WO2023116580A1 - Path switching method and apparatus, network device, and network system - Google Patents

Abstract

A path switching method and apparatus, a network device, and a network system, which relate to the field of communications. The switching method comprises: during the process of a first network device forwarding a message, according to abnormal information sent by a second network device, the first network device can determine that a first path corresponding to the message is in an abnormal state, thereby avoiding the first network device being unable to sense the abnormal state (e.g. congestion or packet loss) of a downstream device. Moreover, an ECMP group determined by the first network device further comprises other equal-cost forwarding paths consistent with a destination identifier of the first path, such as a second path. Therefore, the first network device can also forward the foregoing message by means of the second path, thereby avoiding the message being unable to be forwarded to a device corresponding to the destination identifier, improving the accuracy of data communication, and reducing a transmission delay caused by the state of the path being the abnormal state, thus improving the transmission efficiency of the message.

Description

Path switching method, device, network device, and network system

This application claims the priority of the Chinese patent application with the application number 202111572470.1 and the application name "A Path Adjustment Method and Related Device" submitted to the State Intellectual Property Office on December 21, 2021. The priority of the Chinese patent application with the application number 202210139157.7 and the application name "Path Switching Method, Device, Network Equipment, and Network System" filed with the State Intellectual Property Office on April 15, the entire contents of these applications are incorporated by reference in this application middle.

technical field

The present application relates to the communication field, and in particular to a path switching method, device, network equipment, and network system.

Background technique

The equal-cost multi-path (ECMP) path refers to multiple paths with equal value but different intermediate nodes to the same target Internet Protocol (IP) address or target network segment. When the network device supports the ECMP function, data streams sent to the same target IP address or target network segment can be sent through different paths. Usually, based on the destination IP address and a preset traffic model, the network device performs traffic load sharing on the service data transmitted by the multiple paths included in the ECMP group. However, if the equal-cost forwarding path included in the ECMP group has problems such as congestion or packet loss, the network device cannot perceive the abnormal state of the downstream device in the equal-cost forwarding path, resulting in a longer transmission delay of the equal-cost forwarding path. High, and the routing path of the message cannot be switched to other equal-cost forwarding paths included in the ECMP group. As a result, the transmission delay of service data in the ECMP group increases and the data transmission efficiency decreases. Therefore, how to provide an efficient path switching method when the forwarding path in the ECMP group is abnormal has become an urgent problem to be solved at present.

Contents of the invention

The present application provides a path switching method, device, network equipment, and network system, which solve the problem that when the equivalent forwarding path in the ECMP group is abnormal, the transmission delay of business data in the ECMP group increases and the data transmission efficiency decreases The problem.

This application adopts the following technical solutions.

In the first aspect, the embodiment of the present application provides a path switching method, which can be applied to a network system including a first network device and a second network device, or a network device that supports the path switching method, such as the network device Including chip or chip system. For example, the path switching method is performed by the first network device, and the path switching method includes: the first network device receives the first message, and obtains the corresponding ECMP group according to the destination identifier of the first message, and the ECMP group includes Multiple equivalent forwarding paths corresponding to the destination identifier; furthermore, the first network device determines the first path corresponding to the first message from the aforementioned ECMP group; when the first network device determines the first path according to the abnormal information sent by the second network device When the path is in an abnormal state, the first network device forwards the first message through the second path in the ECMP group.

In the process of the first network device forwarding the message, since the first network device can determine that the first path corresponding to the first message is in an abnormal state according to the abnormal information sent by the second network device, it is avoided that the first network device cannot perceive The abnormal state (such as congestion or packet loss) of the downstream device (such as the second network device), and the ECMP group also includes other equal-cost forwarding paths consistent with the destination identifier of the first path, such as the second path. Therefore, the first network The device can also forward the aforementioned first message through the second path, avoiding that the first message cannot be forwarded to the device corresponding to the destination identifier, improving the accuracy of data communication, and reducing the transmission caused by the abnormal state of the path The time delay improves the transmission efficiency of the first packet.

In an optional example, the abnormal information sent by the second network device indicates that the packet received through the first path affects the forwarding efficiency of the second network device. The forwarding efficiency of the second network device may be represented by any one or a combination of several of the following: bandwidth of the first path, transmission delay of the first path, packet loss rate of data packets in the first path, and the like. When the first network device determines that one or more equal-cost forwarding paths included in the ECMP group are in an abnormal state, it may switch the message originally forwarded by the abnormal equal-cost forwarding path to a normal equal-cost forwarding path in the ECMP group for forwarding, The problem that the first network device cannot perceive the abnormal state of the downstream device is solved, and the transmission efficiency of the data flow in the ECMP group is improved.

In another optional example, the first path is in an abnormal state, including: the first path is marked as abnormal, such as the first path being marked as unavailable; or the first path is deleted, such as the first The network device moves the first path out of the equal-cost forwarding path included in the ECMP group; or the parameters of the first path exceed the set threshold, for example, each equal-cost forwarding path included in the ECMP group corresponds to a path parameter value, in an ECMP In the group, the maximum value and the minimum value of the path parameter value are determined. If the path parameter value corresponding to the first path is set to the path parameter value corresponding to the ECMP group, it is determined that the parameter value of the first path exceeds the set value. threshold, and then determine that the first path is marked as abnormal. In this embodiment, because the network device can determine the equivalent forwarding path in the abnormal state in the ECMP group according to the abnormality information fed back by the downstream device, and furthermore, the network device can use the equivalent forwarding path in the abnormal state to transmit the The message is switched to the normal equal-cost forwarding path (or an idle equal-cost forwarding path), which reduces the transmission delay of the message in the network system and improves the transmission efficiency of the network system.

In yet another optional example, the deletion of the first path includes: invalidating the port provided by the first network device to the first path, or the next-hop route of the second network device in the first path been deleted.

In an optional implementation manner, the first network device forwards the first message through the second path in the ECMP group, including: the first network device offsets the first path by a set value to obtain the second path ; Furthermore, the first network device forwards the first message through the second path. Network devices can determine the equivalent forwarding path of message forwarding by setting the offset value, which improves the accuracy of path adjustment and realizes precise control of data flow when multiple messages belong to the same data flow . For example, the network device or the source device can set an offset value for a specific field, thereby adjusting the internal path selection in the ECMP group, and realizing precise adjustment of the routing path between the source device and the destination device.

In another optional implementation manner, the foregoing path switching method further includes: the first network device receives a second packet, the second packet belongs to the first flow, and the second packet includes an offset value; further, the first A network device forwards the second message through the second path according to the offset value in the second message. Since the first network device quickly adjusts the path in the ECMP group after obtaining the abnormal information, the timeliness of path adjustment in the ECMP group is improved on the basis of improving the data transmission efficiency of the network system.

In a possible situation, the aforementioned offset value is included in the packet header of the second packet.

In a feasible example, if the first packet is a fourth-generation Internet protocol (Internet Protocol version 4, IPv4) packet, the extended option header or the IP ID field included in the IPv4 packet is stored with an offset value . or,

In another feasible example, if the first packet is a sixth-generation Internet Protocol (Internet Protocol version 6, IPv6) packet, the extended options header or flow label (flow label) field included in the IPv6 packet is biased. transfer value.

It should be noted that when the aforementioned first packet and second packet are data packets based on other protocols, the offset value can also be set in other positions of the packet header, which is not limited. In addition, the offset value may be set by the first network device, or the offset value may also be set by the source device communicating with the first network device, which is not limited.

In another optional implementation manner, the first packet includes priority information. The first network device obtains the corresponding equal-cost path load sharing ECMP group according to the destination identifier of the first message, including: step 1, the first network device obtains multiple ECMP groups according to the destination identifier of the first message, and the multiple ECMP groups corresponding to different priorities. Step 2: The first network device determines the ECMP group corresponding to the first message from multiple ECMP groups according to the priority information. Since a network device can provide multiple different ECMP groups for a destination identifier, and these different ECMP groups have different forwarding efficiencies, when the network system needs to serve a large number of users, the network device can give priority to Users with low-latency or large-traffic transmission needs provide ECMP groups with higher forwarding efficiency, so as to realize low-latency and large-traffic transmission of data communication. In addition, for a user, the network device can provide the user with a higher-quality (eg, lower delay and larger bandwidth) ECMP group, which is beneficial to improving the QoE of data communication.

