A distributed routing method for label forwarding paths
By deploying distributed routing modules on network elements, real-time monitoring of path status and autonomous calculation of new paths are achieved, solving the problems of long processing time and slow network fault recovery in traditional rerouting methods, and realizing fast and stable network fault recovery and traffic recovery.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- THE 34TH RES INST OF CHINA ELECTRONICS TECH CORP
- Filing Date
- 2023-11-20
- Publication Date
- 2026-07-21
AI Technical Summary
Traditional rerouting methods only perform calculations when all paths in a protection group fail, which is time-consuming and cannot cope with fault recovery in complex network environments. Furthermore, centralized controllers are prone to becoming isolated, resulting in excessively long network fault recovery times and network congestion caused by synchronization messages.
A distributed routing module is deployed in each network element. By collecting network topology information in real time, monitoring path status, and autonomously calculating new paths in case of failure, the distributed routing module interacts with the topology module to generate and configure new paths, avoiding the involvement of the centralized controller and achieving fast rerouting.
It enables rapid path recovery during network failures, reduces packet loss, improves network stability and reliability, avoids network traffic congestion, and reduces resource synchronization requirements.
Smart Images
Figure CN117376246B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of packet transport network communication routing, specifically to a distributed routing method for label forwarding paths. Background Technology
[0002] Packet transport networks are a new generation of packet transport networks with software control capabilities. The equipment that makes up a packet transport network is called packet transport network equipment, which provides various applications and services in access networks, metropolitan area networks and backbone transport networks for integrated services such as mobile backhaul and enterprise leased lines.
[0003] In intelligent packet transport networks (ISPs), separating network control and data forwarding and implementing network control in software enhances network programmability and flexibility. The SDN controller is a crucial component of the ISP architecture, responsible for managing and controlling network resources.
[0004] When network services are activated, the SDN controller calculates label-switched paths for packet transport network devices based on the overall network topology and sends path configuration commands to these devices via management routing. The controller typically calculates multiple paths for point-to-point connections between devices, forming a path protection group. When a network device's working path fails, it can automatically switch to a protected path within this group, giving the device a degree of self-healing capability in the face of network failures. When a packet transport device receives a data packet, it determines the next-hop address based on the label-switched path. However, in complex network environments, when all paths in a protection group fail, the SDN controller needs to perform rerouting to calculate new paths and reroute the data packet to these new paths to avoid data transmission interruptions.
[0005] Traditional rerouting methods only perform rerouting calculations when all paths in a protection group fail, which is time-consuming and leads to excessively long network fault recovery times. Furthermore, traditional rerouting methods are typically performed by a single master controller, making them unsuitable for complex situations where network faults isolate the controller, causing devices to lose connection with it. Additionally, traditional distributed label allocation methods require resource synchronization between network elements using packets to avoid conflicts; excessive synchronization packets can easily cause congestion in network fault scenarios. Therefore, finding a fast and efficient distributed rerouting method is a pressing issue that needs to be addressed. Summary of the Invention
[0006] The purpose of this invention is to provide a distributed routing method for label forwarding paths that can quickly generate new routes without causing network traffic congestion.
[0007] The technical solutions for achieving the above objectives include the following:
[0008] A distributed routing method for label forwarding paths, characterized by the following steps:
[0009] Step 1: Deploy a routing module in each network element to build a distributed routing module. The distributed routing module interacts with the topology module to collect network topology information in real time.
[0010] Step 2: When the network is activated, the central controller configures the path and protection group to the device management module of the network element. The protection group contains one working path and several protection paths. After the protection group is configured, the device management module of the network element selects the physical forwarding path in the working and protection paths according to the protection group state machine. After the network is activated, the central controller is shut down.
[0011] Step 3: The network element device monitors the status of each path through the OAM protocol. When any path failure is detected, the device management module selects a physical forwarding path within the protection group where the faulty path is located through the protection group state machine. At the same time, it compares its own router-IP with the remote router-IP. When its own router-IP is larger, it generates an inbound / outbound label for the new path based on the obtained protection group ID of the faulty path, and encapsulates the inbound / outbound label into an alarm message and sends it to the distributed routing module.
