A SRv6-based traffic transmission method

Abstract

The application provides a traffic transmission method based on SRv6, comprising the following steps: identifying a source domain entry satellite and a destination domain exit satellite according to a topological relationship of a source gateway station and a destination gateway station; configuring an SID of an End-SPA type in the source gateway station; starting an AI path prediction engine, precomputing an optimal SRv6 Policy path set based on historical ephemeris data, real-time link quality and service SLA demand, and generating an SRv6 Policy path strategy set; distributing the SRv6 Policy path strategy set through a control protocol; configuring a BSID for the SRv6 Policy path, instantiating the BSID and injecting a forwarding table, and meanwhile associating the SRv6 Policy path with an End-SPA endpoint behavior; and forwarding traffic. The traffic transmission method can realize dynamic prediction of a path.

Description

Technical Field

[0001] This invention relates to the field of satellite communications, and more specifically to a traffic transmission method based on SRv6. Background Technology

[0002] Satellite communication is a wide-coverage and flexible wireless communication solution capable of solving communication problems in remote areas and at sea. In recent years, with the rapid development of space and communication technologies, low-Earth orbit (LEO) constellation networking has gradually entered the stage of large-scale application. LEO satellites, being relatively close to the Earth's surface, have advantages such as low latency and suitability for high-speed data transmission. By deploying satellite constellations on a large scale in LEO, high-speed internet access services can be provided. Traditional satellite traffic transmission schemes are mainly based on geostationary orbit satellites, using static routing and predefined ephemeris tables for routing decisions, lacking dynamic optimization capabilities.

[0003] To adapt to the traffic forwarding of large satellite constellations, it is particularly important to design a traffic transmission method based on SRv6. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a traffic transmission method based on SRv6.

[0005] This invention provides an SRv6-based traffic transmission method, comprising: identifying source domain ingress satellites and destination domain egress satellites according to the topological relationship between the source and destination gateway stations; configuring an End-SPA type SID at the source gateway station; activating an AI path prediction engine to pre-calculate the optimal SRv6 Policy path set based on historical ephemeris data, real-time link quality, and service SLA requirements, and generating an SRv6 Policy path policy set; distributing the SRv6 Policy policy set through a control protocol; configuring a BSID for the SRv6 Policy path, instantiating the BSID and injecting it into the forwarding table, and simultaneously associating the SRv6 Policy path with the End-SPA endpoint behavior; and performing traffic forwarding.

[0006] According to one embodiment of the present invention, the step of launching the AI ​​path prediction engine, based on historical ephemeris data, real-time link quality, and service SLA requirements, pre-calculates the optimal SRv6 Policy path set and generates an SRv6 Policy path strategy set, including: processing historical ephemeris data based on the SGP4 model to generate ephemeris forecast time-series features; real-time monitoring of inter-satellite and satellite-to-ground links to extract link KPI feature vectors; parsing service QoS (Quality of Service) requirements and converting them into a weight matrix to output service SLA requirement feature vectors; fusing time-series features, KPI features, and SLA requirement features to train and dynamically update the path decision machine learning model; outputting an SRv6 Policy strategy set containing primary paths, backup paths, and candidate paths; deploying the path decision machine learning model to the real-time inference engine; distributing the SRv6 Policy strategy set through the PCEP protocol, and collecting performance monitoring data to feed back to the inference engine to optimize the full-stack parameters of path decision.

[0007] According to one embodiment of the present invention, the process of fusing temporal features, KPI features, and SLA requirement features to train and dynamically update the path decision machine learning model includes: inputting temporal features, KPI features, and SLA requirement features; aligning the data stream according to UTC timestamps and transforming it to an orbital coordinate system; splicing the temporal features, KPI features, and SLA requirement features, and injecting inter-satellite relative motion vectors and space environment indices; employing a multi-objective deep reinforcement learning (DRL) model, with a loss function including SLA requirement, energy consumption, and stability penalty terms; and optimizing the parameters of the multi-objective deep reinforcement learning (DRL) model based on feedback from acquired performance monitoring data.

[0008] According to one embodiment of the present invention, after the output includes an SRv6 Policy set comprising a primary path, backup paths, and candidate paths, it further includes: verifying the SRv6 Policy set using a digital twin; the verification of the SRv6 Policy set using a digital twin includes: constructing a virtual satellite network; injecting fault scenarios; and evaluating model decisions.