In a possible situation, the priority information is a differentiated services code point (differentiated services code point, DSCP) value. In this embodiment, the network device can provide ECMP groups with different delays and bandwidths for different service types according to different DSCP values, that is, the network device can provide better communication services for differentiated target groups, thereby avoiding Different types of services share the same ECMP group, resulting in problems such as increased processing delay and low data transmission efficiency of the network system. For example, network devices can provide higher-quality communication services for high-priority services and improve QoE.

A second aspect provides a path switching device, the path switching device is applied to a first network device, and the path switching device includes a path switching method for executing the first aspect or any possible implementation manner of the first aspect individual modules. Exemplarily, the path switching device includes: a receiving unit, an acquiring unit, a determining unit, and a forwarding unit.

The receiving unit is used for receiving the first packet, and the first packet belongs to the first stream. The acquiring unit is configured to acquire a corresponding ECMP group according to the destination identifier of the first message, and the ECMP group includes multiple equivalent forwarding paths corresponding to the aforementioned destination identifier. The determining unit is configured to determine the first path corresponding to the first message from the ECMP group, and determine whether the first path is in an abnormal state. Wherein, the determining unit determines that the first path is in an abnormal state according to the abnormal information sent by the second network device. The forwarding unit is configured to forward the first message through the second path in the ECMP group when the determining unit determines that the first path is in an abnormal state.

For the beneficial effects, reference may be made to the description of any implementation manner in the first aspect, and details are not repeated here. The path switching device has the function of implementing the behavior in the method example of any implementation manner in the first aspect above. The functions described above may be implemented by hardware, or may be implemented by executing corresponding software on the hardware. The hardware or software includes one or more modules corresponding to the above functions.

In an optional example, the abnormal information sent by the second network device indicates that the packet received through the first path affects the forwarding efficiency of the second network device.

In another optional example, the abnormal state of the first path includes: the first path is marked as abnormal, or the first path is deleted, or a parameter of the first path exceeds a set threshold.

In yet another optional example, the deletion of the first path includes: invalidating the port provided by the first network device to the first path, or the next-hop route of the second network device in the first path been deleted.

In an optional implementation manner, the forwarding unit is specifically configured to: offset the first path by a set value to obtain the second path; and forward the first packet through the second path.

In an optional implementation manner, the receiving unit is further configured to receive a second packet, the second packet belongs to the first stream, and the second packet includes an offset value. In addition, the forwarding unit is further configured to forward the second message through the second path according to the offset value in the second message.

In a possible situation, the offset value is included in the packet header of the second packet.

In a feasible example, if the first packet is an IPv4 packet, the extended option header or the IP ID field included in the IPv4 packet stores an offset value.

In another feasible example, if the first packet is an IPv6 packet, the extended options header or flow label field included in the IPv6 packet stores an offset value.

In another optional implementation manner, the first packet includes priority information. The obtaining unit is specifically configured to obtain multiple ECMP groups according to the destination identifier of the first message, and the multiple ECMP groups correspond to different priorities. The determining unit is further configured to determine the ECMP group corresponding to the first message from multiple ECMP groups according to the priority information. In one possible situation, the priority information is a DSCP value.

In a third aspect, a network device is provided, including a memory and a processor, wherein: the memory is used to store program codes; and the processor is used to call the program codes to implement the operation steps of the method in any implementation manner in the first aspect.

In a fourth aspect, a network system is provided, including: a first network device and a second network device. In the network system, a first network device receives a first packet, and the first packet belongs to a first flow. Secondly, the first network device obtains the corresponding ECMP group according to the destination identifier of the first packet, and the ECMP group includes multiple equivalent forwarding paths corresponding to the destination identifier. Afterwards, the first network device determines the first path corresponding to the first packet from the ECMP group. Finally, when the first path is in an abnormal state, the first network device forwards the first message to the second network device through the second path in the ECMP group. Wherein, the first network device determines that the first path is in an abnormal state according to the abnormality information sent by the second network device.

In an optional implementation manner, the network system further includes: a source device. The source device sends a second packet to the first network device, the second packet belongs to the first flow, and the second packet includes an offset value. Furthermore, the first network device forwards the second packet to the second network through the second path according to the offset value in the second packet. In a possible situation, the offset value is included in the packet header of the second packet. For the beneficial effects, reference may be made to the description of any implementation manner in the first aspect, and details are not repeated here.

In a fifth aspect, a computer-readable storage medium is provided, in which a computer program or instruction is stored, and when the computer program or instruction is executed by a processor, the operation steps of the method in any implementation manner in the first aspect are implemented.

In a sixth aspect, the present application provides a computer program product, the computing program product includes instructions, and when the computer program product is run on a processor or a computing device, the processor or computing device executes the instructions, so as to realize the first aspect and The method in any possible implementation manner in the first aspect.

In a seventh aspect, the present application provides a chip, including: an interface circuit and a control circuit. The interface circuit is used to receive data from other devices other than the processor and transmit it to the control circuit, or send data from the control circuit to other devices other than the processor. The control circuit is used to implement the logic circuit or execute code instructions The method of any implementation manner in the first aspect. For the beneficial effects, reference may be made to the description of any implementation manner in the first aspect, and details are not repeated here.

On the basis of the implementation manners provided in the foregoing aspects, the present application may further be combined to provide more implementation manners.

Description of drawings

FIG. 1 is a schematic structural diagram of a communication system provided by the present application;

FIG. 2 is a schematic diagram of the architecture of a leaf-spine network provided by the present application;

FIG. 3 is a first schematic flow diagram of a path switching method provided by the present application;

FIG. 4 is a second schematic flow diagram of a path switching method provided by the present application;

FIG. 5 is a third schematic flow diagram of a path switching method provided by the present application;

FIG. 6 is a schematic diagram of the path adjustment process in the ECMP group provided by the present application;

FIG. 7 is a schematic structural diagram of an IPv4 message provided by the present application;

FIG. 8 is a fourth schematic flowchart of a path switching method provided by the present application;

FIG. 9 is a schematic structural diagram of a path switching device provided by the present application;

FIG. 10 is a schematic structural diagram of a network device provided by the present application;

FIG. 11 is a schematic structural diagram of another network device provided by the present application.

Detailed ways

An embodiment of the present application provides a path switching method, the path switching method includes: the first network device receives the first message, and obtains the corresponding ECMP group according to the destination identifier of the first message, and the ECMP group includes the corresponding Multiple equivalent forwarding paths identified by the destination; furthermore, the first network device determines the first path corresponding to the first message from the aforementioned ECMP group; when the first network device determines the first path according to the abnormal information sent by the second network device When in an abnormal state, the first network device forwards the first message through the second path in the ECMP group. In the process of the first network device forwarding the message, since the first network device can determine that the first path corresponding to the first message is in an abnormal state according to the abnormal information sent by the second network device, it is avoided that the first network device cannot perceive The abnormal state (such as congestion or packet loss) of the downstream device (such as the second network device), and the ECMP group also includes other equal-cost forwarding paths consistent with the destination identifier of the first path, such as the second path. Therefore, the first network The device can also forward the aforementioned first message through the second path, avoiding that the first message cannot be forwarded to the device corresponding to the destination identifier, improving the accuracy of data communication, and reducing the transmission caused by the abnormal state of the path The time delay improves the transmission efficiency of the first packet.

In order to make the description of the following embodiments clear and concise, a brief introduction of related technologies is given first.

FIG. 1 is a schematic structural diagram of a communication system provided by the present application. As shown in FIG. 1 , the communication system includes a source device 110 and a destination device 130 , and a network system 120 for providing communication services to the source device 110 and the destination device 130 . The network system 120 may refer to the Internet, a cellular network, a data center network, or other networks. For example, the network system 120 is an Internet service provider (Internet service provider, ISP) network. An ISP is an operator that provides Internet access services, information services, and value-added services to a large number of users.

The network system 120 may include at least one network device, such as network device 121 to network device 126 shown in FIG. 1 . Exemplarily, the network device may be a routing transmission device such as a router or a switch. For example, the network device 121 is a wireless access point (access point, AP), such as a wireless router, and the network device 122 may be a switch connected to the wireless router. In some possible other situations, the network device included in the network system 120 may also be other network devices with a data transmission function, such as a server or a base station device with a data transmission function.

