Deterministic satellite internet routing method and system based on predicted topology

Abstract

The application discloses a kind of certainty satellite internet routing method and system based on predicted topology, mainly solve the problem that the network flooding load of existing constellation topology is high under the scene of high speed change, the stability of deterministic routing is poor and network convergence is slow.The scheme includes: generating a predicted topology file according to the satellite orbit law, each satellite running link state routing protocol and periodically collecting link parameters;Through probe message to establish neighbor relationship, to its announcement collection parameter;According to the predicted topology file, link switching time is generated and flooded link state message, guide the whole network to update link state database;Based on the updated comprehensive link weight and in advance routing pre-computation and backup, and replace routing table at switching time.The application can reduce network flooding load and shorten convergence time, improve the certainty of satellite internet routing and business continuity, and can be used for routing switching and stable forwarding in high dynamic satellite network environment.

Description

Technical Field

[0001] This invention belongs to the field of satellite internet technology, specifically relating to a deterministic routing method and system that can be used for routing switching and stable forwarding in highly dynamic satellite network environments. Background Technology

[0002] In recent years, the demand for high-reliability, low-latency services such as industrial control and vehicle-to-everything (V2X) has been continuously growing. Networks are gradually evolving from the traditional "best-effort" approach towards deterministic networks. Deterministic networks are network technologies that provide deterministic service level guarantees for services within a network. Deterministic service levels typically include deterministic latency limits, low packet loss rates, reduced jitter, and high reliability. Satellite internet refers to internet based on satellite communication. It involves launching a specific number of satellites to form a network, thereby covering the globe and constructing a large satellite system capable of performing information processing, providing communication services such as bandwidth internet access to ground and air terminals. However, with the synergistic evolution of deterministic networks and satellite internet, the high latency and low reliability of traditional satellite networks and the technical complexity of deterministic networks have gradually become key bottlenecks restricting integrated air-space-ground services. Against this backdrop, the proposed concept of deterministic satellite internet leverages the wide coverage of satellite internet and considers its large-scale constellation configuration, highly dynamic topology links, time-varying multidimensional network resources, and random wireless links to provide time-deterministic transmission services globally. This means providing deterministic end-to-end performance guarantees such as latency, jitter, and packet loss rate for time-sensitive applications.

[0003] The key to deterministic satellite internet technology lies in constructing deterministic routes. Existing deterministic routing schemes have the following limitations in the context of satellite internet: (1) Topology is highly time-varying and link duration is limited: Due to the frequent changes in adjacency relationships caused by satellite movement, the available time of links is limited. If the link duration is not explicitly considered in the routing calculation, it is easy to select a path that is about to fail, resulting in topology discovery load. (2) Frequent routing calculations lead to increased network load. In the case of frequent changes in network topology, traditional routing protocols (OSPF, IS-IS) need to perform frequent routing calculations and updates. A large number of control messages generated in the network cause the processing capacity and bandwidth resources of network devices to be consumed in large quantities, affecting the overall performance and stability of the network. Moreover, the network load is related to the scale of the satellite network. When the scale of the satellite network expands, its network flooding load is proportional to the square of the satellite network scale N. (3) Traditional routing protocols cannot respond quickly to topology changes. In the satellite internet scenario, due to the highly dynamic changes in network topology, the network topology information and resource information are not announced in a timely manner at a large spatiotemporal scale. Traditional routing protocols cannot respond quickly and trigger routing calculations, resulting in a slow network convergence speed and affecting the stability of data transmission.

[0004] Patent document 201510667832 discloses a low-overhead dynamic routing method for mobile satellite networks. Based on the orbital patterns of satellites, it pre-calculates the link connectivity relationships between snapshots and distinguishes between predictable and unpredictable changes during network operation. For predictable changes, it uses static processing, while for unpredictable changes, it enables dynamic route updates, thereby reducing control signaling overhead caused by topology changes and improving routing convergence efficiency. However, because this method primarily relies on the predictability of satellite motion to reduce the frequency of link state updates, its routing decisions still revolve around changes in link connectivity. Therefore, the link weight representation is singular, making it difficult to guarantee end-to-end service quality and deterministic service transmission requirements. Furthermore, when unpredictable changes occur, it still triggers link state detection and network-wide updates, causing flooding of link state information. Frequent flooding of control messages in the network significantly increases control plane overhead with network size, further prolonging routing convergence time. Summary of the Invention

[0005] The objective of this invention is to address the shortcomings of the prior art by proposing a deterministic satellite internet routing method based on predictive topology, thereby reducing the network flooding load of deterministic routing, improving routing stability and network convergence speed in dynamic topology environments, and realizing multi-dimensional link weight representation and routing decision-making.

