Network device and communication system

Abstract

The present application relates to the field of communications. Disclosed are a network device and a communication system. The network device specifically comprises: N single boards and M2 switching boards. M nodes and M2 orthogonal connectors are deployed on one single board, the M nodes being respectively in bidirectional connection with M different orthogonal connectors among the M2 orthogonal connectors, and every two nodes of the M nodes being in transmitting-receiving one-way connection by means of two different orthogonal connectors; N orthogonal connectors are deployed on one switching board, the N orthogonal connectors deployed on the switching board being connected in pairs, and the N orthogonal connectors on the switching board being each connected to an orthogonal connector at a same position on each single board of the N single boards.

Description

Network device and communication system

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office on August 8, 2023, with application number 202311003405.6 and application name “A Network Device and Communication System”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of communications, and in particular to a network device and a communication system. Background Art

[0003] A mesh network is a type of network connection in which all nodes are directly connected through physical connections or virtual circuits. Figure 1 illustrates a mesh network connection. For a full-mesh network with N nodes, the total number of connections is [N × (N-1)] ÷ 2.

[0004] As the scale of high-performance computing networks increases, signal frequencies become higher and higher, and the requirements for connection delays between nodes become smaller and smaller; the increase in networking nodes leads to an explosive growth in the number of links in the Mesh network, posing a huge challenge to the design of the Mesh network.

[0005] Currently, the industry's interconnection systems for chassis or cabinets all rely on traditional backplanes—that is, a single printed circuit board (PCB) backplane interconnects the various functional modules. All functional modules are connected to each other via backplane wiring, and the interconnection signals between functional modules are carried by the backplane, enabling direct connectivity between any two nodes. However, as the number of nodes increases, the total number of connections also rises dramatically. Due to the inherent limitations of the PCB, scalability is poor. Furthermore, as interconnection signal rates increase, the insertion loss of the PCB increases, making it difficult to implement a high-speed signal mesh network.

[0006] Another mesh network implementation uses cables (such as optical fibers) to connect nodes. This solution can reduce the excessive insertion loss caused by increasing signal frequencies. Furthermore, the connection channels between nodes are independent, improving engineering reliability and reducing the impact of single-point hardware failures on the entire system. However, as the number of nodes and the total number of connections increases dramatically, this solution not only faces meshing issues, but also the exponential increase in the number of cables, which inevitably leads to uncontrollable problems such as increased costs, signal delays, and poor signal quality.

[0007] Therefore, in scenarios where the network scale grows explosively with the increase in the number of nodes, it is expected to implement a highly reliable, low-latency, and low-maintenance Mesh network within limited hardware resources to meet large-scale networking needs.

[0008] Summary of the Invention

[0009] The present application provides a network device and a communication system to realize a highly reliable, low-latency, and low-maintenance Mesh network to meet the needs of large-scale networking.

[0010] To achieve the above objectives, the present invention adopts the following technical solutions:

[0011] In a first aspect, a network device is provided, which may include: N single boards and Q switch boards. M nodes and Q orthogonal connectors are deployed on a single board, the M nodes are bidirectionally connected to M different orthogonal connectors among the Q orthogonal connectors, and every two nodes among the M nodes are connected through at least one orthogonal connector; N orthogonal connectors are deployed on a switch board, the N orthogonal connectors deployed on a switch board are connected in pairs, and the N orthogonal connectors on a switch board are respectively connected to the orthogonal connectors at the same position on each of the N single boards. M and N are greater than or equal to 2. Q is less than or equal to M 2 .

[0012] Using an orthogonal architecture of single boards and switch boards, a single switch board implements a mesh sub-network between N nodes, splitting the full mesh network of M*N nodes into Q mesh sub-networks between N nodes. Mesh connections are then re-established between the mesh sub-networks through unidirectional connections within the single board. While achieving a full mesh network of M*N nodes, the granularity of the mesh network (sub-network approach) is reduced, avoiding a sharp increase in the number of connections as the number of nodes increases. Furthermore, due to the high scalability and reliability of the orthogonal architecture, low cost, and simple functional maintenance, a highly reliable, low-latency, and low-maintenance mesh network is achieved, meeting the needs of large-scale networking.

