is electrically connected to the I3C_DBG port, and the second interface and the third interface of the first selector 1191 are electrically connected to the first node 1161 and the second node 1162, respectively.

[0350] The first interface of the first selector 1191 is configured to be connected to the second interface of the first selector 1191 or the third interface of the first selector 1191 , so that the first type port of the first management module 1008a can be selectively connected to the first node 1161 or the second node 1162 .

[0351] like Fig.12 As shown, the server 1000 also includes a second selector 1192, and the first interface of the second selector 1192 is configured to be connected to the second interface of the second selector 1192 or the third interface of the second selector 1192. The second management module 1008b includes at least one second-class port. One second-class port is electrically connected to the first interface of the second selector 1192. The second interface of the second selector 1192 is electrically connected to the third node 1163, and the third interface of the second selector 1192 is electrically connected to the fourth node 1164.

[0352] For example, the number of the second type ports can be multiple or more. That is, the first management module 1008a can include one second type port or multiple second type ports. For example, the second type ports can include I3C ports (I3C_MNG ports, I3C_SPD ports, and I3C_SPD ports), PECI ports, and JTAG ports.

[0353] It can be understood that the second type of port is a port that needs to be electrically connected to the third node 1163 and the fourth node 1164, and the second management module 1008b only includes one port. Taking the I3C_MNG port as an example, the second management module 1008b can include one I3C_MNG port, and one I3C_MNG port needs to be electrically connected to the third node 1163 and the fourth node 1164.

[0354] like Fig.12 As shown, there are multiple second selectors 1192. The first interface of one second selector 1192 is electrically connected to the I3C_MNG port, and the second interface and the third interface of the second selector 1192 are electrically connected to the third node 1163 and the fourth node 1164 respectively.

[0355] The first interface of another second selector 1192 is electrically connected to the PECI port, and the second interface and the third interface of the second selector 1192 are electrically connected to the third node 1163 and the fourth node 1164, respectively. The first interface of yet another second selector 1192 is electrically connected to the JTAG port, and the second interface and the third interface of the second selector 1192 are electrically connected to the third node 1163 and the fourth node 1164, respectively. The first interface of yet another second selector 1192 is electrically connected to the I3C_SPD port, and the second interface and the third interface of the second selector 1192 are electrically connected to the third node 1163 and the fourth node 1164, respectively. The first interface of yet another second selector 1192 is electrically connected to the I3C_DBG port, and the second interface and the third interface of the second selector 1192 are electrically connected to the third node 1163 and the fourth node 1164, respectively.

[0356] The first interface of the second selector 1192 is configured to be connected to the second interface of the second selector 1192 or the third interface of the second selector 1192 , so that the second type port of the second management module 1008 b can be selectively connected to the third node 1163 or the fourth node 1164 .

[0357] In some examples, when the server 1000 includes the first selector 1191, the first interface of the first selector 1191 is alternately connected with the second interface of the first selector 1191 and the third interface of the first selector 1191. Or, the first selector 1191 connects the first interface of the first selector 1191 and the second interface of the first selector 1191 based on the request signal of the first node 1161. And, the first selector 1191 connects the first interface of the first selector 1191 and the third interface of the first selector 1191 based on the request signal of the second node 1162.

[0358] It can be understood that the first interface of the first selector 1191 and the second interface of the first selector 1191 and the third interface of the first selector 1191 can be alternately turned on at set intervals, so that the first management module 1008a can be alternately electrically connected to the first node 1161 and the second node 1162.

[0359] Alternatively, when the first selector 1191 receives a request from the first node 1161, the first interface of the first selector 1191 and the second interface of the first selector 1191 may be turned on; and when the first selector 1191 receives a request from the second node 1161, the first interface of the first selector 1191 and the third interface of the first selector 1191 may be turned on.

[0360] By adopting the above configuration, the flexibility of connecting the first type port of the first management module 1008a with the first node 1161 and the second node 1162 can be improved.