[0006] The technical approach of this invention is as follows: By introducing predictive topology, link state flooding is triggered in advance at a certain moment during link switching, reducing the additional overhead caused by topology discovery and lowering the network flooding load. By introducing a multi-dimensional link weight model that includes remaining link bandwidth, propagation delay, and remaining link duration, a unified representation of link load capacity, transmission performance, and link stability is achieved. The optimal path is calculated using Dijkstra's algorithm within the delay jitter range. By completing route calculation in advance within a predetermined timeframe for link switching and rapidly applying the new route during link switching, route stability and rapid network convergence are ensured.

[0007] Based on the above ideas, the technical solution of the present invention includes:

[0008] 1. A deterministic satellite internet routing method based on predicted topology, characterized in that it includes:

[0009] (1) Construct a predictive topology file that includes a set of link switching events based on the satellite orbital operation patterns;

[0010] (2) Each satellite node periodically collects its own direct link parameters;

[0011] (3) Each satellite node's control plane runs the link-state routing protocol, establishes neighbor relationships through probe messages, and announces the parameters it has collected to other satellite nodes through link-state messages, thus completing the construction of its own network topology and initial routing table;

[0012] (4) Each satellite node obtains the link switching event from the prediction topology file. When the satellite in the link switching event is its own node, the node... The system continuously generates link status messages as described in the event (link disconnection / establishment), adds them to the local link status database, and floods them across the entire network.

[0013] (5) Other satellite nodes in During the time slice, the link status messages flooded by the nodes in the link switching event will be received successively, checked, and added to the local link status database.

[0014] (6) Each satellite node calculates the comprehensive weight of the link based on the link parameters in the link state database, performs route calculation based on the comprehensive weight, and generates a route. Routing table at any time And perform a backup;

[0015] (7) When the system time arrives At that time, each satellite node will update its outdated routing table. Switch to the latest routing table and the original routing table delete;

[0016] (8) Each satellite node obtains the next link switching event and repeats (4) to (7) until the prediction topology file ends.

[0017] Furthermore, in step (6), each satellite node calculates the comprehensive weight of the link based on the link parameters in the link state database, which is implemented as follows:

[0018] 6a) Each satellite node obtains the link propagation delay from the link state database. Remaining bandwidth of the link Remaining link duration The maximum and minimum values ​​are found, and these link parameters are normalized to obtain the normalized values. , , :

[0019] 6b) Each satellite node's normalized... , , Linear weighting is performed to obtain the link synthesis weight used to characterize the cost of the transmission edge. .

[0020] Furthermore, in step (6), routing calculation is performed based on the comprehensive weight to generate... Routing table at any time And perform backups, the implementation of which includes:

[0021] 6c) Each satellite node constructs the network topology based on the link state database. ,

[0022] in For a set of nodes, Indicates the first satellite nodes, The total number of nodes;

[0023] For the set of transmission edges, Represents a node and A transmission edge is formed;

[0024] Let be the set of transmission edge costs. Represents a transmission edge The link comprehensive weight is used to characterize the transmission cost of the link at the current moment;

[0025] 6d) Each node is based on the bandwidth constraints of the service to be transmitted. For the remaining transmission bandwidth Bandwidth constraint not met Prune the transmission edges and nodes to obtain the pruned network topology. ;

[0026] 6e) Based on the pruned network topology Perform a minimum-cost path search to obtain a set of end-to-end paths with the minimum total cost. ,in Depend on It consists of several transmission edges;

[0027] 6f) According to Given the propagation delay and transmission delay of each transmission edge, calculate the end-to-end delay of the path set. :

[0028] 6g) Based on the upper limit of the latency of the service to be transmitted When satisfied At that time, the set of paths This constitutes a feasible end-to-end path;

[0029] 6h) Each satellite node will calculate the end-to-end path Build as Routing table that is always in effect And save it in Before that time, the current routing table will still be used for forwarding.

[0030] 2. A deterministic satellite internet routing system based on predicted topology, characterized in that it comprises:

[0031] The predictive topology building module is used to know the changes in links in the network in advance. By analyzing the movement state of nodes and their location information at future times, it can obtain the link switching events in the future, providing precise timing for subsequent link state flooding and deterministic route calculation.

[0032] The link parameter acquisition module is used to periodically collect real-time status parameters of the directly connected links of this node, including at least the link propagation delay, the remaining bandwidth of the link, and the remaining duration of the link. The link parameters are used for link comprehensive weight calculation and serve as an important input for path selection.

[0033] The link state maintenance module is used to establish neighbor relationships between nodes through probe messages and to manage and maintain link state information in a unified manner by combining the predicted topology construction results.

[0034] The link state flooding module is used to synchronously flood link state information in the network and to flood and update the link state in advance based on passive triggering mechanisms or in combination with predicted topology.

[0035] The routing calculation module is used to generate end-to-end paths that take effect at a predetermined time based on the link parameters in the link state database, service bandwidth constraints, and latency upper bound requirements. It also generates path results that meet the deterministic transmission requirements of the service based on the comprehensive link weight, providing a basis for subsequent routing table maintenance.

[0036] The routing table maintenance module is used by nodes to update the current routing table when the network-wide unified time obtained based on time synchronization reaches the predetermined handover time, so as to avoid link interruption caused by link switching during the handover process.

