Multi-node server, out-of-band management controller, and data forwarding method

Abstract

The application discloses a multi-node server, an out-of-band management controller and a data forwarding method. The multi-node server comprises a first node, the first node comprises a BMC, a first PHY chip and a network port, the network port is connected to the first PHY chip, and the first PHY chip is connected to the BMC. The BMC is configured to interact with an external network connected to the network port through the first PHY chip and the network port. By arranging a network signal hardware link between the first PHY chip and the BMC, when a node of the external network accesses the BMC, the access data can reach the BMC after passing through the network port and the first PHY chip, and the path is shorter. Meanwhile, the signal of the first PHY chip can be forwarded to a switching chip through the BMC. In the case that the first PHY chip of all nodes of the multi-node server is connected to the BMC, a switching chip for managing the signal from the first PHY chip is not needed, thereby saving cost and simplifying the path for accessing the BMC of the node in the multi-node server.

Description

Technical Field

[0001] This application relates to the field of server technology, specifically to a multi-node server, an out-of-band management controller, and a data forwarding method. Background Technology

[0002] As the demand for server computing power increases, intensive servers have emerged. Multiple nodes can be deployed on intensive servers, and each node can perform computing as an independent server. These types of servers can also be called multi-node servers.

[0003] In current multi-node server solutions, data exchange between nodes is typically achieved through two switching chips on a management board, referred to as the first switching chip and the second switching chip. Specifically, the first switching chip manages data from the physical layer (PHY) chips of each node, while the second switching chip manages data from the media access control (MAC) chip of each node's baseboard management controller (BMC). When a node needs to access its own or another node's BMC, the access data must pass through the first PHY chip, the first switching chip, and the second switching chip of that node before reaching the MAC chip of the target node's BMC, resulting in a relatively complex path. Summary of the Invention

[0004] This application provides a multi-node server, an out-of-band management controller, and a data forwarding method to simplify the access path to the BMC of a node in a multi-node server, making it easier to access the BMC of each node in the multi-node server.

[0005] In a first aspect, this application provides a data forwarding method applied to a multi-node server. The multi-node server includes a first node, a second node, and a switching chip. The switching chip is connected to the BMC of the first node and the BMC of the second node, respectively. Both the first node and the second node include a first PHY chip and a network port. The network port is connected to the first PHY chip of the node, and the first PHY chip is connected to the BMC of the node. The method includes:

[0006] The BMC of the first node receives the first data through the first PHY chip and the network port of the first node;

[0007] If the destination IP address of the first data indicates that the target node includes the first node, then the BMC of the first node processes the first data; and / or,

[0008] If the target node does not include the first node, the BMC of the first node forwards the first data to the switching chip; if the target node includes the second node, or if the target node does not include any nodes in the multi-node server, and the node in the multi-node server closest to the target node is the second node, the switching chip forwards the first data to the BMC of the second node; if the target node includes the second node, the BMC of the second node processes the first data; and / or, if the target node does not include the second node, the BMC of the second node forwards the first data to the target node through the first PHY chip and the network port of the second node.

[0009] In this application, by setting up a network signal hardware link between the first PHY chip and the BMC of a node in a multi-node server to connect the first PHY chip and the BMC, when a node in an external network accesses the BMC of the first node through the network port of the first node, the first data only needs to pass through the network port and the first PHY chip to reach the BMC of the first node; when a node in an external network accesses the BMC of the second node through the network port of the first node, the first data passes through the first PHY chip, the BMC, and the switching chip of the first node to reach the BMC of the second node, without the need to set up a switching chip for managing the signal from the first PHY chip. This application can save costs, simplify the path to access the BMC, and improve communication efficiency.

[0010] Secondly, this application provides a data forwarding method applied to an out-of-band management controller (BMC) of a first node, wherein the first node is any node in a multi-node server, the first node includes a first PHY chip and a network port, the network port being connected to the physical layer PHY chip of the first port, and the first PHY chip being connected to the BMC; the method includes:

[0011] The system receives first data through the network port and the first PHY chip; if the destination IP address of the first data indicates that the target node includes the first node, the system processes the first data; and / or, if the target node does not include the first node, the system forwards the first data to the target node.

[0012] In one possible implementation, the multi-node server further includes a switching chip; the BMC includes a first media access control (MAC) chip and a second MAC chip; receiving first data through the first PHY chip and the network port includes: the first MAC chip receiving the first data passing through the network port and the first PHY chip; forwarding the first data to the target node includes: establishing a first virtual local area network (VLAN) through the first MAC chip and the second MAC chip; controlling the first MAC chip to send the first data to the second MAC chip through the first VLAN; controlling the second MAC chip to send the first data to the switching chip, so that the switching chip forwards the first data to the target node.

[0013] The first MAC chip is connected to the first PHY chip, and the second MAC chip is connected to the switching chip.

[0014] In this application, by setting the first MAC chip and the second MAC chip in the BMC of the first node, and using the first MAC chip and the second MAC chip to form a first VLAN to transmit the first data, the first data from the external network can directly reach the switching chip through the first PHY chip and the BMC, without the need to set up an additional switching chip to manage the signals from the first PHY chip of each node.

[0015] In another possible implementation, the BMC includes a first media access control (MAC) chip and a second MAC chip; the BMC can receive the first data through the first MAC chip and then send the same data to the switching chip through the second MAC chip.

[0016] In one possible implementation, before controlling the second MAC chip to send the first data to the switching chip, the method further includes: attaching a VLAN tag to the first data; controlling the second MAC chip to send the first data to the switching chip includes: controlling the second MAC chip to send the first data with the VLAN tag attached to it to the switching chip.

[0017] In this application, by attaching VLAN tags to the first data and transmitting different first data in different VLANs, broadcast domains can be isolated, thereby enhancing the security and robustness of the internal network.

[0018] In one possible implementation, after receiving the first data through the first PHY chip and the network port, the method further includes: discarding the first data if the target node does not include the first node.

[0019] Thirdly, this application provides a data forwarding method, which is applied to an out-of-band management controller (BMC) of a first node, wherein the first node is any node in a multi-node server, the first node includes a first port physical layer PHY chip and a network port, the network port being connected to the first PHY chip, and the first PHY chip being connected to the BMC; the multi-node server further includes a switching chip, the switching chip being connected to the BMC; the method includes:

[0020] The system receives second data forwarded by the switching chip; if the destination IP address of the second data indicates that the target node includes the first node, the system processes the second data; and / or, if the target node does not include the first node, the system forwards the second data to the target node through the first PHY chip and the network port.

[0021] In one possible implementation, the BMC includes a first media access control MAC chip and a second MAC chip; receiving the second data forwarded by the switching chip includes: the second MAC chip receiving the second data; forwarding the second data to the target node through the first PHY chip and the network port includes: establishing a first virtual local area network (VLAN) through the first MAC chip and the second MAC chip; controlling the second MAC chip to send the second data to the first MAC chip through the first VLAN; controlling the first MAC chip to send the second data to the first PHY chip, so that the first PHY chip forwards the second data to the target node through the network port.

[0022] In one possible implementation, the second data is data sent to the switching chip by the second node after attaching a VLAN tag; before forwarding the second data to the target node through the first PHY chip and the network port, the method further includes: deleting the VLAN tag from the second data; forwarding the second data to the target node through the first PHY chip and the network port includes: forwarding the second data after deleting the VLAN tag to the target node through the first PHY chip and the network port.

