Four-node through-type server

Abstract

The utility model provides a four-node through-type server which comprises a server device body and a server case, the server case comprises a case shell, a separation structure, a 2 + 2 redundant power supply module, a node server and an IPMI out-of-band node, and the case shell comprises a first side shell, a second side shell, a third side shell and a fourth side shell. The partition structure comprises a first partition plate, a second partition plate, a third partition plate and a fourth partition plate, four groups of power supply modules, a first slot, a second slot, a third slot and a fourth slot are wrapped at the inner end of the 2 + 2 redundant power supply module, and a golden finger server, a first surface and a second surface are arranged at the outer end of each group of power supply modules. The design of a server structure is achieved, on the basis that the node server is installed on a 19-inch rack, the deployment density of the node server can be improved, the occupied space of a case can be reduced, and meanwhile the heat dissipation efficiency can be improved.

Description

TECHNICAL FIELD

[0001] The utility model relates to server structure technical field, specifically, four node through type server. BACKGROUND

[0002] A server is a device that provides computing services. Since the server needs to respond to service requests and process them, in general, the server should have the ability to bear the service and guarantee the service. Due to the need of cloud computing, more and more computing and storage problems can no longer rely on a single server to solve, but need to use cluster servers to solve the computing and data storage needs in the technical environment.

[0003] U is a unit representing the external dimensions of a server. The so-called "1U refers to a server" refers to a server that meets the height of 44.45mm and the width of 19 inches.

[0004] In the prior art, during the use of the server structure, the existing cabinet occupies a large space and the heat dissipation is not comprehensive enough. Therefore, we improve it and propose a four-node through type server. UTILITY MODEL CONTENT

[0005] The utility model aims at the design of the current server structure and provides a four-node through type server cabinet and server equipment, which can reduce the occupied space of the cabinet, improve the heat dissipation system, and provide an IPMI modularization.

[0006] In order to achieve the above utility model purposes, the utility model provides the following technical scheme:

[0007] A four-node through type server is provided to improve the above problems.

[0008] The application is specifically as follows:

[0009] A four-node through type server includes a server equipment body and a server cabinet. The server cabinet includes a cabinet shell, a separation structure, a 2+2 redundant power module, a node server, and an IPMI out-of-band node. The cabinet shell includes a first side shell, a second side shell, a third side shell, and a fourth side shell. The separation structure includes a first separation plate, a second separation plate, a third separation plate, and a fourth separation plate. The inner end of the 2+2 redundant power module is wrapped with four groups of power modules, a first slot, a second slot, a third slot, and a fourth slot. The outer end of each group of power modules is provided with a golden finger server, a first surface, and a second surface.

[0010] As a preferred technical scheme of the application, the separation structure is fixedly connected with the cabinet shell, and the second separation plate and the third separation plate are vertically spaced and fixed in the cabinet shell.

[0011] As the preferred technical scheme of the present application, the first partition plate and the fourth partition plate are respectively fixed transversely between the fourth side shell and the third partition plate and between the third partition plate and the second side shell, and the first partition plate, the second partition plate, the third partition plate, the fourth partition plate, the second side shell and the fourth side shell are divided into four groups of plug-in spaces.

[0012] As the preferred technical scheme of the present application, the second partition plate, the third partition plate and the third side shell and the first side shell are provided with a cavity, the IPMI out-of-band node is wrapped in the cavity, and the rear end of the IPMI out-of-band node is electrically connected with the 2+2 redundant power supply module.

[0013] As the preferred technical scheme of the present application, four groups of plug-in spaces are respectively wrapped with a group of node servers, and the four groups of node servers are electrically connected with the 2+2 redundant power supply module, and the upper and lower of each power supply module are connected in the 2+2 redundant power supply module through the second surface and the first surface.

[0014] As the preferred technical scheme of the present application, the node server is provided from front to back with a heat dissipation air inlet, a PCIe card, a PCIe adapter, a mainboard, a PCIe slot, a graphics card, a bottom plate, a first mounting portion, a fan and a casing.