[0037] Compared with the prior art, the present invention has the following advantages:

[0038] Firstly, this invention introduces predictive topology to trigger link state flooding in advance at a predetermined time of link switching, enabling the network to synchronize topology information before the actual link change, thus avoiding the flooding load generated at the moment of link switching.

[0039] Secondly, this invention introduces a multi-dimensional comprehensive link weight, including remaining link bandwidth, propagation delay, and remaining link duration, during the routing calculation process. This weight provides a unified representation of the link's load capacity, transmission performance, and stability, ensuring that the routing calculation process avoids selecting links that are about to fail while meeting service delay and bandwidth constraints. This effectively avoids topology discovery load caused by routing failures.

[0040] Third, by completing route calculations in advance within the predetermined time of link switching and quickly applying the new route results during link switching, the convergence time caused by network topology changes can be shortened, ensuring the continuous and stable transmission of deterministic services. Attached Figure Description

[0041] Figure 1 is a flowchart of the implementation of the deterministic satellite internet routing method based on predicted topology of the present invention;

[0042] Figure 2 is a schematic diagram of a satellite internet scenario in the method of the present invention;

[0043] Figure 3 is a schematic diagram of a satellite network handover scenario in the method of the present invention;

[0044] Figure 4 is a schematic diagram of network topology construction in the method of the present invention;

[0045] Figure 5 is a block diagram of the deterministic satellite internet routing system based on predictive topology of the present invention. Detailed Implementation

[0046] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0047] Example 1: A Deterministic Satellite Internet Routing Method Based on Predictive Topology

[0048] like Figure 2 As shown, this example's use case comprises a low-Earth orbit (LEO) satellite network, a ground gateway, and ground users. The LEO network is a multi-orbital-plane mesh network. Satellites operate along their respective orbital axes in stable configurations and form a dynamic, interconnected mesh structure via inter-satellite links (ISLs). Within the same orbital plane, stable forward inter-satellite links are established between adjacent satellites, with the link status evolving systematically with the satellite's motion. Within adjacent orbital planes, satellites establish inter-orbit links within specific communication time windows to enable traffic forwarding and routing connectivity between orbits. Ground users generate service traffic, which is accessed through the top satellite to the satellite network and then forwarded to the ground network by the egress satellite. This method ensures deterministic transmission of services within the satellite network.

[0049] Reference Figure 1 The implementation steps of this example include the following:

[0050] Step 1: Construct the prediction topology file.

[0051] 1.1) Based on the six base numbers of each satellite orbit input, calculate the satellite's position from the start time. to any time within the simulation period Spatial location:

[0052] 1.1.1) Let the six elements of the satellite's orbit be... Take the Earth's gravitational constant as ,in For the semi-major axis, For eccentricity, For the track inclination angle, Right ascension of the ascending node, For reference time The angle of the near point;

[0053] 1.1.2) According to The angle of approach at time Calculate at any time The angle of the near point :

[0054] ;

[0055] 1.1.3) Based on the satellite's eccentricity Calculate at any time The second-order approximate near-point angle :

[0056] ;

[0057] 1.1.4) Based on the approximate near-point angle Calculate the true anterior angle Distance within the orbital plane :

[0058] ,

[0059] ,

[0060] 1.1.5) Based on the true anterior angle Distance within the orbital plane Obtain the satellite's orbital plane coordinates. :

[0061] ;

[0062] 1.1.6) Set the orbital plane coordinates By converting to a Cartesian coordinate system using a rotation matrix, we obtain the satellite's position vector in the Cartesian coordinate system. :

[0063] ,

[0064] in, Substitution The rotation matrix around the Z-axis.

[0065] Substitution The rotation matrix around the Z-axis.

[0066] Substitution The rotation matrix around the X-axis;

[0067] 1.2) Define the position vectors of two adjacent satellites as follows: , Calculate the distance between the two positions. :

[0068] ,

[0069] 1.3) Set the communication distance constraint as follows: Determine the positional distance between a satellite and its neighboring satellites. Is it within the communication distance constraint?

[0070] like If the link is broken, the satellite needs to select another available satellite that meets the communication distance requirement for link switching, and proceed to step 1.4).

[0071] like Then the satellite is within normal communication range with neighboring satellites;

[0072] 1.4) From the start time to the end time of the simulation, calculate the inter-satellite link handover status of all adjacent satellites in the entire network during the simulation time, and obtain a series of inter-satellite link handover events. ,in Indicates the first Link switching events Indicates the link switching time. Indicates the source satellite node. Indicates the primary target satellite node. Indicates the second target satellite node;

[0073] 1.5) The link switching events are arranged in chronological order. Form a link switching set This forms the predicted topology file, as shown in Figure 1:

[0074] Table 1. List of events for predicting topology files

[0075]

[0076] 1.6) The predicted topology file is sent to each satellite node via File Transfer Protocol (FTP).

[0077] Step 2: Each satellite node periodically collects its own direct link parameters.

