CN210038641U - a server architecture - Google Patents

Abstract

The application discloses a server architecture, including an external architecture, the external architecture is a cuboid, the external architecture is highly compatible with a double-layer tray structure, and the system air duct of the external architecture is a forward-backward-out type; a first area is used for loading functional modules, The front IO wiring of the functional modules in the first area; the second area is used to load the service components and cooling systems corresponding to the functional modules, the second area is adjacent to the rear wiring board, and the bottom of the second area is preset with A plurality of screw holes for fixing the motherboard; the power supply area includes a fixing frame for fixing the power supply, which is installed at one corner of the rear of the external frame, and the outlet end of the fixing frame is along the long side of the external frame; the first area and In the second area, the installed functional modules and service components can be changed according to actual application scenarios. The partitions are obvious, the framework is concise and clear, and the scalability is strong.

Description

a server architecture

technical field

The utility model relates to the field of computers, in particular to a server architecture.

Background technique

In the era of cloud computing, the marketization of servers according to their functions is becoming more and more obvious, including CVM servers (CVM, CVM Cloud Virtual Machine, cloud virtual machine) focusing on high-performance computing, and PCIe-based storage servers focusing on Hadoop (big data) hot storage. , a heterogeneous acceleration server focusing on AI deep computing.

However, the system architecture requirements of each individual field are completely different. Currently, servers in the market for different application scenarios have separate architecture designs. From the perspective of server development types, the development workload is increased.

Therefore, in order to save development costs and maintenance costs, it is necessary to develop a server architecture with high versatility.

Utility model content

In view of this, the purpose of the present invention is to provide a server architecture with high versatility and reduced research and development costs. Its specific plan is as follows:

A server architecture, comprising an external architecture, the external architecture is divided into a first area, a second area and a power supply area; the first area, the second area and the power supply area are adjacent in sequence;

The external structure is a rectangular parallelepiped, the external structure is highly compatible with a double-layer tray structure, and the system air duct of the external structure is a forward-backward-out type;

The second area and the power supply area are arranged at the rear of the outer frame, and the first area and the second area are separated by a heat dissipation fixing baffle for fixing a heat dissipation device;

The first area is used for loading functional modules, and the functional modules in the first area are pre-IO wiring;

The second area is used to load the service components and the cooling system corresponding to the functional modules, the second area is adjacent to the rear wiring board, and the bottom of the second area is preset with a plurality of screw hole;

The power supply area includes a fixing frame for fixing the power supply, which is installed at one corner of the rear of the outer frame, and the outlet end of the fixing frame is along the long side of the outer frame;

Among them, the air inlet of the external structure is the front, and the air outlet is the rear.

Optionally, three-section detachable slide rails may be provided on both sides of the first area, and upper and lower double-layer trays may be set on the three-section detachable slide rails for carrying functional modules.

Optionally, on the upper and lower double-layer trays in the first area, each layer of trays is placed in a 3x4 form with a hard disk array including 12 hard disks;

The second area is placed with a motherboard for providing computing power to the hard disk, and a heat dissipation system fixed by a heat dissipation fixing baffle is installed between the motherboard in the second area and the first area;

The IOs of the hard disks in the first area are in front, the system IOs in the second area are rear, and a hot-plug interface is provided in the first area to realize hot-plugging of the hard disk arrays of each layer of trays.

Optionally, a two-way computing node is installed in the first area, and one computing node of the two-way computing node is installed on each layer of the upper and lower double-layer trays;

The IO expansion device is installed in the second area, and an intelligent network card slot is reserved, and the first area is adjacent to the heat dissipation system fixed by the heat dissipation fixing baffle in the second area.

Optionally, an integrated GPU board including 8 GPUs is installed in the first area;

The second area is equipped with an IO expansion device for expanding the integrated GPU board IO;

The first area is adjacent to the heat dissipation system fixed by the heat dissipation fixing baffle in the second area.

Optionally, the cooling system includes five fans arranged side by side.