[0015] Compared with the prior art, the present application has the following beneficial effects:

[0016] In the scheme of the present application:

[0017] On the basis of meeting the installation in the 19-inch rack, the deployment density of the node server can be improved, the occupied space of the case can be reduced, and the heat dissipation efficiency can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 An exploded structural schematic view of a four-node through-type server is provided in the present application;

[0019] Figure 2 A structural schematic view of a case shell of a four-node through-type server is provided in the present application;

[0020] Figure 3 A structural schematic view of a server case of a four-node through-type server is provided in the present application;

[0021] Figure 4 A structural size schematic view of a case shell of a four-node through-type server is provided in the present application;

[0022] Figure 5 A structural schematic view of a power supply adapter plate and a power supply module of a four-node through-type server is provided in the present application;

[0023] Figure 6A structural schematic diagram of a node server of a four-node through-type server is provided in the present application.

[0024] Figure 7 A structural size schematic diagram of a node server of a four-node through-type server is provided in the present application.

[0025] Indicated in the figure are:

[0026] 1, server device; 10, server case; 100, case shell; 1001, first side shell; 1002, second side shell; 1003, third side shell; 1004, fourth side shell; 20, node server; 201, bottom plate; 209, first mounting portion; 204, mainboard; 203, display card; 206, PCIe card; 207, PCIe adapter cable; 208, heat dissipation air inlet; 202, PCIe slot; 210, fan; 200, partition structure; 2002, first partition plate; 2003, second partition plate; 2001, third partition plate; 2004, fourth partition plate; 30, 2+2 redundant power module; 300, first slot; 301, second slot; 302, third slot; 303, fourth slot; 31 power module; 5001, gold finger connector; 5002, first surface; 5003, second surface; 205, shell; 50, IPMI out-of-band node; 60, plug-in space. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical scheme and advantages of the embodiments of the present application more clear, the technical scheme in the embodiments of the present application will be described clearly and completely below in combination with the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments.

[0028] Therefore, the following detailed description of the embodiments of the present application is not intended to limit the scope of the claimed present application, but only represents some embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application. It should be noted that, in the case of no conflict, the embodiments in the present application and the features and technical schemes in the embodiments can be combined with each other.

[0029] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0030] In typical IDC (Internet Data Center) server rooms, standard 19-inch rack mount servers are commonly used. Therefore, when selecting server equipment 1, servers with mechanical dimensions conforming to the 19-inch industrial standard are usually chosen. Regarding the chassis height, 4U and 5U chassis are typically used to deploy high-density node servers 20, with 2 to 8 node servers 20 deployed within a single chassis.

[0031] like Figure 1 As shown, this embodiment proposes a four-node through-type server, a four-node through-type server device 1, including a server chassis 10, four node servers 20 and a 2+2 redundant power supply module 30.

[0032] The server chassis 10 has a partition structure 200 inside the chassis shell 100, which divides the internal space of the chassis shell 100 into four pull-out spaces 60 and a cavity, so as to accommodate the pull-out node server 20 and one IPMI out-of-band node 50.

[0033] The four node servers 20 can be installed in two separate insertion spaces 60 via a plug-in method.

[0034] The 2+2 redundant power module 30 also includes four power modules, and the 2+2 redundant power module 30 is electrically connected to the four node servers 20.

[0035] like Figure 2 As shown, as a preferred embodiment, based on the above method, this application embodiment further provides a four-node through-type server chassis 10, including chassis shell 100 and partition structure 200.

[0036] The chassis housing 100 includes a first side shell 1001, a second side shell 1002, a third side shell 1003 and a fourth side shell 1004 connected in sequence. The first side shell 1001 and the third side shell 1003 are arranged opposite to each other, and the second side shell 1002 and the fourth side shell 1004 are arranged opposite to each other.

[0037] The chassis shell 100 has an internal partition structure 200, which consists of two interconnected partition plates, a first partition plate 2002 and a second partition plate 2003. The first partition plate 2002 and the left side shell define one area, the second partition plate 2003 and the right side shell define another area, and a third area is formed between the first partition plate 2002 and the second partition plate 2003. That is, the left and right areas are the same size.

[0038] The third partition plate 2001 is arranged in the cabinet shell 100 and forms an inner partition layer with the first partition plate 2002, thereby subdividing the left side area into two layers. Meanwhile, the third partition plate 2001 is connected with the left side shell and the first partition plate 2002.

[0039] The fourth partition plate 2004 is also arranged in the cabinet shell 100 and forms an inner partition layer with the second partition plate 2003, thereby subdividing the right side area into two layers. The fourth partition plate 2004 is connected with the right side shell and the second partition plate 2003.