[0012] Step 4: After receiving the alarm, the distributed routing module calculates the route of the new path through the entire network topology, encapsulates the generated routing information into a GRPC message, establishes a connection with the device management module of the node on the new path, and closes the connection after sending the GRPC message.
[0013] Step 5: After configuring the new path for the network element, select the physical forwarding path according to the protection group state machine, continuously perform OAM monitoring on all paths, and repeat the above steps.
[0014] Furthermore, in step 1, the distributed routing module connects to the topology module via the TCP protocol to receive network topology messages.
[0015] Furthermore, in step 3, the device management module sends an alarm message to the distributed routing module through the GRPC interface to ensure that there is always a usable path in the protection group, thereby enabling rapid fault recovery.
[0016] Furthermore, in step 3, the alarm message includes the path name, remote IP address, protection group ID, and east-west label.
[0017] Furthermore, in step 4, the distributed routing module initiates an HTTP connection with the device management module of the node on the new path and distributes the GRPC message.
[0018] Compared with the prior art, the present invention has the following advantages:
[0019] 1. Since a centralized controller is not required, the distributed routing calculation module can improve network stability;
[0020] 2. Any path failure will trigger a new path routing calculation, ensuring that there is always a usable backup path on the device, resulting in a lower packet loss rate when the network is restored.
[0021] 3. The packet transport network allocates the label values used for the label switching path itself, without requiring resource synchronization between distributed routing modules. Therefore, it can achieve fast rerouting without increasing the network burden. Attached Figure Description
[0022] Figure 1 This is a schematic diagram illustrating the interaction architecture between the centralized controller and network element devices in an example network system;
[0023] Figure 2 This is a schematic diagram illustrating the interaction architecture between the distributed routing module and other modules in the network system of this embodiment.
[0024] Figure 3 This is a flowchart illustrating how the device management module triggers distributed rerouting calculations when a path failure occurs, as shown in the example.
[0025] Figure 4 This is a flowchart illustrating how the device management module on the new path receives rerouting configurations, as shown in the example. Detailed Implementation
[0026] The technical starting point of this invention is based on the packet transport network device management module reporting faults to the distributed routing module, triggering new route calculation, and realizing the distributed rerouting function of the packet transport network device. The packet transport network device collects network-wide topology information through the topology module; sends detection messages to the peer device according to the OAM protocol to detect path status; based on the path status detection results, it reports faulty paths to the distributed routing module, triggering new route calculation; after the new path calculation is completed, the device receives configuration information from the routing module to switch to the new path. See also... Figures 1 to 4 The entire process of implementing distributed rerouting is as follows:
[0027] 1. When deploying the network system, the centralized controller exists independently, establishing TCP connections with the management module of each network element in the system, and connecting to the topology module to obtain the entire network topology data, such as... Figure 1 As shown; each network element device deploys a distributed routing module, which establishes a connection with the topology module and management module, such as... Figure 2 As shown;
[0028] 2. When network services are activated, the centralized controller calculates path connections between the CE devices on the user side and sends a 1:2 path protection group to the network element devices via the GRPC interface. This group consists of one working path calculated according to the shortest path principle and two protection paths, with weights assigned based on the cost of each protection path. The network element devices receive the configuration from the centralized controller and store the paths sent by the centralized controller into the first three paths of the protection group: the first is the working path, the second is the protection path with higher weight, and the third is the protection path with lower weight. Simultaneously, the devices configure the next hop and label switching model of the switching chips through the management module, completing the path planning configuration. After network service activation is complete, the centralized controller is shut down.
[0029] The path storage data structure is shown in Table 1 below:
[0030] Table 1
[0031] The data structure of the storage protection group is shown in Table 2 below:
[0032] Table 2
[0033] 3. Network element nodes configure their switching chips according to the OAM field issued by the centralized controller and send OAM messages to detect the link status between two network element nodes. When a path failure is detected, automatic switching occurs within the protection group, and an alarm is simultaneously sent to the distributed routing module. Based on the alarm information, the distributed routing module quickly calculates a new path through the entire network topology and connects to the path configuration issued by the network element. See also Figure 3 The specific steps are as follows:
[0034] Step 3-1: When the equipment management module detects a path fault, it automatically switches via the protection group state machine: if the faulty path is the current working path, it switches to a protection path with a lower weight and then jumps to step 3-2; otherwise, the state machine ignores the current fault and jumps directly to step 3-2.
