CN106292911A - A kind of fusion architecture server - Google Patents

Abstract

The invention discloses a converged architecture server, comprising: a computing node, a backplane, an IO/RRC module, an external switch, a fan, and a management device; wherein, the backplane is a slot with a predetermined number of fixed heights, and the The computing nodes are connected to the backplane through slots; the converged architecture server can realize different configurations through hardware reconfiguration by using the slots on the backplane to meet different applications, so that the entire converged architecture server can be flexibly configured.

Description

A Converged Architecture Server

technical field

The present invention relates to the technical field of servers, in particular to a converged architecture server.

Background technique

In recent years, with the rapid development of technologies such as mobile, social media, cloud computing, and big data, the business that servers need to handle has shown explosive growth. The cluster system is widely used in the processing of this kind of business. Using traditional isolated system component clusters, because each system is isolated, each system requires different deployment and management, which causes the company to spend a lot of resources on server deployment, management and maintenance. In order to solve the disadvantages of the traditional cluster implementation, the converged architecture server was born.

Converged architecture servers centralize computing, storage, switching, and management in the system, which greatly simplifies system deployment, management, and maintenance. Converged architecture servers have become the future development trend. Although most applications can now use clusters to solve application expansion and performance improvement, for some key computing and applications, highly reliable and high-performance converged architecture servers are still required. And due to different user applications, the corresponding configuration requirements for the converged architecture server are also different; therefore, how to configure different hardware according to different applications to meet the needs of various services, thereby reducing user costs and increasing system scalability property is a technical problem to be solved by those skilled in the art.

Contents of the invention

The purpose of the present invention is to provide a converged architecture server, which can realize different configurations through hardware reconfiguration by using slots on the backboard, satisfy different applications, and enable the entire converged architecture server to be flexibly configured.

In order to solve the above technical problems, the present invention provides a converged architecture server, including: computing nodes, backplanes, IO/RRC modules, external switches, fans and management devices;

Wherein, the backplane is a predetermined number of slots with fixed heights, and the computing nodes are connected to the backplane through the slots.

Wherein, the management device includes a controller, an internal switch connected to the controller through a network interface, and used to forward the data of the controller through the network interface;

The controller is connected to the backplane through a bus interface, the controller is connected to the external switch through a network interface, and the internal switch is connected to the backplane through a network interface.

Wherein, the computing nodes are half-width computing nodes and/or 4-way computing nodes and/or 8-way computing nodes.

Wherein, it also includes: a storage box used for connecting with the backplane through a slot.

Wherein, it also includes: a GPU expansion box for connecting with the backplane through a slot.

Wherein, the backplane has 8 slots with fixed heights.

Wherein, the number of the external switches is 2.

Wherein, the number of the IO/RRC modules is 2.

Wherein, the number of the fans is 4.

Among them, also include:

The server back box, and the IO/RRC module, the external switch, the fan and the management device are all arranged in the server back box; the server back box is connected to the backplane.

The converged architecture server provided by the present invention includes: a computing node, a backplane, an IO/RRC module, an external switch, a fan, and a management device; wherein, the backplane is a slot with a predetermined number of fixed heights, and the Computing nodes are connected to the backplane through slots;

It can be seen that the converged architecture server can use the slots on the backplane to insert computing nodes with different computing capabilities, such as 4-way computing nodes or 8-way computing points, that is, the expansion of computing functions is realized through hardware reconfiguration to meet the needs of users. Requirements; That is, different configurations can be realized through hardware reconfiguration to meet different applications, so that the entire converged architecture server can be flexibly configured.

Furthermore, the slots on the backplane can be used not only for computing node expansion but also for user storage and GPU expansion, that is, by inserting a storage box into the slot on the backplane to realize storage expansion, and inserting a GPU expansion box to realize GPU expansion; Through the backplane, the hardware reconfiguration configuration of the converged architecture server is flexible and fast, which meets the actual needs of users and improves user experience.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only It is an embodiment of the present invention, and those skilled in the art can also obtain other drawings according to the provided drawings without creative work.

FIG. 1 is a block diagram of a converged architecture server provided by an embodiment of the present invention;

2 is a structural block diagram of an IO/RRC module, an external switch, a fan, and a management device in a converged architecture server provided by an embodiment of the present invention;

Fig. 3 is a structural block diagram of a management device provided by an embodiment of the present invention.

detailed description

The core of the present invention is to provide a converged architecture server. The converged architecture server can realize different configurations through hardware reconfiguration by using the slots on the backboard, satisfying different applications, and enabling the entire converged architecture server to be flexibly configured.

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

Please refer to Fig. 1, Fig. 1 is the block diagram of the structure of the fusion architecture server (which can be referred to as a Tray) provided by the embodiment of the present invention; it can include: computing node 10, backplane 20, IO/RRC module, external switch, fan and management devices;

Wherein, the backplane 20 is a predetermined number of slots with fixed heights, and the computing nodes are connected to the backplane 20 through the slots.

