Integrated monitoring system for operation state of aviation internet based on multi-source data fusion

Abstract

This invention relates to an integrated monitoring system for the operational status of aviation internet based on multi-source data fusion, belonging to the field of aviation internet operational status monitoring technology. The system includes: a multi-source data fusion and unified service module, which integrates heterogeneous data from multiple sources to construct and dynamically update a unified dynamic model of an end-to-end network service chain with flights as the core dimension; a service chain operational status monitoring and tracing module, which performs multi-level indicator monitoring and real-name verification, and realizes root cause tracing and alarm generation in case of anomalies; an adaptive dynamic strategy coordination module, which dynamically generates strategies based on protection rules and comprehensive alarms, and outputs a unified instruction set after arbitration and fusion; and a unified operational status visualization module, which provides integrated display and a one-click operation interface based on models and alarms. This achieves an integrated monitoring closed loop from global situational awareness and intelligent root cause analysis to collaborative strategy handling, significantly improving the operational safety and maintenance efficiency of aviation internet.

Description

Technical Field

[0001] The present invention belongs to the technical field of monitoring the operation status of the aviation Internet, and particularly relates to an integrated monitoring system for the operation status of the aviation Internet based on multi-source data fusion. Background Art

[0002] With the digital transformation of the air transportation industry, the aviation Internet has become the core support for enhancing the travel experience of passengers and the operational efficiency of airlines. The stability of its full-link operation covering satellite communication links, in-flight network devices, cabin wireless access, and passenger terminals is directly related to service quality and flight safety. Currently, the global scale of aviation Internet users continues to expand, and passengers' demand for high-speed network services such as high-definition video and instant messaging has increased sharply. At the same time, airlines need to strictly control the allocation of network resources and ensure the security of operation and maintenance data transmission, which poses high requirements for the integrated, precise, and real-time monitoring of the operation status of the aviation Internet.

[0003] Although existing aviation Internet monitoring technologies already have basic data collection and alarm functions, there are still many defects that need to be solved urgently: First, the multi-source data fusion ability is insufficient. The operation data of the aviation Internet is scattered in the logs of in-flight network devices, satellite communication link data, passenger real-name authentication systems, and flight operation databases. The data formats are heterogeneous, and the transmission protocols vary greatly. Existing solutions mostly adopt the independent processing mode of single data sources and lack a unified fusion framework, resulting in the difficulty of mining data value and the inability to form an understanding of the full-link operation situation. Second, the monitoring levels are fragmented and the coordination is poor. Existing technologies mostly monitor the macro status of flight classes or local parameters of devices separately, and do not realize the linkage analysis of indicators at the flight class, device level, and passenger IP level, making it difficult to accurately locate the root causes of faults in the entire "link-device-terminal" chain.

[0004] Third, the fault tracing efficiency is low. The end-to-end link of the aviation Internet involves multiple complex links such as satellites, in-flight, and cabins. Existing tracing methods rely on manual inspection of each device one by one and lack an intelligent tracing mechanism based on the global topology, resulting in a long time-consuming for fault location and easy delay in the timing of fault handling. Fourth, the policy configuration is static and the adaptability is weak. Existing network service policies are mostly preset fixed rules and cannot be dynamically adjusted according to changes in flight phases, fluctuations in link bandwidth, and passenger access requirements, making it difficult to balance differential service guarantee and efficient resource utilization. Fifth, visualization and operation and maintenance operations are disjointed. Existing monitoring interfaces mostly list data, do not realize the integrated display of the full-link situation, and there is no linkage between the interface operations and the underlying data model, increasing the operation complexity of operation and maintenance personnel. Summary of the Invention

[0005] To solve the above problems existing in the prior art, the present invention provides an integrated monitoring system for the operation status of the aviation Internet based on multi-source data fusion. The object of the present invention can be achieved through the following technical solutions: An integrated monitoring system for aviation internet operation status based on multi-source data fusion includes: The multi-source data fusion and unified service module acquires and merges heterogeneous data from airborne network equipment, satellite communication links, passenger real-name authentication systems, and flight operation databases; dynamically updates the end-to-end network service chain of satellite-airborne-cabin with flight as the core correlation dimension, and constructs a unified dynamic model; the unified dynamic model correlates and maps flight attributes, physical equipment status, IP address resources, passenger session information, and network policies. The service chain operation status monitoring and tracing module, based on the unified dynamic model, calculates and monitors traffic, connection stability and service quality indicators at the flight level, device level and passenger IP level, verifies the compliance of passenger real-name authentication; and when an anomaly is detected, it traces the source to the specific problem link according to the unified dynamic model and generates a comprehensive alarm containing root cause location information. The adaptive dynamic policy coordination module dynamically generates network service policies based on predefined protection rules and the received comprehensive alarms; and performs real-time arbitration and fusion of the network service policies and pre-configured policies to output a unified policy instruction set. The unified operational status visualization module integrates and displays the unified dynamic model, real-time operational status, comprehensive alarms, policy distribution and execution effects, and provides interfaces for policy adjustment, resource isolation or service recovery based on the unified dynamic model and alarms.

[0006] As a preferred technical solution of the present invention, the specific implementation of the flight as the core association dimension includes: maintaining a dynamic data source confidence weight table for each flight; dynamically calculating and adjusting the weight of each data source in data fusion based on the real-time packet loss rate, historical accuracy, and relevance to the current flight phase of the data reported by each data source; and determining the fusion result by performing weighted arbitration based on the data source confidence weight table when different data sources have inconsistent data for the same associated object.

[0007] Specifically, the unified dynamic model includes: a business state graph carrying timestamp attributes in the storage layer; in the business state graph, the edge weights are dynamically calculated based on the real-time topology and traffic data to characterize the propagation intensity of business influence between nodes; when the node state in the unified dynamic model is updated, the associated nodes are traversed along the business state graph to calculate the quantitative impact value of the state change on the associated business indicators.

[0008] Specifically, the satellite-airborne-cabin end-to-end network service chain is represented in the unified dynamic model as a state association topology containing multiple logical segments. The logical segments include at least: a satellite communication segment, an airborne network equipment segment, a cabin wireless access point segment, and a passenger terminal access session segment. The state parameters of each segment are associated, stored, and updated in the unified dynamic model.

[0009] Specifically, the calculation and monitoring of traffic, connection stability, and service quality indicators at the flight level, device level, and passenger IP level are implemented as follows: a hierarchical time window comparison mechanism is adopted to maintain a second-level sliding window, a minute-level aggregation window, and a flight-level benchmark window for each level of indicator; the indicator value of the current second-level sliding window is compared with the statistical baseline of the corresponding minute-level aggregation window; the judgment threshold is dynamically corrected by combining the historical average and fluctuation range of the flight-level benchmark window; if the corrected indicator value exceeds the threshold range, it is judged as abnormal.