[0023] In one possible implementation, the second data includes indication information about the target node attached by the switching chip, which is obtained by the switching chip based on its own routing table and the destination IP address; after receiving the second data forwarded by the switching chip, the method further includes: determining whether the target node is the first node based on the indication information; forwarding the second data to the target node through the first PHY chip and the network port includes: forwarding the second data to the next-hop IP address of the target node in the indication information through the first PHY chip and the network port.

[0024] In this application, the target node is confirmed to be the first node by identifying the destination IP address of the second data through the switching chip, and the indication information is obtained according to the identification result. Finally, the second data with the indication information attached is sent to the first node, so that the first node does not need to identify the destination IP address and can directly process or forward according to the indication information.

[0025] In one possible implementation, after receiving the second data forwarded by the switching chip, the method further includes discarding the second data if the target node does not include the first node.

[0026] Fourthly, this application also provides a multi-node server, which includes a first node, which is any node in the multi-node server. The first node includes a network port, an out-of-band management controller (BMC), and a first port physical layer (PHY) chip. The network port is connected to the first PHY chip, and the first PHY chip is connected to the BMC. The BMC is used to perform data interaction with an external network connected to the network port through the first PHY chip and the network port.

[0027] In one possible implementation, the multi-node server further includes a switching chip, and the BMC further includes a first media access control (MAC) chip and a second MAC chip; the first MAC chip is connected to the first PHY chip, and the second MAC chip is connected to the switching chip; the BMC is also used to form a first virtual local area network (VLAN) through the first MAC chip and the second MAC chip, and to control the first MAC chip to perform data interaction with the second MAC chip through the first VLAN.

[0028] Fifthly, this application also provides an out-of-band management controller for performing the methods described in the second or third aspect.

[0029] In a sixth aspect, this application also provides a computer-readable storage medium including instructions that, when executed on a computer, cause the computer's BMC to perform the methods described in the second or third aspect.

[0030] It should be understood that the beneficial effects of the above aspects can be referenced from each other. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the structure of a multi-node server provided in an embodiment of this application;

[0032] Figure 2 This is a schematic diagram of the structure of another multi-node server provided in an embodiment of this application;

[0033] Figure 3A flowchart illustrating a data forwarding method provided in an embodiment of this application;

[0034] Figure 4 A flowchart illustrating a data forwarding method provided in an embodiment of this application;

[0035] Figure 5 This is a flowchart illustrating another data forwarding method provided in an embodiment of this application. Detailed Implementation

[0036] The embodiments of this application are described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. As those skilled in the art will understand, with the development of technology and the emergence of new scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0037] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0038] In the management scenario of multi-node servers, the management terminal needs to access the multi-node server frequently to monitor and control each node in the multi-node server. For example, it can send query commands to a certain node in the multi-node server to query the central processing unit (CPU) information, query memory information, query device information, or control commands such as powering on, powering off, adjusting hardware power, and ending processes.

[0039] In the existing multi-node server framework, after the command from the management end enters the multi-node server, it needs to pass through a first switching chip that manages the data from the port physical layer (PHY) chips of each node, and a second switching chip that manages the data from the media access control (MAC) chips of the out-of-band management controller (BMC) of each node, before it can reach the BMC of the target node. The access path is relatively complex.

[0040] To address the issue of complex access paths, embodiments of this application provide a multi-node server, an out-of-band management controller, and a data forwarding method to simplify the path to access the BMC of a node in the multi-node server, making it easier for the management terminal to manage the multi-node server.

[0041] It is understood that the management terminal refers to the computing device used to manage and control the multi-node server, such as the main control equipment in the computer room where the multi-node server is located, or a smart terminal near the management personnel. Specifically, the management terminal can be a server, or a personal computer (PC), laptop computer, or tablet computer, etc. The embodiments of this application do not limit the specific form of the management terminal.

[0042] Please see Figure 1 , Figure 1 This is a schematic diagram of the structure of a multi-node server provided in an embodiment of this application. Figure 1 As shown, the multi-node server 100 includes a switching chip 110 and N nodes, where N is greater than 1. Figure 1 In the specific example, each of the N nodes includes a network port, a first port physical layer (PHY) chip, and an out-of-band management controller (BMC).

[0043] The Baseboard Management Controller (BMC) manages the various components within a node, monitoring their operational status and acquiring real-time operating parameters. It's understood that the BMC is also known as the Baseboard Management Controller; different server manufacturers use different names for the BMC, such as iBMC, iLO, and iDRAC.

[0044] Therefore, when the management terminal needs to obtain the running status and related parameters of each node in the multi-node server 100, or to issue management commands to manage and control each node, it needs to access each node through the network.

[0045] Each node's BMC integrates a first media access control (MAC) chip and a second MAC chip. Both the first and second MAC chips can be controlled and managed by the BMC. For example, the BMC can route the first and second MAC chips together to form a virtual local area network (VLAN); or, the BMC can control the data transmission and reception of the first and second MAC chips.

[0046] Each node's network port connects its first PHY chip to an external network, allowing it to receive data from the external network and transmit that data back to the node's first PHY chip. Specifically, this network port connects to the medium dependent interface (MDI) protocol interface of the first PHY chip, and data transmission between the network port and the first PHY chip occurs via the MDI protocol. This network port is the BMC's network port.

[0047] Understandably, this external network could be the Internet or a local area network with 100 connected nodes.

[0048] Optionally, the network port is an RJ45 connector.

[0049] In this configuration, the network signal hardware link on the motherboard of each node connects its own network port, first PHY chip, and first MAC chip, allowing data from the external network to directly reach the first MAC chip through the network port and the first PHY chip. In other words, the BMC can interact with the external network connected to the network port through its own first PHY chip and network port.

[0050] The first PHY chip can be used to convert the communication protocol of the received data. Specifically, it can convert MDI protocol format data into Media Independent Interface (MII) protocol format data, and then send the converted data to the first MAC chip of the local node's BMC. That is, the first PHY chip and the first MAC chip transmit data through the MII protocol. The MII protocol can be either the Reduced Gigabit Media Independent Interface (RGMII) protocol or the Serial Gigabit Media Independent Interface (SGMII) protocol.

[0051] The first MAC chip serves as the external management network port of its BMC. It can receive data sent from the external network through the first PHY chip of this node, and it can also send data to the external network through the first PHY chip.

[0052] Specifically, after the first MAC chip receives data sent by the first PHY chip of this node, the BMC where the first MAC chip is located can identify the destination Internet Protocol (IP) address of the data. When the destination IP address of the data indicates that the target node includes this node, the BMC can directly process the data. When the target node does not include this node, the BMC can route the first MAC chip and the second MAC chip of this node to form a first virtual local area network (VLAN), and then control the first MAC chip to directly forward the data to the second MAC chip through the first VLAN for internal forwarding of the multi-node server 100.

[0053] In this context, the target node refers to the node device pointed to by the destination IP address of the communication data. When the communication data is unicast data, the target node is a single node; when the communication data is multicast data, the target node is a member node in the corresponding group. Therefore, when the destination IP address of the data received by the first MAC chip is a multicast IP address, and the member nodes in the group corresponding to the multicast IP address include the node where the first MAC chip is located, it can be considered that the target node indicated by the destination IP address of the data includes the node where the first MAC chip is located.