Wherein, by setting the slots on the backplane 20 as slots with a fixed height, such a standardized setting can facilitate subsequent hardware reconfiguration of the converged architecture server by inserting different hardware into the slots. And the preset number of slots here can be set by the user according to their actual needs. For example, it can be set to 4 slots or 8 slots under normal circumstances. If the user's business is very large, more numbers can be set. slot. The fixed height is also based on the commonly used hardware heights used by various manufacturers when designing hardware such as computing nodes 10 and storage boxes. The purpose is to facilitate the installation and replacement of hardware such as computing nodes 10 . That is, the fixed height can be designed according to the actual production needs of the manufacturer. The size of the converged architecture server and the height of the inserted hardware should be considered in the design; For half-width computing node slots, two slots can be set in a row, each with a height of 1.25U, as shown in Figure 1. The predetermined number and specific numerical values of the fixed height are not limited here. There is no limitation on the shape of the slot here, as long as the active connection can be realized, the computing node and other hardware connected thereto can be detached and connected.

Wherein, the selection of the computing node 10 here is based on the calculation amount of the actual converged architecture server. In order to control the cost more conveniently and more flexibly select the computing node 10, it is optional here that the computing node 10 is half-width Computing nodes and/or 4-way computing nodes and/or 8-way computing points. The number of half-width computing nodes or 4-way computing nodes or 8-way computing points can also be determined according to the actual situation of the user, and can also be flexibly increased or decreased according to the actual situation to improve resource utilization. The number of half-width computing nodes is 1-8. It should be noted here that a 4-way computing node or 8-way computing node needs to occupy the positions of two half-width nodes, that is, a 4-way computing node or an 8-way computing node is a full-width computing node, and each 4-way computing node or 8-way computing node The computer point is the length of two half-width computing nodes. As shown in Figure 1, if only 4-way computing nodes or 8-way computing nodes are used on the backboard, only four 4-way computing nodes or 8-way computing nodes can be installed. Line up one. That is, according to the actual situation of the user, a 4-channel, 8-channel and other high-reliability systems can be realized by replacing the computing nodes on the backplane to meet the needs of key applications.

That is to say, different hardware can be inserted into the backplane to realize different function expansions. Here, the hardware can be selected according to the actual needs of users to meet the needs of different applications. Preferably, the converged architecture server may further include: a storage box for connecting to the backplane through a slot. Here, setting the storage box to a predetermined height can also realize inserting the storage box on the backplane to expand the local storage of the converged architecture server. Preferably, the converged architecture server may further include: a GPU expansion box for connecting to the backplane through a slot. Here, setting the GPU expansion box to a predetermined height can also realize inserting the GPU expansion box on the backplane to meet the application of high-performance computing.

Specifically, the following specific examples are given to illustrate the above method. Referring to the situation of 8 slots on the backplane listed in Figure 1, the computing nodes on the right side of the first row can be replaced with storage boxes or GPU expansion boxes. , which can meet the needs of cloud computing, virtualization and high-performance computing. When replacing the two half-width computing nodes in the same row in the third row with a computing node that can be expanded to 4-way or 8-way, using 2 computing nodes of this type can realize a 4-way system, and using 4 computing nodes An 8-way system can be realized, so that the system has extremely high reliability and meets the requirements for high-reliability key calculations and applications.

Based on the above technical solution, the converged architecture server provided by the embodiment of the present invention can insert computing nodes with different computing capabilities, such as 4-way computing nodes or 8-way computing nodes, through the slots on the backplane, that is, through hardware reconfiguration The expansion of computing functions meets the actual needs of users; that is, different configurations can be realized through hardware reconfiguration to meet different applications, so that the entire converged architecture server can be flexibly configured.

Furthermore, the slots on the backplane can be used not only for computing node expansion but also for user storage and GPU expansion, that is, by inserting a storage box into the slot on the backplane to realize storage expansion, and inserting a GPU expansion box to realize GPU expansion; Through the backplane, the hardware reconfiguration configuration of the converged architecture server is flexible and fast, which meets the actual needs of users and improves user experience.

Please refer to FIG. 2, and the converged architecture server in FIG. 1 also includes: a server back box 70, and the IO/RRC module 30, the external switch 40, the fan 50 and the management device 60 are all arranged on the server Inside the back box 70 ; the server back box 70 is connected to the backplane 20 .

Optionally, the backplane has 8 slots with fixed heights.

Optionally, the number of the external switches is 2.

Optionally, the number of the IO/RRC modules is 2.

Optionally, the number of the fans is four.

Specifically, when the same component is set in multiples, redundant settings for each component can be realized, preventing the fusion architecture server from working normally when a component fails, thereby improving the reliability and stability of the system. That is, the external switch, management device, and fan are designed redundantly. When a single external switch, management device, and fan is abnormal, the system can still operate normally. Thus the system has extremely high reliability. The management device realizes unified management of the entire Tray, simplifies system management, and makes operation and maintenance more convenient.