[0013] In one possible implementation, Q is M 2 Each two of the M nodes on a single board are connected via at least one orthogonal connector, including: each two of the M nodes on a single board are connected via two different orthogonal connectors for unidirectional transmission and reception. Using the two orthogonal connectors, unidirectional transmission and reception connections are achieved between two nodes at different locations, thereby improving the bandwidth scalability of data transmission between the two nodes.

[0014] In another possible implementation, for the first node and the second node, the first node and the second node are any two nodes deployed on a single board. The first node and the second node are connected in a unidirectional manner through two different orthogonal connectors for sending and receiving, including: a unidirectional connection in the sending direction of the first node to the first orthogonal connector, and a unidirectional connection in the sending direction of the first orthogonal connector to the second node; a unidirectional connection in the sending direction of the second node to the second orthogonal connector, and a unidirectional connection in the sending direction of the second orthogonal connector to the first node. The first orthogonal connector and the second orthogonal connector are different. A unidirectional connection between nodes at two locations is achieved through an orthogonal isolator, thereby avoiding interference when nodes on different boards transmit data and improving the reliability of data transmission.

[0015] In another possible implementation, in a single board, an orthogonal connector for unidirectional connection between two nodes is deployed between orthogonal connectors for bidirectional connection between two nodes.

[0016] In another possible implementation, N boards are parallel to each other, and M 2 Switch boards are parallel to each other, N boards and M 2 The orthogonal architecture improves the scalability and compatibility of network equipment.

[0017] In another possible implementation, the orthogonal connectors on the single board and the switch board are male and female. N orthogonal connectors on a switch board are respectively connected to orthogonal connectors at the same position on each of the N single boards, including: N orthogonal connectors on a switch board are respectively connected to orthogonal connectors at the same position on each of the N single boards.

[0018] In another possible implementation, N nodes at the same location on each of N boards are connected to form a mesh sub-network through the same switch board. The structures of different mesh sub-networks are identical. This reduces the granularity of the mesh network and avoids a sharp increase in the number of connections as the number of nodes increases.

[0019] In another possible implementation, the connection structures of the nodes and the orthogonal connectors in different boards are the same.

[0020] In another possible implementation, the M*N nodes deployed in the network device are connected in pairs. Regarding the third node and the fourth node, the third node and the fourth node are two nodes located in the first position of different boards in the network device. The connection for the third node to send data to the fourth node includes: a bidirectional connection between the third node and the third orthogonal connector in the board to which it belongs, a plug-in connection between the third orthogonal connector and the fourth orthogonal connector on the first switch board, a connection between the fourth orthogonal connector and the fifth orthogonal connector on the first switch board, a plug-in connection between the fifth orthogonal connector and the sixth orthogonal connector on the board to which the fourth node belongs, and a bidirectional connection between the sixth orthogonal connector and the fourth node. The first switch board is used to connect the orthogonal connectors bidirectionally connected to the nodes in the first position of different boards.

[0021] In another possible implementation, the M*N nodes deployed in the network device are connected in pairs. For the third and fourth nodes, which are located at different positions on different boards, the connection for the third node to send data to the fourth node includes: a unidirectional connection in the sending direction between the third node and the seventh orthogonal connector at the second position on the board to which it belongs, a plug-in connection between the seventh orthogonal connector and the eighth orthogonal connector on the second switch board, a connection between the eighth orthogonal connector and the ninth orthogonal connector on the second switch board, a plug-in connection between the ninth orthogonal connector and the tenth orthogonal connector at the second position on the board to which the fourth node belongs, and a unidirectional connection in the sending direction between the tenth orthogonal connector and the fourth node. The second switch board is used to connect the orthogonal connectors at the second positions on different boards.