Optionally, the heat dissipation system and the power supply area share a board, which is used as a power board for power supply.

In the present invention, the server architecture includes an external architecture, and the external architecture is divided into a first area, a second area, and a power supply area; the first area, the second area, and the power supply area are adjacent in sequence; the external architecture is a cuboid, and the height of the external architecture is Compatible with the double-layer tray structure, the system air ducts of the external structure are front and rear; the second area and the power supply area are arranged at the rear of the external structure, and the first area and the second area are fixed by the heat dissipation device for fixing The baffle is separated; the first area is used to load functional modules, and the functional modules in the first area are wired in front of the IO; the second area is used to load the service components and cooling systems corresponding to the functional modules, and the second area is connected to the rear The wiring board is adjacent to each other, and the bottom of the second area is preset with a plurality of screw holes for fixing the main board; the power supply area includes a fixing frame for fixing the power supply, which is installed on one side corner of the rear of the external structure, and the cable outlet of the fixing frame The end is along the long side of the outer frame; wherein, the air inlet of the outer frame is the front part, and the air outlet is the rear part.

The utility model builds a server structure with high flexibility, the first area and the second area can change the installed functional modules and service components according to the actual application scene, the partition is obvious, the framework is concise and clear, and the expansibility is strong, so that a This server architecture can be compatible with a variety of server internal structure designs. At the same time, it adopts a cuboid external structure and adopts a forward-backward-out cooling air duct, which is convenient for multiple servers to be stacked on each other to save space and ensure good cooling capacity.

Description of drawings

In order to more clearly illustrate the embodiments of the present utility model or the technical solutions in the prior art, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description It is only an embodiment of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained according to the provided drawings without creative efforts.

1 is a schematic structural diagram of a server architecture according to an embodiment of the present invention;

2 is a schematic structural diagram of a storage server architecture according to an embodiment of the present invention;

3 is a schematic diagram of the topology of a storage server system according to an embodiment of the present invention;

4 is a schematic structural diagram of a computing server architecture according to an embodiment of the present invention;

5 is a schematic structural diagram of a GPU server architecture according to an embodiment of the present invention;

FIG. 6 is a schematic structural diagram of a fan according to an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present utility model will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present utility model. Obviously, the described embodiments are only a part of the embodiments of the present utility model, rather than all the implementations. example. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present invention.

The embodiment of the present utility model discloses a server architecture, as shown in FIG. 1, including an external architecture, the external architecture is divided into a first area A, a second area B and a power supply area C; the first area A, the second area B and The power supply areas C are adjacent in sequence.

Specifically, the external structure is a rectangular parallelepiped with the middle part running through it for placing the corresponding equipment. The external structure is highly compatible with the double-layer tray structure, which can ensure that there is enough space for expansion, so that the unit server structure has sufficient working capacity.

Among them, the use of a cuboid structure facilitates stacking of multiple server architectures, saves space, and facilitates planning.

Specifically, in order to ensure the high efficiency of heat dissipation, the system air ducts of the external structure are of the forward and backward type, and no additional heat dissipation ports are added around the periphery to ensure that the air ducts will not be scattered.

Specifically, the air inlet of the outer frame is set as the front part, the air outlet is set as the rear part, the second area B and the power supply area C are set at the rear part of the outer frame, and the first area A and the second area B pass through the The heat-dissipating fixing baffles that fix the heat-dissipating device are separated.

It can be understood that the power supply of the power supply area C will provide power to the devices in the first area A and the second area B.

Specifically, the first area A is used to load functional modules. Since the heat dissipation adopts the front-to-rear-out type, the temperature of the rear of the server in the equipment room is relatively high. The IO wiring of the module may need to be rearranged relatively frequently. Therefore, if the IO wiring is rearranged, the maintenance personnel will have a poor working environment when the IO wiring is rearranged. The IO wirings of the functional modules are arranged in front of the maintenance personnel. When the maintenance personnel arrange the IO wirings of the functional modules in the first area A, the temperature of the working environment of the maintenance personnel can be at a relatively good temperature.