[0010] Specifically, the verification of passenger real-name authentication compliance includes: verifying the identity information provided by the passenger when accessing the network against the real-name database in real time, marking devices that fail to verify, have abnormal information, or are unauthenticated, and generating corresponding security alarm events. The security alarm events will trigger the adaptive dynamic policy collaboration module to execute network access interception actions.

[0011] Specifically, the method of tracing back to specific problem links based on the unified dynamic model includes: when an alarm is generated for a monitoring anomaly, the alarm object is traced back along the state association topology according to the mapping relationship in the unified dynamic model, and the passenger session, cabin wireless access point, uplink airborne network equipment and satellite communication link associated with the alarm object are located in sequence.

[0012] Specifically, the predefined guarantee rules include: the core expression form is to ensure that the service target meets the constraint conditions when the trigger condition is met; the adaptive dynamic policy coordination module has a built-in policy compiler that combines the declarative guarantee rules with the real-time running status data in the unified dynamic model to compile and generate a sequence of configuration instructions that can be executed on specific airborne network devices or ground control devices.

[0013] Specifically, the dynamically generated network service strategy includes: after receiving the comprehensive alarm, the adaptive dynamic strategy coordination module matches and instantiates corresponding intervention strategy instructions from the preset emergency response strategy library according to the alarm type and root cause information. The intervention strategy instructions include, but are not limited to, rate limiting or blocking instructions for specific IPs, service restart suggestion instructions for specific devices, or global traffic shaping instructions for link quality.

[0014] Specifically, the real-time arbitration and fusion of network service strategies and pre-configured strategies includes: the core is to solve a constrained multi-objective optimization problem; wherein, the optimization objectives include maximizing passenger service satisfaction, minimizing the number of conflicts between strategies, and balancing network resource utilization; the constraints are the upper limit of the physical resources currently available to the network and the inviolable security policies.

[0015] Specifically, the issuance of the unified policy instruction set is implemented as follows: a tiered activation and monitoring rollback strategy is adopted. For non-critical policy changes, the unified policy instruction set is divided into multiple batches and issued sequentially in a preset order. After each batch of instructions is issued, a preset monitoring period is waited for confirmation that the service chain operation status is stable or abnormal indicators have improved before the next batch of instructions is issued. If the service chain status is detected to be deteriorating, the issuance of subsequent batches of instructions is automatically stopped, and the rollback operation of the issued instructions is performed.

[0016] Specifically, the unified operational status visualization module and the unified dynamic model achieve bidirectional state synchronization and operation mapping, including: Forward mapping: The user's selection, filtering, and drill-down operations in the visual interface are converted into graph query commands or attribute filtering conditions for the unified dynamic model. Reverse synchronization: The state update, alarm generation, and policy activation events in the unified dynamic model drive the style, position, or value update of the corresponding elements in the visualization interface in real time.

[0017] The beneficial effects of this invention are as follows: This system overcomes the challenge of data silos from multiple sources, achieving precise situational awareness across the entire data chain. By constructing a unified data fusion framework, it adapts to the formats and protocols of heterogeneous data from airborne networks, satellite links, passenger registration, and flight operations, enabling efficient integration of multi-source data. Compared to existing independent processing models based on single data sources, this system fully leverages data value, forming a comprehensive operational situational awareness across the entire chain from satellite to airborne to cabin to terminal. This provides complete data support for subsequent monitoring, analysis, and strategy development, enhancing the comprehensiveness and accuracy of operational status perception.

[0018] Breaking down barriers between monitoring levels, this system achieves comprehensive and collaborative monitoring. It constructs a three-tiered, interconnected monitoring system at the flight, equipment, and passenger IP levels, overcoming the limitations of existing technologies that rely on fragmented monitoring layers. Through the coordinated analysis of these three levels of indicators, it can accurately capture operational anomalies along the entire "link-equipment-terminal" chain. This allows for both a comprehensive understanding of the airline's macro-level operational status and a detailed analysis of service quality at individual passenger terminals, providing support for differentiated service guarantees and precise control.

[0019] Improve fault tracing efficiency and shorten response time. This system relies on a global topology to build an intelligent tracing mechanism, replacing the traditional method of manual, device-by-device troubleshooting. When an anomaly is detected, it can quickly trace back along the entire topology to the specific problematic link, identify the root cause of the fault, significantly shorten fault location time, avoid delays in response, effectively improve the operational stability of in-flight internet, and reduce the adverse effects of service interruptions.

[0020] This system achieves dynamic policy adaptation, balancing service assurance and resource utilization. It abandons the existing static preset policy model and can dynamically adjust network service policies based on changes in flight phases, link bandwidth fluctuations, and passenger access demands. Through dynamic policy coordination, it ensures the stable operation of key services such as high-definition video and instant messaging while optimizing network resource allocation, avoiding resource waste, and achieving a precise balance between differentiated service assurance and efficient resource utilization.

[0021] This system achieves visualization and operational integration, reducing operational complexity. It constructs a full-link integrated visualization interface that centrally displays core data such as operational status, alarm information, and policy execution effects, replacing the existing data listing mode. Simultaneously, the interface operation is deeply integrated with the underlying data model, allowing operations personnel to directly perform operations such as policy adjustments and fault handling through the interface, significantly reducing operational complexity and improving operational efficiency. Attached Figure Description

[0022] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0023] Figure 1 This is a flowchart illustrating an integrated monitoring system for the operational status of aviation internet based on multi-source data fusion, according to the present invention. Figure 2 This is a diagram of the unified dynamic model architecture of the present invention. Detailed Implementation

[0024] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0025] Please see Figure 1-2 An integrated monitoring system for aviation internet operation status based on multi-source data fusion includes: An integrated monitoring system for aviation internet operation status based on multi-source data fusion includes: The multi-source data fusion and unified service module acquires and merges heterogeneous data from airborne network equipment, satellite communication links, passenger real-name authentication systems, and flight operation databases; dynamically updates the end-to-end network service chain of satellite-airborne-cabin with flight as the core correlation dimension, and constructs a unified dynamic model; the unified dynamic model correlates and maps flight attributes, physical equipment status, IP address resources, passenger session information, and network policies. The service chain operation status monitoring and tracing module, based on the unified dynamic model, calculates and monitors traffic, connection stability and service quality indicators at the flight level, device level and passenger IP level, verifies the compliance of passenger real-name authentication; and when an anomaly is detected, it traces the source to the specific problem link according to the unified dynamic model and generates a comprehensive alarm containing root cause location information. The adaptive dynamic policy coordination module dynamically generates network service policies based on predefined protection rules and the received comprehensive alarms; and performs real-time arbitration and fusion of the network service policies and pre-configured policies to output a unified policy instruction set. The unified operational status visualization module integrates and displays the unified dynamic model, real-time operational status, comprehensive alarms, policy distribution and execution effects, and provides interfaces for policy adjustment, resource isolation or service recovery based on the unified dynamic model and alarms.