[0078] 2.1) Each satellite node periodically collects parameter information from the four interfaces eth_0, eth_1, eth_2, and eth_3 of its directly connected inter-satellite links. The parameters for each interface include the link propagation delay. Remaining bandwidth of the link and remaining link duration ;

[0079] 2.2) Obtain the round-trip time of the message by sending a Ping packet to each interface and receiving a receipt message. Calculate the link propagation delay for each interface. :

[0080] ;

[0081] 2.3) Based on the total bandwidth of each interface The bandwidth is already being used by the services being transmitted. Calculate the remaining bandwidth for each link. :

[0082] ,

[0083] 2.4) Link switching time based on predicted topology files and the current moment Calculate the remaining available duration of the direct satellite link. :

[0084] .

[0085] Step 3: The control plane of each satellite node runs the link-state routing protocol to complete the network topology and initial routing table construction.

[0086] The link-state routing protocol uses OSPFv3, and each satellite node is configured with a RouterID, the Area, and the interface IPs of the four inter-satellite links. This step is implemented as follows:

[0087] 3.1) Each satellite node periodically sends Hello probe messages on its direct connection interface to discover the Router ID of its directly connected neighbor and establish neighbor relationships;

[0088] 3.3) When each satellite node receives a Hello probe message from a neighboring node, and the neighbor list of the Hello message contains the node's identifier, it determines that the other party is bidirectionally reachable, establishes an adjacency relationship with it, and both parties need to periodically send Hello messages to confirm that the other party is alive.

[0089] 3.4) The link state database (LSDB) is synchronized between adjacent nodes. The relevant link parameters of both parties are synchronized through the exchange of link state information, and the initialization of the local LSDB is completed.

[0090] 3.5) Each satellite node constructs the network topology based on the LSDB and calculates the shortest path from its own node to all nodes in the network according to the Dijkstra algorithm, thus forming the initial routing table.

[0091] Step 4: Each satellite node obtains the link switching event from the predicted topology file and determines its attributes.

[0092] 4.1) Each satellite node obtains the first... Link switching events, such as Figure 3 As shown;

[0093] 4.2) Each satellite node determines whether the satellite node involved in the link handover event is its own node:

[0094] If so, add the event to the event queue of this node, execute step 5, and record its switching time. ;

[0095] Otherwise, it will not participate in this link switching event and will wait to receive the flooded Link Status Message (LSA) from the event node.

[0096] Step 5: The node generates a link state message one moment before the switching time and floods it to the entire network through the local link state database.

[0097] 5.1) This node generates a Link State Message (LSA) based on the type of link switching event, which includes: the local node's Router ID, the peer node's Router ID, the link state change type, and the link switching time. Link propagation delay Remaining bandwidth of the link Remaining link duration Verification parameters such as message sequence number and timestamp;

[0098] 5.2) When the system time arrives... At any given time, this node will flood the LSA to its directly connected neighbor nodes and add the LSA to the local link-state database LSDB.

[0099] Step 6, other satellite nodes in During the time slice, LSA packets flooded by nodes in the link switching event will be received successively, checked, and added to the local LSDB.

[0100] 6.1) Other nodes determine whether they have received the LSA flooded by the nodes during the link handover time:

[0101] If received, perform a validity check on the received link state message, including message validity and integrity, and execute 6.2).

[0102] Otherwise, resend the LSA request within the time window to ensure database synchronization.

[0103] 6.2) Other nodes parse the attribute parameters carried in the message:

[0104] If the link switching time is carried The three elements—the Router ID of the link node, the Router ID of the peer node, and the switching time of the upcoming link switching event—are all related to the timing of the switching event. If the link node Router ID and the peer node Router ID are specified, then proceed to step 6.3).

[0105] If they do not match, the message is deemed invalid and discarded.

[0106] 6.3) Other nodes check their local LSDB for the existence of a corresponding link record based on the link information described in the message, such as the node Router ID and the peer Router ID.

[0107] If the local LSDB does not contain information such as the node Router ID and peer node Router ID for this packet, a new link status entry will be added and the relevant attribute parameters of the packet will be written.

[0108] Otherwise, update the attribute parameters of this link in the local LSDB in sequence: update the propagation delay in the local LSDB. Remaining bandwidth Remaining duration Update the data to the latest message value, update the link state type, and record the timestamp of this update to ensure that the link state database always reflects the latest link state.

[0109] Step 7: Each satellite node calculates the comprehensive weight of the link based on the link parameters in the link state database, performs route calculation based on the comprehensive weight, and generates... Routing table at any time And perform a backup.