[0009] According to one embodiment of the present invention, the construction of the virtual satellite network includes: importing real-time ephemeris data and generating satellite motion trajectories; loading network topology and defining node connection rules and link attributes; and injecting a space environment disturbance model.

[0010] According to one embodiment of the present invention, after configuring a BSID for the SRv6 Policy path, instantiating the BSID and injecting it into the forwarding table, and associating the SRv6 Policy path with the End-SPA endpoint behavior, the process includes: deploying a detection mechanism to detect the primary path and the backup path; the traffic forwarding process includes: using the primary path for traffic forwarding, continuously verifying the reachability and SLA metrics of the backup path; if a primary path failure or SLA exceeding the standard is detected, and the backup path verification passes, then switching to the backup path for traffic forwarding; if the primary path failure is eliminated and stability is restored, then detecting the primary path again; if the primary path passes the continuous stability verification, then switching back to the primary path for traffic forwarding.

[0011] According to an embodiment of the present invention, the deployment detection mechanism, which detects the primary path and backup path, further includes: detecting whether the satellite pointed to by the source gateway station has changed; after detecting whether the satellite pointed to by the source gateway station has changed, the mechanism further includes: if the satellite pointed to by the source gateway station has changed, using the new satellite as the head node, recalculating a new SRv6 Policy path using an AI path prediction engine; if the satellite pointed to by the source gateway station has not changed, determining whether the satellite network involves satellite interconnection in different orbital planes; if it involves satellite interconnection in different orbital planes, the head node derives the topo for the next time slice based on the ephemeris, calculates the candidate path for the next time slice that meets the link SLA, and adds the candidate path for the next time slice as a candidate path for the next time slice to the policy set and injects it into the forwarding table.

[0012] According to one embodiment of the present invention, the step of detecting the main path again if the main path failure is eliminated and stability is restored; and switching back to the main path for traffic forwarding after the main path passes the continuous stability verification, includes: calculating the remaining lifetime of the currently active path based on ephemeris; if the remaining lifetime of the currently active path is not within the optimal lifetime, detecting the status of the candidate path at the next moment; if the status of the candidate path at the next moment is UP, calculating the remaining lifetime of the candidate path at the next moment based on ephemeris; determining whether the remaining lifetime of the candidate path at the next moment is within the optimal lifetime; if the remaining lifetime of the candidate path at the next moment is within the optimal lifetime, using the candidate path at the next moment for traffic forwarding and marking the current path as expired; if the remaining lifetime of the candidate path at the next moment is not within the optimal lifetime, using the current path for traffic forwarding until the remaining lifetime of the current path drops to 2% of the remaining lifetime, and the status of the candidate path at the next moment is UP, then switching to the candidate path at the next moment for traffic forwarding.

[0013] According to one embodiment of the present invention, if the remaining lifetime of the current active path is not within the optimal lifetime, detecting the status of the candidate path at the next moment includes: detecting whether there is a pre-installed candidate path at the next moment; if there is a candidate path, using a lightweight method to detect the status of the candidate path; if there is no candidate path, using an AI path prediction engine to recalculate the path.

[0014] According to one embodiment of the present invention, after the step of if the remaining lifetime of the candidate path at the next moment is not within the optimal lifetime, the method further includes: detecting the current path status; if the current path status is UP, then using the current path for traffic forwarding; if the status of both the primary path and the backup path is DOWN, and the status of the candidate path at the next moment is UP, then switching to using the candidate path at the next moment for traffic forwarding.

[0015] According to the SRv6-based traffic transmission method of the present invention, the traditional ephemeris table static routing is replaced by an AI engine, which enables dynamic prediction of the path.

[0016] It should be understood that the above general description and the following specific embodiments are merely exemplary and illustrative, and do not limit the scope of the invention. Attached Figure Description

[0017] The accompanying drawings, which are part of the specification of this invention, illustrate exemplary embodiments of the invention. The drawings, together with the description in the specification, serve to illustrate the principles of the invention.

[0018] Figure 1 This is a flowchart of a traffic transmission method based on SRv6 according to an embodiment of the present invention. Detailed Implementation

[0019] The features and exemplary embodiments of various aspects of the present invention will now be described in detail. To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present invention and to exemplify the principles of the present invention, and are not configured to limit the present invention. In addition, the structural components in the drawings are not necessarily drawn to scale. For example, the dimensions of some structural components or regions in the drawings may be enlarged for other structural components or regions to aid in the understanding of the embodiments of the present invention.