Specifically, the second area B is used to load the service components and cooling systems corresponding to the functional modules, the second area B is adjacent to the rear wiring board, and the bottom of the second area B is preset with a plurality of screw hole 2.

Specifically, the service component may include an IO expansion device for extending the IO interface for the functional module, and may also be a mainboard for computing that stores data in the computing functional module; The disassembled heat-dissipating fixed partitions are separated, so the space between the first area A and the second area B is variable, and a dual-node layout can be formed between the first area A and the second area B, so that the server becomes a dual-node server.

Specifically, the power supply area C includes a fixing frame for fixing the power supply, which is installed at one side corner of the rear of the outer frame, and the outlet end of the fixing frame is along the long side of the outer frame.

It can be understood that the power supply in the power supply area C has its own cooling fan. In order to be consistent with the heat dissipation direction of the air duct, the outlet end of the fixing frame is along the long side of the external structure, so that the heat dissipation direction of the power supply is the same as forward and backward. The emerging end of the power supply is also conveniently connected to the first area A through the second area B along the long side of the external structure, which reduces the length of the trace and facilitates power supply.

It can be seen that the embodiment of the present invention builds a server architecture with high flexibility. The functional modules and service components installed in the first area A and the second area B can be changed according to actual application scenarios. The partitions are obvious, and the framework is concise and clear. Strong scalability, so that one server architecture can be compatible with a variety of server internal design, at the same time, the use of a rectangular external structure, and the use of front and rear cooling air ducts, which is convenient for multiple servers to be stacked on each other to save space and ensure good heat dissipation. .

The embodiment of the present utility model discloses a specific server architecture. Compared with the previous embodiment, this embodiment further describes and optimizes the technical solution. specific:

Specifically, as shown in FIG. 1 , three-section detachable slide rails 1 can be provided on both sides of the first area A, and upper and lower double-layer trays can be set on the three-section detachable slide rail 1 for carrying functional modules.

Specifically, the upper and lower double-layer trays on the three-segment detachable slide rail 1 can carry two sets of functional modules for pulling movement, which is convenient for maintenance.

Specifically, as shown in FIG. 2 , the server architecture can be provided with storage devices to form a storage architecture. On the basis of the upper and lower double-layer trays in the first area A, each layer of trays is placed in a 3×4 form of a hard disk array 4 including 12 hard disks. ;

The second area B is placed with a motherboard for providing computing power to the hard disk, and a heat dissipation system fixed by the heat dissipation fixing baffle 3 is installed between the motherboard in the second area B and the first area A;

The IOs of the hard disks in the first area A are at the front, and the system IOs in the second area B are located at the rear. The first area A is provided with a hot-plug interface to realize hot-plugging of the hard disk array of each layer of trays.

The hard disks can be SAS/SATA hard disks, the motherboards in the second area B can be dual-channel low-power motherboards that provide computing capabilities, and the power supply PDBs in the power supply area C are shared by all devices.

Specifically, if the heat dissipation system is set at the rear of the external structure, the heat dissipation system is too far from the air inlet, which will cause the wind speed to slow down and the heat dissipation effect to decrease. The first area A is adjacent, which is equivalent to being set in the middle to ensure a certain wind speed and heat dissipation performance.

Further, a schematic diagram of the internal wiring between the first area A and the second area B in the storage architecture is shown in Figure 3. The first area A includes 12 3.5-inch hard drives, the second area B includes 2.5-inch hard drives, and each of the two has 24 hard drives. The CPU0 and CPU1 of the PCIe interface, the integrated PDB and FB, five fans installed on the integrated PDB and FB, the MEZZ Card, and the HBA Card, and the power supply area C includes two CRPS power supplies stacked on top of each other.