[0054] It is understandable that when a node in a multi-node server is the receiver of multicast data, the corresponding multicast address is pre-stored in the node's BMC so that it can identify and process the corresponding multicast message when it receives it.

[0055] The BMC can have a pre-set routing table, which includes the IP address of the second MAC chip of this node. This IP address is assigned to the second MAC chip by the BMC. After the first VLAN is established, the BMC can control the first MAC chip to send the data to the second MAC chip based on this IP address.

[0056] It is understandable that there may be MAC chips other than the first MAC chip and the second MAC chip on the BMC. Therefore, the BMC needs to record the IP addresses assigned to each MAC chip through a routing table. When two or more MAC chips need to communicate with each other, the corresponding IP addresses can be selected through the routing table to form a virtual local area network for data transmission.

[0057] Optionally, the routing table can also be preset in the first MAC chip. When the first MAC chip recognizes that the destination IP address of the data from the first PHY chip is not the IP address of this node, or receives a message from the BMC that the destination IP address of the data is not the IP address of this node, or receives an instruction from the BMC to forward the data to the second MAC chip, the first MAC chip can find the IP address of the second MAC chip according to the routing table and forward it through the first VLAN.

[0058] It is understandable that after completing the data exchange between the first MAC chip and the second MAC chip, the BMC can delete the first VLAN, release the corresponding resources, and reconstruct the first VLAN when data exchange between the first MAC chip and the second MAC chip is required again in the future.

[0059] The second MAC chip, as the internal forwarding network port of its BMC, can receive data forwarded by the switching chip 110 and can also send data to the switching chip 110. For example, it can send data to other nodes of the multi-node server 100 through the switching chip 110.

[0060] Specifically, when the second MAC chip sends data to the switching chip 110, it can attach a VLAN tag to the data so that the switching chip 110 can transmit data through the VLAN indicated by the VLAN tag.

[0061] The switching chip 110 can be used to receive data sent by the second MAC chip of each node of the multi-node server 100, and forward the data according to the information indicating the forwarding object in the data.

[0062] Optionally, the information indicating the forwarding target can be the VLAN tag of the data. The switching chip 110 can obtain the VLAN tag of the data and then forward the data to all nodes in the VLAN indicated by the VLAN tag, or forward it to all VLAN member nodes in the VLAN except for the data source node.

[0063] Optionally, the information indicating the forwarding target can be the destination IP address of the data. The switching chip 110 can identify the destination IP address of the data; then determine the target node based on the destination IP address; when the target node is a node in the multi-node server 100, the switching chip 110 can forward the data to the second MAC chip of the target node; when the target node is a node device outside the multi-node server 100, the switching chip 110 can forward the data to the second MAC chip of the node in the multi-node server 100 closest to the target node.

[0064] The switching chip 110 is equipped with a routing table, which can store the multicast IP addresses corresponding to the multicast data to be received by each node in the multi-node server. Therefore, when the switching chip 110 receives multicast data, it can determine the node in the multi-node server that needs to receive the multicast data as the forwarding object based on the multicast IP address of the multicast data, and then perform the corresponding forwarding.

[0065] The following two examples illustrate the working principle of the multi-node server 100 by accessing the BMC node from the management terminal.

[0066] 1. The management terminal accesses the BMC of node 1 through the network port of node 1.

[0067] The management terminal sends management commands to network port 1 of node 1 via Ethernet. After receiving the management commands, network port 1 sends the management commands to the MDI protocol interface of the first PHY chip 1 of node 1 via the MDI protocol. The first PHY chip 1 converts the management commands from the MDI protocol format to the MII protocol format and sends the converted management commands to the first MAC chip 1 of the BMC1 of node 1 via the MII protocol. After the BMC recognizes that the destination IP address of the management commands is the IP address of node 1, it directly executes the corresponding operation according to the management commands.

[0068] It is understandable that when BMC1 needs to return feedback information according to the management instruction, the feedback information will be returned to the management terminal through the above access path.

[0069] 2. The management terminal accesses the BMC of node 2 through the network port of node 1.

[0070] Please refer to Figure 2 This application also provides a more detailed structural diagram of a multi-node server 100 in this embodiment. For example... Figure 2 As shown, in Figure 1 Based on the multi-node server 100 shown, a second PHY chip is provided between the switching chip 110 and the second MAC chip of each node; the switching chip 110 specifically includes multiple third MAC chips, multiple third PHY chips, a controller and a bus, the multiple third MAC chips and the controller are connected through the bus, and each third PHY chip is connected to a third MAC chip; each second PHY chip is connected to the second MAC chip of a node and one of the third PHY chips in the switching chip 110.

[0071] exist Figure 2In a specific example, the multi-node server 100 includes four nodes and four second PHY chips, while the switching chip 110 includes five third PHY chips and six third MAC chips. It is understood that the multi-node server 100 is housed in a chassis containing a management board, which integrates the switching chip 110 and the second PHY chips. When nodes are to form a multi-node server 100, they need to be inserted into the chassis and connected to the switching chip 110 via the second PHY chips on the management board, thereby enabling data interaction with other nodes of the multi-node server 100.

[0072] Therefore, the number of second PHY chips in the multi-node server 100 is greater than or equal to the number of nodes, or in other words, the number of second PHY chips is equal to the maximum number of nodes in the multi-node server 100.

[0073] The number of the third PHY chip and the third MAC chip in the switching chip 110 is determined at the time of manufacture. The number of the third PHY chips in the switching chip 110 selected and configured in the multi-node server 100 can be greater than or equal to the number of the second PHY chips on the same management board, so that all nodes of the multi-node server 100 are connected to the same switching chip 110. For example Figure 2 As shown, the switching chip 110 is equipped with 5 third PHY chips, allowing a maximum of 5 nodes to connect to the switching chip 110 simultaneously; the number of third MAC chips can be greater than or equal to the number of third PHY chips, such as... Figure 2 As shown, the third MAC chip 6 is used to receive data from the non-PHY chip.

[0074] exist Figure 2 In a specific example, the second PHY chip is used to convert the communication protocol of the received data, converting data in SGMII protocol format to data in MDI protocol format; the third PHY chip is also used to convert the communication protocol of the received data, converting data in MDI protocol format to data in SGMII protocol format. Specifically, the second MAC chip and the second PHY chip communicate via SGMII, the second PHY chip and the third PHY chip communicate via MDI protocol, and the third PHY chip and the third MAC chip communicate via SGMII protocol (not shown in the figure). In other words, the second and third PHY chips are used to perform protocol conversion on the communication data between the second MAC chip and the switching chip 110, ensuring the communication quality between the second MAC chip and the switching chip 110.

[0075] It is understandable that the process of the management command sent by the management terminal from the management terminal to the first MAC chip 1 of node 1 is similar to the embodiment in the previous embodiment where the management terminal accesses the BMC1 of node 1, and will not be described again here.

[0076] After the first MAC chip 1 of node 1 receives the management command, the BMC1 of node 1 can identify that the destination IP address of the management command is the IP address of node 2, that is, not the IP address of node 1. Therefore, the BMC1 can control the first MAC chip 1 to forward the management command to the second MAC chip 1 of node 1. The BMC1 can add a VLAN tag to the data frame of the management command and control the second MAC chip 1 to forward the management command to the third MAC chip 1 through the second PHY chip 1 and the third PHY chip 1.

[0077] The VLAN tag indicates that the member nodes of the VLAN may include nodes 1 to 4.

