Disclosure of Invention
In view of this, the present invention provides a server and a server cabinet, so as to solve the problems of large number of power modules to be supported in the cabinet of the server, and large space and cost occupation in the related art.
In a first aspect, the present invention provides a server comprising:
a chassis;
The at least one double-input power supply module is arranged in the case, and two input ends of the double-input power supply module are respectively connected with one power supply;
The server node is arranged in the chassis, the output end of the double-input power supply module is used for connecting the server node, the power of the double-input power supply module is V, the number of the double-input power supply modules is n, the total power consumption of the server is W, and W is less than or equal to (n-1) V is less than 2W.
The beneficial effects are that: according to the server disclosed by the embodiment of the invention, the server nodes are powered by the double-input power supply modules, and the two input ends of each double-input power supply module are respectively used for being connected with one power supply. Therefore, when any double-input power supply module fails, the other double-input power supply modules can still keep normal power supply, so that the server node can work normally, the total power of the power supply modules is not required to be configured to be twice the power consumption of the server, the number of the power supply modules is reduced, the cost is reduced, and the space is saved.
In an alternative embodiment, (n-2) V < W.
The beneficial effects are that: when the number of the double-input power supply modules is configured according to the proportional relation, whether the input power supply fails or a single double-input power supply module fails can be ensured, the server can keep normal work, the number of the double-input power supply modules required to be used is saved to the greatest extent, and the cost of the server and the space occupied by the double-input power supply modules are reduced.
Assuming that the total power of 5 power modules can satisfy the power supply of the server, when a single input power module is used, it is necessary to configure to n+n redundancy power supply, and thus at least 10 power modules need to be provided. The server in the embodiment of the application only needs to use 6 double-input power supply modules. The cost of four power supply modules is saved, the space occupied by the power supply modules is reduced, and the density of the server is improved.
In an alternative embodiment, the server further comprises:
the power panel, the dual input power module connects in parallel each other, and the output is connected in the power panel, power panel and a plurality of server node communication connection.
The beneficial effects are that: the power supply reliability and stability of the system can be improved by adopting a mode of parallel connection of the double-input power supply modules, and the system can be suitable for occasions with extremely high requirements on power supply stability, such as data centers, enterprise-level servers and the like. The power panel is used as a core power supply component of the server system and is responsible for distributing and regulating power to the server nodes.
In an optional embodiment, an avoidance port is formed at a front window of the chassis, the power panel is arranged in the chassis, a plurality of first interfaces are arranged on the power panel, and the dual-input power module can penetrate through the avoidance port and be connected with the first interfaces in an inserting mode.
The beneficial effects are that:
Through such setting, can realize that dual input power module can take out alone and maintain, can be convenient for dual input power module's maintenance and maintenance. In addition, the design of the case front window avoiding opening reserves a certain space at the case front window, can be used for installing and maintaining internal hardware equipment, and simultaneously provides convenience for heat dissipation. The design also helps to improve the expansibility and future upgrade possibility of the server, so that the user can access partial internal hardware, such as memory strips, expansion cards and the like, without opening the side plate of the case.
In an alternative embodiment, a support structure is disposed in the chassis, and the server further includes:
The first connector is connected to one side of the server node, which is used for being inserted into the avoidance port;
The second connector is connected with the supporting structure and is arranged opposite to the first connector, and the server node can pass through the avoidance port and enables the first connector to be spliced with the second connector;
The centralized management card is arranged at the rear window of the case, and the second connector is connected with the centralized management card.
The beneficial effects are that: through such setting for the server node can take out alone and maintain, can be convenient for the operator overhauls or changes the server.
In an alternative embodiment, the server nodes are arranged in parallel in the width direction of the chassis, the width direction of the server nodes is parallel to the height direction of the chassis, and the dual-input power supply modules are arranged below the server nodes and are arranged in sequence in the width direction of the chassis.
The beneficial effects are that:
The server in the embodiment of the application uses the double-input power module to replace the traditional single-input power module, and the space occupied by the double-input power module is larger than that occupied by the conventional single-input power module because of the addition of an ATS power supply and the addition of an input power supply interface, but the number of the power modules required to be used is correspondingly reduced.