[0022] In a possible implementation, Q is M, and connecting every two of the M nodes on a single board through at least one orthogonal connector includes: connecting every two of the M nodes on a single board through a bidirectional connection through an orthogonal connector.

[0023] In a second aspect, a communication system is provided, comprising at least one network device as described in the first aspect or any possible implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] FIG1 is a schematic diagram of a Mesh network connection;

[0025] FIG2 is a schematic diagram of internal routing of a Mesh network implemented through a PCB backplane;

[0026] FIG3 is a schematic diagram of a hardware connection for implementing a Mesh network through a PCB backplane;

[0027] FIG4 is a schematic diagram of a real engineering scene of a mesh interconnection topology using cable connections;

[0028] FIG5 is a schematic diagram of the internal structure of a network device provided in an embodiment of the present application;

[0029] FIG6a is a schematic diagram of the internal structure of another network device provided in an embodiment of the present application;

[0030] FIG6 b is a schematic diagram of the internal structure of a single board provided in an embodiment of the present application;

[0031] FIG7 is a schematic diagram of the internal structure of another network device provided in an embodiment of the present application;

[0032] FIG8 is a schematic diagram of the internal structure of another network device provided in an embodiment of the present application;

[0033] FIG9 is a schematic diagram of a Mesh network before and after splitting provided in an embodiment of the present application;

[0034] FIG10 is a schematic diagram of the structure of a Mesh sub-network provided in an embodiment of the present application;

[0035] FIG11 is a schematic diagram of the internal structure of another network device provided in an embodiment of the present application;

[0036] FIG12 is a schematic diagram of the structure of another Mesh sub-network provided in an embodiment of the present application. DETAILED DESCRIPTION

[0037] In the embodiments of the present application, in order to clearly describe the technical solutions of the embodiments of the present application, words such as "first" and "second" are used to distinguish between identical or similar items with substantially the same functions and effects. Those skilled in the art will understand that words such as "first" and "second" do not limit the quantity or execution order, and words such as "first" and "second" do not necessarily mean different. There is no order of precedence or priority between the technical features described by "first" and "second".

[0038] In the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner to facilitate understanding.

[0039] In the embodiments of the present application, at least one can also be described as one or more, and multiple can be two, three, four or more, which is not limited in this application.

[0040] In addition, the network architecture and scenarios described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Ordinary technicians in this field can know that with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0041] To facilitate understanding, the terms involved in the embodiments of this application are first explained.

[0042] Board: A physical card that provides functionality within a device. For example, a board might be a line processing unit (LPU) in a router or an interface board in a switch.

[0043] Orthogonal architecture (OD): "Orthogonal" in mathematics refers to the perpendicular relationship between vectors in three-dimensional space. In hardware terms, this refers to a hardware architecture where the circuit boards in a device are orthogonal. For example, in a device that includes circuit boards and switching fabric boards, the orthogonal architecture could be one in which the circuit boards are inserted horizontally and the switching fabric boards are inserted vertically, or vice versa.

[0044] Orthogonal connector: A board-to-board connector that typically uses a plug-in connection method, with male and female connectors. The boards connected by an orthogonal connector have an orthogonal architecture.

[0045] Figure 2 illustrates the internal routing of a mesh network through a PCB backplane. As shown in Figure 2, a line card is connected to the PCB backplane and then, through the routing on the PCB backplane, connects to other line cards, thus enabling connections between nodes within the line card.

[0046] Figure 3 illustrates the hardware connection of a Mesh network using a PCB backplane. As shown in Figure 3, a line card is plugged into the PCB backplane.

[0047] As can be seen from the diagrams in Figures 2 and 3, as the number of nodes increases and the total number of connections also rises sharply, the number of line cards connected to the PCB backplane will increase, and the number of traces in the PCB backplane will also increase. Due to the limitations of the PCB board itself, the scalability is poor; at the same time, as the interconnection signal rate becomes higher and higher, the insertion loss of the PCB board will become larger and larger, making it difficult to implement a high-speed signal Mesh network.

