A multi-chain fusion networking improved active routing method

Abstract

This invention proposes an improved active routing method for multi-chain converged networking, belonging to the field of wireless ad hoc networking technology. In the multi-chain converged networking scenario of this invention, each node carries multiple heterogeneous link interfaces. First, an optimized multi-link state routing protocol is constructed, including maintaining multi-link HELLO messages. Each node selects a multi-link multi-point relay node from all its one-hop symmetric neighbors. Each multi-point relay node sends a multi-link topology control message to construct a global topology graph. For different types of services, routes are calculated starting from each link interface of each node, and a multi-link routing table is constructed. Second, a dynamic routing parameter adjustment strategy and a passive routing-assisted forwarding strategy are set. This invention achieves effective route adaptation for multi-chain converged networking scenarios, ensuring efficient communication for multi-link inputs and multi-service requirements, accelerating route convergence and effectively reducing routing overhead, while also achieving reliable transmission in the event of route failure.

Description

Technical Field

[0001] This invention belongs to the field of wireless ad hoc network technology, specifically relating to a routing method based on the Optimized Multi-Link State Routing (OMLSR) protocol in a multi-link converged networking scenario. Background Technology

[0002] In a multi-chain converged networking communication scenario, each node is assumed to have multiple links (only some links may be active). These links are independent of each other, and their communication distance and link quality vary greatly. Switching data traffic from one link to another is costly.

[0003] Traditional routing protocols mostly consider only single-hop link routing. Even those few protocols that consider the use of multiple links and interfaces between node pairs still operate on the principle of transmitting within a single-link network each time, neglecting cross-link transmission and the integration of cross-link costs with the costs of individual links. While traditional Optimized Link State Routing (OLSR) protocols consider multiple interfaces and links between nodes, they do not actively maintain multi-link information. Furthermore, they do not consider the costs of link switching or diverse service requirements during routing table calculations. Especially when the communication ranges of different links vary significantly, some routes may fail to converge, leading to network performance degradation, network outages, and severely impacting network stability and performance.

[0004] In summary, there is an urgent need for a multi-link routing protocol that can adapt to multi-link converged networking scenarios, dynamically optimize route selection, dynamically adjust route maintenance overhead, and ensure transmission reliability. Summary of the Invention

[0005] To address the issues that traditional routing protocols in multi-link converged network communication scenarios fail to maintain and comprehensively utilize multi-interface and multi-link communication resources, do not actually consider link switching costs and various service traffic requirements, and cannot efficiently adapt to multi-link converged network communication scenarios and ensure reliable transmission, this invention proposes an improved active routing method for multi-link converged networks. It implements an Optimized Multi-Link State Routing (OMLSR) protocol suitable for multi-link converged network scenarios to support efficient data forwarding. It achieves rapid route convergence and reduces overhead through dynamic adjustment of routing parameters, and ensures reliable data packet transmission through passive routing-assisted forwarding technology.

[0006] This invention provides an improved active routing method for multi-chain converged networking, applicable to multi-chain converged networking scenarios. Each communication node in the network has K interfaces for carrying K types of heterogeneous links, where K ≥ 2. The method includes the following steps:

[0007] Step 1: Construct the Optimized Multi-Link State Routing (OMLSR) protocol, including:

[0008] Step 1, maintaining multi-chain HELLO messages, refers to: setting nodes to periodically send HELLO messages from each open interface, with each node maintaining information about one-hop symmetric links, two-hop symmetric links, and symmetric neighbor nodes;

[0009] Step 2: Each node selects a multi-link multi-point relay (MPR) node from all its one-hop symmetric neighbors, then adds the MPR information to the one-hop symmetric link information corresponding to the MPR node, and broadcasts it through a HELLO message. The MPR information includes the one-hop symmetric node being selected as the MPR node by the current node and the link type of the two-hop neighbor links covered by the MPR node. Each MPR node parses the HELLO message to find out that it is the MPR node that is the source node of the HELLO message, records that the source node is the MPR selection node, and obtains the link type of the two-hop neighbor links it covers.

[0010] Each node first forms a one-hop neighbor set based on all its one-hop symmetric neighbor nodes, and forms a two-hop neighbor link set based on all its two-hop symmetric links. Then, it selects an MPR node based on the node's forwarding intention, the quality of heterogeneous links, and the reachability of its one-hop symmetric neighbor nodes. When calculating the reachability of a node, link quality and handover cost are taken into account.