[0026] This embodiment takes the CA1326 flight from Shenzhen to Beijing operated by an A330 aircraft of a certain airline as the application object. The system adopts a distributed deployment architecture of "airborne terminal + ground terminal" and realizes data interaction between the two ends through Ku-band high-throughput satellite link.

[0027] (I) Multi-source data fusion and unified service module 1. Heterogeneous data acquisition technology: Airborne network equipment data: The operating data of the airborne router and cabin AP are collected via SNMPv3 protocol (collection period 10s). Specific parameters include interface traffic, CPU load, memory usage, number of connected users, and the data format is JSON. Satellite communication link data: Real-time data collection via the northbound RESTful API of the satellite modem (collection period 5 seconds) to obtain link bandwidth utilization, packet loss rate, and signal strength (-75dBm~-90dBm is within the normal range). Passenger real-name authentication data: When passengers connect to Wi-Fi, they submit their ID number and name via an HTTPS POST request. The system calls the airline's passenger information system interface to complete real-time verification, and the verification result is synchronized to this module. Flight operation data: retrieved every 30 seconds from the Flight Operations Control (FOC) via FTP protocol, including flight number, real-time location, flight altitude, flight phase (takeoff / level flight / landing), etc.

[0028] 2. Data Fusion Technology: Employing an "ETL cleaning + flight-dimensional correlation fusion" strategy, implemented using the Apache Flink streaming engine: Cleaning phase: Remove null values ​​and outliers (such as abnormal data with a link packet loss rate > 50%), and unify the data timestamp format (UTC+8). Integration phase: Using "flight number + scheduled departure date" as the core association key, establish data mapping relationships, for example, associate the number of access users of cabin AP-03 with the seat area of ​​rows 32-56 of the corresponding flight; Dynamic weighted arbitration: Maintain a confidence weight table for each data source (initial weights: satellite link 0.35, airborne equipment 0.3, real-name system 0.2, FOC 0.15), and dynamically adjust it according to the real-time packet loss rate. When data conflicts occur (such as the deviation between the number of AP access users and the number of real-name verified users > 10%), calculate the fusion result by weighting according to the weights.

[0029] 3. Unified Dynamic Model Construction: Utilizing the Neo4j graph database for storage, a four-level node topology of "Flight-Equipment-IP-Policy" is constructed. Flight node: Attributes include flight number, real-time location, and flight phase; Device node: Attributes include device ID, type, IP, running status, and update timestamp; IP Node: Attributes include IP address, allocation time, associated passenger information, and real-time traffic; Policy node: Attributes include policy ID, type, and scope of application; Node association: Nodes are associated through relationships such as "attribute", "bind", and "application" (e.g., flight-attribute-device, device-bind-IP, IP-application-policy). Updates are made using an event-driven mechanism, and associated nodes are refreshed synchronously when the device status changes.

[0030] (ii) Service chain operation status monitoring and traceability module 1. Calculation and monitoring of three-level indicators: Flight-level metrics: Based on a 5-minute sliding window aggregation calculation, total traffic = traffic from all device interfaces + passenger IP session traffic - redundant forwarding traffic; average connection rate = total traffic ÷ window duration ÷ number of active sessions; network service availability = (total uptime - downtime) ÷ total uptime × 100%. Device-level metrics: Data collected and analyzed via SNMP, the onboard router CPU load threshold is ≤70%, memory utilization is ≤80%, the cabin AP signal strength threshold is ≥-75dBm, and the number of connected users is ≤30 people / AP; Passenger IP-level metrics: Using DPI deep packet inspection technology (sampling interval 1s), we analyze the peak bandwidth of a single IP (30Mbps threshold for economy class) and cumulative throughput, and identify access behavior (e.g., video service corresponding to port 443) through DNS query records and port numbers.

[0031] 2. Real-name authentication compliance verification: Integrates airline real-name database, verification rules: ID number format verification (18-digit regular expression matching) + name and ID number consistency verification. Users who fail the verification or are not authenticated are marked as "illegal access" and a security alarm is generated (alarm level: medium).

[0032] 3. Anomaly tracing technology: Topology reverse tracing based on a unified dynamic model: Example: After detecting a packet loss rate of 15% (threshold ≤ 2%) for IP 192.168.1.45 (passenger seat 42A), an alarm was triggered. Through Cypher query, the access AP-04, uplink router IR8340, and satellite link were located in sequence, and the AP-04 interface failure (root cause) was finally found. The source tracing time was ≤ 2 minutes.

[0033] (III) Adaptive Dynamic Strategy Coordination Module 1. Predefined guarantee rules: written in a declarative policy language, in the format of "when [trigger condition], ensure that [target] meets [constraint]", for example: "when the flight is in cruising phase and the economy class IP bandwidth is >30Mbps, ensure that the IP bandwidth is ≤30Mbps, constraint: does not affect first class service".

[0034] 2. Dynamic policy generation: Based on the Drools rule engine, after receiving the above AP-04 fault alarm, it matches the emergency policy library and generates two policy instructions: "Migrating user traffic from AP-04 to AP-03 and AP-05" and "Restarting AP-04".

[0035] 3. Strategy Arbitration and Integration: Arbitration is carried out according to the priority of "Security Strategy > VIP Strategy > Cabin Strategy > Global Strategy". A weighted multi-objective optimization algorithm is used (optimization objectives: satisfaction 0.4, conflict 0.3, resource utilization 0.3). After integration, a unified instruction set is output and converted into configuration commands that can be executed by airborne equipment.

[0036] (iv) Unified Operation Status Visualization Module 1. Integrated Display Technology: Utilizing a B / S architecture, the front-end is developed based on Vue3 + ECharts, and is divided into 4 main display areas: Global Situation Zone: WebGIS displays the real-time location of flights and the coverage area of ​​satellite links; Topology monitoring area: Visualizes the topology of "satellite-airborne-cabin" equipment, with node colors indicating status (green for normal, yellow for alarm, and red for fault). Alarm list area: Displays alarm information, root cause, and scope of impact by level; Strategy Effect Area: The pie chart shows the percentage of each strategy that is effective, and the line chart shows the bandwidth changes before and after strategy execution.