[0110] Reference Figure 4 The implementation of this step includes:

[0111] 7.1) Each satellite node obtains the link propagation delay from the link state database. Remaining bandwidth of the link Remaining link duration The maximum and minimum values ​​are found, and these link parameters are normalized to obtain the normalized values. , , :

[0112]

[0113] in, Used to characterize the impact of link propagation delay on transmission performance; the smaller the value, the lower the link delay. This is used to characterize the available bandwidth margin of a link; the larger the value, the stronger the service capacity that the link can carry. This parameter characterizes the remaining available duration of a link under the current topology. A larger value indicates higher link stability. The above parameters satisfy... ;

[0114] 7.2) Each satellite node calculates the overall weight for each link based on the link parameters in the LSDB. :

[0115]

[0116] This comprehensive weight Used to uniformly characterize the transmission performance, load capacity, and stability of a link, among which, This is the time delay weighting coefficient, used to adjust... right The intensity of the impact, The larger the value, the more the routing calculation tends to select links with shorter propagation delays in order to reduce end-to-end latency and latency jitter; This is a bandwidth weighting coefficient used for adjustment. right The influence strength of the routing calculation is such that it tends to select links with more remaining bandwidth and lower congestion risk in order to meet the business bandwidth constraints. This is a stability parameter used for adjustment. Comprehensive weight of links The intensity of the impact, The larger the value, the more likely the routing calculation will choose links with a longer remaining available duration, in order to reduce the overhead of rerouting and topology discovery caused by link switching.

[0117] To ensure that the routing calculation results favor links with longer remaining available durations, this example uses, but is not limited to, weighting factors. , , This reduces the overhead of rerouting and topology discovery caused by link switching;

[0118] 7.3) Each satellite node reads link-state entries from the LSDB and constructs a network topology diagram for route calculation. ,in:

[0119] For a set of nodes, Indicates the first satellite nodes, The total number of nodes;

[0120] For the set of transmission edges, Represents a node and A transmission edge is formed;

[0121] Let be the set of transmission edge costs. Represents a transmission edge The link comprehensive weight is used to characterize the transmission cost of the link at the current moment;

[0122] 7.4) Each satellite node determines an IP six-tuple based on the received IP packets, including the source IP address, destination IP address, source port number, destination port number, protocol, and DSCP value;

[0123] 7.5) Based on business The IP six-tuple, a set of constraints for resolving services. ,in, The source node number mapped to the source IP address. The destination node number mapped to the destination IP address. The bandwidth requirement for the message's DSCP value. The upper bound of the delay required by the message's DSCP value. The length of a single packet in the message;

[0124] 7.6) According to Figure 4 Give the specific number of each satellite in the network. For constraint sets Set a specific set of values, where X represents the track number of the node, Y represents the sequence number within the track, and let... Indicates the source node number of the service. Indicates the destination node number of the service. This indicates the bandwidth requirements of the service. This represents the upper bound of the latency required by the business logic. This represents the length of a single packet in the message, and the specific values ​​of these constraint sets are used as... Figure 5 Network topology diagram The input yields a feasible subgraph for path search. ;

[0125] 7.8) Each satellite node is weighted comprehensively. As a cost, in The above performs a business execution path search, with the source node being a satellite ( The destination node is a satellite. Based on the shortest path search algorithm, Shortest weighted path calculation is performed to obtain the smallest set of end-to-end paths that satisfy bandwidth and latency constraints. for:

[0126] ;

[0127] 7.9) For the path set Transmission delay Calculations are performed based on the length of a single packet within the business constraint set. and remaining bandwidth of the link Calculate the transmission delay for each hop of the path. :

[0128] ;

[0129] 7.10) Set of paths Each hop transmission delay With propagation delay Accumulate to obtain the end-to-end delay. :

[0130]

[0131] in, For path set Total number of hops within, For the first Jump node;

[0132] The final end-to-end delay calculated in this example And satisfy The path is determined to be a one-hop feasible end-to-end path;

[0133] 7.11) Each satellite node performs routing table distribution and update for the service flow. Each hop node on the path generates a corresponding service routing table entry to indicate the next hop forwarding node of the service flow and its corresponding outgoing interface, and backs up the routing table entry.

[0134] Step 8, arrive at the global time. At each time point, each satellite node updates its routing table. And obtain the next link switching event.

[0135] 8.1) Each satellite node performs different operations based on its own attributes:

[0136] If it is a node on the forwarding path, then the table entry corresponding to that service will be removed from the forwarding path. Switch to , and delete ;

[0137] Otherwise, retain the original routing table entry and do not perform the switching or deletion operation;

[0138] 8.2) Each satellite node determines from the predicted topology file whether there will be another link handover event:

[0139] If there is another link switching event, repeat steps 4 to 7 to build a new routing table entry;

[0140] Otherwise, the event handling is complete.

[0141] It should be noted that the step numbers in the above embodiments and claims are only for the purpose of clearly and completely describing the implementation schemes of the present invention and for ease of understanding, and their order is not limited.

[0142] Example 2: Deterministic Satellite Internet Routing System Based on Predictive Topology

[0143] Reference Figure 5 This example includes a prediction topology construction module 1, a link parameter acquisition module 2, a link state flooding module 3, a link state maintenance module 4, a route calculation module 5, and a route table maintenance module 6. Specifically, the link state flooding module 3 includes a time synchronization submodule 31 and a prediction-triggered flooding submodule 32; the link state maintenance module 4 includes a consistency verification submodule 41 and a database maintenance submodule 42; and the route calculation module 5 includes a constraint parameter parsing submodule 51, a network topology construction submodule 52, a shortest path search submodule 53, and a route table construction submodule 54.