[0011] Step 3: Each MPR node sends a multi-link topology control (TC) message, including: when sending a TC message, the node adds all one-hop symmetric links between the node and the MPR selected node to the TC message, and sends the TC message out from the interfaces corresponding to the link types of the two-hop neighbor links covered by the current node; after receiving TC messages sent by other nodes, the node records the topology link information in them.

[0012] Step 4: Construct a global topology graph based on all one-hop symmetric links and two-hop symmetric links recorded by the nodes and all link information obtained from TC messages. In the global topology graph, each communication node is divided into K virtual nodes, corresponding to K heterogeneous link interfaces, and a virtual switching link is added between every two virtual nodes. The weight of this link is the link switching cost. In the global topology graph, edges are established according to the links, and link indicators are set according to business requirements. The link cost is calculated by weighted summation of the set link indicators, and the calculated link cost is used as the weight of the edge. For different types of services, the Disktra algorithm is applied to calculate the route starting from each link interface to obtain a multi-link routing table.

[0013] Step 2: Implement a dynamic routing parameter adjustment strategy based on the current network topology changes, including adjusting the sending period of HELLO messages and TC messages;

[0014] Step 3: Execute the passive routing assisted forwarding strategy, which includes: each communication node simultaneously carries the OMLSR protocol and the multi-chain passive routing protocol. The OMLSR protocol is the primary routing protocol. When the primary route fails, a new route discovery process is initiated through the multi-chain passive routing protocol.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0016] (1) The method of this invention is designed for heterogeneous multi-chain converged communication scenarios. It utilizes the OMLSR routing protocol technology, an improvement upon the traditional OLSR protocol, to comprehensively consider multi-chain collaboration, link switching costs, multiple link inputs, and multiple service requirements, achieving efficient routing adaptation for this specific application scenario. This method achieves effective routing adaptation for multi-chain converged networking scenarios, ensuring efficient communication for multiple link inputs and multiple service requirements, accelerating route convergence and effectively reducing routing overhead, while also ensuring reliable transmission in the event of route failure.

[0017] (2) To address the issues of slow initial convergence and high periodic route maintenance overhead in the OMLSR routing protocol, this invention proposes a dynamic route parameter adjustment technique. This technique adjusts the transmission period of the route control packet based on mobility metrics. By shortening the transmission period during the initial route construction phase and when rapid network topology changes lead to frequent route updates, and extending the transmission period when the network topology is stable, this method achieves rapid route convergence and reduces overhead. This invention balances route convergence speed and route maintenance overhead, significantly accelerating route convergence and greatly reducing overhead when the network is stable.

[0018] (3) To address the issue of data packet failure during the active routing convergence phase or in the event of sudden link anomalies, the method of this invention employs passive routing-assisted forwarding technology to ensure reliable transmission. Through a hybrid routing mechanism, while the node carries the OMLSR protocol module, it also carries another passive routing protocol that does not require periodic maintenance. This ensures that data packets can be successfully transmitted via passive routing even when the OMLSR protocol routing fails to converge in the initial stage or when rapid changes in the local network topology cause routing failure. The method of this invention enables passive routing-assisted forwarding when active routing fails to converge or fails, ensuring the successful transmission of highly reliable services. Attached Figure Description

[0019] Figure 1 This is an implementation architecture diagram of the improved active routing method for multi-chain fusion networking of the present invention;

[0020] Figure 2This is a schematic diagram of the network link topology constructed according to an embodiment of the present invention;

[0021] Figure 3 This is a diagram of the routing table structure designed in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. The accompanying drawings have illustrated the explicit embodiments of this disclosure, which will be described in more detail below.

[0023] For heterogeneous network convergence communication scenarios where nodes are equipped with multiple interfaces and multiple heterogeneous links, this invention provides an improved active routing method for multi-link convergence networking. It aims to solve the problem that traditional routing protocols in multi-link convergence networking scenarios only consider single-link information, cannot fully utilize multi-link collaborative communication, and cannot simultaneously meet multiple service requirements. By using OMLSR routing protocol technology as a foundation, dynamic routing parameter adjustment technology, and passive routing assisted forwarding technology as optimizations, it achieves efficient multi-link networking communication in heterogeneous network convergence scenarios.