The source device 110 refers to a device that sends data. For example, the source device 110 is an application server or a storage device that stores data. The source device 110 can use the network system 120 to send data to other devices (such as the The destination device 130) sends the data. The data stored by the source device 110 may include, but is not limited to: audio, video, text, or other types of data. When the data sent by the source device 110 is video, audio or other multimedia files, the source device 110 can transmit these data in a streaming manner, such as a media stream, and the destination device 130 does not download the file before playing. The entire streaming media file of the media stream, so as to reduce the delay of the destination device 130 playing the streaming media corresponding to the media stream, and improve the user experience quality (quality of experience, QoE).

Please continue to refer to FIG. 1, the above-mentioned destination device 130 may refer to a device requesting data, such as the destination device 130 refers to a terminal, and the terminal may be a mobile phone, a tablet computer, a computer with a wireless transceiver function, a personal communication service (personal communication service) , PCS) phones, desktop computers, personal digital assistants (personal digital assistant, PDA), wearable devices, virtual reality (virtual reality, VR) terminal equipment, augmented reality (augmented reality, AR) terminal equipment, industrial control (industrial control) ), wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grid, and transportation safety Wireless terminals, wireless terminals in smart cities, wireless terminals in smart homes, etc.

It should be noted that the above-mentioned source device 110 and destination device 130 are only examples provided by this embodiment, and should not be construed as a limitation to this application. And the quantity is not limited.

During the data transmission process, if the destination IP addresses of multiple packets sent by the source device 110 are the same, in order to improve the transmission efficiency of the network system 120, the routing load sharing technology is generally used to forward the multiple packets. Routing load balancing means that between network nodes (or network devices), there are multiple routing forwarding paths to the same destination, and traffic is shared to multiple routing forwarding paths for forwarding. For example, in the forwarding information base (forwarding information base, FIB) of the network device, if there are multiple entries with the same address and mask, but the next-hop route (such as IP address), outgoing interface, or tunnel ID (tunnel ID ) attributes corresponding to the routing paths of the table entries, and the network device can implement routing load sharing of traffic in the network system based on the multiple routing paths.

In some optional cases, the FIB can also be called a forwarding routing table or a forwarding table. The FIB is stored in the data plane of a router (or other network device), and the entries of the FIB are called forwarding table entries. Each forwarding table entry Both specify a certain destination to be reached, the outbound interface to be passed and the next-hop IP address and other information. In terms of hardware implementation, the data in the FIB (such as the aforementioned forwarding table entries) can be stored in an application-specific integrated circuit (ASIC), which allows network devices (such as routers) to perform data queries in the FIB , a fairly high speed can be achieved. Of course, the storage space that can be used by FIB is limited. Therefore, in large-scale network systems, attention should be paid to the size of the routing table of network equipment. On the premise of ensuring data accessibility, various mechanisms and means should be used to reduce the size of FIB.

As shown in Figure 1, the network system 120 provides multiple different routing paths for data with the same destination IP address, as shown in Figure 1, the source device 110-network device 121-network device 122-network device 126-destination Device 130, source device 110-network device 121-network device 123-network device 126-destination device 130, source device 110-network device 121-network device 124-network device 125-network device 126-destination device 130, or are other possible routing paths etc. Among the multiple routing paths, at least two routing paths with the same cost (cost) but different intermediate nodes may be referred to as mutual equal-cost forwarding paths. Wherein, cost refers to the cost of reaching the destination address indicated by a certain route. Since the destination addresses of multiple equal-cost switching paths are the same, and the costs of these multiple equal-cost switching paths are the same, and the forwarding routes are different, the multiple equal-cost switching paths can realize routing load sharing in the network system 120 . In addition, the foregoing multiple equal-cost forwarding paths may also be referred to as an ECMP group reaching one destination address.

Below, the network system 120 shown in Figure 1 is a spine-leaf network (spine-leaf network) as an example, and the network system is described. 200 includes a plurality of switches, such as spine switch 211 to spine switch 214 , and leaf switch 221 to leaf switch 223 .

A leaf switch (leaf switch) can implement the function of an access switch, and is used to connect physical servers communicating with the leaf-spine network 200, such as servers 231 to 233 shown in FIG. 2 . For example, the leaf switch 221 is connected to the server 231 to connect the server 231 to the leaf-spine network 200 .

A spine switch can implement the function of a core router (CR) for forwarding data from leaf switches. For example, the servers connected to the two leaf switches need to communicate and need to forward data through the spine switch. As shown in FIG. 2, when server 231 and server 232 communicate, data needs to be forwarded through leaf switch 221, leaf switch 222, and a spine switch (such as spine switch 212) communicating with leaf switch 221 and leaf switch 222 at the same time.

In the leaf-spine network 200, data communication can be realized by selecting an equivalent forwarding path based on ECMP between the leaf switch and the spine switch. As shown in Figure 2, the three routing paths (indicated by dotted lines in Figure 2) from the leaf switch 221 to the spine switch 211, the spine switch 212, and the spine switch 213 are each other from the leaf switch 221 to the device indicated by the target IP address (such as a server) 232) equivalent forwarding path.

It is worth noting that in a leaf-spine network, the number of downlink ports of the spine switches determines the number of leaf switches, and the number of uplink ports of the leaf switches determines the number of spine switches. The scale of the leaf-spine network is determined by the number of downlink ports of the spine switches. It is determined jointly with the number of uplink ports of the leaf switch. For more information about the leaf-spine network, reference may be made to related descriptions of common technologies, and details are not repeated here.

It is understandable that scale, virtualization, and cloud computing have become the development direction of the network to integrate information technology (information technology, IT) resources, improve resource utilization efficiency, and reduce maintenance costs. Therefore, a virtual machine (virtual machine, VM) can be deployed in the server shown in FIG. logical computer. A physical device running a virtual machine is called a host machine. For example, the server 231 is the host machine of VM1 and VM2, the server 232 is the host machine of VM3 and VM4, and the server 234 is the host machine of VM5 and VM6.

In a possible example, the computing resource required by the VM comes from a local processor and memory of the server, and the storage resource required by the VM may come from a local hard disk of the server. For example, the virtual machine 231A (the aforementioned VM2) includes a virtual processor 231A1, a virtual memory 231A2, and a virtual network card 231A3, the computing resources required by the virtual processor 231A1 are provided by the processor, and the storage resources required by the virtual memory 231A2 are provided by the memory (or hard disk), and the network resources required by the virtual network card 231A3 are provided by the network card included in the server. In addition, various applications can run in the VM, and users can trigger services such as data access (including reading data, writing data, and migrating data, etc.), streaming media transmission and playback through the applications in the VM.

Please continue to refer to FIG. 2 , when the leaf-spine network 200 provides communication services for any two servers, the leaf-spine network 200 can provide multiple equal-cost forwarding paths with equal value and different intermediate routes to forward data. In addition, the leaf-spine network 200 may also include other routing devices, which are not shown in FIG. 2 . The embodiment of the present application does not limit the number of leaf switches and spine switches included in the leaf-spine network 200 .

The specific implementation manner of the path switching method provided in this embodiment will be described in detail below with reference to the accompanying drawings.

FIG. 3 is a schematic flow diagram of a path switching method provided by the present application. The path switching method can be applied to the communication system shown in FIG. 1 . For example, the path switching method shown in FIG. 3 can be represented by FIG. 2 Any one of the switches included in the illustrated leaf-spine network 200 can be implemented. Here, the source device 300 is realized by the server 231 in FIG. 2 , the network device 310 is realized by the leaf switch 221 in FIG. 2 , and the network devices 321 to 325 are realized by the spine switch in FIG. 2 as an example. In this embodiment, the network device 310 may also be called a first network device, and the network device 321 may also be called a second network device.

As shown in FIG. 3 , the path switching method provided in this embodiment includes steps S310 to S340.

S310. The source device 300 sends the first packet to the network device 310.

Correspondingly, the network device 310 receives the first packet.

Wherein, the first packet belongs to the first stream. The first stream refers to the data stream sent by the source device 300 according to the service request. For example, if the service request is a video playback request, the data stream refers to the video stream, and the first message refers to an image frame data in the video stream. . In addition, the service request may also be an audio playback request, a text data migration request, or other requests, which are not limited.