[0020] The directional terms used in the following description refer to the directions shown in the figures and are not intended to limit the specific structure of the embodiments of the present invention. In the description of the present invention, it should be noted that, unless otherwise stated, the terms "installation," "connection," and "joining" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in the present invention according to the specific circumstances.

[0021] Furthermore, the terms "comprising," "including," "having," or any other variations thereof are intended to cover non-exclusive inclusion, such that a structure or component that includes a list of elements includes not only those elements but also other structural elements that are not expressly listed or inherent to the structure or component. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the article or apparatus that includes the element.

[0022] Spatial relation terms such as "below," "under," "under," "low," "above," "on," and "high" are used for descriptive convenience to explain the positioning of one element relative to a second element, indicating that these terms are intended to cover different orientations of the device, in addition to those different from those shown in the figure. Furthermore, phrases such as "one element on / below another element" can indicate that two elements are in direct contact, or that there are other elements between the two elements. In addition, terms such as "first" and "second" are also used to describe individual elements, areas, parts, etc., without specifically indicating order or sequence, and should not be considered restrictive. Similar terms are used throughout the description to represent similar elements.

[0023] It will be apparent to those skilled in the art that the present invention can be practiced without requiring some of these specific details. The following description of embodiments is merely intended to provide a better understanding of the invention by illustrating examples of the invention.

[0024] Terminology definition:

[0025] SRv6 (Segment Routing IPv6) is a source routing technology based on IPv6 that allows explicit programming of network paths by extending the IPv6 packet header.

[0026] End-SPA type SID (Segment ID) is a 128-bit IPv6 address used in the SRv6 (Segment Routing IPv6) architecture to identify the dynamic next-hop address.

[0027] End-B6 is a behavior directive in SRv6 (Segment Routing IPv6) technology, used to control the forwarding path of data packets.

[0028] PCEP (Path Computation Element Protocol) is a TCP-based application layer protocol defined by the IETF PCE working group for communication between Path Computation Element (PCE) and Path Computation Client (PCC).

[0029] KPI stands for Key Performance Indicator, which is used to quantitatively evaluate network link performance.

[0030] A Service Level Agreement (SLA) is a mutually agreed-upon agreement between a service provider and a user, or between service providers themselves, to ensure the performance and reliability of a service under certain costs.

[0031] Figure 1 This is a flowchart of a traffic transmission method based on SRv6 according to an embodiment of the present invention.

[0032] like Figure 1 As shown, the present invention provides a traffic transmission method based on SRv6, comprising:

[0033] S010: Identify source domain entry satellites and destination domain exit satellites based on the topological relationship between the source and destination gateway stations;

[0034] S020: Configure an End-SPA type SID at the source gateway station;

[0035] S030: Launch the AI ​​path prediction engine to pre-calculate the optimal SRv6 Policy path set based on historical ephemeris data, real-time link quality, and business SLA requirements, and generate the SRv6 Policy path strategy set;

[0036] S040: Distribute the SRv6 Policy set via the control protocol;

[0037] S050: Configure the BSID for the SRv6 Policy path, instantiate the BSID and inject it into the forwarding table, and associate the SRv6Policy path with the End-SPA endpoint behavior;

[0038] S060: Perform traffic forwarding.

[0039] Specifically, End-SPA (End-Space) is a new type of SRv6 SID that indicates the next hop for traffic forwarding is the fixed satellite beam coverage area of ​​the current gateway station. The outgoing interface of this SID actively switches with each time slot, and IGP convergence calculation is not performed when the outgoing interface state remains unchanged. Since routers within the same domain often share the same prefix, only the node changes, by creating a new End-SPA type SID with a dynamic next-hop address and correspondingly modifying the End-B6 SID, the same BSID (representing an SRv6 Policy path to the destination gateway station) can be established for the sky corresponding to different time slots.

[0040] The traffic transmission method provided in this embodiment performs SRv6 traffic forwarding based on BSID, triggers associated End-SPA endpoint behavior, and selects the optimal path for packet encapsulation and transmission according to a pre-calculated policy set. This transmission method replaces traditional ephemeris table static routing with an AI engine, enabling dynamic path prediction. Furthermore, this traffic transmission method uses historical ephemeris data, real-time link quality, and service SLA requirements as inputs to train the AI ​​engine, generating an SRv6 Policy path policy set and effectively improving the real-time performance of path calculation. In addition, the AI ​​engine can intelligently learn when constellation topology changes, ensuring the output of the optimal path. For example, based on orbital dynamics models and historical ephemeris data, it can predict satellite positions and link stability periods over a future period, avoiding signal interruptions during satellite handover.