[0037] 2. Operation Interface Implementation: RESTful API mapping is adopted. After the user clicks the "AP-04 Restart" button, the system automatically assembles SSH commands, sends them through a secure channel, and the operation results are fed back to the interface in real time.

[0038] System operation process example 1. One hour before flight departure: The system loads the basic data of flight CA1326, builds an initial unified dynamic model, and pre-configures the support rules; 2. Level flight phase (altitude 10668m): The data fusion module updates the model in real time, and the monitoring module detects an abnormal packet loss rate of AP-04, generates an alarm, and completes the source tracing. 3. Policy Coordination: Generate and integrate traffic migration and device restart policies, and distribute them to the onboard unit for execution; 4. Operation and maintenance intervention: Ground operation and maintenance personnel confirmed the effect of the policy execution through the visual interface. After AP-04 was restarted, the fault was resolved, and the packet loss rate of passenger IP 192.168.1.45 returned to 0.8%.

[0039] Specifically, the implementation method of using flights as the core association dimension includes: maintaining a dynamic data source confidence weight table for each flight; dynamically calculating and adjusting the weight of each data source in data fusion based on the real-time packet loss rate, historical accuracy, and relevance to the current flight phase (takeoff / climb, level flight, landing / approach) of the data reported by each data source; and determining the fusion result by performing weighted arbitration based on the data source confidence weight table when there are inconsistencies in the data from different data sources for the same associated object.

[0040] Specifically, the unified dynamic model includes: a business state graph carrying timestamp attributes in the storage layer; in the business state graph, the edge weights are dynamically calculated based on the real-time topology and traffic data to characterize the propagation intensity of business influence between nodes; when the node state in the unified dynamic model is updated, the associated nodes are traversed along the business state graph to calculate the quantitative impact value of the state change on the associated business indicators.

[0041] This embodiment takes a domestic flight operated by a certain airline's passenger aircraft as the application object. The unified dynamic model is deployed on the airborne core processing unit (carrying graph database). The core is adapted to the dynamic operation characteristics of the "satellite-airborne-cabin" end-to-end network service chain. The implementation details of the model's core features are described in detail below.

[0042] I. Basic Model Configuration and Business State Diagram Construction In this embodiment, the unified dynamic model is represented in the storage layer as a business state graph carrying timestamp attributes. The definitions of nodes and edges in the graph strictly match the core elements of aviation internet operation. The specific construction rules are as follows: Node Definitions: This includes four core node types, all carrying an update_ts (update timestamp) attribute: ① Flight Node: Attributes include the flight's unique identifier, real-time flight stage (e.g., cruising altitude), and current altitude (e.g., standard cruising altitude). The update_ts is updated synchronously with flight operation data. ② Device Node: Covers onboard routers, cabin APs, satellite modems, etc. Attributes include device type, IP address, and operating status. The update_ts is updated synchronously with the operational data collection cycle. ③ IP Node: Corresponds to the passenger access terminal IP. Attributes include allocation time, associated seat information, and real-time bandwidth. The update_ts is dynamically updated with traffic sampling. ④ Policy Node: Such as cabin speed limit policies. Attributes include the effective scope and bandwidth threshold. The update_ts is the policy issuance time.

[0043] Edge definition: Directed edges are used to represent the business relationships between nodes. Edge types include "Origin" (flight-device), "Access" (device-IP), "Application" (IP-policy), and "Forwarding" (device-device). Each edge carries a weight and update_ts attribute. The initial edge weight is preset based on historical operating data and is dynamically adjusted according to real-time data. The initial business state graph is constructed using graph database query statements.

[0044] II. Implementation of Dynamic Calculation of Edge Weights In this embodiment, edge weights are used to characterize the propagation strength of business impact between nodes. The calculation is based on real-time topology stability and traffic data, and is dynamically updated using a weighted summation algorithm. The core formula is: Edge weight W = α × topology stability coefficient S + β × flow proportion coefficient T Wherein, α and β are weighting coefficients that can be dynamically adjusted according to the flight phase, and the specific calculation process is as follows: Topology stability coefficient S: Calculated based on the historical connectivity and real-time packet loss rate of the links between nodes, quantifying the stability of the links using a preset formula. Taking the "forwarding" edge of "airborne router → cabin AP" as an example, the S value is calculated by combining the historical connectivity and real-time transmission quality data of the link.

[0045] Traffic proportion coefficient T: The proportion of the traffic corresponding to the current edge to the total output traffic of the upstream node, reflecting the weight of the link in the data transmission between upstream and downstream nodes.

[0046] Edge weight calculation: Substitute the values ​​into the formula to calculate the edge weights, and retain the preset precision in the final result. The system collects real-time data at fixed intervals and updates the edge weights and the update_ts attribute synchronously.

[0047] The weight ranges for different types of edges are set differently: the weight range for "Home" edges is high (representing the strong correlation between flights and equipment), the weight range for "Access" edges is medium (representing the moderate correlation between equipment and IP), and the weight range for "Forward" edges is a dynamically fluctuating range (adjusted according to traffic changes).

[0048] III. Calculation of the Correlation Quantification Impact of Node State Updates When the state of any node in the unified dynamic model is updated, the system traverses all associated nodes along the directed edges of the business state graph, and calculates the quantitative impact of the state change on the associated business indicators based on the edge weights. This embodiment takes "increased cabin AP load" as an example to break down the specific implementation process: Node status update trigger: The system collects data through the data acquisition protocol and finds that the load of a certain cabin AP has increased (exceeding the preset threshold). The status is updated to "high load". The load attribute and update_ts attribute of the AP node are updated synchronously, triggering the related impact calculation process.

[0049] Related node traversal: Along the directed edges of the business state graph, traverse upstream the related airborne routers ("forwarding" edges, with corresponding weights) and flight nodes ("home" edges, with corresponding weights), and traverse downstream the related IP nodes ("access" edges, with corresponding weights).

[0050] Quantitative impact value calculation: The core logic is "state change magnitude × edge weight," calculated based on node type. Impact on downstream IP nodes: The quantitative impact value is calculated based on the increase in AP load and the weight of the "access" edge, which leads to a decrease in the real-time bandwidth of IP nodes. The system then updates the bandwidth attributes of the IP nodes synchronously.

[0051] Impact on upstream airborne routers: The quantitative impact value is calculated based on the increase in AP load and the weight of the "forwarding" side, which leads to a decrease in the transmission quality of the interface connected to the AP by the router.