[0040] Part of the first side shell 1001, the first partition plate 2002, part of the third side shell 1003 and the second partition plate 2003 jointly enclose a containing space, and the tail of the IPMI out-of-band node 50 is placed in the containing space, and a power module of the node server 20 is placed between the first partition plate 2002 and the second partition plate 2003.

[0041] According to the embodiment of the present application, the partition structure 200 divides the internal space of the cabinet shell 100 into four plug-in spaces 60 in total, wherein the central space is used to place the IPMI out-of-band node 50 and the power module, and the other four spaces are used to facilitate the node server 20 to be installed in the plug-in space 60 in a plug-in manner. That is, the node server 20 of the embodiment of the present application is a blade node, and the server cabinet 10 of the embodiment of the present application is a cabinet structure that can install four node servers 20. The shell structure of the node server 20 can be changed according to the internal structure of the plug-in space 60, and therefore the arrangement positions of the internal components of the node server 20 can also be designed according to the shell structure of the node server 20. The embodiment of the present application can arrange all the four node servers 20 inside the front panel of the cabinet on the basis that the size of the cabinet shell 100 meets the standard of 19 inches in width, thereby reducing the occupied space of the cabinet and enabling the rack on which the server cabinet 10 is placed to accommodate more node servers 20. It can be understood that the cabinet shell 100 of the embodiment of the present application only limits the width of the cabinet shell 100 to meet the standard of 19 inches, and does not limit the height of the cabinet. Preferably, the height of the cabinet of the embodiment is 4U, and in other embodiments, the height of the cabinet can also be 4.5U or 5U.

[0042] As Figure 3As shown, in a preferred embodiment, based on the above method, the partition structure 200 further includes a first partition plate 2002 and a second partition plate 2003 at the rear of the server. The first partition plate 2002 is disposed inside the chassis shell 100 and extends from the head to the tail of the server, and the two mutually distant side edges of the first partition plate 2002 are respectively connected to the first side shell 1001 and the third side shell 1003; the second partition plate 2003 is disposed inside the chassis shell 100 and is coplanar with the second partition plate 2003, and the two mutually distant side edges of the second partition plate 2003 are respectively connected to the first side shell 1001 and the third side shell 1003.

[0043] In this embodiment, the positions of the first side shell 1001 and the third side shell 1003 can be changed according to the size of the power module. The first side shell 1001 and the third side shell 1003 are preferably arranged in parallel, the first partition plate 2002 is preferably arranged perpendicularly to the first side shell 1001, and the second partition plate 2003 is preferably arranged perpendicularly to the third side shell 1003.

[0044] It is understood that the power module can be directly connected to the first side shell 1001 and the third side shell 1003, or indirectly connected to the first side shell 1001 and the third side shell 1003. That is, the surfaces of the first side shell 1001 and the third side shell 1003 within the receiving space can also be provided with baffles. The baffles can prevent the power module from rubbing against the first side shell 1001 and the third side shell 1003 when it is placed in the receiving space, thereby preventing scratches on the outer surface of the power module. The baffles can also position the power module in the receiving space, so that the power module is stably placed in the receiving space.

[0045] like Figure 4 As shown, in a preferred embodiment, based on the above method, further, along the height direction of the server chassis 10, the height dimension of the second side shell 1002 and the fourth side shell 1004 is D1, the distance from the fourth partition plate 2004 to the first side shell 1001 and to the third side shell 1003 is D2, and the distance from the first partition plate 2002 to the first side shell 1001 is D3, where D2 is less than D1; along the width direction of the server chassis 10, the distance from the first partition plate 2002 to the second partition plate 2003 is D4, the distance from the second side shell 1002 to the fourth side shell 1004 is D5, and the width dimension of the first side shell 1001 and the third side shell 1003 is also D5.

[0046] This application embodiment defines the relationship between D2 and all the slotting spaces 60, so that the partition structure 200 divides the chassis shell 100 into four slotting spaces 60 and two slotting spaces 60D4. This allows node servers 20 with the same external structure to be placed in the slotting spaces 60, reducing the manufacturing cost of node servers 20.

[0047] In some embodiments, D1 is in the range of 177mm, D2 and D3 are in the range of 87mm, and D5 is in the range of 435mm-445mm.