On the basis, the application breaks the limit of the conventional chassis architecture to the width of the server node through reasonable arrangement of the chassis internal architecture, and allows the server node to be set wider, thereby increasing the width of the main board, so that the main board can support more central processing units and memory strips.
In an alternative embodiment, the height of the chassis is h, 300 mm.ltoreq.h.ltoreq.310 mm, the width of the server node is L1, 250 mm.ltoreq.L1.ltoreq.265 mm.
The beneficial effects are that: when the height of the chassis is set to be within the above range, the width of the motherboard of the server node can be configured to be 250mm, and at the same time, each motherboard can support two 600w power cpus and 32 memory banks at maximum, so that the expansibility of the server is greatly improved, the server can provide more powerful computing capacity and rapid data processing speed, and the server can be suitable for a scene requiring processing of a large number of parallel tasks.
Meanwhile, the space below the chassis of the server node can be used for setting the double-input power supply module, and 6 double-input power supply modules with the power of 3300w are calculated to be enough to meet the power supply requirement of the server in a full distribution state, so that the server in the embodiment of the application can simplify the number of the power supply modules on the premise of keeping the power supply capacity and the redundancy capacity of the power supply unchanged, reduce the cost of the power supply modules and support the latest mainboard with the width of 250 mm.
In addition, the server provided by the embodiment of the application can be just suitable for a 19-inch universal cabinet, and the universal model of the server provided by the embodiment of the application does not need a custom cabinet, so that the use cost of the server is greatly reduced.
In an alternative embodiment, the width of the cabinet is 440 mm.ltoreq.L2.ltoreq.450 mm;
The number of server nodes is 8, and the server nodes are uniformly arranged in the case along the width direction of the case.
The beneficial effects are that: when the width of the chassis is in the above range, the dimension of each server node along the thickness direction of the node can be set to be L3, 50mm is less than or equal to L3 and less than or equal to 55mm, in the related art, the dimension of the server node in the high-density server along the thickness direction of the server node is generally 1U (about 44.45 mm), the server node in the embodiment of the application can ensure that a certain gap exists between two adjacent server nodes while ensuring the higher density of the server, and the heat dissipation effect is superior to that of the high-density server in the related art.
In an alternative embodiment, the server further comprises: the input interface is arranged at the bottom of the rear window of the chassis and is opposite to the double-input power supply module, the input end of the input interface is used for being connected with at least two paths of power supplies, and the output end of the input interface is used for being connected with the double-input power supply module.
The beneficial effects are that: the input interface is arranged in the lowest 1U space at the rear side of the case, and since the power input interface is a very important part in the whole system, the arrangement of the input interface at the lowest 1U at the rear side of the case can help to ensure the stability and reliability of the input interface.
In an alternative embodiment, the input interface comprises:
the input interfaces are arranged at intervals along the width direction of the chassis, each pair of input interfaces comprises two input interfaces, each input interface is used for being connected with one input power supply, and each pair of input interfaces is used for being connected with two input interfaces of one double-input power supply module.
The beneficial effects are that: this design ensures that each dual input power module can draw power from two separate PDUs, thereby greatly reducing the risk of system interruption due to single power supply problems.
In an alternative embodiment, the server further comprises:
and the heat radiation module is arranged at the rear window of the chassis and is used for blowing air towards the server node.
The beneficial effects are that: high performance computing devices and servers can generate a significant amount of heat during operation, which can cause the device to overheat if not timely expelled, affecting computing efficiency and hardware life. The centralized fan heat dissipation design can solve the problem in a targeted way, and the fan arranged in the upper area of the rear side can forcedly discharge internal hot air out of the chassis, so that the normal operation temperature of the equipment is maintained.
In an alternative embodiment, the heat dissipation modules are multiple groups and are sequentially arranged along the height direction, and each heat dissipation module comprises a plurality of heat dissipation fans sequentially arranged along the width direction of the chassis.
The beneficial effects are that: by so doing, it is possible to provide a compact structure. The heat radiation fans can be uniformly distributed along the width direction of the cabinet to ensure that heat can be uniformly conducted from the equipment to the air and discharged outside the cabinet through the fans. In addition, the cooling fan covers the position of the machine box corresponding to the server node along the vertical direction, and can effectively dissipate heat of the server node.