In some cases, the destination identifiers of multiple packets belonging to the first flow are the same. If the first stream further includes a second packet, the destination identifier of the second packet is consistent with the destination identifier of the first packet. Exemplarily, the destination identifier refers to the destination IP address or destination MAC address included in the first packet, or the target network segment to which the first packet is to be sent, and the like.

S320, the network device 310 acquires a corresponding ECMP group according to the destination identifier of the first packet.

The ECMP group includes multiple equal-cost forwarding paths corresponding to the destination identifier. As shown in Figure 3, the ECMP group corresponding to the destination identifier of the first message is ECMP group 1, and this ECMP group 1 provides multiple equal-cost forwarding paths to the destination identifier, as shown in Figure 3. Path 1 to Path 3. It should be noted that path 1 to path 3 may refer to different equivalent forwarding paths included in the ECMP group, and their examples in FIG. 3 should not be construed as limiting the present application. In the embodiment of the present application, path 1 may also be called a first path, and path 2 may also be called a second path.

Optionally, the network device 310 may provide multiple ECMP groups for the destination identifier of the first packet, such as ECMP group 1 and ECMP group 2 shown in FIG. 3 .

In a possible situation, the forwarding efficiencies of ECMP group 1 and ECMP group 2 are different. The forwarding efficiency of an ECMP group can be represented by any one or a combination of several of the following: the bandwidth of the equal-cost forwarding path included in the ECMP group, the transmission delay of the equal-cost forwarding path, the packet loss rate of data in the equal-cost forwarding path, etc. .

For example, the bandwidth of the equivalent forwarding path included in ECMP group 1 is different from the bandwidth of the equivalent forwarding path included in ECMP group 2. For example, the bandwidth of the equivalent forwarding path (such as path 2) included in ECMP group 1 is greater than that The bandwidth of the equal-cost forwarding path (such as path 4). In some cases, ECMP group 1 may be called a high-priority ECMP group, and ECMP group 2 may be called a normal ECMP group or a default ECMP group. It should be noted that FIG. 3 is only an example provided by this embodiment, and should not be construed as a limitation to this application.

Since a network device can provide multiple different ECMP groups for a destination identifier, and these different ECMP groups have different forwarding efficiencies, when the network system needs to serve a large number of users, the network device can give priority to Users with low-latency or large-traffic transmission needs provide ECMP groups with higher forwarding efficiency, so as to realize low-latency and large-traffic transmission of data communication. In addition, for a user, the network device can provide the user with a higher-quality (eg, lower delay and larger bandwidth) ECMP group, which is beneficial to improving the QoE of data communication.

Optionally, the above-mentioned first packet includes priority information. The priority information is used to indicate the type of the service corresponding to the first packet. In some feasible implementation manners, different types of services require different ECMP groups for transmitting packets of the services.

In a possible example, the priority information included in the first packet is a DSCP value. For example, if the first message is a data message (IPv4 message) based on the fourth version of the Internet protocol (internet protocol version 4, IPv4), then the DSCP value can be set in the service type (type of service) identification field. For example, the DSCP uses the used 6 bits and the unused 2 bits in the service type identification field to determine a coding value, and uses the coding value to distinguish service priorities.

In a feasible specific example, the DSCP uses 6 bits, and the range of DSCP values is 0 to 63. It is worth noting that each DSCP code value is mapped to a defined hop-by-hop behavior (per hop behavior, PHB) identification code, and the PHB identification code indicates the type of service.

The DSCP value included in the foregoing first packet may be determined by the source device 300 . During the communication process of the first message, the source device 300 enters the DSCP value in the first message, and the intermediate devices (such as terminal devices such as telephones and Windows clients and servers) in the communication process can also input the DSCP value contained in the first message. The data flow of the text (such as the aforementioned first flow) is identified, so as to perform traffic load sharing on the data flow.

In this embodiment, the network device can provide ECMP groups with different delays and bandwidths for different service types according to different priority information, that is, the network device can provide better communication services for differentiated target groups, thereby Avoid different types of services sharing the same ECMP group, resulting in problems such as increased network system processing delay and low data transmission efficiency. For example, network devices can provide higher-quality communication services for high-priority services and improve QoE.

Taking the priority information included in the first packet as an example below, the acquisition of the corresponding ECMP group by the network device 310 according to the destination identifier of the first packet will be described.

First, the network device 310 may acquire multiple ECMP groups according to the destination identifier of the first packet. Wherein, multiple ECMP groups correspond to different priorities. For example, the priority of ECMP group 1 shown in FIG. 3 is higher than the priority of ECMP group 2 .

Second, the network device 310 determines the ECMP group 1 corresponding to the first packet from multiple ECMP groups according to the priority information of the first packet.

In a possible example, the network device 310 may determine the service type corresponding to the first message according to the priority information of the first message, and query the service type corresponding to the service type in an access control list (access control list, ACL). One or more forwarding ports, and then, the network device 310 uses the path corresponding to the one or more forwarding ports as the aforementioned equal-cost forwarding path, and uses the equivalent forwarding path corresponding to the one or more forwarding ports as the ECMP Group 1.

In a possible specific situation, if the priority information is a DSCP value, after the message carrying the DSCP value arrives at the network device 310, the network device 310 queries the ACL based on the DSCP value to determine the outbound interface for forwarding the first message, such as One DSCP value corresponds to one or more outgoing interfaces, and the outgoing interface refers to a forwarding port provided by the network device 310 to the service corresponding to the DSCP value. Furthermore, the network device 310 determines the aforementioned equal-cost forwarding paths through the outgoing interfaces corresponding to the DSCP values, and uses the equal-cost forwarding paths corresponding to all the outgoing interfaces as the aforementioned EMCP group 1 .

It should be noted that FIG. 3 only shows the two ECMP groups provided by the network device 310 for the destination identifier of the first packet, but when the network scale of the network system to which the network device 310 belongs is large enough, for the second The network device 310 may also provide more ECMP groups according to the destination identifier of the message, which is not limited in this application. Conversely, when the network scale of the network system to which the network device 310 belongs is small, the network device 310 may only provide ECMP group 1 or ECMP group 2 shown in FIG. 3 for the destination identifier of the first packet.

Please continue to refer to FIG. 3 , the path switching method provided by the embodiment of the present application further includes steps S330 and S340.

S330. The network device 310 determines the first path corresponding to the first packet from the ECMP group 1.

As shown in FIG. 3 , ECMP group 1 includes three equal-cost forwarding paths, and the first path may refer to path 1 shown in FIG. 3 .

In a first possible example, the manner in which the network device 310 determines the initial path of the first packet may be: the network device 310 randomly selects an equal-cost forwarding path from ECMP group 1 as the first path of the first packet.

In the second possible example, the way the network device 310 determines the initial path of the first packet may be: the network device 310 sorts the multiple equal-cost forwarding paths in ECMP group 1 after numbering, and according to the multiple equal-cost forwarding paths The sequence of paths and the sequence of packets in the data stream match the first path corresponding to the first packet.

In a third possible example, the manner in which the network device 310 determines the initial path of the first packet may be: the network device 310 randomly selects an equal-cost forwarding path from ECMP group 1 as the first path of the first packet.

In a fourth possible example, the manner in which the network device 310 determines the initial path of the first packet may be: the network device 310 parses the information in the packet header of the first packet, and determines that the first packet is in the ECMP Equivalent forwarding paths in group 1. For a specific implementation manner of the fourth possible example, reference may be made to the description of FIG. 5 below, and details are not repeated here.

S340. When the path 1 is in an abnormal state, the network device 310 forwards the first packet through the path 2 included in the ECMP group 1.

Wherein, the network device 310 determines that the path 1 is in an abnormal state according to the abnormality information sent by the network device 321 .

For example, the abnormal information sent by the network device 321 indicates that: the packet received through path 1 affects the forwarding efficiency of the network device 321 . The forwarding efficiency of the network device 321 may be represented by any one or a combination of several of the following: bandwidth of path 1, transmission delay of path 1, packet loss rate of data packets in path 1, and the like.