[0078] The third MAC chip 1 sends the management command to the controller via the bus; after receiving the management command, the controller can execute one of the following three branch schemes:

[0079] 1. The controller can determine that all nodes 2 to 4 of the VLAN network members other than node 1 are forwarding objects based on the VLAN tag of the management command, and then send the management command to the third MAC chip 2 to 4 through the bus; the third MAC chip 2 to 4 forwards the management command to the second MAC chip 2 to 4 through the third PHY chip 2 to 4 and the second PHY chip 2 to 4.

[0080] In this branch scheme, after the second MAC chips 2 to 4 receive the management instruction forwarded by the switching chip 110, they each identify whether the destination IP address of the management instruction is the IP address of their own node. Since the destination IP address of the management instruction is the IP address of node 2, BMC2 of node 2 directly processes the management instruction. BMC3 and BMC4 of nodes 3 and 4 can discard the management instruction if they determine that the destination IP address of the management instruction is not the IP address of their own node.

[0081] 2. The controller determines Node 2 as the target node based on the destination IP address of the management command, and then sends the management command to the third MAC chip 2 via the bus; the third MAC chip 2 forwards the management command to the second MAC chip 2 of Node 2 through the third PHY chip 2 and the second PHY chip 2.

[0082] In this branch scheme, after the second MAC chip 2 receives the management instruction forwarded by the switching chip 110, if the forwarding strategy of the switching chip 110 is to only forward data whose target node is node 2 to the second MAC chip 2, then the BMC2 can directly process the management instruction; if the forwarding strategy of the switching chip 110 is to forward the corresponding data to the second MAC chip 2 when the target node is node 2, or when the node closest to the target node in the multi-node server is node 2, then the BMC2 can identify the destination IP address of the management instruction and verify whether the target node of the management instruction is this node.

[0083] In this embodiment, after the second MAC chip 2 receives the management instruction, since the destination IP address of the management instruction is the IP address of node 2, node 2 can directly process the management instruction.

[0084] 3. The controller determines that Node 2 is the target node based on the destination IP address of the management instruction. At the same time as forwarding the management instruction to the second MAC chip 2 of Node 2, it attaches indication information about the target node to inform Node 2 that the target node of the management instruction is Node 2.

[0085] In this branch scheme, after the second MAC chip 2 receives the management instruction forwarded by the switching chip 110, the BMC2 can directly process the management instruction based on the instruction information.

[0086] It is understandable that when Node 2's BMC2 needs to return feedback information to the management terminal according to the management command, Node 2 can return the feedback information according to the above path, or it can send the feedback information to Node 2's first PHY chip 2 through the first MAC chip 2; the first PHY chip 2 forwards the feedback information to the external network through the network port 2, thereby returning the feedback information to the management terminal.

[0087] It is understandable that not all nodes in the multi-node server 100 need to be equipped with network ports. The management terminal can manage and control all nodes in the multi-node server 100 through the network ports of one or more nodes. For example, when node 2 is not equipped with a network port, node 2's BMC2 needs to return feedback information to the management terminal through the access path in this embodiment.

[0088] It should be noted that, in the specific implementation, the multi-node server 100 can be any of the following: Figure 1 Devices with similar structures. This application does not limit the specific composition of the multi-node server 100 in its embodiments. Furthermore, Figure 1 The structural composition shown does not constitute a limitation on the multi-node server 100, except... Figure 1In addition to the components shown, the multi-node server 100 may include more or fewer components than illustrated, or combine certain components, or have different component arrangements. For example, in some other embodiments, one or more nodes in the multi-node server are equipped with a network port and a first PHY chip for communication with an external network, while other nodes are not equipped with the network port and the first PHY chip, and act as computing nodes to provide computing services to the multi-node server.

[0089] In this embodiment of the application, by setting up a network signal hardware link between the first PHY chip and the BMC of a node in a multi-node server to connect the first PHY chip and the BMC, when a node in an external network accesses the BMC of the first node through the network port of the first node, the first data only needs to pass through the network port and the first PHY chip to reach the BMC of the first node; when a node in an external network accesses the BMC of the second node through the network port of the first node, the first data passes through the first PHY chip and the BMC of the first node, as well as the switching chip, to reach the BMC of the second node, without the need to set up a switching chip for managing the signal from the first PHY chip. This application can save costs, simplify the path to access the BMC, and improve communication efficiency.

[0090] exist Figure 1 Based on the multi-node server shown, please refer to Figure 3 , Figure 3 This is a flowchart illustrating a data forwarding method provided in an embodiment of this application, applied to a multi-node server. The method includes steps 301 to 306.

[0091] The multi-node server includes a first node, a second node, and a switching chip. The switching chip is connected to the BMC of the first node and the BMC of the second node, respectively. Both the first node and the second node include a first PHY chip and a network port. The network port is connected to the first PHY chip of the node, and the first PHY chip is connected to the BMC of the node.

[0092] Step 301: The BMC of the first node receives the first data through the network port and the first PHY chip of the first node.

[0093] The first data can be data whose source Internet Protocol (IP) address is not the IP address of the first node, that is, data from other nodes or devices outside the first node.

[0094] The BMC of the first node and the BMC of the second node both include a first MAC chip and a second MAC chip. The first MAC chip is connected to the first PHY chip of the node, and the second MAC chip is connected to the switching chip.

[0095] The multi-node server is housed in a chassis, with each node located in a different slot within the chassis; alternatively, each node can be housed in its own chassis, and multiple chassis can be housed within a single server rack. The chassis also includes a management board, on which the multi-node server's switching chip is mounted. Each node connects to the switching chip via a second PHY chip on the management board, and the switching chip facilitates data communication between the nodes.

[0096] Before executing step 301, after the BMC is initialized, the BMC can configure the IP address of the first MAC chip as the IP address of the first node.

[0097] The first MAC chip serves as the external management network port of the BMC, and its IP address is used as the IP address of the BMC's external network port, as well as the IP address of the first node. When the management terminal accesses the BMC of a node in the multi-node server, it will use the IP address of the first node as the destination IP address.

[0098] When the second MAC chip of the first node is connected to the switching chip through the second PHY chip set on the management board, the BMC of the first node is connected to the management board. The BMC can obtain the slot information of the slot where the first node is located from the management board, and then configure the IP address for the first MAC chip according to the slot information.

[0099] Optionally, the slot information includes a pre-defined IP address; after obtaining the slot information, the BMC can directly configure the IP address of the first MAC chip to the pre-defined IP address. For example, if the chassis housing a multi-node server has 5 slots, the IP address of the first MAC chip of the node in slot 1 is pre-defined as 10.10.0.101, the IP address of the first MAC chip of the node in slot 2 is pre-defined as 10.10.0.102, and so on.

[0100] Optionally, the slot information includes a slot number, and the BMC can configure the IP address of the first MAC chip according to the slot number and preset configuration rules.

[0101] After the IP address of the first node is configured, the BMC can store the IP address in a preset routing table; when the first data is received, the BMC can identify the destination IP address of the first data; and determine whether the target node indicated by the destination IP address includes the first node itself.

[0102] The routing table can also store the multicast IP addresses corresponding to the multicast data that the first node needs to receive.

[0103] In one possible implementation, after the IP address of the first node is configured, the BMC can send the IP address of the first node and the multicast IP address to the first MAC chip and the second MAC chip, or in response to a request from the first MAC chip and the second MAC chip, send the IP address and the multicast IP address to the first MAC chip and the second MAC chip; after obtaining the IP address, the first MAC chip and the second MAC chip can save the IP address in their respective routing tables.