[0035] Step 3-2: The management module reads the router-IP information of the peer from the path structure and compares it with its own router-IP. If its own router-IP is smaller, the rerouting process is terminated; otherwise, proceed to step 3-3.
[0036] Step 3-3: The management module obtains the protection group ID of the faulty path and generates inbound and outbound labels for the entire path based on the protection group ID. The generation rule is: for eastbound paths (from the local end to the remote end), the inbound and outbound labels are 8000 + (id-1)*2; for westbound paths, the inbound and outbound labels are 8000 + (id*2-1). The remote IP, protection group ID, and east-west labels are encapsulated into an inter-process communication message and transmitted to the distributed routing module. The message definition for the alarm from the management module to the distributed rerouting module is given in Table 3 below:
[0037] Table 3
[0038] The meaning of each part of the message is as follows:
[0039] Name: The name of the path, a unique identifier for each path, expressed as a numeric string such as "1", "2", "3", etc.
[0040] Remote-ip: The router-ip of the remote PE node.
[0041] Tsg-id: The ID of the protection group to which the path belongs. It is a unique identifier for each protection group. The protection group IDs of the two PEs are the same and unique across the entire network.
[0042] Label-east: Eastward label, the outgoing label of the local PE, the eastward outgoing and incoming labels of intermediate nodes, and the incoming label of the peer PE. The calculation method is 8000 + (Tsg-id - 1) * 2. Since Tsg-id is unique across the entire network, the uniqueness of this label value can be guaranteed.
[0043] Label-west: Westward label, the incoming label of the local PE, the westward outgoing and incoming labels of intermediate nodes, and the outgoing label of the peer PE. The calculation method is 8000 + (Tsg-id * 2 - 1). Since Tsg-id is unique across the entire network, the uniqueness of this label value can be guaranteed.
[0044] Steps 3-4: The distributed routing module parses the message to obtain the peer's IP address, searches the topology database for available paths from the local end to the peer, and calculates a new path.
[0045] Steps 3-5: After the new path calculation is complete, the distributed routing module encapsulates the node type, next-hop exit, next-hop MAC address, OAM information, and protection group ID of each node on the new path into a GRPC message, initiates an HTTP connection to the management module of the nodes on the new path, distributes the message, and closes the connection after message distribution is complete.
[0046] Table 4 defines the configuration messages sent from the distributed routing module to the management module:
[0047] Table 4
[0048] The meaning of each part of the message is as follows:
[0049] Name: The name of the path, a unique identifier for each path, expressed as a numeric string such as "1", "2", "3", etc.
[0050] Type: Node type, 0 indicates that the network element is a PE (edge) node, 1 indicates a P (intermediate switch) node.
[0051] lsp-east: Path information. If the node type is pe, it represents the inbound and outbound labels and next-hop information. If the node type is p, it represents the eastbound inbound and outbound labels and next-hop information.
[0052] lsp-west: This field is ignored by PE nodes. P nodes represent eastward inbound and outbound labels and next-hop information. Tsg-id: Protection group ID. The management module retrieves the protection group based on this ID and then adds the new path to this protection group.
[0053] Oam: OAM configuration information; this field is ignored in p nodes.
[0054] The data domains Lsp-east and lsp-west are defined as shown in Table 5 below:
[0055] Table 5
[0056] In-label: The incoming label. When a PE node receives a data packet with this label value, it performs a label removal operation. When a P node receives a data packet with this label value, it performs a label swap operation, replacing the label with an out-label.
[0057] Out-label: The PE node assigns this label value to unlabeled packets sent by the CE device, while the P node exchanges the label value of packets with in-labels for out-label values.
[0058] Port: Outgoing packets are sent out through this port after tag and MAC address editing. Incoming packets will only undergo tag lookup if they are sent through this port.