In the path switching method provided by the embodiments of the present application, during the process of forwarding packets by the first network device (such as the above-mentioned network device 310), since the first network device can ) to determine that the first path corresponding to the first message is in an abnormal state, thus avoiding the problem that the first network device cannot perceive the abnormal state (such as congestion or packet loss) of the downstream device. Moreover, the first network device can also forward the aforementioned first message through other equivalent forwarding paths in the ECMP group that are consistent with the destination identifier of the first path, such as the second path, so as to avoid that the first message cannot be forwarded to The device corresponding to the destination identifier improves the accuracy of data communication, reduces the transmission delay caused by the abnormal state of the path, and improves the transmission efficiency of the first message.

On the basis of the network device 310 and the network device 321 shown in FIG. 3, the embodiment of the present application provides a possible implementation method for the network device to determine that the equivalent forwarding path is in an abnormal state, as shown in FIG. 4, which is Flowchart 2 of a path switching method provided in this application, port 1 to port 3 are the ports of the equivalent forwarding path provided by ECMP group 1 for the destination identifier of the first message, for example, network device 310 through port 1 A path 1 for transmitting the first packet is established with the network device 321 .

It should be noted that FIG. 4 only shows path 1 included in ECMP group 1, but network device 310 may also establish other equivalent forwarding paths consistent with the destination identifier of path 1 with other network devices. For example, the network device 310 establishes a path 2 for transmitting packets with the aforementioned network device 322 through port 2;

An analysis engine for detecting the forwarding efficiency of the network device 321 may be deployed in the network device 321. The analysis engine may refer to a processor included in the network device 321, or may refer to a virtual machine or a container provided by the network device 321 using virtualization technology. (container) or management node (manage node), and the forwarding efficiency of the network device 321 is detected by the virtual machine, container or management node.

As shown in FIG. 4, the path switching method provided by the embodiment of the present application includes steps S410 to S450.

S410, the network device 310 sends the first packet to the network device 321.

S420, the network device 321 determines whether the forwarding efficiency of the network device 321 is abnormal.

For example, the network device 321 judges whether the port connected to the port 1 in the network device 321 is valid, and if the port connected to the port 1 is valid, it indicates that the forwarding efficiency of the network device 321 is normal; if the port connected to the port 1 is invalid, it indicates The forwarding efficiency of the network device 321 is abnormal.

In another example, the network device 321 determines whether the message queue to be processed in the network device 321 is congested. Among them, the message queue includes a group of queue pairs (queue pair, QP), and QP includes a sending queue (send queue, SQ) and a receiving queue (receive queue, RQ). For example, the message queue used to send data in a network device is a sending Queue, the message queue used to receive data in the network device is the receive queue. For more information about the message queue, reference may be made to related descriptions of common technologies, which will not be repeated here.

For another example, the network device 321 judges the packet loss rate in the historical statistics period. The packet loss rate refers to the ratio of the number of lost data packets in data communication to the total number of sent data packets. For example, network device 321 can judge whether the forwarding efficiency of network device 321 is abnormal according to whether the packet loss rate of path 1 is greater than a threshold within 10 seconds. If the packet loss rate is greater than or equal to 1‰, then determine the forwarding efficiency of network device 321 An exception occurs.

In addition, the network device 321 may also detect the forwarding efficiency of the network device 321 according to the data flow in the path 1, or the transmission delay of the path 1.

For example, if the data flow in the path 1 is greater than or equal to the first threshold, it is determined that the forwarding efficiency of the network device 321 is abnormal; if the data flow in the path 1 is smaller than the first threshold, then it is determined that the forwarding efficiency of the network device 321 is in a normal state .

As another example, if the transmission delay of path 1 is greater than or equal to the second threshold, it is determined that the forwarding efficiency of the network device 321 is abnormal; if the transmission delay of path 1 is less than the second threshold, it is determined that the forwarding efficiency of the network device 321 is normal state. For example, the second threshold is 10 milliseconds (milli second, ms), 15 ms, or other time.

It should be noted that the above-mentioned threshold of packet loss rate, the first threshold corresponding to the data flow, and the second threshold corresponding to the transmission delay can be preset, or can be based on the processing capability of the network equipment and the network provided by the network system. If the resource size or other information is determined, it will not be limited.

If the forwarding efficiency of the network device 321 is not abnormal, step S430 is executed.

S430. The network device 321 forwards the first packet to other devices.

The other device may refer to the device indicated by the next routing address after the network device 321 in path 1 . The other device may be the destination device indicated by the destination identifier of the first packet, or may be another network device in path 1 that forwards the first packet.

If the forwarding efficiency of the network device 321 is abnormal, step S440 is executed.

S440, the network device 321 sends abnormality information to the network device 310.

Exemplarily, after determining that the forwarding efficiency of the network device 321 is abnormal, the network device 321 generates an abnormality information, and sends the abnormality information to the network device 310 . For the specific content of the abnormal information, reference may be made to the relevant description of S340 above, and details are not repeated here.

S450, the network device 310 determines that the path 1 is in an abnormal state according to the abnormality information sent by the network device 321.

Optionally, path 1 is in an abnormal state, which may include, but is not limited to: path 1 is marked as abnormal, or path 1 is deleted, or a parameter of path 1 exceeds a set threshold. The foregoing abnormal state will be described as an example below.

In the first possible situation, path 1 is marked as abnormal means: path 1 is marked as abnormal. For example, the path 1 is marked as unavailable.

In the second possible situation, path 1 is marked as abnormal means: path 1 is deleted. Exemplarily, the network device 310 moves the path 1 out of the equal-cost forwarding paths included in the ECMP group 1 . For example, the port 1 provided by the network device 310 to the path 1 is disabled, or the next-hop route (or IP address) of the network device 321 in the path 1 is deleted or cancelled, and so on.

In the third possible situation, the marking of path 1 as abnormal means that the parameter of path 1 exceeds a set threshold. For example, each equivalent forwarding path included in ECMP group 1 corresponds to a path parameter value. In an ECMP group, the maximum value and minimum value of the path parameter value are determined. If the path parameter value corresponding to path 1 is set to If it does not belong to the path parameter value corresponding to ECMP group 1, it is determined that the parameter of path 1 exceeds the set threshold, and then it is determined that path 1 is marked as abnormal.

In this embodiment, because the network device can determine the equivalent forwarding path in the abnormal state in the ECMP group according to the abnormal information fed back by the downstream device (such as the network device 321 ). Furthermore, the network device can switch the packets that need to be transmitted by the equal-cost forwarding path in the abnormal state to the normal equal-cost forwarding path (or an idle equal-cost forwarding path), which reduces the transmission of packets in the network system. The time delay improves the transmission efficiency of the network system.

It should be noted that the above three possible situations are only examples provided by this embodiment, and should not be construed as limiting the present application.

In the path switching method provided in the above embodiments, the network device 310 determines path 1, and forwards packets through path 2 when path 1 is in an abnormal state. Here, a possible implementation is provided, as shown in FIG. 5 . 5 is the third schematic flow diagram of a path switching method provided in this application. The path switching method can be executed by the above-mentioned network device 310. The determination process of the equivalent forwarding path in the path switching method includes the following steps 1 to 5.

In step 1, the network device 310 selects a hash factor according to the configuration information of ECMP group 1, and extracts data in fields related to path selection in the message as input information for the path calculation process.

The configuration information of the ECMP group 1 may include but not limited to: source IP address, destination IP address, source MAC address, virtual MAC address, virtual local area network (virtual local area network, VLAN), transport layer source port, transport layer destination port, One or more information in IP protocol, flow label and physical source port.

Step 2, the network device 310 selects the corresponding algorithm, the scrambling item (salt value), and the effective bit selection according to the processing capability of the processor included in the network device 310 (the general calculation value selects the required bits, 8bit), and outputs the corresponding The hash calculation result. For example, the algorithm is a cyclic redundancy check (cyclic redundancy check, CRC) algorithm, specifically, the CRC32 algorithm.

Step 3, the network device 310 takes the modulus of the available path N to obtain the initial path modulo result according to the hash calculation result, selects the equivalent forwarding path (such as path 1) corresponding to the initial path modulo result in ECMP group 1, and forwards the message arts. Wherein, N refers to the number of equal-cost forwarding paths included in the ECMP group 1.