[0052] Impact on flight nodes: The quantitative impact value is calculated based on the increase in AP load and the weight of the "home" edge, which leads to a decrease in the availability index of flight-level cabin network services. If the preset threshold is not exceeded, no emergency alarm will be triggered.

[0053] Impact Result Storage and Application: All quantified impact values ​​are associated with the impact_value attribute of the corresponding node, and historical version snapshots (including the status before and after the update, impact value, and timestamp) are also retained; if the impact value exceeds the preset threshold, the service chain operation status monitoring and tracing module will generate an alarm.

[0054] This embodiment verifies the core characteristics of the unified dynamic model through specific business state graph construction, edge weight calculation logic, and node update linkage process. It realizes the dynamic correlation representation and quantitative impact assessment of the aviation Internet operation status, providing accurate model support for subsequent anomaly tracing and strategy coordination.

[0055] Specifically, the satellite-airborne-cabin end-to-end network service chain is represented in the unified dynamic model as a state association topology containing multiple logical segments. The logical segments include at least: a satellite communication segment, an airborne network equipment segment, a cabin wireless access point segment, and a passenger terminal access session segment. The state parameters of each segment are associated, stored, and updated in the unified dynamic model.

[0056] Specifically, the calculation and monitoring of traffic, connection stability, and service quality indicators at the flight level, device level, and passenger IP level are implemented as follows: a hierarchical time window comparison mechanism is adopted to maintain a second-level sliding window, a minute-level aggregation window, and a flight-level benchmark window for each level of indicator; the indicator value of the current second-level sliding window is compared with the statistical baseline of the corresponding minute-level aggregation window; the judgment threshold is dynamically corrected by combining the historical average and fluctuation range of the flight-level benchmark window; if the corrected indicator value exceeds the threshold range, it is judged as abnormal.

[0057] This embodiment uses domestic routes operated by a certain airline's passenger aircraft as the application object. It employs a layered time window comparison mechanism to dynamically calculate and monitor traffic, connection stability, and service quality indicators at the flight, equipment, and passenger IP levels. The system deploys a window management engine in the onboard core processing unit to uniformly schedule the creation, updating, and data comparison of second-level sliding windows, minute-level aggregation windows, and flight-level baseline windows. The implementation process of each level of indicator is described in detail below.

[0058] I. Overall Deployment of Layered Time Windows In this embodiment, the hierarchical time window comparison mechanism adopts a three-level collaborative architecture of "real-time acquisition - mid-term aggregation - historical benchmark". The core positioning and association logic of each window is as follows: Second-level sliding window: Independent second-level sliding windows are configured for flight-level, equipment-level, and passenger IP-level indicators to collect raw operational data in real time. The windows slide and update according to a preset second-level step size to ensure the real-time nature of the data and retain the raw data within the latest preset time period. Minute-level aggregation window: Corresponds one-to-one with the second-level sliding window. It aggregates and calculates the raw data of the corresponding second-level window according to the preset minute-level cycle to generate a statistical baseline (including statistics such as mean and variance) as an intermediate reference standard for real-time data. Flight-level benchmark window: Stores historical indicator data for the same period of the flight (such as the same route and the same flight phase), generates historical averages and fluctuation ranges, and serves as a historical reference for dynamic threshold correction. The window data is updated synchronously as the flight phase progresses.

[0059] Each window uses timestamps to align data, ensuring that indicator data at the same time can be compared across windows; the window management engine monitors the running status of each window in real time to ensure the temporal consistency of data collection, aggregation, and comparison.

[0060] II. Implementation of Layered Window Comparison for Indicators at Each Level (a) Monitoring of flight-level indicators Flight-level metrics focus on the overall operational status of the entire flight journey, with core monitoring indicators in three categories: traffic, connection stability, and service quality. The specific application process of the tiered window is as follows: Second-level sliding window acquisition: Real-time acquisition of raw data such as traffic data, connection establishment success rate, and data transmission latency of the entire flight link. The window slides in second-level steps and continuously retains the latest raw data for the preset duration, providing basic data support for real-time monitoring. Minute-level aggregation window calculation: After each preset minute period, the raw data within the second-level window is aggregated to generate a minute-level statistical baseline, such as calculating the average traffic, average connection stability, and service quality compliance rate within the preset minutes. Flight-level benchmark window adaptation: Extract flight-level indicator data from the same historical period (such as the same level flight phase) of the flight, generate historical averages and fluctuation ranges, and use them as a reference for the current threshold correction. Comparison and threshold correction: The real-time indicator value of the current second-level sliding window is compared with the statistical baseline of the corresponding minute-level aggregation window. Combined with the historical average and fluctuation range of the flight-level benchmark window, the anomaly judgment threshold is dynamically corrected (e.g., the threshold range is appropriately expanded when the historical fluctuation is large, and the threshold range is narrowed when the fluctuation is small). Anomaly detection: If the corrected real-time indicator value exceeds the threshold range, it is determined to be an abnormal flight-level indicator, triggering the corresponding level of alarm prompt.

[0061] (II) Equipment-level indicator monitoring Device-level metrics target core devices such as airborne routers, cabin access points, and satellite modems, focusing on traffic, connection stability, and quality of service indicators related to device operation (such as device port traffic, access user connection stability, and device forwarding latency). The tiered window comparison process is as follows: Second-level sliding window acquisition: Each core device is configured with an independent second-level sliding window to collect raw data of various operating indicators of the device in real time. The window slides in second-level steps to ensure real-time capture of changes in the device's operating status. Minute-level aggregation window calculation: According to a preset minute cycle, the raw data within the second-level window of a single device is aggregated to generate the minute-level statistical baseline of the device, such as the average port traffic, average connection stability, and average forwarding latency of the device. Flight-level benchmark window adaptation: Extract the operation index data of similar equipment in the same historical period of the flight, generate historical average and fluctuation range, and combine them with the performance parameters of the equipment itself to provide a reference for threshold correction; Comparison and threshold correction: The real-time indicator value of the device's current second-level window is compared with its own minute-level statistical baseline, and the threshold is dynamically corrected in combination with historical reference data of the flight-level benchmark window to ensure that the threshold is adapted to the device's operating characteristics and flight scenarios; Anomaly detection: If the corrected real-time indicator value exceeds the threshold range, it is determined to be a device-level anomaly. The abnormal device and corresponding indicator are accurately located, and a device-level alarm is generated.