In an alternative embodiment, the server further comprises an expansion card detachably arranged at the rear window of the chassis, and the expansion card is in communication connection with the server node.
The beneficial effects are that: the expansion card can be used for increasing expansion equipment of original equipment functions, and provides additional expansion capability, thereby being beneficial to improving the performance and efficiency of the data center.
In an alternative embodiment, the expansion card includes a bus module;
the bus module is inserted into the heat dissipation module, and at least one heat dissipation fan is arranged on two sides of the bus module; and/or the number of the groups of groups,
The bus module comprises a shell and a pair of bus network cards arranged in the shell, wherein the shell is detachably connected to the rear window of the case, and each bus network card is in communication connection with one server node.
The beneficial effects are that: through the arrangement, the heat dissipation fan can play a certain heat dissipation role on the bus module, and on the basis, because the heat dissipation modules are multiple groups and are sequentially arranged along the vertical direction, different kinds of expansion cards can be inserted into each group of heat dissipation modules, and therefore the expansibility of the server is greatly improved.
In a second aspect, the present invention also provides a server cabinet, including:
a cabinet;
The server provided in the first aspect of the present invention is disposed in a cabinet.
The beneficial effects are that: the server cabinet of the second aspect of the present invention includes or uses the server related to the first aspect of the present invention, and therefore has the advantage that when any input power fails and cannot supply power, the corresponding dual-input power module can automatically switch and connect the other input power to the server node, so that normal operation is maintained. Therefore, when any double-input power supply module fails, the other double-input power supply modules can still keep normal power supply, so that the server node can work normally, the total power of the power supply modules is not required to be configured to be twice the power consumption of the server, the number of the power supply modules is reduced, the cost is reduced, and the space is saved.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Typical server node layouts may be: 1. the structure with horizontally arranged nodes is that two nodes with 19 inch and half width are left and right placed in the space of each U of a 19 inch cabinet, the width of a node main board 301 is about 210mm, and the support form is 2U4 nodes or 6U12 nodes; 2. the node vertical structure is, for example, 4U is selected as the node height, the width of the node main board 301 is about 160mm, 8 nodes are placed side by side, the node height is about 1U-1.25U, and the support form is 4U8 nodes or 4U10 nodes.
In addition, the node front-end and the node back-end bring different architectures, and the layout of the nodes and the power modules of the typical server node 3 may also be as follows:
1. The nodes are arranged on the front side of the case 1, and the power supply module is arranged on the rear side of the case 1;
2. the node is placed at the rear side of the case 1, the power module is also placed at the rear side of the case 1, and the node further comprises different layouts such as the power module is placed inside the node and the power module is placed on the side face of the node.
In the related art, the power modules used by the server are all single-input power modules, taking a processor platform with 2U4 nodes as a main stream as an example, 4 power modules are needed, 12V outputs of the power modules are connected together to form N+N, namely 2+2 power supply redundancy, and the power consumption of the whole system does not exceed the maximum power supply capacity of 2 power modules. The cabinet is externally connected with two paths of 220V AC input power supplies, each path of AC input power supply is connected to 2 power supply modules, when the single path of 220V AC input power supply of the server room fails and cannot supply power, the corresponding 2 power supply modules cannot supply power, the other path of 220V AC input power supply keeps normal power supply, 2 power supply modules can keep normal power supply, and the server nodes keep normal work; when a single power module fails and cannot provide power, the rest power modules still keep normal power supply, and the server node keeps normal work.
However, as the slot size of the CPU (Central Processing Unit, the CPU 305) increases, when each server CPU is full of 16 memory banks 306, the motherboard 301 needs to have a width of 250mm to put down the CPU and the memory devices, and none of the existing blade server architectures support the motherboard 301 with this width. And with the increase of CPU power, taking an 8-node vertical architecture as an example, each node supports two 600W CPUs and 32 memory banks 306 (and adopts liquid cooling heat dissipation), the power consumption of the whole machine can provide enough power supply capacity by using 5 3200W power supply modules to work simultaneously, and for an N+N power supply redundant system, the system needs to support 10 power supply modules at maximum, and occupies a large space and cost.
Embodiments of the present invention are described below with reference to fig. 1 to 6.