[0104] Optionally, when the first MAC chip or the second MAC chip receives the first data, the first MAC chip can obtain the destination IP address from the first data and compare the destination IP address with the IP address of the first node in its own routing table and the multicast IP address to confirm whether the target node corresponding to the first data includes the first node.

[0105] In the above scheme, when the destination IP address is the IP address of the first node, or when the destination IP address is the multicast IP address, and the target node includes the first node, the BMC of the first node executes step 302; when the destination IP address is neither the IP address of the first node nor the multicast IP address, and the target node does not include the first node, the BMC of the first node executes step 303.

[0106] Step 302: If the destination IP address of the first data indicates that the target node includes the first node, the BMC of the first node processes the first data.

[0107] When the first node BMC confirms that the first node is the target node of the first data, it can perform corresponding operations based on the content of the first data.

[0108] For example, when the first data is a notification message, the BMC can record the content of the notification message or modify the local data record according to the notification message; when the first data is a control command, the BMC can control the hardware in the first node according to the control command and adjust the relevant hardware parameters; when the first data is a query command, the BMC can obtain the latest hardware information and feed the hardware information back to the access end; when the first data is a synchronization command, the BMC can send multicast messages about synchronization data to other nodes in the multi-node server through the switching chip.

[0109] Step 303: If the target node does not include the first node, the BMC of the first node forwards the first data to the switching chip.

[0110] In this case, when the target node corresponding to the first data does not include the first node, the first node may be a transit node in the path of the first data to the target node; at this time, the BMC of the first node can forward the first data to the switching chip, and the switching chip decides whether to forward the first data further or to discard the first data.

[0111] In one possible implementation, after the first MAC chip of the first node receives the first data, the BMC of the first node can form a first VLAN through the first MAC chip and the second MAC chip of the BMC; then control the first MAC chip to forward the first data to the second MAC chip through the first VLAN; and then control the second MAC chip to forward the first data to the switching chip.

[0112] The BMC of the first node has a pre-set routing table, which includes the IP addresses assigned by the BMC to the first MAC chip and the second MAC chip. When the BMC needs to control the data interaction between the first MAC chip and the second MAC chip, the BMC can form a first VLAN based on their IP addresses and transmit data through the first VLAN.

[0113] In one possible implementation, the BMC first attaches a VLAN tag to the first data; then controls the second MAC chip to forward the first data with the attached VLAN tag to the switching chip, so that the switching chip transmits the first data in the VLAN indicated by the VLAN tag.

[0114] The VLAN indicated by the VLAN tag can be formed before the second MAC chip forwards the first data. The member nodes of the VLAN may include the second MAC chips of some nodes or all nodes in a multi-node server.

[0115] Specifically, the first node can request the master node in the multi-node server to establish the VLAN before the second MAC chip forwards the first data; or it can collaboratively establish the VLAN after communicating with other nodes. It is understood that the first node can determine the corresponding communication scenario based on the data type, data format, or communication protocol of the first data, thereby establishing different VLANs to transmit the first data to the target node corresponding to the first data.

[0116] For example, when the first data is a synchronization command of the Intelligent Platform Management Interface (IPMI), the synchronization command needs to be sent to other nodes in the multi-node server. At this time, the first node can request that the second MAC chips of all nodes in the multi-node server be divided into the same VLAN; and then broadcast the first data in the VLAN through the switching chip.

[0117] For example, when the first data is related to a certain service, the first node can request that the nodes related to the service in the multi-node server be assigned to the same VLAN, and then broadcast the first data in the VLAN through the switching chip.

[0118] In one possible implementation, after the second MAC chip forwards the first data to the switching chip, the BMC of the first node can delete the VLAN indicated by the VLAN tag to free up resources.

[0119] Understandably, when the first node is the master node in the multi-node server, the first node can create or delete VLANs on its own.

[0120] Step 304: If the target node includes the second node, or if the target node does not include any nodes in the multi-node server, and the second node is the node in the multi-node server that is closest to the target node, the switching chip forwards the first data to the BMC of the second node.

[0121] The second node can be any node in the multi-node server other than the first node.

[0122] In one possible implementation, the switching chip can identify the destination IP address of the first data, and then determine whether the target node includes nodes in the multi-node server based on the destination IP address; if the target node does not include nodes in the multi-node server, the switching chip can determine the node in the multi-node server that is closest to the target node based on its own routing table and the destination IP address.

[0123] In cases where the destination IP address matches the IP address of the second node, or the node indicated by the destination IP address includes the second node, the switching chip can determine that the target node includes the second node, and thus forward the first data to the second node.

[0124] In cases where the target node does not include any nodes in the multi-node server, the switching chip can determine the optimal path for the first data from the multi-node server to the target node indicated by the destination IP address based on its own routing table, and then determine the node closest to the target node from among the nodes belonging to the multi-node server based on the optimal path.

[0125] If the target node does not include any nodes in the multi-node server, and the node in the multi-node server that is closest to the target node is the second node, the switching chip can forward the first data to the second node.

[0126] When the switching chip cannot determine the target node in its own routing table, the switching chip can discard the first data.

[0127] In one possible implementation, the first data is the first data after the first node has been tagged with a VLAN tag; the switching chip can broadcast the first data in the VLAN indicated by the VLAN tag according to the VLAN tag.

[0128] By delegating the identification and routing steps to nodes in a multi-node server, the computing power of the switching chip can be saved, and the working efficiency of the switching chip can be improved.

[0129] In another possible implementation, the first data is the first data after the first node has been tagged with a VLAN tag; the switching chip can traverse the destination IP address of the first data from the list of IP addresses corresponding to the networking nodes of the VLAN indicated by the VLAN tag; if the destination IP address exists in the list of IP addresses, the switching chip can determine whether the node corresponding to the destination IP address includes the node in the multi-node server based on the traversal result.

[0130] When the traversal result is not empty, the switching chip can determine that the target node includes the node corresponding to the IP address in the traversal result.

[0131] When the traversal result is empty, the switching chip can determine in its own routing table that the node corresponding to the shortest path to the destination IP address is the node in the multi-node server that is closest to the target node.

[0132] When the switching chip cannot determine the second node, it can discard the first data.

[0133] Optionally, when forwarding the first data to the second node, the switching chip may attach indication information about the target node corresponding to the first data to the first data to inform the second node whether it is the target node of the first data; if the second node is not the target node, the switching chip may also attach the next-hop IP address of the first data after the second node to the target node to the indication information.

[0134] The indication information is obtained by the switching chip based on its own routing table and the destination IP address of the first data. Specifically, when the switching chip identifies that the target node includes a second node based on the destination IP address, the switching chip can obtain the indication information that the target node includes a second node; when the switching chip identifies that the target node does not include any node in the multi-node server based on the destination IP address and its own routing table, and the second node is the node in the multi-node server that is closest to the target node, the switching chip can further determine the IP address of the next-hop node to which the first data leads to the target node, and append the next-hop IP address and the information that the target node does not include a second node as indication information to the first data.

[0135] The switching chip can obtain the optimal path between the first data from the multi-node server to the target node based on its own routing table and the destination IP address of the first data, and determine the node in the multi-node server that is closest to the target node based on the optimal path; the IP address of the next-hop node leading to the target node is the next-hop IP address of the node closest to the target node in the optimal path.