[0048] Figure 4 illustrates a real-world engineering scenario for a mesh interconnect topology using cable connections. As shown in Figure 4, numerous cables are located on the back of the device, leading to inevitable interference between cables and uncontrolled cable length and layout, which can lead to serious and uncontrolled network issues. As the number of nodes and the total number of connections increase dramatically, the number of cables on the back of the device will also increase exponentially, increasing costs. This surge in cables also leads to uncontrollable issues such as signal latency and quality degradation, which in turn reduces reliability.

[0049] Based on this, the present application provides a network device that adopts an orthogonal architecture of a single board and a switch board. A switch board realizes a Mesh sub-network between N nodes, thereby splitting the full Mesh network of M*N nodes into Q Mesh sub-networks between N nodes. The Mesh sub-networks then re-establish Mesh connections through unidirectional connections in the single board. While realizing a full Mesh network of M*N nodes, the granularity of the Mesh network (sub-network method) is reduced, avoiding a sharp increase in the number of connections after the number of nodes increases. In addition, due to the high scalability and reliability of the orthogonal architecture, low cost, and simple functional maintenance, a highly reliable, low-latency, and low-maintenance Mesh network is realized, which can meet the needs of large-scale networking.

[0050] The solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0051] The solution provided in this application can be applied to a communication system including multiple nodes. The multiple nodes form a Mesh network to realize the functions of the communication system.

[0052] Among them, the node can be a chip or functional module deployed on a single board, the communication system can be a network device, and the network device can be a router or a switch, or other forms, which is not limited in the embodiments of the present application.

[0053] The following embodiments of the present application describe the solution provided by the present application from the internal structure of the network device.

[0054] 5 , an embodiment of the present application provides a network device, which may include N single boards 501 and Q switch boards 502. M and N are greater than or equal to 2.

[0055] As shown in FIG5 , M nodes 5011 and Q orthogonal connectors 5012 are deployed on a single board 501 .

[0056] The M nodes 5011 in a single board 501 are bidirectionally connected (T+R connection) to M different orthogonal connectors 5012 among the Q orthogonal connectors 5012 in the single board 501. Every two nodes in the M nodes 5011 in a single board 501 are connected via at least one orthogonal connector 5012.

[0057] In one possible implementation, as shown in FIG5 , Q is M 2 , every two nodes among the M nodes on a single board 501 are connected through at least one orthogonal connector, including: every two nodes among the M nodes on a single board 501 are connected through two different orthogonal connectors for sending and receiving unidirectional connections.

[0058] Among them, the unidirectional connection in the sending direction is a T connection, and the connection in the receiving direction is an R connection.

[0059] N orthogonal connectors 5021 are deployed on a switch board 502. Each of the N orthogonal connectors 5021 deployed on a switch board 502 is connected to each other (not shown in FIG5 ). The N orthogonal connectors 5021 on a switch board 502 are connected to the orthogonal connector 5012 at the same position on each of the N single boards 501.

[0060] In one possible implementation, as shown in FIG6a , Q is M, and every two of the M nodes on a single board 501 are connected via at least one orthogonal connector. This includes: every two of the M nodes on a single board 501 are bidirectionally connected via one orthogonal connector. In this case, Q can be M(M+1) / 2.

[0061] Of course, Q can also be other values, which is not limited in the embodiments of the present application.

[0062] It should be noted that the units with the same structure in FIG. 5 or FIG. 6 a are the same units and are not marked one by one in FIG. 5 or FIG. 6 a .