[0361] In the case where the server 1000 includes the second selector 1192, the first interface of the second selector 1192 is alternately connected with the second interface of the second selector 1192 and the third interface of the second selector 1192. Alternatively, the second selector 1192 connects the first interface of the second selector 1192 and the second interface of the second selector 1192 based on the request signal of the third node 1163. Furthermore, the second selector 1192 connects the first interface of the second selector 1192 and the third interface of the second selector 1192 based on the request signal of the fourth node 1164.

[0362] It can be understood that the first interface of the second selector 1192, the second interface of the second selector 1192, and the third interface of the second selector 1192 can be alternately turned on at set intervals, so that the second management module 1008b can be alternately electrically connected to the third node 1163 and the fourth node 1164.

[0363] Alternatively, when the second selector 1192 receives a request from the third node 1163, the first interface of the second selector 1192 and the second interface of the second selector 1192 may be turned on; and when the second selector 1192 receives a request from the second node 1161, the first interface of the second selector 1192 and the third interface of the second selector 1192 may be turned on.

[0364] Such configuration can improve the flexibility when the second type port of the second management module 1008 b is connected to the third node 1163 and the fourth node 1164 .

[0365] In some examples, such as Fig.12 As shown, the server 1000 further includes a detection module 1210. The detection module 1210 is electrically connected to the management module 1008, and the detection module 1210 is electrically connected to the first node 1161, the second node 1162, the third node 1163, and the fourth node 1164. The detection module 1210 is used to detect the hardware status of at least one of the first node 1161, the second node 1162, the third node 1163, and the fourth node 1164, and feedback to the management module 1008 electrically connected to the node when the hardware status of the node is abnormal.

[0366] For example, the detection module 1210 may include an SPI flash memory for storing a CPU BIOS (basic inputoutput system). The detection module 1210 may be electrically connected to the first management module 1008a and the second management module 1008b. When the hardware state of the first node 1161 is abnormal, the detection module 1210 can feed back the abnormality of the first node 1161 to the first management module 1008a. When the hardware state of the second node 1162 is abnormal, the detection module 1210 can feed back the abnormality of the second node 1162 to the first management module 1008a. When the hardware state of the third node 1163 is abnormal, the detection module 1210 can feed back the abnormality of the third node 1163 to the second management module 1008b. When the hardware state of the fourth node 1164 is abnormal, the detection module 1210 can feed back the abnormality of the fourth node 1164 to the second management module 1008b.

[0367] In some examples, when the hardware status of the node is normal, the detection module 1210 can also feedback the hardware status of the node normally to the management module 1008 electrically connected to the node.

[0368] By setting up the detection module 1210, the management module 1008 can obtain the abnormal state of the detection module 1210, thereby improving the reliability of the server 1000.

[0369] In some examples, such as Fig.12 As shown, the detection module 1210 includes a first detection module 1211, a second detection module 1212, a third detection module 1213 and a fourth detection module 1214. The first detection module 1211 is electrically connected to the first management module 1008a and the first node 1161, and is used to detect the hardware status of the first node 1161, and to feed back to the first management module 1008a when the hardware status of the first node 1161 is abnormal. The second detection module 1212 is electrically connected to the first management module 1008a and the second node 1162, and is used to detect the hardware status of the second node 1162, and to feed back to the first management module 1008a when the hardware status of the second node 1162 is abnormal.

[0370] The third detection module 1213 is electrically connected to the second management module 1008b and the third node 1163, and is used to detect the hardware status of the third node 1163, and to provide feedback to the second management module 1008b when the hardware status of the third node 1163 is abnormal. The fourth detection module 1214 is electrically connected to the second management module 1008b and the fourth node 1164, and is used to detect the hardware status of the fourth node 1164, and to provide feedback to the second management module 1008b when the hardware status of the fourth node 1164 is abnormal.

[0371] For example, Fig.12 As shown, the server 1000 may include a sixth selector 1196. It can be understood that the first interface of the sixth selector 1196 is configured to be connected to the second interface of the sixth selector 1196 or the third interface of the sixth selector 1196. The sixth selector 1196 includes a first sixth selector 1196a, a second sixth selector 1196b, a third sixth selector 1196c, and a fourth sixth selector 1196d.