[0144] The working principle of the entire system is as follows:

[0145] The predictive topology construction module 1 is used to deduce the establishment, disconnection and switching relationships of inter-satellite links in advance based on the motion state and future position information of satellite nodes, form a time-ordered link switching event sequence, and output the link switching events to the link status flooding module 3.

[0146] The link parameter acquisition module 2 is used to periodically collect the status of the direct links of this node, and at least obtain real-time parameters such as link propagation delay, remaining bandwidth and remaining duration of the link, and output the collection results to the link status flooding module 3.

[0147] The link status flooding module 3 is used to propagate the collected link parameters and link status information that is about to switch within the network range, so that each satellite node can obtain a consistent link status view. Among them, the time synchronization submodule 31 is used to synchronize the local clock of each node at the millisecond level and output the unified clock to the prediction trigger flooding submodule 32. The prediction trigger flooding submodule 32 actively triggers the link status flooding within the time margin before the actual link change according to the link switching event, and outputs the flooded message to the link status maintenance module 4.

[0148] The link state maintenance module 4 is used to establish neighbor relationships between nodes by periodically sending and receiving probe messages, and to manage and maintain link state information in a unified manner by combining the link switching events output by the predictive topology construction module. Among them, the consistency verification submodule 41 performs correctness and version consistency verification on the received link state messages before route calculation to avoid erroneous path calculations, and outputs the correctly verified messages to the link state database maintenance submodule 42. The link state database maintenance submodule 42 checks whether there are missing or abnormal entries in the link state database. When a missing or abnormal entry is found, it requests the corresponding neighbor node to resend the link state information until the database is complete and reliable, and outputs the maintained link state database to the route calculation module 5.

[0149] The routing calculation module 5 is used to generate an end-to-end path that takes effect at a predetermined handover time, based on the converged link state database and in conjunction with the bandwidth constraints and end-to-end latency upper bound requirements of the service. Specifically: the constraint parameter parsing submodule 51 parses the bandwidth requirements and latency constraints of the service to form a constraint set for path calculation, and outputs it to the network topology construction submodule 52; the network topology construction submodule 52 prunes links that do not meet the service constraints based on resource conditions such as remaining link bandwidth, and constructs a feasible network topology based on the comprehensive link weights, outputting the pruned network topology to the shortest path search submodule 53; the shortest path search submodule 53 executes the minimum cost path algorithm on the feasible subgraph to obtain an end-to-end path that meets the constraints, and outputs the end-to-end path to the routing table construction submodule 54; the routing table construction submodule 54 binds the end-to-end path with the corresponding link handover time, generates candidate routing table entries, and outputs them to the routing table maintenance module 6.

[0150] The routing table maintenance module 6 is used to switch the currently effective routing table to the generated candidate routing table and delete the original old routing table or its corresponding old entries to avoid link interruption caused by the failure of the old path during the link switching process, and to ensure that deterministic services can be forwarded continuously and reliably before and after the switching.

[0151] It should be noted that the above functional modules can be implemented, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, they can be implemented, in whole or in part, as program instruction products. A program instruction product includes one or a set of program instructions. When the program instructions are loaded and executed on a computer, all or part of the above-mentioned clustered processes or functions will be implemented. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The program instructions can be stored in a computer-readable and writable storage medium, or transferred from one computer's readable and writable storage medium to another.

[0152] In this embodiment, the direct coupling or communication connection between the modules can be achieved through indirect coupling or communication connection via interfaces, devices, or modules. The functional modules and sub-modules in this embodiment can dynamically reside within a single processing unit, or each module can exist physically independently, or two or more modules can dynamically reside within a single processing unit. When these dynamic components are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable and writable storage medium. This storage medium can be a memory, disk, or optical disc, etc.

[0153] The above descriptions are merely two specific examples of the present invention and do not constitute any limitation on the present invention. Obviously, those skilled in the art, after understanding the content and principles of the present invention, may make various modifications and changes in form and details without departing from the principles and structure of the present invention. For example, step one and the predicted topology construction module 1 can adopt a method based on orbit parsing, learning based on historical link data, or extrapolation based on real-time ephemeris, and are not limited to the method of direct derivation based on satellite orbit operation laws as described in this embodiment; the control plane routing protocol of step two and the link state flooding module 3 can adopt the OSPF protocol, or other link state routing protocols such as IS-IS or BGP-LS, and is not limited to the single link state routing protocol described in this embodiment; step 7 and the network topology construction submodule 52 can adopt linear weighting, nonlinear mapping, machine learning models, or multi-objective optimization methods, and are not limited to the specific weight function form described in this embodiment; the route calculation in step 7 and the shortest path search submodule can be based on the Dijkstra algorithm, or other path search algorithms such as Bellman-Ford and A*, and is not limited to the specific shortest path algorithm described in this embodiment. However, these modifications and alterations based on the ideas of this invention are still within the scope of protection of the claims of this invention.