[0063] Figure 6b illustrates the internal architecture of the board 501. In the board 501 shown in Figure 6b, every two nodes among the M nodes on the board 501 are connected in a unidirectional manner via two different orthogonal connectors. As shown in Figure 6b, among the M nodes 5011 deployed in a single board 501, any two nodes are recorded as the first node 5011-1 and the second node 5011-2. For the first node 5011-1 and the second node 5011-2, the first node 5011-1 and the second node 5011-2 send and receive unidirectional connections through two different orthogonal connectors 5012, specifically including: a unidirectional connection in the sending direction of the first node 5011-1 to the first orthogonal connector 5012-1 (connection a in Figure 6b), and a unidirectional connection in the sending direction of the first orthogonal connector 5012-1 to the second node 5011-2 (connection b in Figure 6b); a unidirectional connection in the sending direction of the second node 5011-2 to the second orthogonal connector 5012-2 (connection c in Figure 6b), and a unidirectional connection in the sending direction of the second orthogonal connector 5012-2 to the first node 5011-1 (connection d in Figure 6b).

[0064] The first orthogonal connector 5012 - 1 and the second orthogonal connector 5012 - 2 are two different orthogonal connectors in a single board 501 .

[0065] For example, in the network device shown in FIG5 , the connection between any two nodes 5011 in a single board 501 may be the same as the internal architecture shown in FIG6 b .

[0066] For example, in the single board 501, a unidirectional / bidirectional connection can be realized by software.

[0067] Specifically, in the single board 501, which orthogonal connector 5012 the node 5011 is connected to, and which orthogonal connector 5012 it is bidirectionally connected to, and which orthogonal connector 5012 it is unidirectionally connected to, can all be configured according to actual needs. The embodiment of the present application is not limited to this. In different single boards 501, the connection structure of the node 5011 and the orthogonal connector 5012 is the same.

[0068] For example, in a single board 501, an orthogonal connector 5012 that is unidirectionally connected between two nodes 5011 can be deployed between the orthogonal connectors that are bidirectionally connected between two nodes. In the internal structure of the single board 501 as shown in Figure 6b, the first orthogonal connector 5012-1 and the second orthogonal connector 5012-2 are located between the orthogonal connector 5012 that is bidirectionally connected between the first node 5011-1 and the second node 5011-2.

[0069] Furthermore, as shown in FIG5 or FIG6a, N single boards 501 are parallel to each other, Q switch boards 502 are parallel to each other, and the N single boards and the Q switch boards are orthogonally connected to form an orthogonal architecture.

[0070] It should be noted that in the network device illustrated in FIG5 or FIG6a, N single boards are located on one side of Q switch boards. This is for illustrative purposes only and does not constitute a specific limitation. In actual applications, N single boards may also be located on multiple sides of the Q switch boards, and this embodiment of the present application is not limited to this.

[0071] Furthermore, in the network device illustrated in FIG5 or FIG6a , the orthogonal connector 5012 on the single board 501 and the orthogonal connector 5021 on the switch board 502 are male and female. The N orthogonal connectors 5021 on a switch board 502 are pluggably connected to the orthogonal connector 5012 at the same position on each of the N single boards 501.

[0072] In the network device shown in FIG5 or FIG6a, N nodes 5011 at the same position on each of the N boards 501 are connected to form a Mesh sub-network through the same switching board 502, and the structures of different Mesh sub-networks are the same.

[0073] Specifically, since the connection structure between the node 5011 and the orthogonal connector 5012 in each single board 501 is the same, the nodes 5011 at the same position in each single board 501 are bidirectionally connected to the orthogonal connector 5012 at the same position; the orthogonal connectors 5012 at the same position on N single boards 501 are connected to the N orthogonal connectors 5021 in the same switching board 502, so that the nodes 5011 at the same position in the N single boards 501 are all connected to the same switching board 502; since the N orthogonal connectors 5021 in the switching board 502 are connected in pairs, the N nodes 5011 at the same position on each single board 501 in the N single boards 501 are connected to form a Mesh sub-network through the same switching board 502.

[0074] Through the network device shown in FIG5 or FIG6a, the M*N nodes deployed in the network device are connected to each other, that is, a Mesh network composed of M*N nodes is realized.