[0062] (III) Monitoring of Passenger IP-level Indicators Passenger IP-level metrics focus on the access experience of a single passenger terminal, with core monitoring of terminal IP traffic, connection stability, and service quality indicators (such as real-time traffic per IP, connection interruption frequency, and service access latency). The tiered window comparison process is as follows: Second-level sliding window data acquisition: Each passenger terminal IP connected to the cabin network is configured with an independent second-level sliding window to collect raw data of various access indicators of a single IP in real time. The window slides in second-level steps to track the terminal access status in real time. Minute-level aggregation window calculation: According to a preset minute cycle, the raw data within a second-level window of a single IP is aggregated to generate a minute-level statistical baseline for that IP, such as average traffic, average connection stability, and average service access latency of a single IP. Flight-level baseline window adaptation: Extract historical access index data of terminal IPs of passengers in the same cabin class on the flight, generate the average mean and fluctuation range of the same cabin class, and use it as a reference for threshold correction; Comparison and threshold correction: The real-time indicator value of the current single IP second-level window is compared with its own minute-level statistical baseline, and the threshold is dynamically corrected in combination with the historical reference data of the same cabin class at the flight level, taking into account individual access differences and the overall service standards of the cabin class. Anomaly detection: If the corrected real-time indicator value exceeds the threshold range, it is determined that the passenger's IP-level indicator is abnormal, and a terminal-level alarm is generated to provide a basis for precise operation and maintenance intervention.

[0063] III. Example of a Complete Anomaly Detection Process Taking "abnormal cabin AP equipment-level connection stability index" as an example, the complete anomaly judgment process of layered time window comparison is demonstrated: A sliding window, updated in seconds, collects raw data on the connection stability of a cabin AP in real time, with the window continuously updating in seconds-level increments. Once the preset minute cycle is reached, the minute-level aggregation window aggregates the raw data within the second-level window of the AP to generate a minute-level connection stability statistical baseline. The flight-level baseline window extracts the connection stability data of the same type of AP during the historical level flight phase of the flight, and generates the historical average and fluctuation range. The system compares the real-time connection stability value of the current AP second-level window with the minute-level statistical baseline, finds that the real-time value deviates from the baseline, and then dynamically corrects the connection stability judgment threshold by combining the historical average and fluctuation range of the flight-level benchmark window. The system determined that the corrected real-time connection stability value exceeded the threshold range, and therefore identified the cabin AP connection stability as abnormal, generating a device-level alarm and simultaneously pushing it to the unified operational status visualization module.

[0064] This embodiment achieves accurate calculation and dynamic monitoring of three-level indicators through the coordinated operation of layered time windows. It not only ensures the rapid capture of real-time anomalies, but also improves the accuracy of anomaly judgment and scenario adaptability by correcting thresholds through historical benchmark data, thus providing reliable monitoring support for the stable operation of aviation internet.

[0065] Specifically, the verification of passenger real-name authentication compliance includes: verifying the identity information provided by the passenger when accessing the network against the real-name database in real time, marking devices that fail to verify, have abnormal information, or are unauthenticated, and generating corresponding security alarm events. The security alarm events will trigger the adaptive dynamic policy collaboration module to execute network access interception actions.

[0066] Specifically, the method of tracing back to specific problem links based on the unified dynamic model includes: when an alarm is generated for a monitoring anomaly, the alarm object is traced back along the state association topology according to the mapping relationship in the unified dynamic model, and the passenger session, cabin wireless access point, uplink airborne network equipment and satellite communication link associated with the alarm object are located in sequence.

[0067] Specifically, the predefined guarantee rules include: the core expression form is to ensure that the service target meets the constraint conditions when the trigger condition is met; the adaptive dynamic policy coordination module has a built-in policy compiler that combines the declarative guarantee rules with the real-time running status data in the unified dynamic model to compile and generate a sequence of configuration instructions that can be executed on specific airborne network devices or ground control devices.

[0068] This embodiment uses an in-flight internet scenario operated by a certain airline's passenger aircraft as the application object. The predefined guarantee rules built into the adaptive dynamic policy coordination module are written in a declarative policy language, and the core follows the expression form of "when [trigger condition], ensure that [service target] meets [constraint condition]". The policy compiler parses these rules and associates them with real-time running status data in the unified dynamic model to complete the transformation from abstract rules to a sequence of executable configuration instructions for the device. The implementation process of each core component is described in detail below.

[0069] I. Design and Classification of Predefined Guarantee Rules In this embodiment, the predefined assurance rules are designed around the three core requirements of aviation internet: "service quality assurance, equipment operation safety, and compliance management." All rules are expressed in a unified declarative format, facilitating parsing and dynamic adaptation by the policy compiler. Typical rule examples and classifications are as follows: (a) Rules for Service Quality Assurance Rule 1: When the number of passenger terminals accessing a certain cabin class reaches a preset density threshold and the service quality index of the corresponding cabin wireless access device is lower than the benchmark value, ensure that the core network service access latency of the passenger terminal in that cabin class meets the preset latency threshold. The constraint is that the service quality standards already in effect for other cabin classes are not reduced.

[0070] Rule 2: When [the flight enters the cruising phase and the satellite communication link bandwidth is sufficient], ensure that [the high-definition video service bandwidth for high-priority passengers] meets [the preset guaranteed bandwidth], with the constraint that [the occupied bandwidth does not exceed the preset proportion of the total available bandwidth of the link].

[0071] (ii) Equipment operation safety rules Rule 3: When the CPU load or memory usage of the airborne core network equipment continuously exceeds the preset safety threshold, ensure that the operating load of the equipment falls back to within the safety threshold. The constraint is to prioritize ensuring that the transmission link of flight-related critical business data is not interrupted.

[0072] (III) Compliance and Control Rules Rule 4: When [a passenger terminal attempts to access the cabin network before completing real-name authentication], ensure that [the terminal's network access permissions] meet the condition of [only opening the real-name authentication service interface and restricting access to other network resources], with the constraint that [the access bandwidth of the real-name authentication interface is not restricted, ensuring a smooth authentication process].

[0073] All of the above rules are stored in the system's rule base, and can be dynamically invoked and prioritized according to flight type and flight phase. Parameters such as "preset threshold" and "benchmark value" in the rules can be linked to the historical operation data of the unified dynamic model for adaptive calibration.

[0074] II. Core Workflow of the Strategy Compiler The policy compiler is the core component that connects declarative protection rules with device-executable instructions. Its core logic is "rule parsing - real-time data association - instruction compilation - format adaptation," and the specific workflow is as follows: (a) Rule parsing phase After receiving predefined guarantee rules from the rule base, the policy compiler uses a parser to break down the core elements of the rules, namely "triggering condition," "service goal," and "constraint," and constructs an element mapping table. For example, when parsing rule 1, it breaks it down into: triggering condition "dense passenger access in cabin class + substandard service quality of access equipment," service goal "passenger access latency to core services in cabin class," and constraint condition "do not reduce the service quality of other cabin classes." At the same time, it marks the parameter types corresponding to each element (such as access quantity, service quality index, latency threshold, etc.).