[0024] like Figure 1 As shown, the technical architecture of this invention mainly consists of two parts: the OMLSR routing protocol technology and the dynamic adjustment of routing parameters and passive route-assisted forwarding technology. The OMLSR routing protocol technology maintains the entire network link topology for multi-chain convergence scenarios and constructs a routing table based on input links and tasks, providing a foundation for multi-chain convergence collaborative communication. The dynamic adjustment of routing parameters and passive route-assisted forwarding technology optimize routing on this basis, improving route convergence speed, ensuring reliable transmission, and reducing routing overhead by adjusting routing parameters and using passive route-assisted forwarding.

[0025] like Figure 2 As shown in the multi-chain converged networking scenario of this invention embodiment, each communication node in the network carries three heterogeneous links, labeled as A, B, and C, with corresponding chain numbers of chain 1, chain 2, and chain 3. The quality of different links varies, assuming a chain quality order of chain 1 > chain 2 > chain 3. Each node may only partially activate these three chains. These links are independent of each other, with significant differences in communication distance and link quality. Switching data traffic from one link to another incurs a high cost, thus creating virtual link costs between the node's three interfaces. The routing technology of this invention maintains the multi-chain topology of the entire network and establishes the optimal routing path based on business requirements and the interfaces of the input links. Figure 3 The multi-link routing table shown enables efficient communication.

[0026] First, to adapt to multi-chain converged network communication scenarios, a routing protocol suitable for multi-chain scenarios needs to be designed and implemented. This invention proposes the OMLSR routing protocol technology. By improving the maintenance mechanism of multi-link HELLO messages and TC (Topology Control) messages for the characteristics of multi-chain network communication, improving the Multipoint Relay (MPR) node selection mechanism, and improving the routing table structure and calculation method for multi-chain inputs and multi-service requirements, a hybrid routing table is finally constructed for different input interfaces and different service requirements. The OMLSR routing protocol technology of this invention mainly consists of four parts: multi-chain HELLO message maintenance, multi-chain MPR node set selection, multi-chain TC message maintenance, and multi-link routing table calculation, which are described in detail below.

[0027] (1) Multi-chain HELLO message maintenance. In a multi-chain scenario, each node carries multiple interfaces (e.g., three, some of which are enabled), and there are multiple heterogeneous links between neighboring nodes. Compared with the traditional OLSR protocol, the OMLSR protocol needs to actively maintain multiple links. In order to actively discover neighbors and maintain multiple link information, the OMLSR protocol sets each enabled interface of the node to periodically send a HELLO message. The HELLO message contains the link interface that sent the message and the link information of the current node. The link information includes all one-hop link information of the current node. Each node maintains information on one-hop symmetric links, two-hop symmetric links, and symmetric neighbor nodes.

[0028] After receiving a HELLO message, a neighboring node establishes the corresponding link and one-hop neighbor entry based on the link interface information transmitted in the HELLO message. It then determines whether the link and one-hop neighbor entry are symmetric or asymmetric based on the link information of the neighboring nodes carried in the HELLO message, and obtains the remaining symmetric two-hop neighbor nodes. This two-hop neighbor node information contains multiple link information, i.e., multiple links containing the same one-hop node or multiple links containing multiple different one-hop nodes. These links are specified by the interface addresses at both ends of the link. Each node may contain multiple interface addresses, which are mapped to the node's primary address.

[0029] If nodes 1 and 2 are within each other's communication range and receive a HELLO message from the other's corresponding interface (e.g., interface 1), then a symmetric link is established between interface 1 of node 1 and interface 1 of node 2. Two neighboring nodes can be considered symmetric neighbors when any symmetric link is established between them. Similarly, when a symmetric link between two neighboring nodes is broken, the corresponding neighbor entry will only be deleted if no other symmetric link exists. Nodes record information about one-hop symmetric neighbors and one-hop symmetric links. For a one-hop symmetric neighbor, there exists at least one symmetric link with the same link interface.

[0030] If there is a symmetrical link between node 1 and node 2, and a symmetrical link between node 2 and node 3, then node 1 and node 3 are considered two-hop symmetrical nodes. It should be noted that the two-hop symmetrical link between node 1 and node 3 includes information from multiple links. For example, if a hop node is a two-hop symmetrical link between node 2, then the two-hop symmetrical link between node 1 and node 3 is formed by the one-hop symmetrical link between node 1 and node 2 and the one-hop symmetrical link between node 2 and node 3. If node 4 is also a one-hop symmetrical node between node 1 and node 3, then the two-hop symmetrical link between node 1 and node 3 also includes a two-hop symmetrical link where a hop node is node 4.