[0372] Continue to refer to Fig.12 , the first interface of the first sixth selector 1196a is electrically connected to the first detection module 1211, the second interface of the first sixth selector 1196a is electrically connected to the first node 1161, and the third interface of the first sixth selector 1196a is electrically connected to the QSPIO0 port of the first management module 1008a. In this way, the first detection module 1211 can be selectively connected to the first node 1161 and the first management module 1008a, so that when the hardware status of the first node 1161 is abnormal, feedback can be given to the first management module 1008a.

[0373] The first interface of the second sixth selector 1196b is electrically connected to the second detection module 1212, the second interface of the second sixth selector 1196b is electrically connected to the second node 1162, and the third interface of the second sixth selector 1196b is electrically connected to the QSPIO1 port of the first management module 1008a. In this way, the second detection module 1212 can be selectively connected to the second node 1162 and the first management module 1008a, so that when the hardware status of the second node 1162 is abnormal, feedback can be given to the first management module 1008a.

[0374] The first interface of the third sixth selector 1196c is electrically connected to the third detection module 1213, the second interface of the third sixth selector 1196c is electrically connected to the third node 1163, and the third interface of the third sixth selector 1196c is electrically connected to the QSPIO0 terminal of the second management module 1008b. In this way, the third detection module 1213 can be selectively connected to the third node 1163 and the second management module 1008b, so that when the hardware status of the third node 1163 is abnormal, feedback can be given to the second management module 1008b.

[0375] The first interface of the fourth sixth selector 1196d is electrically connected to the fourth detection module 1214, the second interface of the fourth sixth selector 1196d is electrically connected to the fourth node 1164, and the third interface of the fourth sixth selector 1196d is electrically connected to the QSPIO1 terminal of the second management module 1008b. In this way, the fourth detection module 1214 can be selectively connected to the fourth node 1164 and the second management module 1008b, so that when the hardware status of the fourth node 1164 is abnormal, it can be fed back to the second management module 1008b.

[0376] It can be understood that the detection module 1210 is set to include a first detection module 1211, a second detection module 1212, a third detection module 1213 and a fourth detection module 1214, so that the hardware status of the first node 1161, the second node 1162, the third node 1163 and the fourth node 1164 can be detected respectively, reducing the mutual influence when detecting the hardware status of each node and improving the reliability of the server 1000.

[0377] In some examples, the first management module 1008a may upgrade the programs of the first detection module 1211 and the second detection module 1212 , and the second management module 1008b may upgrade the programs of the third detection module 1213 and the fourth detection module 1214 to improve the reliability of the server 1000 .

[0378] In some examples, such as Fig. 9 and Fig.10 As shown, the server 1000 includes a first switch unit 1171, a second switch unit 1172, a third switch unit 1173, and a fourth switch unit 1174. The first switch unit 1171 is electrically connected to the first node 1161, the second switch unit 1172 is electrically connected to the second node 1162, the third switch unit 1173 is electrically connected to the third node 1163, and the fourth switch unit 1174 is electrically connected to the fourth node 1164.

[0379] The switch unit may be disposed on the first protection plate 1301 connected to the node plate 1002. For example, the switch unit may be a UID button, wherein the UID button may be used as a positioning indicator in addition to enabling the node. That is, after the user triggers the UID button, a UID light will be lit on the corresponding node to help the user locate the node or the server 1000 where the node is located.

[0380] For example, the first switch unit 1171 may be electrically connected to the first node plate 1002a of the first node plate group 1150a, so that the first switch unit 1171 can control the operation of the first node 1161. The second switch unit 1172 may be electrically connected to the second node plate 1002b of the first node plate group 1150a, so that the second switch unit 1172 can control the operation of the second node 1162. The third switch unit 1173 may be electrically connected to the first node plate 1002a of the second node plate group 1150b, so that the third switch unit 1173 can control the operation of the third node 1163. The fourth switch unit 1174 may be electrically connected to the second node plate 1002b of the second node plate group 1150b, so that the fourth switch unit 1174 can control the operation of the fourth node 1164.