Step 4, when the path determined in step 3 is in an abnormal state, the network device 310 adds an offset factor to the path calculation result of the message to adjust the path modulo result of the message. The offset factor can also be called an offset value.

In step 5, the network device 310 determines the packet forwarding path (such as the aforementioned path 2) according to the adjusted path modulo result.

As shown in Figure 6, Figure 6 is a schematic diagram of the path adjustment process in the ECMP group provided by this application, the destination device 330 can be implemented by the destination device 130 shown in Figure 1, or by any server shown in Figure 2 accomplish. Path 1 to path 3 shown in FIG. 6 are three equivalent forwarding paths for communication between network device 310 and destination device 330. If network device 310 determines that the path (such as path 1) of the message from source device 300 is abnormal state, the network device 310 adds an offset factor to the modulo result of the initial path of the message. For example, if the value corresponding to the offset factor is 2, the network device 310 switches the equivalent forwarding path of the message from path 1 to path 3.

In a possible situation, the network device 310 may also offset the first path (path 1 shown in FIG. 6 ) corresponding to the packet (such as the first packet in the foregoing embodiment) by a set value , to obtain a second path (path 2 shown in FIG. 6 ); furthermore, the network device 310 forwards the message through the second path.

For example, each path corresponds to a set path value, and the network device 310 offsets the set path value of path 1, so that the offset path value corresponds to the path value of path 2, and then, the network device 310 The path 2 is searched according to the path value corresponding to the path 2, and the first packet is sent through the path 2. For example, the path value of path 1 is "0001", and the path value of path 2 is "0002". If the first message needs to be switched from path 1 to path 2, the above offset factor (or offset value) can be "0001", the network device 310 obtains the path value of the path 2 according to the offset factor and the path value of the path 1, and then, the network device 310 forwards the first message through the path 2.

In this embodiment, the network device may add an offset factor to the path modulo result when the equivalent forwarding path corresponding to the path modulo result is in an abnormal state according to the path modulo result obtained by path calculation, so that Realize the value set by offsetting the equivalent forwarding path (such as the first path) corresponding to the path modulo result to obtain the second path. In addition, the network device can also use the offset path modulo result to forward the message corresponding to the second path, so as to realize the forwarding of the message and improve the forwarding efficiency of the network system.

In the path switching method provided by the embodiments of the present application, in the process of path calculation, no matter whether the equal-cost forwarding path included in the ECMP group is abnormal, the network device does not need to renumber the equal-cost forwarding path. When the path is in an abnormal state, the network device will offset the abnormal path by the set value, so as to determine that the path for forwarding the message is another normal path, and forward the message through the normal path, avoiding the time-lapse of message transmission. The problem of high delay is solved, and the data transmission efficiency of the network system is improved.

For above-mentioned message below, a kind of feasible implementation mode is provided, as shown in Figure 7, Fig. 7 is the structural representation of a kind of IPv4 message that the present application provides, IPv4 message comprises message header (header) and For the data part, in combination with the IPv4 message shown in FIG. 7 , Table 1 below provides descriptions of some fields included in the message header of the IPv4 message.

Table 1

Wherein, the service type field may also be referred to as a differentiated service field. The service type field can be used to store the priority information provided by the foregoing embodiments, such as the DSCP value.

The identification field in the message header of the IPv4 message may also be referred to as the IP ID field of the IPv4 message. The IP ID field can be used to store the offset value provided by the foregoing embodiments. For example, since the length of the IP ID field is 8 bits, you can select the head 4 bits or tail 4 bits of the IP ID field, or any continuous 4 bits in the middle to set the offset value. For example, if the header 4bit of the IP ID field of the message is set to "0x02", then when the equivalent forwarding path corresponding to the path calculation result of the message is path 1 (in an abnormal state), since the offset value is "0x02", the actual forwarding routing path of the packet will be shifted to path 3, which prevents the packet from being forwarded through an abnormal path and improves the efficiency of packet forwarding.

In addition, the option field is also called the extended option header of the packet header, and the extended option header can also be used to store the offset value provided by the foregoing embodiments.

In the process of data communication, if the network system needs to transmit a large amount of data, messages belonging to the same data flow can be adjusted from the abnormal path to the normal path included in the ECMP group, realizing the timeliness of path adjustment in the ECMP group. In addition, because the network device can determine the equivalent forwarding path of message forwarding by setting the offset value, the accuracy of path adjustment is improved, and when multiple messages belong to the same data flow, the data flow is realized. precise regulation.

As shown in FIG. 8 , FIG. 8 is a fourth schematic flowchart of a path switching method provided in the present application. The path switching method provided in the embodiment of the present application further includes the following steps S810 and S820 .

S810. The source device 300 sends the second packet to the network device 310.

Correspondingly, the network device 310 receives the second packet.

Wherein, the second packet and the first packet in the foregoing embodiment both belong to the first flow, and the second packet includes an offset value. The destination identifiers of multiple packets belonging to the same data flow are the same, for example, the destination identifier of the first packet and the destination identifier of the second packet are the same.

In a feasible implementation manner, the offset value is included in the packet header of the second packet.

In an example, if the second packet is an IPv4 packet, the offset value can be set in the IP ID field or option field included in the packet header.

In another example, if the second packet is an IPv6 packet, the offset value may be set in a flow label (flow label) field or an option field included in a packet header of the IPv6 packet.

It should be noted that when the aforementioned first packet and second packet are data packets based on other protocols, the offset value can also be set in other positions of the packet header, which is not limited.

In a first optional implementation manner, the offset value included in the second packet is set by the source device 300 . For example, when the network device 310 determines that the path 1 for transmitting the first packet is in an abnormal state, the source device 300 obtains the abnormality information that the path 1 is in an abnormal state, and sends other packets of the first flow (such as the above-mentioned During the transmission of the second packet), offset values are added to other packets of the first stream.

In some optional situations, each path included in the ECMP group has its own path value (or serial number, serial number, path flag, etc.), such as the path value of path 1 is "0001", and the path value of path 2 is "0002 ", if the second message needs to be switched from path 1 to path 2, the above-mentioned offset value can be determined by the difference between the path value of path 2 and the path value of path 1, such as the offset value is " 0001". In addition, the offset value may also be determined by the difference between the path value of any other normal path (such as path 3 ) included in ECMP group 1 and the path value of path 1 . It should be noted that this example uses decimal as an example to illustrate the path value of the path, but in other optional situations, the path value and offset value of the path can also be in binary, octal, hexadecimal system or other bases are not limited.

In a second optional implementation manner, the offset value included in the second packet is set by the network device 310 . For example, when the network device 310 determines that path 1 for transmitting the first packet is in an abnormal state, the network device 310 sets offset values for packets of all data flows that path 1 needs to transmit.

It is worth noting that a data flow includes many packets, and different offset values can be added to multiple packets so that the multiple packets can be allocated to different equal-cost forwarding paths for transmission, reducing the cost of the data flow. Transmission delay, improve the overall transmission efficiency of data flow.

S820. The network device 310 forwards the second packet through path 2 according to the offset value in the second packet.

As an example, assume that in the preset traffic load sharing policy, multiple packets belonging to the first flow are all forwarded by path 1 included in ECMP group 1. After the network device 310 determines that path 1 is in an abnormal state, the source device 300 (or the network device 310) adds an offset value indicating an equivalent forwarding path to other packets (such as the second packet) belonging to the first flow. For example, if the offset value between path 2 and path 1 is 1, and the offset value between path 3 and path 1 is 2, then if the second packet is about to be forwarded through path 2, the network device 310 will The offset value added to the message is 1; if the second message is about to be forwarded through path 3, the network device 310 adds an offset value of 2 to the second message. It should be understood that the above-mentioned specific value of the offset value is only an example provided by this embodiment, and the offset value between any two equivalent forwarding paths may also be 2, 3 or other values, etc., which is not limited.

In the path switching method provided by the embodiment of this application, when the first network device (such as the above-mentioned network device 310) determines that one or more equivalent forwarding paths included in the ECMP group are in an abnormal state, The packet forwarded by the path is switched to the normal equal-cost forwarding path in the ECMP group for forwarding, which improves the transmission efficiency of the data flow in the ECMP group.