[0031] (2) Selection of Multi-Chain MPR Node Set. MPR nodes are used to build the communication backbone network and reduce topology message flooding. In multi-chain converged networking scenarios, due to the large differences in communication quality and range among various heterogeneous links, multi-chain information needs to be comprehensively considered during the selection of MPR nodes. The construction process of the MPR node set in the scenario of this invention is as follows:

[0032] Step 11. Obtain the complete set of one-hop neighbors and the set of two-hop neighbor links for the current node: The set of one-hop neighbors is the set of primary addresses of all one-hop symmetric neighbors, and the set of two-hop neighbor links is the set of two-hop symmetric links corresponding to all two-hop symmetric neighbor nodes, not just primary addresses. This means that each two-hop symmetric neighbor node contains multiple link information, resulting in multiple links with the same one-hop neighbor node or multiple links with multiple different one-hop neighbor nodes. In the set of two-hop neighbor links, detect two-hop neighbor nodes whose primary address is also a one-hop neighbor and exclude these links. However, in multi-link networking scenarios, there may be situations where a two-hop path is superior to a one-hop path. For example, both two-hop paths may be connected through the highest quality chain 1, while the one-hop path may only be connected through the lowest quality chain 2 or chain 3. In this case, the two-hop neighbor link corresponding to this two-hop neighbor node is retained.

[0033] Step 12. Select all neighbor nodes with a forwarding intention of ALWAYS (always willing to forward): All one-hop symmetric neighbor nodes with a forwarding intention of ALWAYS are directly selected as MPR nodes, and the two-hop symmetric links they cover are removed from the two-hop neighbor link set. At the same time, other links of lower quality of the two-hop neighbor nodes corresponding to these two-hop symmetric links are also removed. These are collectively referred to as "removing covered two-hop neighbor links".

[0034] For a node selected as an MPR, if the current node can reach the two-hop neighbor corresponding to the two-hop neighbor link through the MPR node, that is, the MPR node is an interface of a two-hop neighbor link, then it means that the MPR node covers the two-hop neighbor link.

[0035] Step 13. Select a unique covering node: If there exists a two-hop symmetric neighbor node, and all its corresponding two-hop symmetric links correspond to only one one-hop neighbor node, that is, the one-hop neighbor node uniquely undertakes the communication between the current node and the two-hop neighbor node, then this unique covering one-hop neighbor node is selected as the MPR node, and the covering two-hop neighbor links are deleted.

[0036] Step 14. Select neighbor nodes with higher reachability: Calculate the reachability of each remaining one-hop neighbor node, i.e., the weighted sum of the number of two-hop symmetric links corresponding to that one-hop neighbor node. Because heterogeneous links exist and switching between them incurs significant costs, different two-hop symmetric links should be assigned different weights. Additionally, since the links between the MPR node and the current node will form the backbone network, heterogeneous links between the MPR node and the current node are also assigned different weights. In the current scenario, the link quality is chain 1 > chain 2 > chain 3. Assume the following weights are assigned to one-hop neighbor links: chain 1 is assigned a weight of 12, chain 2 a weight of 6, and chain 3 a weight of 2. The specific link weights can be set according to the actual scenario. In this embodiment, different weights are pre-assigned to different chains based on communication quality; the better the link quality, the higher the weight. Further weights are assigned to two-hop neighbor links: if the two-hop neighbor link is chain 1 and the one-hop neighbor link contains chain 1, the weight of the two-hop neighbor link is 12; otherwise, the weight is 4. If the two-hop neighbor link is chain 2 and the one-hop neighbor link contains chain 2, the weight is 6; otherwise, the weight is 2. If the two-hop neighbor link is chain 3 and the one-hop neighbor link contains chain 3, the weight is 2; otherwise, the weight is 1. The specific link weight allocation needs to be specified according to the actual heterogeneous link situation in different scenarios. The reachability of this one-hop symmetric neighbor node is obtained by summing the weights of the one-hop neighbor link corresponding to each one-hop symmetric neighbor and its corresponding non-deleted two-hop neighbor link. The selection of the MPR node set is completed by starting with the one-hop neighbor node with the highest reachability and deleting the covered two-hop neighbor links. This process continues until all two-hop symmetric links in the two-hop neighbor link set have been deleted.