[0048] The height and width of the case shell 100 are limited in this embodiment, so that the parameters of the case shell 100 meet the installation parameters of a 19-inch rack while maintaining the height of the overall case at 4U. Preferably, the case height of this embodiment is 4U, and the setting position of the partition structure 200 is limited in parameters, so that the case is a 4U4-node server case 10 while also having a high case space utilization rate. Considering the convenience of equipment installation and maintenance, the product weight of the server case 10 of this embodiment is lighter and occupies less space, so that more server cases 10 can be deployed in the same rack, improving maintenance convenience and reducing maintenance costs.

[0049] On the other hand, referring to Figure 1 This embodiment provides a server device 1, which includes a server case 10, four node servers 20, an IPMI out-of-band node 50, and a 2+2 redundant power supply module 30.

[0050] The server case 10 is provided with a partition structure 200 in the case shell 100, which divides the internal space of the case shell 100 into four plug-in spaces 60 and two accommodation spaces D4, so as to place the plug-in installed node servers 20, IPMI out-of-band nodes 50, and power supply modules.

[0051] The four node servers 20, IPMI out-of-band nodes 50, and power supply modules can be respectively installed in the two plug-in spaces 60 and the accommodation spaces D4 via the plug-in mode. The four node servers 20 are installed in the plug-in mode, so that each node server 20 can work independently.

[0052] 2+2 redundant power module 30 is arranged at the tail of accommodating space D4, and the 2+2 redundant power module 30 is electrically connected with the two node servers 20. The 2+2 redundant power module is a power supply that can be used on the node server 20, which is composed of four identical power supplies, and the four power supplies collectively realize power supply for the node server 20, which can meet the power supply and power consumption demand of the node server 20. The 2+2 redundant power module 30 can control the load balancing of the power supply, when one of the power supplies fails, the 2+2 redundant power module 30 enables another power supply to take over the power supply work of the failed power supply, avoiding the situation that the node server 20 stops working due to the failure of the power supply; after replacing the failed power supply, the 2+2 redundant power module 30 enables the two stable working power supplies to work cooperatively. The arrangement of the 2+2 redundant power module 30 provides stable current for the operation of the node server 20, and ensures the stability of the server system.

[0053] On the basis that the size of the case shell 100 meets the 19-inch width, the two node servers 20 can be arranged in the case in a cross manner, reducing the occupied space of the case, so that the rack in which the server case 10 is placed can accommodate more node servers 20. Further, the case shell 100 of the embodiment of the application only limits that the width of the case shell 100 should meet the 19-inch standard, and does not limit the height of the case, preferably, the height of the case of the embodiment is 4U, and in other embodiments, the height of the case can also be 5U, 6U, etc.

[0054] Based on the embodiment of the application, the server device 1 of the embodiment of the application can accommodate two node servers 20 and a 2+2 redundant power module 30. The independent operation of the two node servers 20 and the arrangement of the 2+2 redundant power module 30 enable the server device 1 of the embodiment to meet the 19-inch rack installation parameters while ensuring the stability of the operation of the server device 1.

[0055] As shown in Figure 5 As a preferred embodiment, on the basis of the above-mentioned mode, further, the server device 1 further comprises a power adapter plate. The power module comprises a plate body and a first slot 300, a second slot 301, a third slot 302 and a fourth slot 303 arranged on the plate body and arranged opposite on the plate body, and the slots are located in the four plug-in spaces 60, so that the four node servers 20 are respectively in hot plug with the four slots.

[0056] 2+2 redundant power module 30 also includes four power modules, one end of each power module is provided with a hot plug gold finger connector 31 power module; 5001, each power module has a first surface 5002 and a second surface 5003 disposed opposite to each other, the gold finger connector 31 power module; 5001 is closer to the second surface 5003 than the first surface 5002, two power modules are installed in the server rear end of the accommodating space D4, four first surfaces 5002 are away from each other and four second surfaces 5003 are arranged opposite to each other, so as to arrange the slots on the two opposite surfaces of the board.

[0057] It can be understood that the power adapter board is provided for the power supply of the power module to the node server 20. On this basis, the embodiment of the application arranges four first slots on the two opposite surfaces of the power adapter board, which can be hot plugged with the node server 20 in the plug-in space 60, and the four slots can be hot plugged with the gold finger connector 31 power module; 5001 of the power module in the second accommodating space. This design makes the four power modules can be electrically connected with the power adapter board, that is, using a power adapter board can realize the connection between two node servers 20 and two power modules, thereby simplifying the complexity of circuit design.