[0041] According to one embodiment of the present invention, an AI path prediction engine is activated, and based on historical ephemeris data, real-time link quality, and service SLA requirements, the optimal SRv6 Policy path set is pre-calculated, and an SRv6 Policy path strategy set is generated, including:

[0042] S030.10: Based on the SGP4 model, process historical ephemeris data to generate ephemeris forecast time series features;

[0043] S030.20: Real-time monitoring of inter-satellite and satellite-to-ground links, and extraction of link KPI feature vectors;

[0044] S030.30: Parse the service QoS (Quality of Service) requirements and convert them into a weight matrix, outputting the service SLA requirement feature vector;

[0045] S030.40: Integrate time-series features, KPI features, and SLA requirement features to train and dynamically update the path decision machine learning model;

[0046] S030.50: Outputs an SRv6 Policy set containing the primary path, backup path, and candidate path;

[0047] S030.60: Deploy the path decision machine learning model to the real-time inference engine;

[0048] S030.70: Distribute the SRv6 Policy set via the PCEP protocol and collect performance monitoring data to feed back to the inference engine to optimize the full-stack parameters of path decision-making.

[0049] In this embodiment, the inference engine is an onboard module that calculates path decisions in real time on the satellite. Historical ephemeris data includes position and velocity vectors, etc. Link KPI feature vectors may include latency, packet loss rate, and bandwidth utilization. This traffic transmission method continuously optimizes the path decision machine learning model, feature strategy, and weight matrix by feeding performance monitoring data (e.g., packet loss rate, latency) back to the inference engine, effectively improving response speed. For example, the actual latency of traffic forwarding can be compared with the predicted latency; if the error exceeds a threshold, model retraining is triggered, effectively addressing the high dynamism of satellite networks.

[0050] Business SLA requirement characteristics are structured, quantified metrics extracted from the original Quality of Service (QoS) requirements of users or the service layer. In S030.30, a multi-objective optimization model can be constructed, taking a weight matrix representing importance as input and outputting a structured feature vector for training a path decision machine learning model. For example, a latency weight of 0.6, a jitter weight of 0.4, and a packet loss rate weight of 0 can be assigned to voice services, while a bandwidth weight of 0.7 and a packet loss rate weight of 0.3 can be assigned to video services.

[0051] Priorities can be set for primary and backup paths based on SLA parameters. For example, paths with latency ≤ 5ms are designated as primary paths, and paths with latency ≤ 10ms are designated as backup paths.

[0052] According to one embodiment of the present invention, a path decision machine learning model is trained and dynamically updated by integrating time-series features, KPI features, and SLA requirement features, including:

[0053] S030.40.1: Input time series features, KPI features and SLA requirement features, align the data stream according to UTC timestamps and transform it to the orbital coordinate system;

[0054] S030.40.2: Combine time-series features, KPI features, and SLA requirement features, and inject inter-satellite relative motion vectors and space environment indices to form training data;

[0055] S030.40.3: Employ a multi-objective deep reinforcement learning (DRL) model, with the loss function including SLA requirement, energy consumption, and stability penalty terms; train the multi-objective deep reinforcement learning (DRL) model using the training data output from step S030.40.2;

[0056] S030.40.4: Optimize the parameters of the multi-objective deep reinforcement learning (DRL) model based on the feedback from step S030.70.

[0057] Satellite network routing requires simultaneous optimization of multiple conflicting objectives, including Service Level Agreement (SLA) requirements (e.g., latency, bandwidth, packet loss rate), energy consumption constraints (e.g., satellite battery life protection), and stability requirements. In this embodiment, the loss function parameters are adjusted using SLA and energy consumption data actually monitored in step S030.70, thereby driving the update of the parameters of the multi-objective DRL model. This traffic transmission method, through a multi-objective DRL model and a closed-loop feedback mechanism, can solve the problem of satellite network routing needing to optimize multiple conflicts simultaneously, while meeting the requirements of SLA assurance, onboard energy consumption, and routing stability.