[0381] In some examples, when the duration of the first switch unit 1171 being triggered is within the first duration range, the first node 1161 works alone, and the first management module 1008a manages the first node 1161. When the duration of the second switch unit 1172 being triggered is within the second duration range, the second node 1162 works alone, and the first management module 1008a manages the second node 1162. When the duration of the third switch unit 1173 being triggered is within the third duration range, the third node 1163 works alone, and the second management module 1008b manages the third node 1163. When the duration of the fourth switch unit 1174 being triggered is within the fourth duration range, the fourth node 1164 works alone, and the second management module 1008b manages the fourth node 1164.

[0382] For example, the switch unit can be triggered by pressing. Alternatively, the switch unit can be triggered by touching or the like. The first switch unit 1171, the second switch unit 1172, the third switch unit 1173, and the fourth switch unit 1174 can be triggered in the same way or in different ways.

[0383] When the duration of the first switch unit 1171 being triggered is within the first duration range, it indicates that the first node 1161 is an independent node, the first node 1161 works independently, and the first management module 1008a can monitor and manage the first node 1161. For example, the first duration range may be 4s (unit: seconds) to 8s. For example, when the duration of the first switch unit 1171 being triggered is 6s, the first node 1161 works independently, and the first management module 1008a manages the first node 1161.

[0384] When the duration of the second switch unit 1172 being triggered is within the second duration range, it indicates that the second node 1162 is an independent node, the second node 1162 works independently, and the first management module 1008a can monitor and manage the second node 1162. For example, the second duration range may be 4s (unit: seconds) to 8s. For example, when the duration of the second switch unit 1172 being triggered is 6s, the second node 1162 works independently, and the first management module 1008a manages the second node 1162.

[0385] When the duration of the third switch unit 1173 being triggered is within the third duration range, it indicates that the third node 1163 is an independent node, the third node 1163 works independently, and the second management module 1008b can monitor and manage the third node 1163. For example, the third duration range may be 4s (unit: seconds) to 8s. For example, when the duration of the third switch unit 1173 being triggered is 6s, the third node 1163 works independently, and the second management module 1008b manages the third node 1163.

[0386] When the duration of the fourth switch unit 1174 being triggered is within the fourth duration range, it indicates that the fourth node 1164 is an independent node, the fourth node 1164 works independently, and the second management module 1008b can monitor and manage the fourth node 1164. For example, the fourth duration range may be 4s (unit: seconds) to 8s. For example, when the duration of the fourth switch unit 1174 being triggered is 6s, the fourth node 1164 works independently, and the second management module 1008b manages the fourth node 1164.

[0387] The first duration range, the second duration range, the third duration range and the fourth duration range may be the same or different. The embodiment of the present application does not further limit the values ​​of the first duration range, the second duration range, the third duration range and the fourth duration range.

[0388] It can be understood that when the duration when the first switch unit 1171 is triggered is within the first duration range, the duration when the second switch unit 1172 is triggered is within the second duration range, the duration when the third switch unit 1173 is triggered is within the third duration range, and the duration when the fourth switch unit 1174 is triggered is within the fourth duration range, the first management module 1008a manages the first node 1161 and the second node 1162, the second management module 1008b manages the third node 1163 and the fourth node 1164, and the server 1000 operates in four-node mode, that is, four separate nodes.

[0389] When the duration of at least one of the first switch unit 1171 and the second switch unit 1172 being triggered is within the fifth duration range, the first node 1161 and the second node 1162 are connected in series to form a first node group 1160a, and the main management module manages the first node group 1160a. When the duration of at least one of the third switch unit 1173 and the fourth switch unit 1174 being triggered is within the sixth duration range, the third switch unit 1173 and the fourth switch unit 1174 are connected in series to form a second node group 1160b, and the main management module manages the second node group 1160b.

[0390] For example, after the first node board 1002a and the second node board 1002b of the first node board group 1150a are connected in series through the high-speed cable 1006, at least one of the first switch unit 1171 and the second switch unit 1172 may be triggered. When the duration of at least one of the first switch unit 1171 and the second switch unit 1172 being triggered is within the fifth duration range, the UPI bus between the first node 1161 and the second node 1162 is enabled, so that the first node 1161 and the second node 1162 can be connected in series to form the first node group 1160a. The main management module can monitor and manage the first node group 1160a.