[0037] Step 15. MPR Node Selection Notification: Each node adds MPR information to the one-hop symmetric link information corresponding to its calculated MPR node. This information includes the fact that the one-hop symmetric node has been selected as the MPR node by the current node, as well as the link types of the two-hop neighbor links covered by this MPR node. This information is then broadcast to the corresponding MPR node via a HELLO message. When an MPR node receives this HELLO message, it determines that it is the MPR node from which the HELLO message originated. It then records the source node as the selected MPR node and parses the two-hop neighbor link types as the basis for the TC message. Recording the link types of the two-hop symmetric links covered by the MPR node ensures that when forwarding the TC message, the MPR node only needs to forward it from the ports corresponding to these types, guaranteeing that the TC message is broadcast to every node in the entire network.

[0038] (3) Each MPR node periodically maintains multi-chain TC messages. When a node sends a TC message, if the current node has an MPR selection node, it adds all one-hop symmetric links between the current node and the MPR selection node to the TC message, and sends it only from the output interface corresponding to the link type covering the two-hop neighbor links recorded in the record, so that the TC message can be transmitted to all two-hop neighbor nodes covered by the current node, thereby reducing unnecessary flooding. When a node receives a TC message sent by another node, it records the topology link information in it. If this node is an MPR node, it forwards the received TC message from the output interface corresponding to the link type covering the two-hop neighbor links recorded in its own record.

[0039] (4) Calculate the multi-link routing table, including the following steps 21 and 22.

[0040] Step 21. Global Topology Construction: When calculating the routing table, the first step is to integrate all one-hop symmetric links and two-hop symmetric links recorded by the nodes, along with all link information obtained from TC messages, into a global topology graph. The two endpoints of each link are used as nodes in the topology graph, links are used as edges, and the link cost Q is used as the edge weight. The link cost Q is calculated by comprehensively considering the number of route hops, link hold time, link quality such as packet loss rate, latency, bandwidth, and congestion indicators. The link hold time T is calculated by assuming the time required to predict the link breakage between the current node and its neighbors based on the movement status of neighboring nodes is... The upper limit threshold for link hold time is Then T is calculated as follows:

[0041] .

[0042] Here, min() means finding the minimum value.

[0043] The packet loss rate P reflects the quality of the current link. The lower the packet loss rate, the higher the reliability of the link.

[0044] Latency D also reflects the quality of the current link. If the latency is high, service packets cannot be transmitted in a timely manner. The calculation formula is as follows:

[0045] ;

[0046] in, For the current link latency, This represents the maximum delay threshold.

[0047] Bandwidth B also reflects the quality of the current link. The larger the bandwidth, the higher the data transmission rate. The calculation formula is as follows:

[0048] ;

[0049] in For the current link latency, This represents the maximum delay threshold.

[0050] The link congestion level C is determined by the proportion of packets occupying the buffer queue of the next-hop node. Let the total length of the node's buffer queue be... The total length occupied by the data packet is The formula for calculating the link congestion level C is as follows:

[0051] .

[0052] Therefore, the cost Q for each link is calculated as follows: The first item corresponds to a jump. ,satisfy .

[0053] To illustrate the multiple input / output interfaces of each node and the switching costs between links of different interfaces, such as Figure 2 As shown, this invention splits each communication node into three virtual sub-nodes, each corresponding to a heterogeneous link interface address. An edge representing a virtual switching link is added between each pair of virtual nodes. The weight of the edge is set to 4 or according to the actual scenario. The weight is used to represent the cost of link switching.

[0054] Step 22. Calculate the shortest path and update the routing table: Assume there are three types of service data with different Quality of Service (QoS) requirements, and assign weights to the link costs. Because the paths differ, three routing entries are calculated for each target node and each type of service data. Furthermore, when data packets enter from different link interfaces, the starting point at the current node differs, resulting in different calculated routing paths. Therefore, three routing entries are calculated for each link's input interface. Finally, nine routing entries are calculated for each target node using the Dijisktra algorithm, such as... Figure 3 As shown.

[0055] When a node sends a data packet as a source address, it first parses its QoS requirements, finds the three routing entries corresponding to its QoS requirements (i.e., the routing entries corresponding to the three link interfaces), compares the path costs of the routing entries corresponding to the three interfaces, and selects the path with the lowest cost to transmit the data packet. When a node forwards a data packet as a relay node, it only needs to select the routing entry corresponding to the input interface of the data packet to forward the data packet.