[0059] DMAC: Outgoing packets will have their destination MAC field edited to DMAC, with the value being the router-mac of the next-hop device.
[0060] The Oam data domain is defined as shown in Table 6 below:
[0061] Table 6
[0062] Meg-id: oam group ID. Each path's meg-id is unique across the entire network.
[0063] Mep-id: This segment's mep-id is the same as the remote rmep-id.
[0064] Rmep-id: Remote MEP-id, same as the remote MEP-id.
[0065] cc-en: 0 indicates that cc messages are enabled, 1 indicates that cc messages are allowed.
[0066] cc-interval: CC message transmission period, 1 represents 3.3ms, 2 represents 10ms, 3 represents 100ms, 4 represents 1 second, 5 represents 10 seconds, 6 represents 1 minute, and 7 represents 10 minutes.
[0067] Steps 3-6: The distributed routing module calculates the new path;
[0068] 4. The device receives messages and configures paths, switching protection paths according to the protection group status. See [link / reference]. Figure 4 The specific steps are as follows:
[0069] Step 4-1: The device management module on the new path receives and parses the path configuration message via the GRPC interface.
[0070] Step 4-2: Configure the path according to the path type. If it is a PE node (edge node), configure the inbound / outbound label, next hop, and oam; if it is a P node (intermediate node), configure the eastbound / outbound label, eastbound next hop, westbound / outbound label, and westbound next hop.
[0071] Step 4-3: Locate the protection group based on the protection group ID, and store the new path configured by the distributed routing module in the fourth path of the protection group. If the protection group already has a fourth path, delete the existing fourth path and configure the new path.
[0072] Step 4-4: The new path is added to the state machine of the trigger protection group. The state machine will only select the fourth path to forward the service data if the first three paths are unavailable.
[0073] 5. After rerouting, continuously monitor the path status in the protection group through the OAM protocol. If there is a path failure, repeat the above process to achieve rapid recovery of network failures through distributed rerouting.
[0074] The routing method in this embodiment features technological advancements such as lower resource consumption, faster route reconstruction, and greater stability and reliability.
Claims
1. A distributed routing method for label forwarding paths, characterized in that, The steps include the following: Step 1: Deploy a routing module in each network element to build a distributed routing module. The distributed routing module interacts with the topology module to collect network topology information in real time. Step 2: When the network is activated, the central controller configures the path and protection group to the device management module of the network element. The protection group contains one working path and several protection paths. After the protection group is configured, the network element device management module selects the physical forwarding path and shuts down the central controller. Step 3: The network element device monitors the status of each path through the OAM protocol. When any path failure is detected, the device management module selects a physical forwarding path within the protection group where the faulty path is located through the protection group state machine. At the same time, it compares its own router-IP with the remote router-IP. When its own router-IP is larger, it generates an inbound / outbound label for the new path based on the obtained protection group ID of the faulty path, and encapsulates the inbound / outbound label into an alarm message and sends it to the distributed routing module. Step 4: After receiving the alarm, the distributed routing module calculates the route of the new path through the entire network topology, encapsulates the generated routing information into a GRPC message, establishes a connection with the device management module of the node on the new path, and closes the connection after sending the GRPC message. Step 5: After configuring the new path for the network element, select the physical forwarding path according to the protection group state machine, continuously perform OAM monitoring on all paths, and repeat the above steps.
2. The distributed routing method for label forwarding paths according to claim 1, characterized in that, In step 1, the distributed routing module connects to the topology module via the TCP protocol and receives network topology messages.
3. The distributed routing method for label forwarding paths according to claim 1, characterized in that, In step 3, the device management module sends an alarm message to the distributed routing module through the gRPC interface.
4. The distributed routing method for label forwarding paths according to claim 1, characterized in that, In step 3, the alarm message includes the path name, remote IP address, protection group ID, and east-west label.
5. The distributed routing method for label forwarding paths according to claim 1, characterized in that, In step 4, the distributed routing module initiates an HTTP connection with the device management module of the node on the new path and distributes the GRPC message.