[0136] Step 305: If the target node includes a second node, the BMC of the second node processes the first data.

[0137] In this process, the BMC of the second node receives the first data from the switching chip through the second MAC chip of the BMC.

[0138] In one possible implementation, after the BMC of the second node receives the first data, the BMC can identify the destination IP address of the first data and determine whether the corresponding target node includes the second node based on the destination IP address; if yes, the BMC of the second node directly processes the first data; if no, the BMC of the second node executes step 306.

[0139] In cases where the second node is the target node corresponding to the first data, or the nearest relay node leading to the target node, the switching chip will forward the first data to the second node. Therefore, after receiving the first data, the BMC of the second node needs to further confirm whether the second node is the target node or the nearest relay node based on the destination IP address of the first data, and then perform subsequent processing.

[0140] It is understandable that the distances between the first and second nodes and nodes in the external network can be different. The external network can use multi-node servers as intermediaries to obtain a better or optimal routing path.

[0141] Where the destination IP address of the first data is the same as the IP address of the second node, or the destination IP address is used as a multicast IP address to indicate that the target node includes the second node, the BMC of the second node can directly process the external data.

[0142] In one possible implementation, the first data includes indication information attached to the switching chip regarding the target node corresponding to the first data; when the indication information indicates that the second node is the target node, the BMC of the second node can directly process the first data.

[0143] In another possible implementation, when the target node corresponding to the first data includes the second node, the switching chip forwards the first data to the second node; in this case, the second node can directly process the first data after receiving it.

[0144] Specifically, the way the BMC of the second node processes the first data is similar to the way the BMC of the first node processes the first data in step 302, and will not be described again here.

[0145] Step 306: If the target node does not include the second node, the second node forwards the first data to the target node through the first PHY chip and network port of the second node.

[0146] In this process, the BMC of the second node can first form a first VLAN through the first MAC chip and the second MAC chip in the BMC, and then control the second MAC chip to forward the first data to the first MAC chip through the first VLAN; then control the first MAC chip to forward the first data to the target node in the external network through the first PHY chip and the network port of the second node.

[0147] Specifically, before the first MAC chip forwards the first data to the target node, the BMC of the second node can delete the VLAN tag from the first data and then forward the first data after deleting the VLAN tag to the target node.

[0148] Optionally, the BMC of the second node is configured with a routing table. The BMC can determine the IP address of the next-hop node on the route path for transmitting the first data to the target node based on the routing table and the destination IP address, and then forward the first data to that IP address.

[0149] Understandably, when the BMC of the second node cannot determine the next-hop node, the second MAC chip can return an error message to the source IP address of the second data, such as returning an IP address access failure, or discarding the external data.

[0150] In another possible implementation, the first data includes indication information about the target node corresponding to the first data attached to the switching chip; when the indication information indicates that the second node is not the target node, the BMC of the second node can obtain the next-hop IP address of the first data to the target node from the indication information, and forward the first data to the next-hop IP address through the first PHY chip and network port of the second node.

[0151] In this application, by setting up a network signal hardware link between the first PHY chip and the BMC of each node in a multi-node server to connect the first PHY chip and the BMC, when an external network node accesses the BMC of the first node through the network port of the first node, the first data only needs to pass through the network port and the first PHY chip to reach the BMC of the first node; when an external network node accesses the BMC of the second node through the network port of the first node, the first data only needs to be forwarded by a single switching chip to reach the BMC of the second node. Compared with the existing solution that requires at least two switching chips, this application can save costs, simplify the path to access the BMC, and improve communication efficiency.

[0152] Please see Figure 4 This application embodiment also provides a flowchart of another data forwarding method. This method is applied to the BMC of a first node, which is any node in a multi-node server. The first node includes a first PHY chip and a network port. The network port is connected to the first PHY chip, and the first PHY chip is connected to the BMC. Specifically, this method includes steps 401 to 403.

[0153] Step 401: The BMC of the first node receives the first data through the network port and the first PHY chip of the first node.

[0154] Before receiving the first data, the BMC of all nodes in the multi-node server can be configured with external management network ports and internal forwarding network ports.

[0155] In this system, the external management network port of each node's BMC is the first MAC chip of that BMC. The BMC can configure an IP address for the first MAC chip, and the IP address of the first MAC chip will serve as the external IP address of the BMC, as well as the node IP address of the node where the BMC is located.

[0156] Specifically, the implementation method for configuring the IP address of the first MAC chip is the same as... Figure 3 The implementation of configuring the first MAC chip in step 301 of the illustrated embodiment is similar and will not be described in detail here. For example, in... Figure 2 In the multi-node server shown, each node's BMC can configure the IP address of the first MAC chip according to the slot where the node is located. The IP address of the first MAC chip 1 of node 1 is 10.10.0.101, and the IP address of the first MAC chip 3 of node 3 is 10.10.0.103.

[0157] In this configuration, the internal forwarding network port of each node's BMC is its second MAC chip. The BMC can assign an IP address to this second MAC chip, or specifically, configure the IP address based on the slot where the node is located. For example, in... Figure 2 In the multi-node server shown, the IP address of the second MAC chip 1 of node 1 is 10.10.1.101, and the IP address of the second MAC chip 3 of node 3 is 10.10.1.103.

[0158] In one possible implementation, after the internal forwarding ports of all nodes are configured in the BMC, the BMC of each node can add VLAN tags to the internal forwarding ports according to the preset configuration, thereby dividing the internal forwarding ports of each node into different VLANs to transmit different types of data.

[0159] Specifically, multi-node servers can divide VLANs according to business attributes, for example in Figure 2 In the multi-node server shown, nodes 1 and 2 are responsible for the same service, while nodes 3 and 4 are responsible for another service. The BMCs of nodes 1 and 2 can add VLAN tag 4090 to the internal forwarding network port, which is the corresponding second MAC chip; the BMCs of nodes 3 and 4 can add VLAN tag 4091 to the internal forwarding network port.

[0160] Specifically, a multi-node server can first assign all nodes to the same VLAN as an internal forwarding network. After a node receives data that requires special forwarding, it can then request the master node or collaborate with other nodes to create the corresponding VLAN to transmit the data.

[0161] For example in Figure 2In the multi-node server shown, nodes 1 to 4 can add VLAN tag 4094 to their internal forwarding ports, allowing data between nodes and some data from the external network to be transmitted via VLAN 4094. When node 1, acting as the master node, receives data related to a specific service, node 1 can delete VLAN 4094, assigning the internal forwarding ports of nodes 1 and 2 responsible for that service to VLAN 4090. It can be understood that during this process, the BMCs of nodes 1 to 4 delete VLAN tag 4094 from their internal forwarding ports, while the BMCs of nodes 1 and 2 add VLAN tag 4090 to their internal forwarding ports.

[0162] In another possible implementation, when each node of the multi-node server determines that data needs to be forwarded through the internal forwarding network port, it determines the corresponding VLAN to be built based on information such as the data type, data format or communication protocol of the data to be forwarded, and then builds the VLAN by requesting the master node or by cooperating with other nodes.

[0163] In another possible implementation, the multi-node server can be pre-configured with a VLAN configuration policy, which configures different VLANs at different times to support the corresponding services.

[0164] It is understandable that the VLAN ID range is from 0 to 4095, with VLAN 0 and VLAN 4095 reserved for protocol use. Therefore, a later ID can be used to avoid occupying the IDs of other VLANs.