According to an embodiment of the present invention, in one aspect, there is provided a server including a chassis 1, a dual input power module 2, and a server node 3. At least one dual-input power module 2 is disposed in the chassis 1, and two input ends of the dual-input power module 2 are respectively connected with one input power. At least one server node 3 is arranged within the chassis 1. The output end of the double-input power supply module 2 is used for being connected with a server node 3, the power of the double-input power supply module 2 is V, the number of the double-input power supply modules 2 is n, the total power consumption of the server is W, and W is less than or equal to (n-1) V is less than 2W.
According to the server disclosed by the embodiment of the invention, the server node 3 is powered by the double-input power supply modules 2, and the two input ends of each double-input power supply module 2 are respectively used for being connected with one power supply, so that when any input power supply fails and cannot supply power, the corresponding double-input power supply module 2 can be automatically switched and the other input power supply is connected with the server node 3, and therefore normal operation is maintained. Therefore, only one redundant power supply module is needed to be added, when any double-input power supply module 2 fails, the other double-input power supply modules 2 can still keep normal power supply, so that the server node 3 can work normally, the total power of the power supply modules is not needed to be configured to be twice the power consumption of the server, the number of the power supply modules is reduced, the cost is reduced, and the space is saved.
Each dual-input power module 2 is internally provided with an ATS (Auto-Transfer-Switch) switching circuit, can be simultaneously connected with two paths of 220VAC inputs, can automatically Switch to the other path of 220VAC input to supply power when a single path of 220V AC input is lost, has a switching time of less than 10ms, and keeps the voltage of the external output 12V unchanged in a normal fluctuation range during the switching, and does not influence the server node 3.
In one embodiment, (n-2) V < W.
When the number of the dual input power modules 2 is configured according to the above-mentioned proportional relation, it is possible to ensure that the server can maintain normal operation regardless of whether the input power fails or whether a single dual input power module 2 fails, and to maximally save the number of dual input power modules 2 to be used, so as to reduce the cost of the server and the space occupied by the dual input power modules 2.
Assuming that the total power of 5 power modules can satisfy the power supply of the server, when a single input power module is used, it is necessary to configure to n+n redundancy power supply, and thus at least 10 power modules need to be provided. The server of the embodiment of the application only needs to use 6 double-input power supply modules 2. The cost of four power supply modules is saved, the space occupied by the power supply modules is reduced, and the density of the server is improved.
In one embodiment, the server further comprises a power panel 4. The dual-input power modules 2 are connected in parallel, and the output ends of the dual-input power modules are connected to a power panel 4, and the power panel 4 is in communication connection with a plurality of server nodes 3.
The power supply reliability and stability of the system can be improved by adopting the mode of connecting the double-input power supply modules 2 in parallel, and the system can be suitable for occasions with extremely high requirements on power supply stability, such as data centers, enterprise-level servers and the like. The power panel 4 is used as a core power supply component of the server system and is responsible for distributing and adjusting power to the server node 3. Comprising a plurality of power module interfaces and a power input interface 303 of the server node 3.
Preferably, in the embodiment of the present application, the power panel 4 is disposed at the bottom of the chassis 1, so that the wiring and the ventilation of the interior of the chassis 1 can be facilitated.
The communication connection may be a signal connection or a wire or cable connection.
Illustratively, in the present embodiment, the power panel 4 is optionally connected to the server node 3 through a bus bar, or is optionally connected to the server node 3 through a conventional cable, so as to provide 12V power for the server node 3.
In one embodiment, as shown in fig. 2, a first connector 302, a power input interface 303 and a power output interface 304 are disposed at the rear end of the server node 3, where the first connector 302 is used to connect a centralized line card, and the power input interface 303 and the power output interface 304 are connected to the power board 4 through a bus bar or a cable.
In one embodiment, as shown in fig. 4, an avoidance port 101 is formed at a front window of the chassis 1, the power panel 4 is disposed in the chassis 1, a plurality of first interfaces 401 are disposed on the power panel 4, and the dual-input power module 2 can pass through the avoidance port 101 to be plugged with the first interfaces 401.
Through such setting, can realize that dual input power module 2 can take out alone and maintain, can be convenient for dual input power module 2's overhaul and maintain. In addition, the design of the front window avoidance port 101 of the case 1 reserves a certain space at the front window of the case 1, can be used for installing and maintaining internal hardware equipment, and simultaneously provides convenience for heat dissipation. This design also helps to improve server scalability and future upgrade possibilities, allowing users to access internal hardware such as memory bank 306, expansion cards, etc. without opening the side panels of chassis 1.