Claims

1. A deterministic satellite internet routing method based on predicted topology, characterized in that, include: (1) Construct a predictive topology file that includes a set of link switching events based on the satellite orbital operation patterns; (2) Each satellite node periodically collects its own direct link parameters; (3) Each satellite node's control plane runs the link-state routing protocol, establishes neighbor relationships through probe messages, and announces the parameters it has collected to other satellite nodes through link-state messages, thus completing the construction of its own network topology and initial routing table; (4) Each satellite node obtains the link switching event from the prediction topology file. When the satellite in the link switching event is its own node, the node... The system continuously generates link status messages as described in the event (link disconnection / establishment), adds them to the local link status database, and floods them across the entire network. (5) Other satellite nodes in During the time slice, link state messages flooded by nodes in the link switching event will be received successively, checked, and added to the local link state database. (6) Each satellite node calculates the comprehensive weight of the link based on the link parameters in the link state database, performs route calculation based on the comprehensive weight, and generates a route. Routing table at any time And perform a backup; (7) When the system time arrives At that time, each satellite node will update its outdated routing table. Switch to the latest routing table and the original routing table delete; (8) Each satellite node obtains the next link switching event and repeats (4) to (7) until the prediction topology file ends.

2. The method according to claim 1, characterized in that, The construction of the prediction topology file in (1) includes: 1a) Obtain the orbital parameters or ephemeris data of each satellite and calculate the satellite position and link distance; 1b) Calculate the link establishment / disconnection interval between every two satellites based on inter-satellite geometric visibility, pointing constraints, or communication range settings; 1c) Formed according to the link establishment / disconnection interval Link switching event set ,in To switch the link number, To switch time, For source node satellites, The first satellite before the switchover. This is the second satellite after the switch; 1d) Arrange the events in the link switching event set from earliest to latest according to the switching time to form a prediction topology file, and distribute it to each satellite node.

3. The method according to claim 1, characterized in that, In (2), each satellite node periodically collects its own direct link parameters, including: remaining link bandwidth, propagation delay, and remaining link duration, wherein: The remaining bandwidth of the link is determined by the total bandwidth of the interface. The bandwidth is already being used by the services being transmitted. The remaining bandwidth is obtained: ; The propagation delay is the round-trip time of the packet obtained by sending a Ping packet and receiving a receipt message through the interface. Calculate the link propagation delay for each interface. : ; The remaining duration of the link is calculated by each satellite node based on the predicted topology file, from the current time to the link switching time. Remaining transmission time Assuming the current time is Calculate the remaining duration of the link. : 。 4. The method according to claim 1, characterized in that: The establishment of neighbor relationships through probe messages in (3) is used to confirm the reachability between satellite nodes and to provide a basis for subsequent flooding of link status messages; In (3), completing the construction of each network topology means constructing the network topology based on the local link state database; In (3), the initial routing table construction refers to obtaining the initial routing table information of the network by running the shortest path algorithm on the network topology.

5. The method according to claim 1, characterized in that: In (4), each satellite node obtains the link switching event from the prediction topology file, which includes the first... Link, switching time Source node satellite The first satellite before the switch The second satellite after the switch , represented as ; The link status message in (4) refers to the direct link status of this node used to flood other nodes, including the remaining bandwidth of the link, the link propagation delay, the remaining duration of the link, and the connection status. The reserved time margin in (4) This includes the latency of link state flooding and the latency of route calculation, which are used to ensure that each satellite node completes link state synchronization and route calculation before the link switch occurs.

6. The method according to claim 1, characterized in that, The other satellite nodes mentioned in (5) The link status messages received successively within the time slice will be checked, and the implementation includes: 5a) Perform a validity check on the link state messages to verify their validity and integrity; 5b) Parse the attribute parameters carried in the message, including the source node, destination node, and handover time of the link; 5c) Compare the handover times in the link handover events. Source node satellite The first satellite before the switch The second satellite after the switch The message is checked for consistency with the attribute parameters carried in it, and messages with consistent parameters are added to the link-state database; otherwise, the message is discarded.