[0058] For example, when a satellite's battery is depleted by more than 80%, a secondary penalty can be applied to prevent specific satellites from overloading, thereby extending the constellation's lifespan.

[0059] According to one embodiment of the present invention, after outputting the SRv6 Policy set including the primary path, backup paths, and candidate paths, the method further includes:

[0060] S030.51: Validate the SRv6 Policy set using digital twins.

[0061] The SRv6 Policy set was validated using digital twins, including:

[0062] S030.51.1: Construct a virtual satellite network;

[0063] S030.51.2: Injecting fault scenarios;

[0064] S030.51.3: Evaluation model decision.

[0065] In this embodiment, if the verification passes, proceed to step S030.60; if the verification fails, return to step S030.40 to retrain the new model. For example, in a fault scenario, if the SLA compliance rate is greater than a threshold, it means the verification has passed, and only then can step S030.60 proceed. This traffic transmission method uses digital twin verification to verify path connectivity and avoid single points of failure. Simultaneously, it ensures the SLA compliance rate and guarantees that latency, packet loss rate, bandwidth, etc., meet business requirements.

[0066] According to one embodiment of the present invention, constructing a virtual satellite network includes:

[0067] S030.51.1.1: Import real-time ephemeris data and generate satellite motion trajectories;

[0068] S030.51.1.2: Load network topology, define node connection rules and link attributes;

[0069] S030.51.1.3: Injection of space environment disturbance model.

[0070] In this embodiment, for example, the space environment disturbance model may include an atmospheric attenuation model, an ionospheric scintillation model, a high-energy particle radiation model, etc.

[0071] According to one embodiment of the present invention, configuring a BSID for an SRv6 Policy path, instantiating the BSID and injecting it into the forwarding table, and then associating the SRv6 Policy path with the End-SPA endpoint behavior, includes:

[0072] S051: Deploy a detection mechanism to detect the primary and backup paths;

[0073] S060: Perform traffic forwarding, including:

[0074] S060.10: Use the primary path for traffic forwarding and continuously verify the reachability and SLA metrics of the backup path;

[0075] S060.20: If a primary path failure or SLA exceedance is detected, and the backup path verification is successful, then switch to the backup path for traffic forwarding;

[0076] S060.30: If the main path failure is eliminated and stability is restored, the main path is tested again; if the main path passes the continuous stability verification, the traffic is switched back to the main path for forwarding.

[0077] In this embodiment, the reachability of the backup path can be verified through a BFD session, and the SLA metrics of the backup path can be measured through TWAMP. For example, if the packet loss rate of the primary path is >1% or the latency is >20ms, traffic forwarding is switched to the backup path. In step S060.20, if a primary path failure or SLA exceeding the limit is detected, and the current state of the backup path is verified to be stable and meet the service requirements, traffic forwarding is switched to the backup path; if the primary path recovers, the primary path is detected again. For example, if the primary path remains stable for 30 seconds and the SLA metrics are compliant, traffic forwarding is switched back to the primary path.

[0078] According to an embodiment of the present invention, S051: Deploy a detection mechanism to detect the primary path and the backup path, and further include:

[0079] Detect whether the direction of the source signal station towards the satellite has changed.

[0080] After detecting whether the direction the source gateway station is pointing at the satellite has changed, the following steps are included:

[0081] S052: If the satellite pointed to by the source gateway changes, a new SRv6 Policy path is recalculated using the new satellite as the head node and the AI ​​path prediction engine (i.e., steps S030-S050 are repeated).

[0082] If the source signal station does not point to the satellite, proceed to step S053;

[0083] S053: Determine whether the satellite network involves interconnection of satellites in different orbital planes;

[0084] If satellite interconnection is involved in different orbital planes, the head node obtains the topo of the next time slice based on the ephemeris, calculates the candidate path for the next time slice that meets the link SLA, and adds the candidate path for the next time slice as a candidate path for the next time slice (i.e. as a backup candidate) to the policy set and injects it into the forwarding table.

[0085] If satellite interconnection across different orbital planes is not involved, proceed to step S060.

[0086] The traffic transmission method in this embodiment ensures data traffic forwarding by actively switching paths within the SRv6 Policy, effectively reducing the performance pressure caused by link calculation and link allocation during topology switching. Furthermore, by dividing time slices using ephemeris data, for large satellite constellations with different orbits, when the source and destination nodes remain unchanged, it ensures smooth switching of traffic paths across multiple time slices for the same source and destination nodes, effectively reducing route awareness. For example, detection mechanisms may include primary / backup path BFD / TWAMP monitoring, source satellite topology change detection, different orbit link status monitoring, and ephemeris validity period alarms.