In addition, because the equivalent forwarding path of message forwarding can be determined by setting the offset value, the accuracy of path adjustment is improved, and the accuracy of data flow is realized when multiple messages belong to the same data flow. regulation. For example, the network device or the source device can set an offset value for a specific field, thereby adjusting the internal path selection in the ECMP group, and realizing precise adjustment of the routing path between the source device and the destination device.

In the case that the network device does not need to modify the number of paths taken modulo in the ECMP group, or in the case that the network device does not need to adjust all the equal-cost forwarding paths in the ECMP group, the network device can use the minimum cost (such as the aforementioned overhead value) ) solves the problem of path abnormality (such as congestion, packet loss or high delay) in the ECMP group, and improves the data transmission efficiency of the network system.

Finally, since the network device quickly adjusts the path in the ECMP group after obtaining the abnormal information, the timeliness of path adjustment in the ECMP group is also improved on the basis of improving the data transmission efficiency of the network system.

In the path switching method provided by the above-mentioned embodiments of the present application, multiple equal-cost forwarding paths included in the ECMP group are used as an example for illustration, but the path switching method provided by the present application can also be applied to other routes that are similar to routing load sharing. Scenarios, such as the path communication scenario implemented by multiple network devices based on the Trunk interface, will not be described in detail here.

It can be understood that, in order to implement the functions in the foregoing embodiments, the network device includes hardware structures and/or software modules corresponding to each function. Those skilled in the art should easily realize that the present application can be implemented in the form of hardware or a combination of hardware and computer software with reference to the units and method steps of the examples described in the embodiments disclosed in the present application. Whether a certain function is executed by hardware or computer software drives the hardware depends on the specific application scenario and design constraints of the technical solution.

As shown in FIG. 9, FIG. 9 is a schematic structural diagram of a path switching device provided by the present application. The path switching device 900 can be used to implement the functions of the network device in the above method embodiment, so it can also implement the above method embodiment. have beneficial effects. In the embodiment of the present application, the path switching device 900 may be a network device, such as any network device, the source device 110 or the destination device 130 shown in FIG. 1 above, or any switch shown in FIG. 2 Or the server, or the source device 300 shown in FIG. 3 or any network device, such as the network device 310 or the network device 321, etc., may also be a module (such as a chip) applied to the network device.

As shown in FIG. 9 , the path switching apparatus 900 includes: a receiving unit 910 , an acquiring unit 920 , a determining unit 930 and a forwarding unit 940 . The path switching apparatus 900 is used to implement the functions of the network equipment in the method embodiments shown in the above figures. In a possible example, the path switching device 900 is used to implement the above path switching method, and the functions of components of the path switching device 900 will be described in detail below.

The receiving unit 910 is configured to receive a first packet, where the first packet belongs to the first flow.

The acquiring unit 920 is configured to acquire a corresponding ECMP group according to the destination identifier of the first message, and the ECMP group includes multiple equivalent forwarding paths corresponding to the aforementioned destination identifier.

The determining unit 930 is configured to determine the first path corresponding to the first packet from the ECMP group, and determine whether the first path is in an abnormal state. Wherein, the determining unit 930 determines that the first path is in an abnormal state according to the abnormal information sent by the second network device.

The forwarding unit 940 is configured to forward the first message through the second path in the ECMP group when the determining unit 930 determines that the first path is in an abnormal state.

When the path switching device 900 is used to realize the function of the network device 310 in the method embodiment shown in FIG. 940 is used to execute S340.

When the path switching device 900 is used to realize the function of the network device 310 in the method embodiment shown in FIG. S430.

When the path switching device 900 is used to implement the method embodiment shown in FIG. 5 : the receiving unit 910 , the acquiring unit 920 , the determining unit 930 and the forwarding unit 940 are used to coordinately execute steps 1 to 5 shown in FIG. 5 .

When the path switching apparatus 900 is used to implement the function of the network device 310 in the method embodiment shown in FIG. 8 , the receiving unit 910 is used to perform S810, and the forwarding unit 940 is used to perform S820.

Regarding the beneficial effect when the path switching device 900 executes the operation steps of the method in the embodiment provided in the preceding figures, reference may be made to the description of the foregoing embodiment, and details are not repeated here.

When the path switching device 900 implements any of the path switching methods shown in the preceding figures through software, the path switching device 900 and its modules may also be software modules. The above-mentioned path switching method is realized by calling the software module by the processor. The processor can be a central processing unit (central processing unit, CPU), a specific application integrated circuit (application-specific integrated circuit, ASIC) implementation, or a programmable logic device (programmable logic device, PLD), and the above-mentioned PLD can be a complex program Logic device (complex programmable logical device, CPLD), field programmable gate array (field programmable gate array, FPGA), general array logic (generic array logic, GAL) or any combination thereof.

For a more detailed description of the above-mentioned path switching apparatus 900, reference may be made to relevant descriptions in the embodiments shown in the preceding figures, and details are not repeated here.

When the path switching device 900 is implemented by hardware, the hardware may be implemented by a processor or a chip. The chip includes an interface circuit and a control circuit. The interface circuit is used to receive data from devices other than the processor and transmit it to the control circuit, or send data from the control circuit to devices other than the processor.

The control circuit implements the method in any possible implementation manner in the foregoing embodiments through a logic circuit or by executing code instructions. For the beneficial effects, reference may be made to the description in any of the foregoing embodiments, and details are not repeated here.

Optionally, the path switching apparatus 900 may also be implemented by a network device, as shown in FIG. 10 , which is a schematic structural diagram of a network device provided in this application. The network device 1000 includes a communication interface 1010 and a processor 1020 . It can be understood that the communication interface 1010 may be a transceiver or an input/output (input/output, I/O) interface. Optionally, the network device 1000 may further include a memory 1030 for storing instructions executed by the processor 1020 or storing input data required by the processor 1020 to execute the instructions, or storing data generated by the processor 1020 after executing the instructions. The network device 1000 may refer to the network device 310 in the foregoing embodiments.

When the network device 1000 is used to implement the embodiments shown in the foregoing figures, the processor 1020, the communication interface 1010 and the memory 1030 can also cooperate to implement various operations in the path switching method performed by each network device and source device in the network system step. The network device 1000 may also perform the functions of the path switching apparatus 900 shown in FIG. 9 , which will not be described in detail here.

In the embodiment of the present application, the specific connection manner among the communication interface 1010, the processor 1020, and the memory 1030 is not limited.

The memory 1030 can be used to store software programs and modules, such as program instructions/modules corresponding to the path switching method provided in the embodiment of the present application, and the processor 1020 executes various functional applications by executing the software programs and modules stored in the memory 1030 and data processing. The communication interface 1010 can be used for signaling or data communication with other devices. In this application, the network device 1000 may have multiple communication interfaces 1010 .

The communication interface 1010 can be used to provide the forwarding ports of multiple equal-cost forwarding paths included in the ECMP group 1, such as port 1 to port 3 shown in FIG. port 2 is the forwarding port provided by the network device 310 for path 2 in the foregoing embodiment, and port 3 is the forwarding port provided by the network device 310 for path 3 in the foregoing embodiment.

The present application also provides another possible network device, as shown in FIG. 11 , which is a schematic structural diagram of another network device provided by the present application. The network device 1100 includes: a processor 1110, a processor 1120 and an I/O O interface 1130. For example, the network device 1100 may refer to the network device 321 or the network device 322 in the foregoing embodiments, or other downstream devices of the network device 310 .

The I/O interface 1130 can be used to receive messages from other devices. The network device 1100 may include one or more I/O interfaces 1130, which is not limited.

An analysis engine may be provided in the processor 1120. For example, the analysis engine refers to a virtual machine, container (container) or management node (management node) provided by the network device 1100 using virtualization technology, and is detected by the virtual machine, container or management node. The forwarding efficiency of the network device 1100.

By arranging two different processors in the network device 1100, one of the processors 1120 is used to detect the forwarding efficiency of the network device 1100, which is beneficial to quickly obtain the abnormal information of the network device 1100, so that the processor 1110 takes the abnormal information Forward to the upstream device of network device 1100. Exemplarily, when the network device 1100 refers to the above-mentioned network device 321, the upstream device of the network device 1100 may be the network device 310. When the network device 310 obtains the abnormal information fed back by the network device 1100, the network device 310 and Flows between network devices 1100 are switched from one equal-cost forwarding path to other equal-cost forwarding paths.

