Based on the neighbor connection point table in the ribp data file, unilateral external links are discovered. Then, the downstream connection points notify the upstream and downstream connection points of that unilateral external link according to the uploaded ulp/ulbp (unilink packet/unilink broadcast packet) message format. When the upstream node receives this information, it feeds back to the downstream node through the ul_ack message. After receiving the ribp message, the key information in the message is parsed, and the local neighbor node table is compared. Finally, the ulp/ulbp message is sent or the local routing table is updated. The system receives the ribp message to obtain the source node number, item information, etc., and then queries the local neighbor node information. If there is no information about this connection point in the system software, the unilateral external link notification mechanism is initiated. According to whether there is router information in the system software, it is determined whether to convert to the ulp or ulbp message format. If the system has the node information, the corresponding link status is updated, the routing table is updated based on the link information, and then all path information is traversed to update the routing information to each destination node in the routing table and update the underlying forwarding table information.

1. The one-way link discovery mechanism is a type of network topology.

Each connection point regularly broadcasts ribp packets. If connection point D receives the ribp packet sent by A, it confirms that the external link from A to D is functioning. At the same time, for D, it will place the information of which nodes can reach this link into the neighbor node table of the ribp packet. That is, in this topology structure, D will place node A in its ribp packet's neighbor node table. Similarly, for the A node, since it does not receive the ribp packet sent by D, D is not its neighbor node. Therefore, when D receives the ribp packet sent by A, it cannot find its own existence in the adjacent node table of this packet. Based on this, the D node determines that the link from A to D is a one-way link.

The node A in the one-way link A→D is called the upstream node, and node D is the downstream node. When the connection point between the upstream and downstream detects the existence of this one-way external link, it must notify this information to the upstream and downstream connection points, so that the upstream and downstream connection points can utilize the external link A→D.

2. Analysis of Simulation Experiment Results

The dynamic network topology obtained by arbitrarily converting the self-organizing network simulation service platform in the article was used for functional testing. The characteristics differences between the TDFP routing protocol and the traditional self-organizing network routing protocol were compared and analyzed based on three key performance parameters: sorting transmission rate, convergence time, and router cost.

3. Simulation results of group transmission rate

The sorting transmission rate refers to the ratio of the number of sorted statistics received at the destination connection point to the number of sorted statistics transmitted from the source connection point. This value describes the packet loss rate of the ad hoc network and reflects the completeness and correctness of the routing protocol. The simulation scenario of this group of experiments is a rectangular area of 1500 m × 300 m, with a simulation time of 900 seconds and 50 simulation nodes. The movement model of the nodes is set as the random waypoint model, with a moving speed of 360 m/s. The routing protocol performance is simulated under different network topology change rates. The communication source is a connection point with a constant bit rate (CBR) for uniform motion, and its sorting speed is 4 sorts per second. There are 30 communication sources, and the routing protocol characteristics under different data traffic are simulated.

Among them, the node movement model is a random point model. This model includes a pause time parameter, which is the duration for the node to stay at the target position after reaching it. In this group of experiments, the pause times selected were 0 s, 30 s, 60 s, 120 s, 300 s, 600 s, and 900 s. A pause time of 0 s describes the continuous movement of the node, while a pause time of 900 s describes the node being in a stationary state. In any point entity model, the fewer the termination time, the higher the wireless communication network of the connection point; conversely, the lower the wireless communication network of the connection point.

As shown in the figure, when the node mobility is high, that is, the pause time is less than 300 seconds, the packet delivery rate of the DSDV protocol is poor. The reason is that the DSDV protocol only saves one route to the destination node. When this route is disconnected, the nodes will lose a large number of packets. The packet delivery rate performance of the other three protocols is better and always remains above 0.95. Among them, the on-demand router AODV protocol has excellent characteristics in the natural environment with high wireless communication network. However, in the natural environment with low wireless communication network, it has certain differences from the TDFP routing protocol designed and conceived in this paper. The reason is that the TDFP protocol belongs to the table-driven routing protocol at the lowest level and regularly broadcasts the router message format to maintain the accuracy of the router. However, the AODV protocol only maintains the router when there is a necessary communication. If a router error is detected, the statistics data before the router repair will be immediately discarded.

4. Simulation results of convergence time

Only the pause time can be adjusted to 0, 20, 40, 60, 80, 100, 120, 140, 160, and 180 intervals. As shown in the figure, when the node mobility is high, that is, the pause time is less than 100 seconds, the convergence time of the DSDV protocol is longer, while the convergence time of the AODV protocol is shorter. The reason is that most of the routing messages of the DSDV protocol are sent periodically. When this route fails, the node still needs to wait for the sending period to arrive. However, the AODV protocol, when it detects a route failure, will immediately notify the source node to rebuild the route. Therefore, in a high mobility environment, its convergence time is short. On the contrary, when the node mobility is low, that is, the pause time is greater than 120 seconds, the convergence time of the AODV protocol is longer, while the convergence time of the DSDV protocol is shorter. The overall characteristics of the TDFP routing protocol designed in the article are close to those of the two. The reason is that the TDFP protocol belongs to the table-driven routing protocol at the lowest level, and regularly broadcasts the router message format to maintain the accuracy of the router. At the same time, the article adds a convergence mechanism to it. If the router is invalid, it will immediately pass this message to the necessary re-calculated connection point, thereby improving the sensitivity of the routing protocol to network topology changes.

5. Simulation results of routing costs

The simulation parameters are shown in the table, and the simulation results are presented in the figure. The routing overheads of the four protocols vary significantly. The TDFP protocol and the CGSR protocol have the smallest overheads, while the DSDV protocol has the largest overhead. The AODV protocol's overhead changes significantly with the variation of node mobility. The routing overhead of the DSDV protocol is huge because it belongs to a planar structure routing protocol and always periodically sends routing messages to neighboring nodes. In scenarios with high node mobility, the DSDV protocol sends complete routing update messages; in scenarios with low node mobility, the DSDV protocol sends partially growing routing update messages, and thus its curve shows a downward trend as node mobility decreases. The AODV protocol is a typical on-demand routing protocol. In scenarios with high connection point wireless communication networks, the AODV protocol often needs to rebuild invalid routers, so the router cost is high; in scenarios with low connection point wireless communication networks, the AODV protocol only needs to upload Hello information from time to time to maintain the routers, so the router cost is relatively low. The routing overheads of the TDFP protocol and the CGSR protocol designed in this paper are very low, and they gradually decrease as node mobility decreases. In scenarios with high node mobility, the TDFP protocol not only needs to periodically send routing messages but also triggers the convergence mechanism mentioned above when the route fails, further accelerating the route convergence speed; in scenarios with low node mobility, the TDFP protocol only needs to periodically maintain the routing connection, and the routing overhead remains basically unchanged.

Conclusion

Based on the simulation results, the following can be calculated: In large-scale ad hoc networks, the TDFP routing protocol, compared with traditional ad hoc network routing protocols, can achieve excellent improvement in terms of enhancing the statistical data sorting transmission rate, reducing the Internet convergence time, and reducing the router cost.