[0056] After the above steps, the OMLSR routing protocol can efficiently adapt to multi-chain converged networking communication scenarios and various business needs, ensuring reliable data transmission.

[0057] Dynamic routing parameter adjustment and passive route-assisted forwarding technologies optimize the OMLSR routing protocol, primarily including dynamic routing parameter adjustment and passive route-assisted forwarding. Dynamic routing parameter adjustment achieves rapid route convergence and reduces routing overhead by dynamically adjusting the sending cycles of HELLO and TC messages. Passive route-assisted forwarding, by incorporating another passive routing protocol, addresses the issue of data packet failure during active route convergence or in the event of sudden link anomalies, thus achieving reliable transmission.

[0058] (5) Implement a dynamic routing parameter adjustment strategy, which mainly involves dynamically adjusting the sending period of HELLO and TC messages in the OMLSR protocol. When the network is initially established or the topology changes rapidly, reduce the sending period of HELLO and TC messages to speed up route convergence; when the route converges successfully and the network topology changes slowly, increase the sending period of HELLO and TC messages to reduce routing overhead.

[0059] Step 31. Dynamically adjust the HELLO message sending period. This embodiment of the invention sets a minimum value for the HELLO message sending period. maximum value Initially, the HELLO message sending cycle is set to the minimum value. .

[0060] The previous sending cycle is calculated each time a HELLO message is sent. Number of newly added one-hop neighbor links per unit time Number of deleted one-hop neighbor links The number of one-hop neighbor links that change with link symmetry The sum of the total number of one-hop neighbor links The ratio of mobility The calculation formula is as follows:

[0061] ;

[0062] This results in a new HELLO message sending cycle. The calculation formula is as follows:

[0063] .

[0064] Step 32. Dynamically adjust the TC message sending period. This embodiment sets a minimum value for the TC message sending period. maximum value The initial TC message sending period is set to... When a node is added or deleted from the MPR set selected by the node, the current TC message sending period is immediately adjusted to the minimum value. Let the time of the last TC message being sent be... The current time is Then the next TC message sending time The calculation formula is as follows:

[0065] The `max()` function calculates the maximum value.

[0066] If no MPR selection set was added or deleted during the previous TC message sending cycle, then the next TC message sending cycle will be set to [a later date]. .

[0067] (6) Implement a passive routing-assisted forwarding strategy. To address the issue of slow convergence of the OMLSR protocol during the initial stage of the network or when the topology changes rapidly, which leads to packet transmission failure, a hybrid routing technology is adopted. By incorporating another multi-chain passive routing protocol, active routing forwarding is assisted, thereby improving the packet delivery rate.

[0068] Step 41. Configure a single node in the network to carry both OMLSR and multi-chain passive routing protocols: At the network layer, the IPv4 module integrates OMLSR and multi-chain passive routing protocols, and the different protocols are invoked through the datagramLocalOutHook or datagramPreRoutingHook functions. Because the multi-chain passive routing protocol does not require periodic maintenance, both protocols can be enabled simultaneously.

[0069] Step 42. Multi-chain passive routing protocol-assisted forwarding: When the overall network topology changes slowly, the OMLSR protocol is used as the main route. Each time a data packet is sent or forwarded, the routing table calculated by the OMLSR protocol is queried first. If there is a valid route path to the target node, the data packet is directly transmitted using this route path; if there is no valid route path to the target node, another multi-chain passive routing protocol, such as the multi-chain AODV routing protocol, is used to start a new route discovery process.

[0070] By additionally equipping each node with a multi-chain passive routing protocol that does not require periodic maintenance, when the OMLSR routing table cannot find an available route due to the discovery of routes not converging in time or sudden link anomalies, this passive routing is immediately used for route discovery to ensure reliable data transmission.

[0071] In summary, this invention successfully addresses the problem of multi-heterogeneous link cooperative communication in multi-chain heterogeneous network convergence scenarios by proposing and implementing the OMLSR routing protocol technology. This technology can efficiently maintain global link topology information and maintain multiple routing paths to meet the QoS requirements of various types of service data and the switching costs of heterogeneous links, ensuring efficient data transmission. Simultaneously, by adaptively adjusting the control packet sending period of the OMLSR protocol and employing hybrid routing technology, it effectively reduces routing control overhead and routing convergence time, improving packet delivery rate. The modular design and scalability of this invention allow it to be customized and extended according to different multi-chain application scenarios and business needs, demonstrating broad application potential.