[0075] The following describes the connections in a Mesh network of M*N nodes using the network device shown in FIG5 as an example. The connections in a Mesh network of M*N nodes in the network device shown in FIG6a will not be described in detail.

[0076] 5 , any two nodes 5011 located in the same board 501 are connected via a unidirectional connection in the sending and receiving directions between the two nodes 5011 in the board 501 to achieve connection between the two nodes 5011 .

[0077] For any two nodes 5011 located on different boards 501 in the network device shown in Figure 5, they are recorded as the third node 5011-3 and the fourth node 5011-4. The following describes the specific content of the connection between the third node 5011-3 and the fourth node 5011-4 in the Mesh network of the network device shown in Figure 5 with reference to the accompanying drawings.

[0078] In one possible implementation, if the third node 5011-3 and the fourth node 5011-4 are located at the same position (referred to as the first position) in different boards 501, the connection for the third node 5011-3 to send data to the fourth node 5011-4 may be the link illustrated in the internal structure of the network device as shown in FIG7 , which may specifically include: a bidirectional connection between the third node 5011-3 and the third orthogonal connector 5012-3 in the board 501 to which it belongs, a plug-in connection between the third orthogonal connector 5012-3 and the fourth orthogonal connector 5021-4 on the first switch board 502-1, a connection between the fourth orthogonal connector 5021-4 and the fifth orthogonal connector 5021-5 on the first switch board 502-1, a plug-in connection between the fifth orthogonal connector 5021-5 and the sixth orthogonal connector 5012-6 on the board 501 to which the fourth node 5011-4 belongs, and a bidirectional connection between the sixth orthogonal connector 5012-6 and the fourth node 5011-4. The first switching board 502 - 1 is used for connecting the orthogonal connector 5012 bidirectionally connected to the nodes at the first position in different boards.

[0079] It should be noted that the connection for transmitting data from the fourth node 5011-4 to the third node 5011-3 is the same as the connection indicated by the dotted line in Figure 7 above when the third node 5011-3 and the fourth node 5011-4 are located at the same position in different boards 501.

[0080] In another possible implementation, if the third node 5011-3 and the fourth node 5011-4 are located at different positions in different boards 501, the connection for the third node 5011-3 to send data to the fourth node 5011-4 may be a link illustrated in the internal structure of the network device as shown in FIG8 , which may specifically include: a unidirectional connection in the sending direction between the third node 5011-3 and the seventh orthogonal connector 5012-7 at the second position in the board 501 to which it belongs, and a unidirectional connection in the sending direction between the seventh orthogonal connector 5012-7 and the seventh orthogonal connector 501 The second switch board 502-2 is used to connect the orthogonal connectors 5012 in the second position on different boards 501.

[0081] It should be noted that the connection for transmitting data from the fourth node 5011-4 to the third node 5011-3 can be implemented by the third switching board 502-3 (used to connect the orthogonal connector 5012 at the third position in different boards 501, where the third position is the position of the orthogonal connector 5012 for unidirectional transmission of the fourth node 5011-4) when the third node 5011-3 and the fourth node 5011-4 are located at different positions on different boards 501, and no further description will be given. Refer to the connection illustrated in FIG. 8 above.

[0082] As can be seen from Figure 7, nodes 5011 at the same position in different boards 501 are connected through the same switch board 502. As can be seen from Figure 8, nodes 5011 at different positions in different boards 501 are connected to the same switch board 502 by connecting to the orthogonal connector 5012 at the same position in different boards 501. Therefore, a switch board 502 in the network device realizes a mesh sub-network between N nodes, and then connects M nodes through the connection within the board 501. 2 Mesh sub-networks are connected to form a full-Mesh network of M*N nodes. It should be understood that through the architecture of the network device shown in Figure 5, the full-Mesh network of M*N nodes is split into M 2 Mesh sub-network (Mesh sub-network with N nodes). The Mesh network before and after the split is shown in Figure 9. The full-Mesh network with M*N nodes has many connections and is complex. 2 Mesh sub-networks reduce network granularity and provide flexible networking.