Preferably, as shown in fig. 3 and 4, handles 201 are further disposed on the front side of the dual input power module 2 in a one-to-one correspondence manner, and the handles 201 can facilitate the operator to pull and maintain the dual input power module 2.
In one embodiment, a support structure is provided within the chassis 1, and the server further includes a first connector 302, a second connector, and a centralized management card. Wherein the first connector 302 is connected to a side of the server node 3 for insertion into the dodge port 101. The second connector is connected to the support structure and is disposed opposite to the first connector 302, and the server node 3 can pass through the escape opening 101 and plug the first connector 302 into the second connector. The centralized management card is arranged at the rear window of the case 1, and the second connector is connected with the centralized management card.
In the description of the present application, it should be understood that the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or number of technical features indicated. In the description of the present application, the meaning of "plurality" is two or more unless specifically defined otherwise.
By such arrangement, the server node 3 can be independently pulled out for maintenance, and the operator can be facilitated to overhaul or replace the server. Wherein the server node 3 is preferably a blade node. The blade server, due to its modular design, can integrate more server nodes 3 in a limited space, providing a higher computational density. Each blade has independent computing power, supports a multi-core processor and a high-speed network interface, and is suitable for executing massive parallel computing tasks.
Preferably, the centralized management card is disposed in the lowest 1U at the rear side of the chassis 1, and is capable of centralized management of a plurality of nodes in the chassis 1, including monitoring the status of servers, performing remote management tasks, managing power and fans, and the like. Preferably, the central management card is connected to the connector inside the chassis 1, typically by a cable or PCB board, so as to implement management of the server node 3. By the arrangement, the efficiency and the reliability of server management can be improved, and meanwhile, the maintenance cost is reduced.
Wherein, the supporting structure can be a middle bracket fixedly arranged inside the case 1.
In one embodiment, as shown in fig. 2, the front of the server node 3 can support a bus network card 701, a hard disk, and management IOs, and the back of the server node 3 can be used to set up the power connectors and the first connector 302. Preferably, the front side of the server node 3 supports management of IOs, and the blade server is a high-density and high-efficiency server form, and the design concept is that rapid data transmission and low-delay communication are realized through an integrated IO interface and high-speed interconnection technology. In a blade server, front-end management IOs involve not only traditional network and storage IOs, but also hot plug support for blade modules, power management, and co-operation with other blades or management nodes.
In one embodiment, as shown in fig. 3, the server nodes 3 are arranged in parallel in the width direction of the chassis 1, the width direction of the server nodes 3 is parallel to the height direction of the chassis 1, and the dual input power modules 2 are arranged below the server nodes 3 and are arranged in sequence in the width direction of the chassis 1.
The server in the embodiment of the application uses the dual-input power module 2 to replace the traditional single-input power module, and the space occupied by the dual-input power module 2 is larger than that occupied by the conventional single-input power module because of the addition of an ATS power supply and the addition of an input power supply interface, but the number of the power modules required to be used is correspondingly reduced, and on the basis of the server again, the space in the case 1 is reasonably utilized by adopting a framework with vertically arranged server nodes, so that the space occupied by the dual-input power module is reduced while the sufficient power supply capability of the dual-input power module 2 is ensured, and more space in the case 1 can be used for setting the server nodes 3.
On the basis, the application breaks the limit of the conventional chassis 1 architecture to the width of the server node 3 through reasonable arrangement of the internal architecture of the chassis 1, allows the server node 3 to be set wider, and increases the width of the motherboard 301, so that the motherboard 301 can support more central processors 305 and memory banks 306.
In one embodiment, the height of the chassis 1 is h, 300 mm.ltoreq.h.ltoreq.310 mm, the width of the server node 3 is L1, 250 mm.ltoreq.L1.ltoreq.265 mm.
When the height of the chassis 1 is set to be within the above range, the width of the motherboard 301 of the server node 3 can be configured to be 250mm, and at this time, each motherboard 301 can support two 600w power cpus 305 and 32 memory banks 306, so that the expansibility of the server is greatly improved, so that the server can provide more powerful computing power and rapid data processing speed, and the server can be suitable for a scenario requiring processing of a large number of parallel tasks.