[0165] The external management network cable of the multi-node server is connected to the network port of one or more nodes of the multi-node server so that the external network can access the BMC of the multi-node server through the network cable and the one or more network ports.

[0166] Upon receiving the first data, the BMC in the first phase can identify whether the target node indicated by the destination IP address of the first data includes the first node, in order to determine the subsequent processing method.

[0167] Step 402: If the destination IP address of the first data indicates that the target node includes the first node, the BMC of the first node processes the first data.

[0168] The specific implementation method of step 402 is as follows: Figure 3 The implementation of step 302 in the illustrated embodiment is similar and will not be described in detail here.

[0169] Step 403: If the target node does not include the first node, the BMC of the first node forwards the first data to the target node.

[0170] When the BMC of the first node needs to forward the first data to the target node, it means that the first node may be a transit node in the path of the first data to the target node, and the path through the first node and the switching chip is a better or optimal path. In this case, the BMC of the first node can forward the first data to the switching chip so that the switching chip can forward the first data to the target node.

[0171] The fact that the external network sends the first data to the first node indicates that the target node corresponding to the first data includes the first node, or that the communication path forwarded by the multi-node server where the first node is located is shorter.

[0172] When the target node does not include the first node, the target node may be a node outside the multi-node server or another node in the multi-node server other than the first node. Therefore, the first node can forward the first data to the switching chip, and the switching chip will make the next forwarding judgment.

[0173] In some other embodiments, if the target node does not include the first node, the BMC of the first node can directly discard the first data.

[0174] The specific implementation method of step 403 is as follows: Figure 3 The implementation of step 303 in the illustrated embodiment is similar and will not be described in detail here.

[0175] For example, for such Figure 2 In the multi-node server shown, node 1 receives the first data, and the destination IP address of the first data is 10.10.0.103, which is not the IP address of node 1, 10.10.0.101. Therefore, BMC1 can forward the first data to the switching chip, and the switching chip can confirm the destination of the first data.

[0176] Specifically, BMC1 can first divide the first MAC chip 1 and the second MAC chip 1 into a VLAN based on the IP address 10.10.0.101 of the first MAC chip 1 and the IP address 10.10.1.101 of the second MAC chip 1, forming a first VLAN. Then, BMC1 controls the first MAC chip 1 to forward the first data to the second MAC chip 1 through the first VLAN. Then, BMC1 adds a VLAN tag 4094 to the first data to identify that the first data is transmitted through VLAN 4094. Finally, BMC1 controls the second MAC chip 1 to forward the first data with the added VLAN tag to the switching chip.

[0177] It is understandable that after receiving the first data, the switching chip can forward the first data to node 3 based on the destination IP address or the VLAN tag. The specific implementation method is the same as... Figure 3 The implementation of step 304 in the illustrated embodiment is similar and will not be repeated here.

[0178] In this embodiment, by configuring an external management network port for each node's BMC, the first data from the external network can be directly transmitted to the node's BMC through the node's network port and the first PHY chip; and by configuring an internal forwarding network port and building an internal forwarding network through VLAN tags, the process of cross-node access by the management end can be completed through a single switching chip, simplifying the access path and improving the efficiency of external network access to the node BMC in a multi-node server.

[0179] Please see Figure 5 This application also provides a flowchart of another data forwarding method. This method is applied to the BMC of a first node, which is any node in a multi-node server. The first node includes a first PHY chip and a network port. The network port is connected to the first PHY chip, and the first PHY chip is connected to the BMC. Specifically, this method includes steps 501 to 503.

[0180] Step 501: The BMC of the first node receives the second data forwarded by the switching chip.

[0181] The second data can be data sent or forwarded by the second node to the switching chip. For example, the second node receives data through its network port and the first PHY chip, and then forwards the data to the first node through the switching chip; or it can be communication data between the second node and the first node. Specifically, the second node can be any node in the multi-node server other than the first node.

[0182] The multi-node server also includes a switching chip, which is connected to the second MAC chip of the BMC of each node of the multi-node server; that is, the BMC receives the second data through its second MAC chip.

[0183] Before receiving the first data, the BMC of all nodes in the multi-node server can be configured with external management network ports and internal forwarding network ports.

[0184] The specific implementation method for configuring the external management network port and the internal forwarding network port of the BMC is similar to the implementation method for configuring the external management network port and the internal forwarding network port of the BMC in step 401, and will not be repeated here.

[0185] Upon receiving the second data, the BMC of the first node can determine whether the target node corresponding to the second data includes the first node, in order to determine the subsequent processing method.

[0186] Specifically, the BMC of the first node can identify the destination IP address of the second data; if the destination IP address is the IP address of the first node, or the destination IP address is the multicast IP address corresponding to the multicast data to be received by the first node, then the BMC of the first node can determine that the target node includes the first node, and thus execute step 502; if the destination IP address is not the IP address of the first node or the multicast IP address, then the BMC of the first node can determine that the target node does not include the first node, and thus execute step 503.

[0187] In some other embodiments, the second data includes indication information about the target node of the second data attached to the switching chip; at this time, the first node can obtain the indication information; if the indication information indicates that the target node includes the first node, the BMC of the first node can perform step 502; if the indication information indicates that the target node does not include the first node, the BMC of the first node can perform step 503.

[0188] Step 502: If the destination IP address of the second data indicates that the target node includes the first node, the BMC of the first node processes the second data.

[0189] The specific implementation method of step 502 is as follows: Figure 3 The implementation of step 305 in the illustrated embodiment is similar and will not be described in detail here.

[0190] Step 503: If the target node does not include the first node, the BMC of the first node forwards the second data to the target node through the first PHY chip and network port of the first node.

[0191] The fact that the switching chip forwards the second data to the first node indicates that the target node corresponding to the second data includes the first node, or that the first node is the node in the multi-node server closest to the target node. Therefore, when the target node does not include the first node, the target node is a node outside the multi-node server, and the first node can forward the second data to the target node through the first node's first PHY chip and network port.

[0192] The specific implementation method of step 503 is as follows: Figure 3 The implementation of step 306 in the illustrated embodiment is similar and will not be described in detail here.

[0193] For example, the first node is as follows Figure 2In the multi-node server shown, node 3 has an IP address of 10.10.0.103. When the destination IP address of the second data is neither the IP address of node 3 (10.10.0.103) nor the multicast IP address corresponding to the multicast data to be received by node 3, BMC3 can forward the second data to the target node indicated by the destination IP address.

[0194] Specifically, BMC3 can first classify the first MAC chip 3 and the second MAC chip 3 into the same VLAN based on the IP address of the first MAC chip 3 (10.10.0.103) and the IP address of the second MAC chip 3 (10.10.1.103), forming a first VLAN; then control the second MAC chip 3 to forward the second data to the first MAC chip 3 through the first VLAN; finally, BMC3 controls the first MAC chip 3 to forward the second data after deleting the VLAN tag to the destination IP address through the first PHY chip 3 and network port 3.

[0195] The second data may include a VLAN tag attached to the second node, and the BMC3 may remove the VLAN tag before forwarding the second data to the destination IP address.

[0196] Optionally, the IP address of the next-hop node for forwarding the second data by BMC3 can be determined based on its own routing table and the destination IP address.

[0197] Optionally, the second data includes indication information about the target node of the second data attached to the switching chip; the BMC3 can obtain the next-hop IP address of the second data to the target node from the indication information.