[0058] Further, since the internal wiring layout of the power adapter board is not sufficient to support the opposite arrangement of the slots at the same height on the power adapter board, that is, the projection of the slot on the power adapter board cannot overlap. To solve this problem, the embodiment limits the gold finger connector 31 power module; 5001 is closer to the second surface 5003 than the first surface 5002, that is, the embodiment arranges the four power modules in a central symmetric manner, which avoids the projection of the first slot 300 and the second slot 301 on the power adapter board from overlapping, so that a power adapter board can realize the connection between two node servers 20 and two power modules.

[0059] The hot plug refers to the behavior of the node server 20, the IPMI out-of-band node 50 and the power module being inserted from the outside or being pulled out from the system without turning off the power supply of the system, which does not affect the normal operation of the system. In the embodiment, the four node servers 20 can be hot-plugged with the four slots respectively, so that when there is a damaged node server 20 in the system during startup, pulling out the damaged node server 20 will not affect the normal operation of the system. In addition, one end of each power module is provided with a gold finger connector 31 power module; 5001, the hot plug of the gold finger connector 31 power module; 5001 can enable the power module to be replaced at any time while maintaining the voltage stability of the entire power supply system, thereby ensuring the normal operation of other power modules in the modular power supply system. In the embodiment, the node servers 20 are cross-set and the power modules are arranged in a mirror structure, which facilitates the maintenance of the node servers 20 and the timely replacement of the power modules.

[0060] As shown in Figure 6 As a preferred embodiment, on the basis of the above-mentioned mode, each node server 20 further includes a housing 205, and the housing 205 includes a bottom plate 201 and a first mounting portion 209.

[0061] In the height direction of the server case 10, the slot of the first mounting portion 209 is lower than the height of the fan 210, so that the fan 210 is embedded in the exhaust port at the tail of the insertion space 60; in the length direction of the server case 10, the gold finger at the tail of the server power supply is smaller than the length of the housing 205, so that the end of the housing 205 in the insertion direction of the node server 20 forms an avoidance space for avoiding the space protruding from the fan 210.

[0062] In the embodiment, each node server 20 further includes a mainboard 204, a graphics card 203 and at least one PCIe card 206.

[0063] The size of the mainboard 204 is ITX specification, and the mainboard 204 has a CPU installation area and a PCIe slot 202 area, and the CPU installation area and the PCIe slot 202 area are respectively at both ends of the housing 205. Preferably, since the mainboard 204 of ITX specification has a relatively small volume and good scalability, the size specification of the mainboard 204 of the embodiment is the ITX specification with good compatibility, and the use of the mainboard 204 of ITX specification can also reduce the maintenance cost of the customer.

[0064] When the height of the server case 10 is 4U, the number of graphics cards 203 is at most 4, the graphics cards 203 are arranged in the tail end, and the graphics cards 203 are electrically connected with the mainboard 204.

[0065] PCIe card 206 is at least a network card. PCIe card 206 is located on the front panel and is electrically connected to the motherboard 204.

[0066] Understandably, in some embodiments, a fan 210 is also provided at the rear of the graphics card 203. The fan 210 provides auxiliary cooling for the node server 20. When the node server 20 is working, the heat generated inside the chassis 205 is quickly dissipated by the fan 210. The fan 210 also dissipates the heat generated by the motherboard 204 during operation to the area of ​​the graphics card 203, thereby achieving a cooling effect and meeting the requirements of the node server 20 for high-power, long-term operation. In addition, since the four-node server 20 is independently installed inside the chassis 100, the airflow channels inside the chassis 205 of the four-node server 20 are independent and the airflow channel positions are reasonably designed. Therefore, the operation of the four-node server 20 is not affected by heat dissipation, which is conducive to the stability of the system's high-power operation.

[0067] In other embodiments, when the height of the server chassis 10 is greater than 4U, a larger CPU cooler and more graphics cards 203 can be used inside the server chassis 10. Therefore, there is no constraint on the number of graphics cards 203.

[0068] In some embodiments, each node server 20 further includes multiple PCIe adapter cables 207, and the graphics card 203 and PCIe card 206 are connected to the PCIe slot 202 in the PCIe slot 202 area via the PCIe adapter cables 207.