[0087] Specifically, to clarify the preceding and following statements in this application, the lifecycle is described as follows: A satellite in one orbit may link with two satellites in different orbits. The lifecycle of the first satellite is defined as the period from linking with the first satellite to disconnecting it. When the second satellite links, the first satellite may not be interrupted at this point. The optimal lifecycle can be 80% of the entire lifecycle for lifecycle management of dynamic topo paths.

[0088] According to one embodiment of the present invention, if the main path failure is eliminated and stability is restored, the main path is detected again; if the main path passes the continuous stability verification, after switching back to the main path for traffic forwarding, the process includes:

[0089] S060.40: Calculate the remaining survival time of the current activity path based on the ephemeris;

[0090] S060.50: If the remaining lifespan of the current active path is not within the optimal lifespan, check the status of the candidate paths at the next moment;

[0091] S060.60: If the state of the candidate path at the next moment is UP, then calculate the remaining survival time of the candidate path at the next moment based on the ephemeris.

[0092] S060.70: Determine whether the remaining lifespan of the candidate path at the next moment is within the optimal lifespan;

[0093] If the remaining lifetime of the candidate path at the next time step is within the optimal lifetime, proceed to step S060.80; if the remaining lifetime of the candidate path at the next time step is not within the optimal lifetime, proceed to step S060.90.

[0094] S060.80: Use the candidate path in the next time step for traffic forwarding and mark the current path as expired;

[0095] S060.90: Use the current path for traffic forwarding until the remaining lifetime of the current path approaches a critical point (e.g., drops to 2% of the remaining lifetime), and when the status of the candidate path at the next time step is UP, switch to the candidate path at the next time step for traffic forwarding.

[0096] In the traffic transmission method provided in this embodiment, for the SRv6 Policy path corresponding to the BSID, the forwarding path of the next time slice is calculated, the lifetime of the current path and the path of the next time slice is recorded, and the path status of the next time slice is verified. When the path of the next time slice is available and its lifetime is good, the system actively switches to the next time slice, completing the smooth switching of the onboard router path in a large constellation network. This traffic transmission method makes full use of the multi-path state of the low-Earth orbit satellite constellation network when establishing a topology with other orbits, converting the result of dynamic network topology changes into lifetime, reducing the link instability time in the early stage of link establishment and when the link is about to be lost, and increasing link reliability. This traffic transmission method analyzes the cross-orbit scenario of the constellation network, fully considers the link switching scenario of low-Earth orbit constellation satellites communicating with other orbits, and has the advantages of fast path switching and low requirements for satellite equipment performance. This traffic transmission method has significant advantages in terms of adaptability and feasibility, and is very suitable for path optimization in large low-Earth orbit satellite constellations. Furthermore, the inter-satellite path switching method of this traffic transmission method can effectively reduce the packet loss rate and make the path switching smoother compared to the ordinary link disconnection and reconstruction method.

[0097] In this embodiment, if the remaining lifetime of the primary path is not within its optimal lifespan, traffic forwarding is switched to the backup path. If the remaining lifetimes of both the primary and backup paths are not within their optimal lifespans, the status of the candidate paths at the next time step is checked. If the remaining lifetime of the candidate paths at the next time step is within their optimal lifespan, then the candidate paths at the next time step are used for traffic forwarding.

[0098] For example, if the remaining lifetime of the candidate path at the next moment is not within the optimal lifetime, traffic is forwarded using the current path until the remaining lifetime of the current path drops to 2% of the remaining lifetime and is not lower than the minimum time threshold (e.g., the minimum time threshold is 50ms). If the status of the candidate path at the next moment is UP, and the UP status of the candidate path at the next moment is maintained for a set time (e.g., 10ms), then traffic is switched to the candidate path at the next moment.

[0099] According to one embodiment of the present invention, if the remaining lifetime of the current active path is not within the optimal lifetime, detecting the state of the candidate path at the next moment includes:

[0100] S060.50.1: Detect whether there is a candidate path for the next time step of the pre-installed installation;

[0101] S060.50.2: If a candidate path exists, use a lightweight method to detect the status of the candidate path;

[0102] S060.50.3: If no candidate path exists, recalculate the path using the AI ​​path prediction engine, i.e., return to step S030.