In another application scenario of the embodiment of the present invention, the server architecture may be set as a computing architecture. Referring to FIG. 4 , two-way computing nodes are installed in the first area A, and two upper and lower trays are installed on each tray. A computing node 6 of the road computing node;

The IO expansion device is installed in the second area B, and an intelligent network card slot is reserved. The first area A is adjacent to the heat dissipation system fixed by the heat dissipation fixing baffle in the second area B.

Specifically, the IO expansion device in the second area B can expand the IO interface of each computing node in the first area A. At the same time, an intelligent network card can also be installed in the second area B, so that the dual-channel computing nodes can be connected through the intelligent network card. to the network.

In another application scenario of the embodiment of the present invention, the server architecture may be set as a GPU architecture. Referring to FIG. 5 , an integrated GPU board 5 including 8 GPUs is installed in the first area A;

The second area B is installed with an IO expansion device for expanding the IO of the integrated GPU board;

The first area A is adjacent to the heat dissipation system fixed by the heat dissipation fixing baffle in the second area B.

Further, as shown in FIG. 6 , the above-mentioned cooling system may include 5 fans arranged side by side, and the cooling system may share a board card with the power supply area C, which is used as a power supply board for power supply, shortening the connection between the first area A and the first area C. The wiring distance of the power supply lines for the power supply in the second area B, while ensuring the reliability of the power supply of the cooling system.

The above description of the disclosed embodiments enables any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (7)

1. A server architecture is characterized by comprising an external architecture, wherein the external architecture is divided into a first area, a second area and a power supply area; the first area, the second area and the power supply area are adjacent in sequence;

the external framework is a cuboid, the height of the external framework is compatible with a double-layer tray structure, and a system air duct of the external framework is forward and backward;

the second area and the power supply area are arranged at the rear part of the external framework, and the first area and the second area are separated by a heat dissipation fixing baffle plate for fixing a heat dissipation device;

the first area is used for loading functional modules, and IO wiring is arranged in front of the functional modules in the first area;

the second area is used for loading service components and a heat dissipation system corresponding to the functional module, the second area is adjacent to the rear wiring board, and a plurality of screw holes for fixing the mainboard are preset at the bottom of the second area;

the power supply area comprises a fixing frame for fixing a power supply, the fixing frame is arranged at one side corner of the rear part of the external framework, and a wire outlet end of the fixing frame is arranged along the long edge of the external framework;

wherein, the air inlet of the external framework is the front part, and the air outlet is the rear part.

2. The server architecture according to claim 1, wherein three detachable sliding rails can be disposed on two sides of the first region, and an upper tray and a lower tray can be disposed on the three detachable sliding rails for carrying the functional module.

3. The server architecture of claim 2, wherein each tray of the upper and lower double-layered trays of the first area is provided with a hard disk array comprising 12 hard disks in a format of 3x 4;

a mainboard for providing computing power for a hard disk is placed in the second area, and a heat dissipation system fixed through a heat dissipation fixing baffle is installed between the mainboard in the second area and the first area;

the IO of hard disk in the first region is preposed, the system IO of the second region is postpositioned, and a hot plug interface is arranged in the first region to realize the hot plug of the hard disk array of each layer of tray.

4. The server architecture according to claim 2, wherein two-way computing nodes are installed in the first area, and one computing node of the two-way computing nodes is installed on each layer of trays on the upper and lower layers of trays;

IO expansion equipment is installed in the second area, an intelligent network card slot is reserved, and the first area is adjacent to a heat dissipation system fixed in the second area through a heat dissipation fixing baffle plate.

5. The server architecture of claim 1, wherein the first region is populated with an all-in-one GPU board comprising 8 GPUs;

IO expansion equipment for expanding IO of the integrated GPU board is installed in the second area;

the first area is adjacent to the heat dissipation system fixed in the second area through the heat dissipation fixing baffle.

6. The server architecture according to any one of claims 1 to 5, wherein the heat dissipation system comprises 5 fans arranged side by side.

7. The server architecture according to any one of claims 1 to 5, wherein the heat dissipation system and the power supply area share a board as a power board for supplying power.

Images