[0072] Except for the technical features described in the specification, all other technologies are known to those skilled in the art. Descriptions of well-known components and technologies are omitted in this invention to avoid redundancy and unnecessary limitation. The embodiments described above do not represent all embodiments consistent with this application. Various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this invention are still within the protection scope of this invention.

Claims

1. An improved active routing method for multi-chain converged networking, used in multi-chain converged networking scenarios, wherein each communication node in the network has K interfaces for carrying K types of heterogeneous links, where K ≥ 2; characterized in that, The method includes the following steps: Step 1: Construct the Optimized Multi-Link State Routing (OMLSR) protocol, including: Step 1, maintaining multi-chain HELLO messages, refers to: setting nodes to periodically send HELLO messages from each open interface, with each node maintaining information about one-hop symmetric links, two-hop symmetric links, and symmetric neighbor nodes; Step 2: Each node selects a multi-link multi-point relay (MPR) node from all its one-hop symmetric neighbors, then adds the MPR information to the one-hop symmetric link information corresponding to the MPR node, and broadcasts it through a HELLO message. The MPR information includes the one-hop symmetric node being selected as the MPR node by the current node and the link type of the two-hop neighbor links covered by the MPR node. Each MPR node parses the HELLO message to find out that it is the MPR node that is the source node of the HELLO message, records that the source node is the MPR selection node, and obtains the link type of the two-hop neighbor links it covers. Each node first forms a one-hop neighbor set based on all its one-hop symmetric neighbor nodes, and forms a two-hop neighbor link set based on all its two-hop symmetric links. Then, it selects an MPR node based on the node's forwarding intention, the quality of heterogeneous links, and the reachability of its one-hop symmetric neighbor nodes. When calculating the reachability of a node, link quality and handover cost are taken into account. Step 3: Each MPR node sends a multi-link topology control (TC) message, including: when sending a TC message, the node adds all one-hop symmetric links between the node and the MPR selected node to the TC message, and sends the TC message out from the interfaces corresponding to the link types of the two-hop neighbor links covered by the current node; after receiving TC messages sent by other nodes, the node records the topology link information in them. Step 4: Construct a global topology graph based on all one-hop symmetric links and two-hop symmetric links recorded by the nodes and all link information obtained from TC messages. In the global topology graph, each communication node is divided into K virtual nodes, corresponding to K heterogeneous link interfaces, and a virtual switching link is added between every two virtual nodes. The weight of this link is the link switching cost. In the global topology graph, edges are established according to the links, and link indicators are set according to business requirements. The link cost is calculated by weighted summation of the set link indicators, and the calculated link cost is used as the weight of the edge. For different types of services, the Disktra algorithm is applied to calculate the route starting from each link interface to obtain a multi-link routing table. Step 2: Implement a dynamic routing parameter adjustment strategy based on the current network topology changes, including adjusting the sending period of HELLO messages and TC messages; Step 3: Execute the passive routing assisted forwarding strategy, which includes: each communication node simultaneously carries the OMLSR protocol and the multi-chain passive routing protocol. The OMLSR protocol is the primary routing protocol. When the primary route fails, a new route discovery process is initiated through the multi-chain passive routing protocol.

2. The method according to claim 1, characterized in that, Step 1 includes: the HELLO message contains all one-hop link information of the current node and the link interface for sending the message; maintaining multi-link HELLO messages includes: when a node receives a HELLO message sent by a neighbor node, it establishes the corresponding link and one-hop neighbor node entry according to the link interface of the HELLO message, determines whether the link with the corresponding interface of the neighbor node is a symmetrical link according to the link information of the neighbor node carried in the HELLO message, and obtains two-hop symmetrical neighbor nodes and two-hop symmetrical links.