[0391] It can be understood that the fifth duration range is different from the first duration range, the second duration range, the third duration range and the fourth duration range. For example, the value of the fifth duration range can be greater than 0s and less than 4s. For example, when the duration of at least one of the first switch unit 1171 and the second switch unit 1172 being triggered is 3s, the first node 1161 and the second node 1162 are connected in series to form a first node group 1160a.

[0392] When the first node 1161 and the second node 1162 are connected in series to form the first node group 1160a, if the duration of the first switch unit 1171 being triggered is within the first duration range, or the duration of the second switch unit 1172 being triggered is within the second duration range, the UPI bus between the first node 1161 and the second node 1162 fails, and the first node 1161 and the second node 1162 work independently.

[0393] For example, after the first node board 1002a and the second node board 1002b of the second node board group 1150b are connected in series through the high-speed cable 1006, at least one of the third switch unit 1173 and the fourth switch unit 1174 is triggered. When the duration of at least one of the third switch unit 1173 and the fourth switch unit 1174 being triggered is within a sixth duration range, the UPI bus between the third node 1163 and the fourth node 1164 is enabled, so that the third node 1163 and the fourth node 1164 can be connected in series to form the second node group 1160b. The main management module can monitor and manage the second node group 1160b.

[0394] It can be understood that the sixth duration range has different values ​​from the first duration range, the second duration range, the third duration range and the fourth duration range. The fifth duration range and the sixth duration range may have the same value or may be different values.

[0395] For example, the value of the sixth duration range can be greater than 0s and less than 4s. For example, when the duration of at least one of the third switch unit 1173 and the fourth switch unit 1174 being triggered is 3s, the third node 1163 and the fourth node 1164 are connected in series to form a second node group 1160b.

[0396] When the third node 1163 and the fourth node 1164 are connected in series to form the second node group 1160b, if the duration for which the third switch unit 1173 is triggered is within the third duration range, or the duration for which the fourth switch unit 1174 is triggered is within the fourth duration range, the UPI bus between the third node 1163 and the fourth node 1164 fails, and the third node 1163 and the fourth node 1164 work independently.

[0397] When the duration when at least one of the first switch unit 1171 and the second switch unit 1172 is triggered is within the fifth duration range, and the duration when at least one of the third switch unit 1173 and the fourth switch unit 1174 is triggered is within the sixth duration range, the main management module manages the first node group 1160a and the second node group 1160b, and the server 1000 operates in dual-node mode, i.e., dual-path and two-node.

[0398] It can be understood that by setting the switch unit to be electrically connected to the node, the working mode (dual-node mode or four-node mode) of the server 1000 can be switched through different triggering durations of the switch unit. The operation is simple and the convenience of operation and maintenance is improved.

[0399] For example, when the duration of the first switch unit 1171 being triggered is within the first duration range, the PCIE interface, eSPI interface, NCSI interface, I2C interface, I3C interface (I3C_MNG port, I3C_SPD port and I3C_SPD port), PECI interface, and JTAG interface of the first management module 1008a are all electrically connected to the first node 1161. When the duration of the second switch unit 1172 being triggered is within the second duration range, the PCIE interface, eSPI interface, NCSI interface, I2C interface, I3C interface I3C_MNG port, I3C_SPD port and I3C_SPD port, PECI interface, and JTAG interface of the first management module 1008a are all electrically connected to the second node 1162.

[0400] When the duration of the third switch unit 1173 being triggered is within the first duration range, the PCIE interface, eSPI interface, NCSI interface, I2C interface, I3C interface (I3C_MNG port, I3C_SPD port and I3C_SPD port), PECI interface and JTAG interface of the second management module 1008b are all electrically connected to the third node 1163. When the duration of the fourth switch unit 1174 being triggered is within the second duration range, the PCIE interface, eSPI interface, NCSI interface, I2C interface, I3C interface I3C_MNG port, I3C_SPD port and I3C_SPD port, PECI interface and JTAG interface of the second management module 1008b are all electrically connected to the fourth node 1164.