[0083] For example, the network device shown in FIG5 implements M 2 The structure of a Mesh sub-network can be shown in Figure 10. As shown in Figure 10, the network architecture includes M 2 A switching board 502 and its connected orthogonal connector 5012 and node 5011 constitute a sub-network plane corresponding to the switching board 502 .

[0084] For the switching board 502 of the orthogonal connector 5012 connected to the fourth position (the position of the orthogonal connector 5012 bidirectionally connected to the node 5011) in each single board 501, the corresponding sub-network plane is: the Mesh sub-network bidirectionally connected to the node 5011 at the fourth position.

[0085] For example, the subnetwork plane kM illustrated in FIG10 is the mesh subnetwork between nodes kM in each board 501. The subnetwork plane kM+1 illustrated in FIG10 is the mesh subnetwork between nodes kM+1 in each board 501. The subnetwork plane kM+M-1 illustrated in FIG10 is the mesh subnetwork between nodes kM+M-1 in each board 501. Here, k = 0, 2, 3, ..., N-1.

[0086] By analogy, the sub-network planes shown in FIG10 are not described one by one.

[0087] For the switching board 502 of the orthogonal connector 5012 connected to the fifth position (the position of the orthogonal connector 5012 where the node 5011 is unidirectionally connected in the sending direction) in each single board 501, the corresponding sub-network plane is: the Mesh sub-network between the node 5011 connected to the fifth position in the sending direction and the node 5011 connected to the fifth position in the receiving direction.

[0088] Exemplarily, the sub-network plane kM->kM+1 illustrated in FIG10 is a Mesh sub-network in which the node kM in each board 501 sends data to the node kM+1 in each board 501 .

[0089] The sub-network plane kM<-kM+1 shown in Figure 10 is a Mesh sub-network in which the node kM+1 in each board 501 sends data to the node kM in each board 501. Here, k=0, 2, 3, ..., N-1.

[0090] The solution provided in this application is described in detail below through specific examples.

[0091] Assuming that the network device includes N=16 boards and each board includes M=2 nodes, the mesh network of M*N=32 nodes can be split into four mesh sub-networks each including 16 nodes.

[0092] The connection structure of the network device may be shown in Figure 11. The principle of Figure 11 is the same as that of Figure 5 and will not be repeated here.

[0093] The connection of the Mesh sub-network in the network device can be shown in Figure 12. The four sub-network planes shown in Figure 12 are:

[0094] Subnetwork plane 2j: Mesh subnetwork between nodes 2j on each board, where j = 0, 2, 3, ..., 15.

[0095] Subnetwork plane 2j->2j+1: A mesh subnetwork in which node 2j on each board sends data to node 2j+1 on each board.

[0096] Subnetwork plane 2j<-2j+1: Node 2j+1 on each board, a mesh subnetwork that sends data to node 2j on each board.

[0097] Subnetwork plane 2j+1: Mesh subnetwork between nodes 2j+1 on each board.

[0098] It should be noted that the above embodiments are examples of the structure of the network device and do not constitute a specific limitation.

[0099] The present application provides a communication system, which includes at least one network device described in the above embodiment.

[0100] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented using software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the process or function described in the embodiments of the present application is performed in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device or other programmable device. The computer program or instruction can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer program or instruction can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, such as a floppy disk, a hard disk, or a tape; it can also be an optical medium, such as a digital video disc (DVD); it can also be a semiconductor medium, such as a solid state drive (SSD). The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present application, and such modifications or substitutions should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A network device, characterized in that: include: N boards, M nodes and Q orthogonal connectors are deployed on one board, the M nodes are bidirectionally connected to M different orthogonal connectors among the Q orthogonal connectors, and every two nodes among the M nodes are connected by at least one orthogonal connector; M and N are greater than or equal to 2, and Q is less than or equal to M 2 ; There are Q switching boards, N orthogonal connectors are deployed on one switching board, two of the N orthogonal connectors deployed on one switching board are connected, and the N orthogonal connectors on one switching board are respectively connected to the orthogonal connectors at the same position on each of the N single boards.