[0075] (II) Real-time data association stage The compiler queries real-time runtime status data in the unified dynamic model through a data interface to complete the association and matching between parsed elements and actual runtime data. Specifically, this includes: Retrieve data related to trigger conditions: such as querying the current number of passengers accessing each cabin class and the real-time service quality indicators (access latency, packet loss rate, etc.) of the corresponding cabin wireless access devices. Obtain data related to constraints: such as querying the actual values ​​of current service quality indicators for other cabin classes, to ensure that the subsequently generated strategies do not exceed constraints; Obtain device-related data: such as querying the device type, communication protocol, and management address of the target cabin wireless access device, to provide a basis for command format adaptation.

[0076] If the real-time data in the unified dynamic model does not meet the triggering conditions, the compiler terminates the compilation process of the current rule and returns a "rule not triggered" status; if the triggering conditions are met, the compiler enters the instruction compilation stage.

[0077] (III) Instruction compilation and format adaptation stage Based on the parsed rule elements and associated real-time data, the compiler generates initial policy instructions in conjunction with a pre-defined device instruction template library. Then, it adapts and optimizes these instructions according to the protocol type and instruction format requirements of the target device, and finally outputs a sequence of directly executable configuration instructions.

[0078] The instruction template library covers instruction format specifications for mainstream airborne network equipment (such as cabin wireless access equipment and airborne routers) and ground control equipment, and supports dynamic template matching based on equipment type. Examples include bandwidth control instruction templates for cabin wireless access equipment and access control list configuration templates for airborne routers.

[0079] III. Complete Compilation Process Example Taking the compilation and execution of rule 1 as an example, the complete process from rule triggering to instruction generation is fully demonstrated: Rule Invocation and Parsing: The system invokes Rule 1 based on the current flight operation phase; the policy compiler parses the rule to obtain the triggering conditions, service objectives, and constraints, and marks parameters such as the number of accesses to be associated, service quality, and latency threshold. Real-time data association: The compiler queries the unified dynamic model to confirm that the number of passengers accessing a certain cabin class has reached the preset dense threshold, and the access latency of the corresponding cabin wireless access device has exceeded the benchmark value. At the same time, it obtains that the current service quality indicators of other cabin classes are all in compliance status. Initial instruction generation: Based on service quality assurance requirements, generate initial policy instructions to "optimize channel allocation for cabin radio access equipment and limit bandwidth for non-core services"; Format adaptation: Based on the communication protocol of the target cabin wireless access device, the initial command is adapted to the configuration command format supported by the device, and a command sequence (including channel adjustment command, bandwidth limitation command, command execution order identifier, etc.) is generated. Command output: The compiler outputs the adapted configuration command sequence to the adaptive dynamic policy coordination module, which then sends it to the target airborne network device for execution via a secure communication channel.

[0080] This embodiment verifies the feasibility of predefined protection rules and the core functionality of the policy compiler by visualizing the design of declarative protection rules and deconstructing the policy compilation process. This design achieves precise adaptation between abstract rules and actual device operation. Simultaneously, relying on real-time data association through a unified dynamic model, it ensures that the generated configuration instruction sequence has scenario adaptability and execution effectiveness, providing reliable technical support for dynamic policy management of aviation internet.

[0081] Specifically, the dynamically generated network service strategy includes: after receiving the comprehensive alarm, the adaptive dynamic strategy coordination module matches and instantiates corresponding intervention strategy instructions from the preset emergency response strategy library according to the alarm type and root cause information. The intervention strategy instructions include, but are not limited to, rate limiting or blocking instructions for specific IPs, service restart suggestion instructions for specific devices, or global traffic shaping instructions for link quality.

[0082] Specifically, the real-time arbitration and fusion of network service policies and pre-configured policies includes: the core is to solve a constrained multi-objective optimization problem; wherein, the optimization objectives include maximizing passenger service satisfaction, minimizing the number of policy conflicts, and balancing network resource utilization; the constraints are the upper limit of the physical resources currently available to the network and inviolable security policies (including passenger real-name authentication compliance policies and malicious access blocking policies).

[0083] Specifically, the issuance of the unified policy instruction set is implemented as follows: a tiered activation and monitoring rollback strategy is adopted. For non-critical policy changes, the unified policy instruction set is divided into multiple batches and issued sequentially in a preset order. After each batch of instructions is issued, a preset monitoring period is waited for confirmation that the service chain operation status is stable or abnormal indicators have improved before the next batch of instructions is issued. If the service chain status is detected to be deteriorating, the issuance of subsequent batches of instructions is automatically stopped, and the rollback operation of the issued instructions is performed.

[0084] This embodiment uses the cabin network service optimization scenario of a passenger flight operated by a certain airline as the application object. For non-critical policy changes (such as cabin wireless access equipment channel optimization, dynamic adjustment of passenger terminal bandwidth, etc.), the issuance of the unified policy instruction set adopts a hierarchical activation and monitoring rollback strategy. Through the closed-loop logic of "batch splitting - sequential issuance - monitoring verification - stable re-issuance / deterioration rollback", the stability of the service chain operation is ensured during the policy change process. The implementation process of each core link is described in detail below.

[0085] I. Overall Architecture for Tiered Activation and Monitoring Rollback In this embodiment, the unified policy instruction set distribution system consists of an instruction splitting module, a batch scheduling module, a status monitoring module, and a rollback execution module. These modules work together to implement tiered activation and monitoring rollback functions. Instruction splitting module: For the unified policy instruction set corresponding to non-critical policy changes, it is split into multiple independent batches according to the principle of "from small to large impact scope and from low to high correlation". Each batch contains a set of logically related policy instructions, and the preset issuance order of each batch is marked. Batch scheduling module: According to the preset order, it sends batch instructions to airborne network equipment or ground control equipment through a secure communication link, and records the issuance time, target equipment and other information of each batch instruction simultaneously; Status monitoring module: After each batch of instructions is issued, a preset monitoring cycle is started to collect service chain operation status data in real time (such as equipment operating load, passenger terminal connection stability, network service quality indicators, etc.) and compare it with the baseline status before the instructions were issued; Rollback Execution Module: Pre-stores the reverse recovery instructions corresponding to each batch of instructions. If the service chain status is detected to deteriorate, it immediately triggers the subsequent batch issuance abort process and executes the rollback operation of the issued instructions to restore the stable state before the policy change.