7. The method according to claim 1, characterized in that, The step (6) involves each satellite node calculating the comprehensive weight of the link based on the link parameters in the link state database. This calculation includes: 5a) Each satellite node obtains the link propagation delay from the link state database. Remaining bandwidth of the link Remaining link duration The maximum and minimum values ​​are found, and these link parameters are normalized to obtain the normalized values. , , : ， in, Used to characterize the impact of link propagation delay on transmission performance; the smaller the value, the lower the link delay. This value is used to characterize the available bandwidth margin of a link; the larger the value, the stronger the service capacity that the link can carry. This parameter characterizes the remaining available duration of a link under the current topology. A larger value indicates higher link stability. The above parameters satisfy... ; 5b) Each satellite node pairs the normalized data. , , Linear weighting is performed to obtain the link synthesis weight used to characterize the cost of the transmission edge. : ， ， in, This is the time delay weighting coefficient, used to adjust... right The intensity of the impact, The larger the value, the more the routing calculation tends to select links with lower propagation latency in order to reduce end-to-end latency and latency jitter; This is a bandwidth weighting coefficient used for adjustment. right The influence strength of the routing calculation is such that it tends to select links with more remaining bandwidth and lower congestion risk in order to meet the business bandwidth constraints. This is a stability parameter used for adjustment. Comprehensive weight of links The intensity of the impact, The larger the value, the more likely the routing calculation will choose links with a longer remaining available duration, in order to reduce the overhead of rerouting and topology discovery caused by link switching.

8. The method according to claim 1, characterized in that, In step (6), routing calculation is performed based on the comprehensive weight to generate... Routing table at any time And perform backups, the implementation of which includes: 6c) Each satellite node constructs the network topology based on the link state database. , in For a set of nodes, Indicates the first satellite nodes, The total number of nodes; For the set of transmission edges, Represents a node and A transmission edge is formed; Let be the set of transmission edge costs. Represents a transmission edge The link comprehensive weight is used to characterize the transmission cost of the link at the current moment; 6d) Each node determines the bandwidth constraint of the service to be transmitted based on the message six-tuple information. And the upper limit of business latency For the remaining transmission bandwidth Bandwidth constraint not met Prune the transmission edges and nodes to obtain the pruned network topology. ; 6e) Based on the pruned network topology Perform a minimum-cost path search to obtain a set of end-to-end paths with the minimum total cost. ,in Depend on It consists of several transmission nodes; 6f) According to Given the propagation delay and transmission delay of each transmission edge, calculate the end-to-end delay of the path set. : ， Among them, the Transmission delay of each transmission edge With the remaining bandwidth of this link Related, assuming a single data packet size is It is represented as: ； 6g) Based on the upper limit of the latency of the service to be transmitted When satisfied At that time, the set of paths This constitutes a feasible end-to-end path; The satellite nodes along the 6h path will calculate the end-to-end path. Build as Routing table that is always in effect And save it in Before that time, the current routing table will still be used for forwarding.

9. A deterministic satellite internet routing system based on predictive topology, characterized in that, include: The predictive topology building module is used to know the changes in links in the network in advance. By analyzing the movement state of nodes and their location information at future times, it can obtain the link switching events in the future, providing precise timing for subsequent link state flooding and deterministic route calculation. The link parameter acquisition module is used to periodically collect real-time status parameters of the directly connected links of this node, including at least the link propagation delay, the remaining bandwidth of the link, and the remaining duration of the link. The link parameters are used for link comprehensive weight calculation and serve as an important input for path selection. The link state maintenance module is used to establish neighbor relationships between nodes through probe messages and to manage and maintain link state information in a unified manner by combining the predicted topology construction results. The link state flooding module is used to synchronously flood link state information in the network and to flood and update the link state in advance based on passive triggering mechanisms or in combination with predicted topology. The routing calculation module is used to generate an end-to-end path that takes effect at a predetermined time based on the link parameters in the link state database, service bandwidth constraints, and latency upper bound requirements. It also generates path results that meet the deterministic transmission requirements of the service based on the comprehensive link weight, providing a basis for subsequent routing table maintenance. The routing table maintenance module is used by nodes to update the current routing table when the network-wide unified time obtained based on time synchronization reaches the predetermined handover time, so as to avoid link interruption caused by link switching during the handover process.

10. The system according to claim 1, characterized in that, The link state flooding module includes: The time synchronization submodule is used to perform millisecond-level time synchronization of the local clocks of each node within the network, providing a consistent time base across the entire network. The prediction-triggered flooding submodule is used to trigger link state flooding within the time slice before the actual link change, based on the link switching events output by the prediction topology construction module, so that the network can complete state convergence before the link switch. The link state maintenance module includes: The consistency verification submodule is used to verify the correctness and version consistency of link state messages before route calculation, so as to avoid path calculation deviations. The link-state database maintenance submodule is used to check whether there are any missing or abnormal entries in the link-state database. If there are any missing entries, it will re-request the neighboring nodes to send the link state.

11. The system according to claim 1, characterized in that, The routing calculation module includes: The constraint parameter parsing submodule is used to parse the bandwidth constraints and latency upper bound QoS requirements of the service, and form a constraint set for path calculation; The network topology construction submodule is used to prune links that do not meet service constraints based on resource conditions such as remaining bandwidth, and to construct a feasible network topology by combining the comprehensive weight of the links. The shortest path search submodule is used to perform the minimum cost path algorithm on the feasible subgraph and output the end-to-end path that satisfies the constraints. The routing table construction submodule is used to bind the end-to-end path with the predetermined switching time and generate corresponding candidate routing table entries.

Images