Optionally, the management device 60 includes a controller, and an internal switch connected to the controller through a network interface, configured to forward data of the controller through the network interface;

The controller is connected to the backplane through a bus interface, the controller is connected to the external switch through a network interface, and the internal switch is connected to the backplane 20 through a network interface.

Specifically, the type and model of the network interface and bus interface here can be selected according to the actual situation of the user's specific application of the fusion architecture server; optionally, the bus interface is an SMBUS interface and a GPIO interface, and the network interface is a gigabit network interface.

That is, please refer to FIG. 3 , which is a structural block diagram of a specific management device 60 ; the management device 60 is composed of a controller 61 and an internal switch 62 . The controller 61 outputs the SMBUS interface and the GPIO interface to the backplane 20 for connecting to computing, storage, switching, and I/O modules in the converged architecture server. The internal switch 62 outputs a gigabit network interface to the backplane 20 for connecting to computing, storage, switching, and I/O modules in the converged architecture server. The controller 61 is connected to the internal switch 62 through the Gigabit network interface, and the network data of the controller 61 is forwarded through the internal switch 62, so as to achieve a large number of computing, storage, switching, and I/O modules in the converged architecture server through the network interface. data exchange needs. At the same time, the controller 61 outputs a network interface for connecting to the external switch 40, so as to realize remote management of all servers in the computer room. That is to say, the controller 61 is connected to the external switch 40 through the Ethernet interface to realize remote management, and it is possible to use a computer to manage all the servers in the computer room anytime and anywhere.

Wherein, the external switch and the internal switch here are both switches, and they are only distinguished by internal and external, so as to facilitate the distinction of which position the switch is, and to facilitate understanding.

In the following, FIG. 1 is taken as a specific example to illustrate that the converged architecture server is referred to as a server, and a converged architecture server is called a Tray, and the composition of the Tray is shown in FIG. 1 . A Tray consists of computing nodes, backplanes, IO/RRC modules, switches (external switches), management devices, and fans. The height of the entire Tray is 5U, and each computing node is 1.25U. A Tray can accommodate up to eight half-width computing nodes, two IO/RRC modules, two dual-port 10G switches, two management devices, and four dual-rotor fans. Each part of the server adopts a modular design, which can realize the replacement of each module under the condition of uninterrupted power supply, so as to ensure that the system can still work normally when a certain part of the server is abnormal. At the same time, the switch, management device, and fan of the server are all redundantly designed. When a single switch, management device, and fan are abnormal, the server can still operate normally. Thus the server has extremely high reliability. The management device behind the Tray realizes unified management of the entire Tray, simplifies system management, and makes operation and maintenance more convenient. The biggest advantage of this server is that it can realize 4-way, 8-way and other high-reliability systems by replacing computing nodes to meet the needs of key applications. At the same time, computing nodes are replaced with storage boxes to expand local storage to meet the needs of cloud computing and virtualization applications. Furthermore, by expanding the computing node into a GPU expansion box, it is used to meet the application of high-performance computing. The whole system can be flexibly configured, and different configurations can be realized through hardware reconfiguration to meet different applications.

The converged architecture server provided by the present invention has been introduced in detail above. In this paper, specific examples are used to illustrate the principle and implementation of the present invention, and the descriptions of the above embodiments are only used to help understand the method and core idea of the present invention. It should be pointed out that for those skilled in the art, without departing from the principle of the present invention, some improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims (10)

1. a fusion architecture server, it is characterised in that including: calculate node, backboard, IO/RRC module, external switch, Fan and managing device；

Wherein, described backboard is the slot with predetermined number level altitude, and described calculating node is by slot and the described back of the body Plate is connected.

2. fusion architecture server as claimed in claim 1, it is characterised in that described managing device includes controller, with institute State the internal switch that controller is connected by network interface, for being forwarded the number of described controller by described network interface According to；

Described controller is connected with described backboard by EBI, and described controller is by network interface and described outside exchange Machine is connected, and described internal switch is connected with described backboard by network interface.

3. fusion architecture server as claimed in claim 2, it is characterised in that described calculating node is that half-breadth calculates node And/or 4 tunnel calculate node and/or No. 8 computer points.

4. the fusion architecture server as described in any one of claim 1-3, it is characterised in that also include:

For the storage box being connected with described backboard by slot.

5. the fusion architecture server as described in any one of claim 1-3, it is characterised in that also include:

For the GPU extension box being connected with described backboard by slot.

6. fusion architecture server as claimed in claim 5, it is characterised in that described backboard has inserting of 8 level altitudes Groove.

7. fusion architecture server as claimed in claim 6, it is characterised in that the number of described external switch is 2.

8. fusion architecture server as claimed in claim 7, it is characterised in that the number of described IO/RRC module is 2.

9. fusion architecture server as claimed in claim 8, it is characterised in that the number of described fan is 4.

10. fusion architecture server as claimed in claim 9, it is characterised in that also include:

Server rear cabinet, and described IO/RRC module, described external switch, described fan and described managing device be respectively provided with In described server rear cabinet；Described server rear cabinet is connected with described backboard.