[0103] Lightweight detection is an efficient and low-overhead path state verification mechanism. In this embodiment, lightweight detection of candidate path states enables rapid verification of candidate path states (e.g., connectivity and basic SLA metrics). For example, microsecond-level bidirectional forwarding detection can be used for millisecond-level connectivity checks.

[0104] According to one embodiment of the present invention, if the remaining lifetime of the candidate path at the next time step is not within the optimal lifetime, the method further includes:

[0105] S081: Detect the current path status;

[0106] S082: If the current path status is UP, proceed to step S060.90;

[0107] If both the primary and backup paths are in the DOWN state, and the candidate path for the next time step is in the UP state, then switch to using the candidate path for the next time step for traffic forwarding.

[0108] In this embodiment, an alarm is triggered if both the primary path and the backup path are in a DOWN state, and the candidate path in the next moment is also in a DOWN state. For example, if the packet loss rate is ≤0.5% and the latency is ≤200ms, the path state is UP; if BFD fails to detect three times consecutively, the path state is DOWN.

[0109] According to one embodiment of the present invention, the head node derives the topo of the next time slice based on the ephemeris and calculates the candidate path for the next time step that meets the link SLA, including:

[0110] Calculate the primary candidate path and the backup candidate path for the next time step, and set their corresponding priorities.

[0111] Configure the primary candidate path for the next time step with the same BSID as the SRv6 Policy path;

[0112] The candidate path for the next time step is associated with End-SPA, and the alternative candidate path for the next time step remains in an inactive state.

[0113] According to one embodiment of the present invention, detecting the state of a candidate path at the next time step includes:

[0114] The SRv6 Path Verify is used to detect the status of the candidate path for the next time step (i.e., the path for the next time slice obtained in step S053).

[0115] In this embodiment, when there are enough BFD sessions, BFD can also be used to detect the status of candidate paths in the next time slot. In this traffic transmission method, the on-board candidate path design adopts the next time slot alternative path strategy, and uses pathverify or BFD to detect the connectivity status of the path, which improves the satellite's ability to cope with sudden satellite node or link failures and other problems, thereby improving the reliability and disaster recovery of the low-Earth orbit satellite constellation network.

[0116] The above embodiments of the present invention can be combined with each other and have corresponding technical effects.

[0117] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A traffic transmission method based on SRv6, characterized in that, include: Identify source domain entry satellites and destination domain exit satellites based on the topological relationship between source and destination gateway stations; Configure an End-SPA type SID at the source gateway station; wherein, the End-SPA type means that the next hop for traffic forwarding is the fixed satellite beam coverage area covered by the current gateway station; the outgoing interface of the End-SPA type SID actively switches with time slots; at the same time, the same BSID is established for the sky corresponding to the gateway station for different time slots; Launch the AI ​​path prediction engine to pre-calculate the optimal SRv6 Policy path set based on historical ephemeris data, real-time link quality, and business SLA requirements, and generate the SRv6 Policy path strategy set. Distribute SRv6 Policy sets via control protocols; Configure the BSID for the SRv6 Policy path, instantiate the BSID and inject it into the forwarding table, and associate the SRv6 Policy path with the End-SPA endpoint behavior; Traffic forwarding.

2. The traffic transmission method according to claim 1, characterized in that, The AI ​​path prediction engine is launched to pre-calculate the optimal SRv6 Policy path set based on historical ephemeris data, real-time link quality, and service SLA requirements, and to generate an SRv6 Policy path strategy set, including: Historical ephemeris data is processed based on the SGP4 model to generate ephemeris forecast time series features; Real-time monitoring of inter-satellite and satellite-to-ground links, and extraction of link KPI feature vectors; Analyze the service QoS requirements and transform them into a weight matrix, then output the service SLA requirement feature vector; By integrating time-series features, KPI features, and SLA requirement features, a path decision machine learning model is trained and dynamically updated. The output includes a set of SRv6 policy strategies for the primary path, backup path, and candidate paths; Deploy path decision machine learning models to the real-time inference engine; The SRv6 Policy set is distributed via the PCEP protocol, and performance monitoring data is collected and fed back to the inference engine to optimize the full-stack parameters for path decision-making.