[0401] In some examples, such as Fig.12 As shown, the server 1000 further includes a first hanging ear 1401 , a second hanging ear 1402 , a third selector 1193 and a fourth selector 1194 .

[0402] The first interface of the third selector 1193 is electrically connected to the USB port of the first hanging ear 1401, and the second interface of the third selector 1193 and the third interface of the third selector 1193 are electrically connected to the USB port of the first management module 1008a and the USB port of the second management module 1008b respectively.

[0403] For example, the first management module 1008a and the second management module 1008b support PCIE expansion of two USB3.0 signals. For example, two USB ports can be set on the first mounting ear 1401, and any USB port can be optionally electrically connected to the first management module 1008a or the second management module 1008b through the third selector 1193.

[0404] It can be understood that the USB port of the first hanging ear 1401 can be optionally electrically connected to the first management module 1008a or the second management module 1008b, so that external components can be electrically connected to the first management module 1008a or the second management module 1008b through the USB port of the first hanging ear 1401, thereby improving the convenience of connecting the first management module 1008a and the second management module 1008b with external components.

[0405] In some examples, a UID button may be provided on the first hanging ear 1401. By operating the UID button, the management module 1008 electrically connected to the USB port can be switched, thereby improving the convenience of operation and maintenance.

[0406] Continue to refer to Fig.12 The first interface of the fourth selector 1194 is electrically connected to the DP port of the second hanging ear 1402, and the second interface of the fourth selector 1194 and the third interface of the fourth selector 1194 are electrically connected to the DP port of the first management module 1008a and the DP port of the second management module 1008b respectively.

[0407] For example, the first management module 1008a and the second management module 1008b support DP signals. For example, a DP port may be provided on the second mounting ear 1402 , and the DP port may be electrically connected to the first management module 1008a or the second management module 1008b through the fourth selector 1194 .

[0408] It can be understood that the DP port of the second hanging ear 1402 can be optionally electrically connected to the first management module 1008a or the second management module 1008b, so that external components can be electrically connected to the first management module 1008a or the second management module 1008b through the DP port of the second hanging ear 1402, thereby improving the convenience of connecting the first management module 1008a and the second management module 1008b with external components.

[0409] In some examples, a UID button may be provided on the second hanging ear 1402. By operating the UID button, the management module 1008 electrically connected to the DP port can be switched, thereby improving the convenience of operation and maintenance.

[0410] In some examples, such as Fig.12 As shown, a first network port is provided on the first hanging ear 1401, and the first network port is electrically connected to the SGMII port of the first management module 1008a. For example, the first network port can be an RJ45 network port. It can be understood that the first node 1161 and the second node 1162 can be managed through the first network port.

[0411] The second hanging ear 1402 is provided with a second network port, and the second network port is electrically connected to the SGMII port of the second management module 1008b. For example, the second network port can be an RJ45 network port. It can be understood that the third node 1163 and the fourth node 1164 can be managed through the second network port.

[0412] It is understandable that other interfaces may also be provided on the first hanging ear 1401 and the second hanging ear 1402 , and the embodiments of the present application do not further limit this.

[0413] In summary, the embodiments of the present application have at least the following beneficial effects:

[0414] In the embodiment of the present application, the node board group 1150 and / or the functional module 1003 are pluggable and connected to the management board 1100, so that the management board 1100, the node board group 1150 and the functional module 1003 can be physically independent of each other, and they are electrically connected through connectors or cable connections to achieve communication with each other. In this way, the node board group 1150 and / or the functional module 1003 can be decoupled from the management board 1100 separately.

[0415] The first zone A is close to the first side 1001a of the chassis 1001, and the node board group 1150 and / or at least one functional module 1003 arranged in the first zone A are connected to the management board 1100 in a pluggable manner, so that the node board group 1150 and / or at least one functional module 1003 can be pulled out or pushed in from the first side 1001a of the chassis 1001. The third zone C is close to the second side 1001b of the chassis 1001, and the node board group 1150 and / or at least one functional module 1003 arranged in the third zone C are connected to the management board 1100 in a pluggable manner, so that the node board group 1150 and / or at least one functional module 1003 can be pulled out or pushed in from the second side 1001b of the chassis 1001.