[0069] This embodiment connects the graphics card 203 and PCIe card 206 to the PCIe slot 202, allowing devices such as a 10 Gigabit Ethernet card to be installed in the PCIe slot 202, thus meeting the customer's network transmission needs. Using a PCIe adapter cable 207 to connect the graphics card 203 and other PCIe devices avoids direct connection between the PCIe devices and the motherboard 204, reducing the installation height of the ITX motherboard 204 within the PCIe slot 202 area.

[0070] like Figure 7 As shown, as a preferred embodiment, based on the above method, the size range of the chassis 205 of the node server 20 is further limited.

[0071] Along the height direction of the housing 205, the height dimension is D10, and the value of D10 is D10≥87mm;

[0072] Along the width direction of the housing 205, the width dimension is D11, and the value of D11 ranges from 179mm;

[0073] The length dimension D12 is in the range of 698mm along the length direction of the shell 205.

[0074] In combination with the size structure of the server case 10, the shell 205 can have a thinner structure by limiting the size of the shell 205. The four-node server 20 can be inserted into the insertion space 60, improving the space utilization of the server case 10. The structural size of the shell 205 meets the installation conditions of the 19-inch rack, so that more node servers 20 can be deployed in the same rack.

[0075] The above embodiments are only used to illustrate the technical solutions described in the present application and are not limited to the technical solutions described in the present application. Although the present application has been described in detail with reference to the above embodiments, the present application is not limited to the above specific embodiments. Therefore, any modification or equivalent replacement of the present application; all technical solutions and improvements that do not deviate from the spirit and scope of the present application are included in the scope of the claims of the present application.

Claims

1. A four-node through-type server, comprising a server device (1) body and a server chassis (10), characterized in that, The server chassis (10) includes a chassis shell (100), a partition structure (200), a 2+2 redundant power supply module (30), a node server (20), and an IPMI out-of-band node (50). The chassis shell (100) includes a first side shell (1001), a second side shell (1002), a third side shell (1003), and a fourth side shell (1004). The partition structure (200) includes a first partition plate (2002), a second partition plate (2003), a third partition plate (2001), and a fourth partition plate (2004). The inner end of the 2+2 redundant power supply module (30) is wrapped with four sets of power modules (31), a first slot (300), a second slot (301), a third slot (302), and a fourth slot (303). The outer end of each set of power modules (31) is provided with a gold finger server (5001), a first surface (5002), and a second surface (5003).

2. A four-node interconnected server according to claim 1, characterized in that, The partition structure (200) is fixedly connected to the chassis shell (100), and the second partition plate (2003) and the third partition plate (2001) are fixedly fixed inside the chassis shell (100) at vertical intervals.

3. A four-node interconnected server according to claim 2, characterized in that, The first partition plate (2002) and the fourth partition plate (2004) are respectively horizontally fixed between the fourth side shell (1004) and the third partition plate (2001), and between the third partition plate (2001) and the second side shell (1002). The first partition plate (2002), the second partition plate (2003), the third partition plate (2001), the fourth partition plate (2004), the second side shell (1002), and the fourth side shell (1004) are separated into four sets of insertion and removal spaces (60).

4. A four-node interconnected server according to claim 3, characterized in that, A cavity is provided between the second partition plate (2003), the third partition plate (2001), the third side shell (1003), and the first side shell (1001). The IPMI out-of-band node (50) is enclosed in the cavity, and the rear end of the IPMI out-of-band node (50) is electrically connected to the 2+2 redundant power supply module (30).

5. A four-node interconnected server according to claim 4, characterized in that, Each of the four sets of insertion spaces (60) encloses a set of node servers (20). The four node servers (20) are electrically connected to the 2+2 redundant power supply module (30). The upper and lower parts of each set of power supply modules (31) are connected to the 2+2 redundant power supply module (30) through the second surface (5003) and the first surface (5002).

6. A four-node interconnected server according to claim 5, characterized in that, The node server (20) is provided with a heat dissipation air inlet (208), a PCIe card (206), a PCIe adapter cable (207), a motherboard (204), a PCIe slot (202), a graphics card (203), a base plate (201), a first mounting part (209), a fan (210), and a chassis (205) from front to back.

Images