3. The traffic transmission method according to claim 2, characterized in that, The method of integrating time-series features, KPI features, and SLA requirement features to train and dynamically update the path decision machine learning model includes: Input time-series features, KPI features, and SLA requirement features, align the data stream according to UTC timestamps, and transform it to the orbital coordinate system; The training data is formed by splicing together time-series features, KPI features, and SLA requirement features, and injecting relative motion vectors between satellites and space environment indices. A multi-objective deep reinforcement learning (DRL) model is adopted, and the loss function includes SLA requirement, energy consumption, and stability penalty terms; the multi-objective deep reinforcement learning (DRL) model is trained using the training data; The parameters of the multi-objective deep reinforcement learning (DRL) model are optimized based on feedback from the collected performance monitoring data.

4. The traffic transmission method according to claim 2, characterized in that, The output, after containing the SRv6 Policy set of primary path, backup path, and candidate path, also includes: The SRv6 Policy set was validated using digital twins, including: Build a virtual satellite network; Inject fault scenarios; Evaluation model decision.

5. The traffic transmission method according to claim 4, characterized in that, The construction of the virtual satellite network includes: Import real-time ephemeris data to generate satellite motion trajectories; Load the network topology and define node connection rules and link attributes; Inject space environment disturbance model.

6. The traffic transmission method according to claim 2, characterized in that, The process of configuring a BSID for the SRv6 Policy path, instantiating the BSID and injecting it into the forwarding table, and then associating the SRv6 Policy path with the End-SPA endpoint behavior includes: Deploy a detection mechanism to detect both the primary and backup paths; The traffic forwarding includes: Traffic is forwarded using the primary path, and the reachability and SLA metrics of the backup path are continuously verified. If a primary path failure or SLA exceeding the limit is detected, and the backup path passes verification, then switch to the backup path for traffic forwarding. If the main path failure is resolved and stability is restored, the main path is tested again; if the main path passes the continuous stability verification, traffic is switched back to the main path for forwarding.

7. The traffic transmission method according to claim 6, characterized in that, The deployment detection mechanism, which detects both the primary and backup paths, also includes: Detect whether the direction the source signal station is pointing at the satellite has changed; If the satellite pointed to by the source gateway changes, a new SRv6 Policy path is recalculated using the AI ​​path prediction engine with the new satellite as the head node. If the source gateway station does not point to a satellite, determine whether the satellite network involves interconnection of satellites in different orbital planes; If satellite interconnection is involved in different orbital planes, the head node obtains the topo of the next time slice based on the ephemeris, calculates the candidate path for the next time slice that meets the link SLA, and adds the candidate path for the next time slice as a candidate path for the next time slice to the policy set and injects it into the forwarding table.

8. The traffic transmission method according to claim 6, characterized in that, If the main path failure is resolved and stability is restored, the main path is tested again; if the main path passes the continuous stability verification, the process of switching back to the main path for traffic forwarding includes: Calculate the remaining survival time of the current activity path based on the ephemeris; If the remaining lifespan of the current active path is not within the optimal lifespan, check the status of the candidate paths at the next moment; If the candidate path's state is UP at the next moment, then calculate the remaining survival time of the candidate path at the next moment based on the ephemeris. Determine whether the remaining lifespan of the candidate path at the next moment is within the optimal lifespan; If the remaining lifespan of the candidate path at the next time step is within the optimal lifespan, then the candidate path at the next time step is used for traffic forwarding, and the current path is marked as expired; If the remaining lifetime of the candidate path at the next moment is not within the optimal lifetime, the current path is used for traffic forwarding until the remaining lifetime of the current path drops to 2% of the remaining lifetime, and the status of the candidate path at the next moment is UP, then the traffic is switched to the candidate path at the next moment.

9. The traffic transmission method according to claim 8, characterized in that, The step of detecting the status of candidate paths at the next moment if the remaining lifetime of the current active path is not within the optimal lifetime includes: Detect whether there are candidate paths for the next time step that are pre-installed; If candidate paths exist, use lightweight detection to check the status of the candidate paths; If no candidate path exists, the path is recalculated using the AI ​​path prediction engine.

10. The traffic transmission method according to claim 8, characterized in that, If the remaining survival time of the candidate path at the next moment is not within the optimal lifespan, the following steps are also included: Detect the current path status; If the current path status is UP, then the current path will be used for traffic forwarding; If both the primary and backup paths are in the DOWN state, and the candidate path for the next time step is in the UP state, then switch to using the candidate path for the next time step for traffic forwarding.

Images