2. The network device according to claim 1, characterized in that: Q is M 2 ; Every two of the M nodes are connected via at least one orthogonal connector, including: every two of the M nodes are connected via two different orthogonal connectors for sending and receiving unidirectional connections.

3. The network device according to claim 2, characterized in that: For a first node and a second node, the first node and the second node are any two nodes deployed on one of the single boards, and the first node and the second node are unidirectionally connected through two different orthogonal connectors for transmission and reception, including: The first node sends a unidirectional connection in a direction to the first orthogonal connector, and the first orthogonal connector sends a unidirectional connection in a direction to the second node; The second node sends a unidirectional connection in a direction to the second orthogonal connector, and the second orthogonal connector sends a unidirectional connection in a direction to the first node; Wherein, the first orthogonal connector is different from the second orthogonal connector.

4. The network device according to claim 2 or 3, characterized in that: In one of the single boards, an orthogonal connector for unidirectional connection between two nodes is disposed between the orthogonal connectors for bidirectional connection between the two nodes.

5. The network device according to any one of claims 1 to 4, characterized in that: The N single boards are parallel to each other, the Q switch boards are parallel to each other, and the N single boards are orthogonally connected to the Q switch boards.

6. The network device according to any one of claims 1 to 5, characterized in that: The orthogonal connectors in the single board and the switch board are male and female to each other; The N orthogonal connectors on one of the switch boards are respectively connected to the orthogonal connectors at the same position on each of the N boards, including: The N orthogonal connectors on one of the switch boards are plug-connected with the orthogonal connectors at the same position on each of the N single boards.

7. The network device according to any one of claims 1 to 6, characterized in that: The N nodes at the same position on each of the N boards are connected to form a Mesh sub-network through the same switching board, and the structures of different Mesh sub-networks are the same.

8. The network device according to any one of claims 1 to 7, characterized in that: In different boards, the connection structure between the node and the orthogonal connector is the same.

9. The network device according to claim 2, characterized in that: The M*N nodes deployed in the network device are connected in pairs; For a third node and a fourth node, the third node and the fourth node are two nodes located on different boards in the network device, and the connection in which the third node sends data to the fourth node includes: The bidirectional connection between the third node and the third orthogonal connector in the single board to which it belongs, the plug-in connection between the third orthogonal connector and the fourth orthogonal connector on the first exchange board, the connection between the fourth orthogonal connector and the fifth orthogonal connector on the first exchange board, the plug-in connection between the fifth orthogonal connector and the sixth orthogonal connector on the single board to which the fourth node belongs, and the bidirectional connection between the sixth orthogonal connector and the fourth node; wherein the third node and the fourth node are located at the first position in different single boards, and the first exchange board is used to connect the orthogonal connectors bidirectionally connected to the nodes at the first position in different single boards; or, The third node is connected to the seventh orthogonal connector at the second position of the single board to which it belongs, the unidirectional connection in the sending direction, the plug-in connection between the seventh orthogonal connector and the eighth orthogonal connector on the second switching board, the connection between the eighth orthogonal connector and the ninth orthogonal connector on the second switching board, the plug-in connection between the ninth orthogonal connector and the tenth orthogonal connector at the second position on the single board to which the fourth node belongs, and the unidirectional connection in the sending direction between the tenth orthogonal connector and the fourth node; wherein, The third node and the fourth node are located at different positions in different boards, and the second switching board is used to connect the orthogonal connectors at the second positions in different boards.

10. A communication system, characterized in that: The method comprises at least one network device according to any one of claims 1 to 9.

Images