Meanwhile, the space below the chassis 1 of the server node 3 can be used for setting the dual-input power supply module 2, and 6 dual-input power supply modules 2 with power of 3300w are calculated to be enough to meet the power supply requirement of the server in a full distribution state, so that the server in the embodiment of the application can simplify the number of power supply modules on the premise of keeping the power supply capacity and redundancy capacity unchanged, reduce the cost of the power supply modules, and support the latest motherboard 301 with the width of 250 mm.
In addition, the server provided by the embodiment of the application can be just suitable for a 19-inch universal cabinet, and the server provided by the embodiment of the application has good universality, and does not need a custom cabinet, so that the use cost of the server is greatly reduced.
In a preferred embodiment, the height of the cabinet 1 is 7U.
It should be noted that in the embodiment of the present application, when all the server nodes 3 are in the full configuration state, 6 dual-input power modules 2 are required to meet the power supply requirement of the server. However, in the actual use process, the configuration of each server node 3 is increased or decreased according to the use requirement of the user, so that an operator can calculate the actual power consumption of the server and adjust the number of the dual-input power modules 2 according to the actual power consumption of the server.
In one embodiment, the width of the chassis 1 is 440mm < L2 < 450mm, and the number of the server nodes 3 is 8, and the server nodes are uniformly arranged in the chassis 1 along the width direction of the chassis 1.
When the width of the chassis 1 is within the above range, the dimension of each server node 3 in the thickness direction of the node can be set to L3, 50mm is less than or equal to L3 is less than or equal to 55mm, in the related art, the dimension of the server node 3 in the thickness direction of the server node 3 in the high-density server is generally 1U (about 44.45 mm), the server node 3 of the embodiment of the present application can ensure a certain gap between two adjacent server nodes 3 while ensuring a higher density of the server, and the heat dissipation effect is superior to that of the high-density server in the related art.
In a preferred embodiment, the width of the chassis 1 is 10U, and the size of each server node 3 in its thickness direction is 1.25U.
In a preferred embodiment, each server node 3 occupies a width of 1.25U along the width direction of the chassis 1, so that the heat dissipation effect of the server can be ensured while the density of the server is ensured.
In one embodiment, the server further comprises an input interface 5. The input interface 5 is arranged at the bottom of the rear window of the case 1 and is opposite to the double-input power module 2, the input end of the input interface 5 is used for being connected with at least two paths of power supplies, and the output end of the input interface 5 is used for being connected with the double-input power module 2.
Preferably, the input interface 5 is disposed in the lowest 1U space at the rear side of the cabinet 1, and since the power input interface 303 is a very important part of the entire system, the placement of the input interface at the lowest 1U at the rear side of the cabinet 1 can help to ensure stability and reliability thereof.
In one embodiment, the input interface 5 comprises a plurality of pairs of input interfaces 5. Each pair of input interfaces 5 comprises two input interfaces 5, each input interface 5 is used for being connected with one input power supply, and each pair of input interfaces 5 is used for being connected with two input interfaces 5 of one double-input power supply module 2.
As shown in fig. 6, the bottom of the rear window of the chassis 1 is provided with 12 AC input interfaces 5, each AC input interface 5 can be connected to one 220V AC, each dual-input power module 2 is connected to two AC input interfaces 5 on the rear side of the chassis 1 through a circuit board and a cable, each cabinet normally provides 2 220V AC input PDUs (Power Distribution Unit ), the 1 st AC input interface 5 of each dual-input power module 2 is connected to the 1 st 220V AC input interface 5 of the cabinet, and the 2 nd AC input interface 5 of each power module is connected to the 2 nd 220V AC input interface 5 of the cabinet. This design ensures that each dual input power module 2 can draw power from two separate PDUs, thereby greatly reducing the risk of system interruption due to a single power problem.
In addition, the dual-input power supply is positioned in the server, and the server externally adopts the universal 220V input interface 5, so that the universality is good.