3. The method according to claim 1, characterized in that, In step 2, the step of a communication node selecting a multi-link multi-point relay (MPR) node from all its one-hop symmetric neighbors includes: (11) The one-hop neighbor set records the main addresses of all one-hop symmetric neighbor nodes of the current node; each node contains K interface addresses and one main address; if the main address of a certain two-hop symmetric neighbor node is in the one-hop neighbor set, check the quality of the two-hop symmetric link between the two-hop symmetric neighbor node and the current node. If the quality is better than the one-hop symmetric link between the two-hop symmetric neighbor node and the current node, retain the corresponding two-hop symmetric link; otherwise, delete the corresponding two-hop symmetric link. (12) Select all one-hop symmetric neighbor nodes with forwarding intention of ALWAYS in the one-hop neighbor set as MPR nodes, delete the two-hop symmetric links covered by the MPR nodes from the two-hop neighbor link set, and delete the two-hop symmetric links with worse quality between the two-hop neighbor nodes corresponding to the deleted two-hop symmetric links than the deleted links. (13) If all two-hop symmetric links between the current node and a certain two-hop symmetric neighbor node communicate through the same one-hop symmetric neighbor node, then select the one-hop symmetric neighbor node as the MPR node, delete the two-hop symmetric links covered by the MPR node from the two-hop neighbor link set, and delete the two-hop symmetric links between the two-hop neighbor nodes corresponding to the deleted two-hop symmetric links whose quality is worse than the deleted links. (14) Calculate the reachability of each remaining one-hop symmetric neighbor node in the current one-hop neighbor set. Starting from the largest reachability, select the corresponding one-hop symmetric neighbor node as the MPR node in turn. Delete the two-hop symmetric link covered by the MPR node from the two-hop neighbor link set. At the same time, delete the two-hop symmetric links with worse quality between the two-hop neighbor nodes corresponding to the deleted two-hop symmetric links. Continue until all two-hop symmetric links in the two-hop neighbor link set have been deleted, and the selection of the MPR node set is completed. The reachability of a one-hop symmetric neighbor node is obtained by summing the weights of the one-hop symmetric link corresponding to that neighbor node and the weights of the two-hop symmetric links covered in the current two-hop neighbor link set; the weights of the links are set according to the quality of the heterogeneous links, and the better the link quality, the higher the weight.

4. The method according to claim 1, characterized in that, Step 3 further includes: when a node receives a TC message from another node, it determines whether it has an MPR selection node. If it does, it sends the received TC message out from the interface corresponding to the link type that covers the two-hop neighbor link recorded by itself.

5. The method according to any one of claims 1 to 4, characterized in that, In step 4, one way to calculate the link cost is to calculate the link cost Q by comprehensively considering the normalized route hop count, link hold time T, packet loss rate P, latency D, bandwidth B, and node congestion level C, as shown below: ； The first item corresponds to a jump. It is the weight of the i-th link metric; The link hold time T is calculated as follows: ; This is the upper limit threshold for link hold time. The time required for the link between the current node and its neighboring nodes to break, as predicted by the motion states of the neighboring nodes. The link congestion level C is calculated from the ratio of packets occupied in the buffer queue of the next-hop node of the link.

6. The method according to claim 5, characterized in that, In step 4, when calculating the multi-link routing table, the weights of the link metrics are assigned differently when calculating the link cost based on the service quality requirements of different services. For each service, a routing table entry is calculated for each link interface of the node. When a node sends a data packet as a source address, it first parses to obtain the service quality required by the service, looks up the corresponding routing entry, compares the path costs of the routing entries corresponding to different link interfaces, and selects the path with the lowest cost to transmit the data packet; when a node acts as a relay node, it forwards the data packet according to the routing entry corresponding to the link interface into which the data packet is input.

7. The method according to claim 1, characterized in that, In step two, upper and lower thresholds are set to characterize the speed of network topology changes. When a network topology change is detected to exceed the upper threshold, it indicates a fast change, and the sending period of HELLO and TC messages is reduced. When a network topology change is detected to be below the lower threshold, it indicates a slow change, and the sending period of HELLO and TC messages is increased.

8. The method according to claim 1 or 7, characterized in that, Step two includes: (1) The way to adjust the HELLO message sending period is to set the minimum value of the HELLO message sending period. and maximum value The initial HELLO message sending cycle is set to the minimum value; each time a HELLO message is sent, the previous sending cycle is counted. Number of newly added one-hop neighbor links per unit time Number of deleted one-hop neighbor links The number of one-hop neighbor links that change with link symmetry Calculate mobility as follows: ； in, This indicates the total number of current one-hop neighbor links; Further update the HELLO message sending cycle as follows: ； (2) The way to adjust the TC message sending period is to set the minimum value of the TC message sending period. and maximum value The initial TC message sending period is set to the minimum value; when a node is added or deleted from the MPR node set, the current TC message sending period is adjusted to the minimum value; the last TC message sending time is set to... The current time is Then the next TC message sending time The calculation is as follows: ； If the MPR node set of a node has not changed during the previous TC message sending cycle, then the next TC message sending cycle will be set to the maximum value.

Images