[0416] For example, by pulling out, the node board group 1150 and / or at least one functional module 1003 arranged in the first zone A can be decoupled from the management board 1100, and at this time, the node board group 1150 and / or at least one functional module 1003 arranged in the first zone A can be separated from the chassis 1001, thereby being exposed; by pushing in, the node board group 1150 and / or at least one functional module 1003 can be arranged in the first zone A and plugged into the management board 1100.

[0417] Similarly, by pulling out, the node board group 1150 and / or at least one functional module 1003 arranged in the third zone C can be decoupled from the management board 1100, and at this time, the node board group 1150 and / or at least one functional module 1003 arranged in the third zone C can be separated from the chassis 1001, thereby being exposed; by pushing in, the node board group 1150 and / or at least one functional module 1003 can be arranged in the third zone C and plugged into the management board 1100.

[0418] When the node board group 1150 and / or at least one functional module 1003 arranged in the first zone A and the third zone C need to be repaired or replaced, there is no need to open the upper cover of the chassis 1001, but the node board group 1150 and / or at least one functional module 1003 can be repaired by pulling out and pushing in, thereby realizing the repair of the server 1000 on both sides (front and rear sides) of the server 1000, thereby improving the convenience of operation and maintenance, thereby improving the operation and maintenance efficiency.

[0419] In addition, for rack-mounted servers, the node board group 1150 and / or at least one functional module 1003 can be inspected and repaired by pulling out and pushing in, without removing the server 1000 from the cabinet, thereby improving the convenience of operation and maintenance and thus improving the efficiency of operation and maintenance.

[0420] The management module 1008 is arranged on the management board body 1110 of the management board 1100, and the multiple node boards 1002 of the node board group 1150 are respectively plugged into the management board 1100, and the multiple functional modules 1003 are respectively plugged into the management board 1100, so that the multiple node boards 1002 of each node board group 1150 can be maintained separately, and each functional module 1003 (such as IO card, GPU card or DPU card, etc.) can be maintained separately, thereby improving the convenience and flexibility of operation and maintenance.

[0421] The first management module 1008a and the second management module 1008b are electrically connected so that the first management module 1008a and the second management module 1008b can communicate with each other without signal convergence on the management board body 1110 and without other logic devices (such as complex programmable logic devices, CPLD) participating in the communication between the first management module 1008a and the second management module 1008b.

[0422] In this way, the communication path can be shortened and the efficiency of information interaction between the first management module 1008a and the second management module 1008b can be improved.

[0423] The first management module 1008a is electrically connected to the first node 1161 and the second node 1162, so that the first management module 1008a can monitor and manage at least one of the first node 1161 and the second node 1162. The second management module 1008b is electrically connected to the third node 1163 and the fourth node 1164, so that the second management module 1008b can monitor and manage at least one of the third node 1163 and the fourth node 1164.

[0424] In this way, two management modules 1008 can manage four nodes (one management module 1008 manages two nodes), and there is no need to set the number of management modules 1008 to be the same as the number of nodes, which reduces the number of management modules 1008, simplifies the structure of the server 1000, and reduces the cost of the server 1000. In addition, the management scheme of the server 1000 for multiple nodes is simplified.

[0425] 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 server, characterized in that: include: A chassis, comprising a first area and a second area; along a first direction, the first area is closer to a first side of the chassis than the second area; A management board is arranged in the second area; the management board is provided with a management module for managing the operation of the server; and, A node board group and at least one functional module; the node board group includes a plurality of node boards arranged in a stacked manner; a business processing module for processing business is arranged on the node board, and the functional module is used to cooperate with the business processing module to realize data transmission or processing; The node board group and / or at least one of the functional modules are arranged in the first area and are pluggably connected to the management board.