As an alternative embodiment, in an embodiment not shown in the drawings, two aviation sockets are disposed at the bottom of the rear window of the chassis 1, each aviation socket is connected to one input power, and each aviation socket is connected to one input interface 5 of the multiple dual-input power modules 2. Through such setting, when the input of arbitrary alternating current suddenly breaks off or loses efficacy, two input power module 2 can switch to another way alternating current automatically, and use the aviation socket of heavy current can reduce the required quantity of the input interface 5 that uses of server, and then reduces the redundancy of system.
In one embodiment, the server further comprises a heat sink module 6. The heat dissipation module 6 is disposed at a rear window of the chassis 1 for blowing air toward the server node 3. High performance computing devices and servers can generate a significant amount of heat during operation, which can cause the device to overheat if not timely expelled, affecting computing efficiency and hardware life. The centralized fan heat dissipation design can solve the problem in a targeted manner, and the fan arranged in the upper area of the rear side can forcedly discharge the internal hot air out of the chassis 1, so that the normal operation temperature of the equipment can be maintained.
In one embodiment, the heat dissipation modules 6 are multiple groups and are sequentially arranged along the height direction, and each heat dissipation module 6 includes a plurality of heat dissipation fans 601 sequentially arranged along the width direction of the chassis 1.
By so doing, it is possible to provide a compact structure. The heat radiation fan 601 can be uniformly distributed along the width direction of the cabinet 1 to ensure that heat can be uniformly conducted from the equipment into the air and discharged outside the cabinet 1 through the fan. Further, the heat radiation fan 601 vertically covers the position of the cabinet 1 vertically corresponding to the server node 3, and can effectively radiate heat to the server node 3.
Wherein, the heat dissipation module 6 is optionally drawably disposed at the rear window of the server, thereby facilitating the maintenance and replacement of the heat dissipation module 6 by an operator. Specifically, the heat dissipation module 6 may include a fan frame and a plurality of heat dissipation fans 601, where the fan frame is connected to a rear window of the chassis 1, and the plurality of heat dissipation fans 601 are respectively disposed on the fan frame in a drawable manner, so that an operator can conveniently overhaul or replace the heat dissipation fans 601.
Wherein, radiator fan 601 optional is through cable or female row and centralized management card and power strip 4 communication connection to carry out centralized management through centralized management card to radiator fan 601, carry out centralized power supply to radiator fan 601 through power strip 4, through so setting, can improve efficiency and the reliability of server management, also reduced the maintenance cost simultaneously.
In one embodiment, the cooling fan 601 is a high performance fan capable of sufficiently cooling the server node 3.
In one embodiment, the server further comprises an expansion card detachably arranged at the rear window of the chassis 1, and the expansion card is in communication connection with the server node 3.
The expansion card is preferably, but not limited to, a network card or a hard disk, etc., and the expansion card refers to an expansion device inserted into a computer or other electronic devices to increase the functions of the original devices. Common expansion cards include graphics cards, sound cards, network cards, and the like. They are connected to the motherboard 301 through interfaces such as PCI, PCI Express, etc., to provide additional functions or enhance existing functions. The expansion card can be connected with the server node 3 through a built-in cable fixed inside the chassis 1, and the design of the rear PCIe network card and the Multi-Host network card can provide additional expansion capability under the condition of not occupying a front expansion slot. The design is particularly suitable for a high-density server with limited space, can effectively utilize the space at the rear part of the chassis 1, and simultaneously supports interconnection among a plurality of CPUs or systems through a Multi-Host technology, thereby improving the performance and efficiency of a data center.
In one embodiment, the expansion card includes a bus module 7;
the bus module 7 is inserted into the heat dissipation module 6, and at least one heat dissipation fan 601 is disposed on two sides of the bus module 7.
By means of the arrangement, the heat dissipation fan 601 can play a certain heat dissipation role on the bus module 7, and on the basis, because in the embodiment of the application, the heat dissipation modules 6 are multiple groups and are sequentially arranged along the vertical direction, different kinds of expansion cards can be inserted into each group of heat dissipation modules 6, and therefore the expansibility of the server is greatly improved. Illustratively, in the embodiment shown in FIG. 6, the expansion card may also be selected to include other types of expansion cards to meet the needs of different users.
Preferably, an expansion card slot is further formed on the fan frame, and the expansion card is retractably arranged on the expansion card slot.