It can be understood that the processor 1120 in the embodiment of the present application may be a CPU, a neural processing unit (neural processing unit, NPU) or a graphics processing unit (graphic processing unit, GPU), and may also be other general-purpose processors, digital Signal processor (digital signal processor, DSP), ASIC, FPGA or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. A general-purpose processor can be a microprocessor, or any conventional processor.

The method steps in the embodiments of the present application may be implemented by means of hardware, or may be implemented by means of a processor executing software instructions. Software instructions can be composed of corresponding software modules, and software modules can be stored in random access memory (random access memory, RAM), flash memory, read-only memory (read-only memory, ROM), programmable read-only memory (programmable ROM) , PROM), erasable programmable read-only memory (erasable PROM, EPROM), electrically erasable programmable read-only memory (electrically EPROM, EEPROM), register, hard disk, mobile hard disk, CD-ROM or known in the art any other form of storage medium. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. Of course, the storage medium may also be a component of the processor. The processor and storage medium can be located in the ASIC. In addition, the ASIC can be located in a network device or a terminal device. Certainly, the processor and the storage medium may also exist in the network device or the terminal device as discrete components.

In each embodiment of the present application, if there is no special explanation and logical conflict, the terms and/or descriptions between different embodiments are consistent and can be referred to each other, and the technical features in different embodiments are based on their inherent Logical relationships can be combined to form new embodiments.

In this application, "at least one" means one or more, and "multiple" means two or more. "And/or" describes the association relationship of associated objects, indicating that there may be three types of relationships, for example, A and/or B, which can mean: A exists alone, A and B exist at the same time, and B exists alone, where A, B can be singular or plural. In the text description of this application, the character "/" generally indicates that the contextual objects are an "or" relationship; in the formulas of this application, the character "/" indicates that the contextual objects are a "division" Relationship.

It can be understood that the various numbers involved in the embodiments of the present application are only for convenience of description, and are not used to limit the scope of the embodiments of the present application. The size of the serial numbers of the above-mentioned processes does not mean the order of execution, and the execution order of each process should be determined by its functions and internal logic.

Claims (24)

1. A path switching method, characterized in that the method is performed by a first network device, and the method includes:

   receiving a first packet, where the first packet belongs to the first flow;

   Obtaining a corresponding equal-cost path load sharing ECMP group according to the destination identifier of the first message, where the ECMP group includes multiple equal-cost forwarding paths corresponding to the destination identifier;

   determining a first path corresponding to the first packet from the ECMP group;

   When the first path is in an abnormal state, forward the first message through a second path in the ECMP group;

   Wherein, the first network device determines that the first path is in an abnormal state according to the abnormality information sent by the second network device.
2. The method according to claim 1, wherein the abnormal information sent by the second network device indicates that the packet received through the first path affects the forwarding efficiency of the second network device.
3. The method according to claim 1 or 2, wherein the abnormal state of the first path includes: the first path is marked as abnormal, or the first path is deleted, or the first The parameter of the path exceeds the set threshold.
4. The method according to claim 3, wherein the deleting of the first path comprises: invalidating the port provided by the first network device to the first path, or, the second network The next-hop route of the device in the first path is deleted.
5. The method according to any one of claims 1 to 4, characterized in that,

   The forwarding of the first message through the second path in the ECMP group includes:

   offsetting the first path by a set value to obtain the second path;

   Forwarding the first packet through the second path.
6. The method according to any one of claims 1 to 4, wherein the method further comprises:

   receiving a second packet, where the second packet belongs to the first flow, and the second packet includes an offset value;

   Forwarding the second packet through the second path according to the offset value in the second packet.
7. The method according to claim 6, wherein the offset value is included in a header of the second packet.
8. The method according to claim 7, characterized in that,

   If the first message is a fourth-generation Internet Protocol IPv4 message, the extended option header or the Internet Protocol Identification IP ID field included in the IPv4 message stores the offset value; or,

   If the first packet is a sixth-generation Internet protocol IPv6 packet, the offset value is stored in an extended option header or a flow label field included in the IPv6 packet.
9. The method according to any one of claims 1 to 8, wherein the first message includes priority information;

   The obtaining the corresponding equal-cost path load sharing ECMP group according to the destination identifier of the first message includes:

   Acquire multiple ECMP groups according to the destination identifier of the first message, and the multiple ECMP groups correspond to different priorities;

   Determining an ECMP group corresponding to the first packet from the plurality of ECMP groups according to the priority information.
10. The method according to claim 9, wherein the priority information is a Differentiated Services Code Point (DSCP) value.
11. A path switching device, characterized in that the device is applied to a first network device, and the device includes:

    a receiving unit, configured to receive a first packet, where the first packet belongs to the first stream;

    An acquiring unit, configured to acquire a corresponding ECMP group according to the destination identifier of the first message, where the ECMP group includes multiple equivalent forwarding paths corresponding to the destination identifier;

    A determining unit, configured to determine a first path corresponding to the first message from the ECMP group, and determine whether the first path is in an abnormal state;

    A forwarding unit, configured to forward the first message through a second path in the ECMP group when the determining unit determines that the first path is in an abnormal state;

    Wherein, the determining unit determines that the first path is in an abnormal state according to the abnormal information sent by the second network device.
12. The apparatus according to claim 11, wherein the abnormal information sent by the second network device indicates that the packet received through the first path affects the forwarding efficiency of the second network device.
13. The device according to claim 11 or 12, wherein the abnormal state of the first path includes: the first path is marked as abnormal, or the first path is deleted, or the first The parameter of the path exceeds the set threshold.
14. The apparatus according to claim 13, wherein the deletion of the first path comprises: invalidating the port provided by the first network device to the first path, or the second network The next-hop route of the device in the first path is deleted.
15. The device according to any one of claims 11 to 14, wherein the forwarding unit is specifically used for:

    offsetting the first path by a set value to obtain the second path;

    Forwarding the first packet through the second path.
16. Apparatus according to any one of claims 11 to 14, characterized in that

    The receiving unit is further configured to receive a second message, the second message belongs to the first flow, and the second message includes an offset value;

    The forwarding unit is further configured to forward the second message through the second path according to the offset value in the second message.
17. The device according to claim 16, wherein the offset value is included in a packet header of the second packet.
18. The device according to claim 17, wherein, if the first message is a fourth-generation Internet protocol IPv4 message, the extended option header or the Internet protocol identifier IP ID field included in the IPv4 message is stored the above offset value; or,

    If the first packet is a sixth-generation Internet protocol IPv6 packet, the offset value is stored in an extended option header or a flow label field included in the IPv6 packet.
19. The device according to any one of claims 11 to 18, wherein the first message includes priority information;

    The acquiring unit is specifically configured to acquire multiple ECMP groups according to the destination identifier of the first message, and the multiple ECMP groups correspond to different priorities;

    The determining unit is further configured to determine the ECMP group corresponding to the first message from the plurality of ECMP groups according to the priority information.
20. The device according to claim 19, wherein the priority information is a Differentiated Services Code Point (DSCP) value.
21. A network device, characterized in that it includes a memory and a processor, wherein:

    The memory is used to store program codes;

    The processor is configured to invoke the program code to implement the method described in any one of claims 1-10.
22. A network system, characterized by comprising: a first network device and a second network device;

    The first network device receives a first packet, and the first packet belongs to a first flow;

    The first network device obtains a corresponding equal-cost path load sharing ECMP group according to the destination identifier of the first message, and the ECMP group includes a plurality of equal-cost forwarding paths corresponding to the destination identifier;

    The first network device determines a first path corresponding to the first message from the ECMP group;

    When the first path is in an abnormal state, the first network device forwards the first message to the second network device through a second path in the ECMP group;

    Wherein, the first network device determines that the first path is in an abnormal state according to the abnormality information sent by the second network device.
23. The system according to claim 22, further comprising: a source device;

    The source device sends a second packet to the first network device, the second packet belongs to the first flow, and the second packet includes an offset value;

    The first network device forwards the second packet to the second network through the second path according to the offset value in the second packet.
24. The system according to claim 23, wherein the offset value is included in a packet header of the second packet.

Images