2. The server according to claim 1, characterized in that: The chassis further includes a third area, and along the first direction, the second area is arranged between the first area and the third area, and the third area is closer to the second side of the chassis than the second area; The node board group is arranged in the first area and / or the third area; when the node board group is arranged in the first area, each node board in the node board group can be respectively pulled out or pushed in from the first side of the chassis by being pluggable connected to the management board respectively; when the node board group is arranged in the third area, each node board in the node board group can be respectively pulled out or pushed in from the second side of the chassis by being pluggable connected to the management board respectively; and / or, At least one of the functional modules is arranged in the first area and / or the third area; when the functional module is arranged in the first area, it can be pulled out or pushed in from the first side of the chassis through a pluggable connection with the management board; when the functional module is arranged in the third area, it can be pulled out or pushed in from the second side of the chassis through a pluggable connection with the management board.

3. The server according to claim 2, characterized in that: A length of the first region along the first direction is equal to a length of the third region along the first direction.

4. The server according to claim 2 or 3, characterized in that: A first opening is formed on the first side of the chassis, and the node plate group and / or the functional module arranged in the first area are configured to be pulled out or pushed in from the first opening; A second opening is formed on the second side of the chassis, and the node plate group and / or the functional module disposed in the third area are configured to be pulled out or pushed in from the second opening.

5. The server according to any one of claims 2 to 4, characterized in that: The chassis further comprises a plurality of partitions, wherein the partitions extend along the first direction; The plurality of partitions include a first partition and a second partition; the first partition is disposed in the first region and divides the first region into a plurality of first sub-regions; the second partition is disposed in the third region and divides the third region into a plurality of third sub-regions; Wherein, the position of the partition along the second direction is adjustable; the second direction intersects with the first direction.

6. The server according to any one of claims 2 to 5, characterized in that: The management board comprises a management board body and a connector module, the management board body and the connector module are electrically connected, and the management module is arranged on the management board body; The connector module is arranged on the management board body; or, the connector module is arranged on one side of the management board body along the first direction and is adjacent to the management board body; The connector module comprises a fixing plate and a connector mounted on the surface of the fixing plate; the number of the connectors is multiple, and the node plate / or the functional module is configured to be plugged with at least one of the connectors; The number of the connector modules is two, and the two connector modules are spaced apart along the first direction; Of the two connector modules, the connector module close to the first side of the chassis is the first connector module, and the connector module close to the second side of the chassis is the second connector module; The connector opening of the first connector module faces the first side of the chassis, and the connector opening of the second connector module faces the second side of the chassis.

7. The server according to claim 6, characterized in that: The connector includes a high-speed connector, and at least one of the node board group and the functional module is electrically connected to the high-speed connector for transmitting high-speed signals.

8. The server according to claim 6 or 7, characterized in that: The connector includes a low-speed connector, which is electrically connected to the management board body, and at least one of the node board group and the functional module is pluggably connected to the low-speed connector for transmitting low-speed signals.

9. The server according to any one of claims 6 to 8, characterized in that: It also includes a power supply module, the power supply module includes a power supply output interface; the node board includes a first power supply input interface, and the functional module includes a second power supply input interface; The connector includes a first power connector, a second power connector and a third power connector; the third power connector is electrically connected to the power output interface, the first power connector is electrically connected to one of the first power input interface and the second power input interface, and the second power connector is electrically connected to the other of the first power input interface and the second power input interface.

10. The server according to any one of claims 1 to 9, characterized in that: The management module includes a first management module and a second management module; the first management module and the second management module are electrically connected; The server includes a first node, a second node, a third node and a fourth node; the first node and the second node are electrically connected to the first management module respectively, and the third node and the fourth node are electrically connected to the second management module respectively; The functional module is used to cooperate with at least one of the first node, the second node, the third node and the fourth node to perform data transmission or processing; One of the first management module and the second management module is used to manage the functional module; When the server is in a first working state, the first node, the second node, the third node and the fourth node are configured to work independently; The first management module is used to manage at least one of the first node and the second node; And / or, the second management module is used to manage at least one of the third node and the fourth node.

Images