As an alternative embodiment, the heat dissipation module 6 may also be a liquid cooling module, where the liquid cooling module includes a plurality of cold plates, and the plurality of cold plates and the server nodes 3 are alternately arranged, so that heat dissipation can be performed on servers on two sides, and by using the server architecture of the embodiment of the present application, it can be ensured that a gap exists between two adjacent server nodes 3 along the width direction of the chassis 1, and the gap can be used to accommodate the cold plates, so that the heat dissipation effect of the servers is improved by using the cold plates, and the PUE value of the liquid cooling system is generally lower than that of the air cooling system, which means that under the same heat dissipation effect, the liquid cooling system can save more energy sources, and reduce the overall energy consumption of the data center.
The bus module 7 includes a housing and a pair of bus network cards 701 disposed in the housing, the housing being detachably connected to the rear window of the chassis 1, each bus network card 701 being communicatively connected to one of the server nodes 3.
Through the arrangement, every two expansion cards can be used as a unit and synchronously plugged and unplugged, and an operator can conveniently overhaul or replace the expansion cards. As shown in fig. 6, in a preferred embodiment, the bus modules 7 have 4 groups, each heat dissipation module 6 can be provided with two groups of bus modules 7 at intervals, and 8 bus network cards 701,8 of bus network cards 701 in the 4 groups of bus modules 7 can be connected to one server node 3 respectively and connected to the power panel 4 through cables or circuit boards.
The bus network card 701 may be connected to the server node 3 and the power panel 4 through a patch card or a cable.
Preferably, in an embodiment not shown in the drawings, the bus network card 701 is connected to the server node 3 and the power board 4 through a riser card, and by this arrangement, not only more expansion slot space can be provided: allowing users to install more expansion cards, can also increase the functions and performance of the server, and reduce the complexity of hardware wiring: and the connection between the hardware components is simpler and more orderly, so that the probability of faults is reduced. Providing better heat dissipation effect: by vertically installing the expansion card, the air flow is smoother, the heat dissipation is facilitated, and the system is suitable for different server configurations. Also, because different servers may have different expansion card requirements and space constraints, server riser often provides multiple types of slots and connection interfaces to accommodate different server configurations.
According to an embodiment of the present invention, in another aspect, there is also provided a server cabinet, including a cabinet and a server provided in the first aspect of the present invention, where the server is disposed in the cabinet.
The server cabinet according to the second aspect of the present invention includes or uses the server according to the first aspect of the present invention, and therefore has the advantage that when any one of the input power sources fails and cannot supply power, the corresponding dual-input power source module 2 can automatically switch and connect the other input power source with the server node 3, thereby maintaining normal operation. Therefore, only one redundant power supply module is needed to be added, when any double-input power supply module 2 fails, the other double-input power supply modules 2 can still keep normal power supply, so that the server node 3 can work normally, the total power of the power supply modules is not needed to be configured to be twice the power consumption of the server, the number of the power supply modules is reduced, the cost is reduced, and the space is saved.
Preferably, in an embodiment of the present application, the server rack is a 19 inch rack.
In summary, the server of the first aspect and the server cabinet of the second aspect of the present invention have the following advantages:
The invention newly designs a blade server architecture, a front double-input power supply is introduced into the blade server, the front double-input power supply occupies the lowest 1U space independently, the power supply redundancy requirement of 5+5=10 single-input power supplies is reduced to the power supply redundancy requirement of only 5+1=6 double-input power supplies, and a large amount of power supply module cost is saved for the blade server. And the 6U-height vertical nodes are adopted, 8 nodes are transversely arranged, each blade node can support a main board 301 with the width of 250mm, 2 intel and AMD (advanced data processing) latest and maximum packaging CPU (Central processing Unit) and 32 memory banks 306 can be supported by a single node, the height of each blade node is 1.25U, and the heat dissipation condition is superior to that of a conventional 1U-height high-density server architecture.
In the present application, unless explicitly specified and limited otherwise, the terms "mounted," "connected," "secured," and the like are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communicated with the inside of two elements or the interaction relationship of the two elements. The specific meaning of the above terms in the present application can be understood by those of ordinary skill in the art according to the specific circumstances.
Although embodiments of the present invention have been described in connection with the accompanying drawings, various modifications and variations may be made by those skilled in the art without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope of the invention as defined by the appended claims.