[0198] In this embodiment, by configuring an external management network port and an internal forwarding network port for each node's BMC, and by forming an internal forwarding network through the internal forwarding network ports of different nodes, the process of nodes in the external network accessing other IP addresses in the external network through this multi-node server can be completed through a single switching chip, thereby improving communication efficiency.

[0199] This application embodiment also provides an out-of-band management controller, which is used to perform the above-described... Figure 4 or Figure 5 The method in the illustrated embodiment.

[0200] This application also provides a computer-readable storage medium including instructions that, when executed on a computer, cause the computer's BMC to perform the above-described actions. Figure 4 or Figure 5 The method in the illustrated embodiment.

[0201] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments of this application.

[0202] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0203] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0204] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0205] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0206] If the aforementioned function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application embodiment, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A data forwarding method, characterized in that, The method is applied to a multi-node server, which includes a first node, a second node, and a switching chip. The switching chip is connected to the out-of-band management controller (BMC) of the first node and the BMC of the second node, respectively. Both the first node and the second node include a first port physical layer (PHY) chip and a network port. The network port is connected to the first PHY chip of the node, and the first PHY chip is connected to the BMC of the node. The BMC includes a first media access control (MAC) chip and a second MAC chip. The switching chip includes multiple third PHY chips. A second PHY chip is provided between the switching chip and each second MAC chip. Each second PHY chip is connected to a second MAC chip and one of the third PHY chips of the switching chip, wherein the number of third PHY chips is greater than or equal to the number of second PHY chips. The second PHY chips are used to convert SGMII protocol format data to MDI protocol format data, and the third PHY chips are used to convert MDI protocol format data to SGMII protocol format data. The method includes: The BMC of the first node receives the first data through the network port of the first node and the first PHY chip; If the destination IP address of the first data indicates a target node that includes the first node, the BMC of the first node processes the first data; and / or, If the target node does not include the first node, the BMC of the first node forwards the first data to the switching chip; If the target node includes the second node, or if the target node does not include any nodes in the multi-node server, and the node in the multi-node server closest to the target node is the second node, the switching chip forwards the first data to the BMC of the second node. In the case where the target node includes the second node, the BMC of the second node processes the first data; and / or, If the target node does not include the second node, the BMC of the second node forwards the first data to the target node through the first PHY chip and the network port of the second node.

2. A data forwarding method, characterized in that, The method is applied to the out-of-band management controller (BMC) of a first node, which is any node in a multi-node server. The multi-node server further includes a switching chip. The first node includes a first port physical layer (PHY) chip and a network port. The network port is connected to the first PHY chip, and the first PHY chip is connected to the BMC. The BMC includes a first media access control (MAC) chip and a second MAC chip. The switching chip includes multiple third PHY chips. A second PHY chip is disposed between the second MAC chip and the switching chip. Each second PHY chip is connected to one second MAC chip and one of the third PHY chips of the switching chip. The number of third PHY chips is greater than or equal to the number of second PHY chips. The second PHY chips are used to convert SGMII protocol format data to MDI protocol format data, and the third PHY chips are used to convert MDI protocol format data to SGMII protocol format data. The method includes: First data is received through the network port and the first PHY chip; If the destination IP address of the first data indicates a target node that includes the first node, process the first data; and / or, If the target node does not include the first node, the first data is forwarded to the target node.

3. The method according to claim 2, characterized in that, Receiving first data through the network port and the first PHY chip includes: The first MAC chip receives the first data that has passed through the network port and the first PHY chip; Forwarding the first data to the target node includes: A first virtual local area network (VLAN) is constructed using the first MAC chip and the second MAC chip. Control the first MAC chip to send the first data to the second MAC chip through the first VLAN; The second MAC chip is controlled to send the first data to the switching chip, so that the switching chip forwards the first data to the target node.

4. The method according to claim 3, characterized in that, Before controlling the second MAC chip to send the first data to the switching chip, the method further includes: VLAN tags are appended to the first data; The step of controlling the second MAC chip to send the first data to the switching chip includes: The second MAC chip is controlled to send the first data with the VLAN tag attached to it to the switching chip.

5. A data forwarding method, characterized in that, The method is applied to the out-of-band management controller (BMC) of a first node, which is any node in a multi-node server. The first node includes a first port physical layer PHY chip and a network port. The network port is connected to the first PHY chip, and the first PHY chip is connected to the BMC. The BMC includes a first media access control (MAC) chip and a second MAC chip. The multi-node server also includes a switching chip, which includes multiple third PHY chips. The switching chip is connected to the BMC. A second PHY chip is disposed between the second MAC chip and the switching chip. Each second PHY chip is connected to one second MAC chip and one of the third PHY chips of the switching chip. The number of third PHY chips is greater than or equal to the number of second PHY chips. The second PHY chips are used to convert SGMII protocol format data to MDI protocol format data, and the third PHY chips are used to convert MDI protocol format data to SGMII protocol format data. The method includes: Receive the second data forwarded by the switching chip; If the destination IP address of the second data indicates a target node that includes the first node, then process the second data; and / or, If the target node does not include the first node, the second data is forwarded to the target node through the first PHY chip and the network port.

6. The method according to claim 5, characterized in that, The receipt of the second data forwarded by the switching chip includes: The second MAC chip receives the second data; The forwarding of the second data to the target node via the first PHY chip and the network port includes: A first virtual local area network (VLAN) is constructed using the first MAC chip and the second MAC chip. Control the second MAC chip to send the second data to the first MAC chip through the first VLAN; The first MAC chip is controlled to send the second data to the first PHY chip, so that the first PHY chip forwards the second data to the target node through the network port.

7. The method according to claim 5 or 6, characterized in that, The second data includes indication information about the target node attached by the switching chip, which is obtained by the switching chip based on its own routing table and the destination IP address; After receiving the second data forwarded by the switching chip, the method further includes: Determine whether the target node includes the first node based on the indication information; The forwarding of the second data to the target node via the first PHY chip and the network port includes: The second data is forwarded to the next-hop IP address of the target node via the first PHY chip and the network port.

8. A multi-node server, characterized in that, The multi-node server includes a first node, which is any node in the multi-node server, and the multi-node server also includes a switching chip. The first node includes a network port, an out-of-band management controller (BMC), and a first port physical layer (PHY) chip. The network port is connected to the first PHY chip, and the first PHY chip is connected to the BMC. The BMC is used to interact with the external network connected to the network port through the first PHY chip and the network port. The BMC includes a first media access control (MAC) chip and a second MAC chip. The switching chip includes multiple third PHY chips. The first MAC chip is connected to the first PHY chip, the second MAC chip is connected to the switching chip, and a second PHY chip is disposed between the second MAC chip and the switching chip. Each second PHY chip is connected to one second MAC chip and one of the third PHY chips of the switching chip. The number of third PHY chips is greater than or equal to the number of second PHY chips. The second PHY chip is used to convert data in SGMII protocol format to data in MDI protocol format, and the third PHY chip is used to convert data in MDI protocol format to data in SGMII protocol format.

9. The multi-node server according to claim 8, characterized in that, The BMC is also used to form a first virtual local area network (VLAN) through the first MAC chip and the second MAC chip, and to control the first MAC chip to perform data interaction with the second MAC chip through the first VLAN.

10. An out-of-band management controller, characterized in that, The out-of-band management controller is used to perform the method described in any one of claims 2 to 7.

Images