[0086] II. Implementation Details of Core Components (I) Principles of Strategy Instruction Set Splitting and Batch Division For non-critical policy changes, the instruction splitting module follows the core principle of "minimal impact and gradual implementation" to divide the changes into batches. The specific splitting logic is as follows: Prioritize breaking down instructions that have a localized impact: Instructions that only affect a single cabin area, a single type of equipment, or a small group of passengers are grouped into earlier batches to avoid a chain reaction caused by large-scale instruction changes; Grouping by business relevance: Instructions with unrelated business logic are grouped into different batches to reduce the impact of a single batch of instruction execution failures spreading. Mark batch dependencies: If subsequent batch instructions need to be based on the execution results of previous batch instructions, then the dependency order between batches should be clearly marked to ensure the logical continuity of the delivery.

[0087] Example: For non-critical policy changes such as "cabin network channel optimization + dynamic adjustment of bandwidth for economy class passengers", the instruction set is split into 3 batches: Batch 1 is "AP channel adjustment instruction for a single cabin area", Batch 2 is "AP channel adjustment instruction for other cabin areas", and Batch 3 is "bandwidth adjustment instruction for economy class passenger terminals". The preset issuance order is Batch 1 → Batch 2 → Batch 3.

[0088] (II) Batch Issuance and Monitoring Verification Process The core process for tiered activation is "single batch distribution - monitoring and verification - stable re-deployment", specifically implemented as follows: The batch scheduling module sends batch 1 instructions (single cabin area AP channel adjustment instructions) to the target equipment in a preset order, and records the "successful sending" status after the instructions are sent. The preset monitoring cycle is started, and the status monitoring module collects the operating status (such as channel occupancy rate and equipment load), connection stability (such as connection interruption frequency) and service quality (such as access latency) of the AP in the cabin area in real time. After the monitoring period ends, the system automatically compares the monitoring indicators with the baseline values ​​before the instruction is issued: if there are no abnormalities in the indicators, the service chain is running stably, or the abnormal indicators (such as excessive latency caused by the original channel congestion) are improved, then the system determines that "verification is passed" and the batch scheduling module triggers the issuance of batch 2 instructions. Repeat the above process to complete the issuance and monitoring verification of batch 2 (AP channel adjustment instructions for other cabin areas). After successful verification, trigger the issuance of batch 3 (bandwidth adjustment instructions for economy class passenger terminals).

[0089] (III) Deterioration of Status and Implementation of Rollback Operation The core of the rollback strategy is "real-time monitoring - rapid abort - precise recovery", which is implemented as follows: Deterioration criteria: Predefined multi-dimensional thresholds. If any of the following conditions occur during the monitoring period, the service chain is judged to have deteriorated, including: the equipment load exceeds the safety threshold, the frequency of passenger terminal connection interruptions increases sharply, and core service quality indicators (such as access latency) decrease significantly and exceed the acceptable range. Abort and rollback trigger: If a state deterioration is detected within the monitoring period after a batch of instructions is issued (e.g., after batch 2 is issued, a sudden increase in load occurs in an AP in a cabin area), the system immediately performs two operations: First, the batch scheduling module suspends the issuance process of the subsequent batch (batch 3) to avoid the impact from spreading; second, the rollback execution module calls the reverse recovery instruction corresponding to batch 2 to restore the AP channel adjusted in that batch to its state before the change. Rollback Verification and Alarm: After the rollback operation is completed, the system restarts short-term monitoring to confirm that the service chain status has returned to the baseline stable level; at the same time, an "abnormal rollback of policy change" alarm is generated and pushed to the unified operation status visualization module to prompt operation and maintenance personnel to investigate the cause.

[0090] III. Example of a complete implementation process Taking the issuance of the "cabin network channel optimization + dynamic adjustment of bandwidth for economy class passengers" strategy instruction set as an example, the complete process of tiered activation and monitoring rollback is demonstrated. Command splitting: The policy command set is split into 3 batches, with the issuance order specified as Batch 1 (single-area AP channel adjustment) → Batch 2 (other area AP channel adjustment) → Batch 3 (economy class bandwidth adjustment). Batch 1 Issuance and Verification: Issuing Batch 1 instructions initiates the monitoring cycle, verifies that the single-area AP is operating stably and that passenger connections are normal, and triggers Batch 2 issuance; Batch 2 Issuance and Verification: Issue Batch 2 instructions. If monitoring detects a sudden increase in AP load (state deterioration) in a certain area, immediately stop Batch 3 issuance, execute Batch 2 rollback operation, and restore the AP channels in that area to their original state. Follow-up processing: After the rollback, monitoring confirmed that the status was stable, and an alarm was generated to prompt the operation and maintenance personnel to investigate the cause of the sudden increase in AP load. After the problem was resolved, the distribution of batch 2 and subsequent batches was replanned.

[0091] This embodiment achieves the safe and smooth implementation of non-critical policy changes through the implementation of tiered activation and monitoring rollback strategies. It not only ensures the gradual advancement of policy optimization, but also effectively avoids service chain operation risks that may be caused by policy changes through real-time monitoring and rapid rollback mechanisms, thereby improving the reliability and stability of aviation internet policy management.

[0092] Specifically, the unified operational status visualization module and the unified dynamic model achieve bidirectional state synchronization and operation mapping, including: Forward mapping: The user's selection, filtering, and drill-down operations in the visual interface are converted into graph query commands or attribute filtering conditions for the unified dynamic model. Reverse synchronization: The state update, alarm generation, and policy activation events in the unified dynamic model drive the style, position, or value update of the corresponding elements in the visualization interface in real time.

[0093] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. An integrated monitoring system for the operational status of an aeronautical internet based on multi-source data fusion, characterized in that, The application relates to a flight network service management system, which comprises the following modules: a multi-source data fusion and unified service module, which acquires and fuses heterogeneous data from airborne network equipment, a satellite communication link, a passenger real-name authentication system and a flight operation database; dynamically updates a satellite-airborne-cabin end-to-end network service chain with a flight as a core correlation dimension, and constructs a unified dynamic model; the unified dynamic model correlates and maps flight attributes, physical equipment states, IP address resources, passenger session information and network policies; a service chain operation state monitoring and tracing module, which calculates and monitors flight class, equipment class and passenger IP class traffic, connection stability and service quality indexes based on the unified dynamic model, and checks passenger real-name authentication compliance; when an exception is detected, the unified dynamic model is traced to a specific problem link to generate a comprehensive alarm containing root cause positioning information; an adaptive dynamic policy coordination module, which dynamically generates network service policies according to predefined guarantee rules and the received comprehensive alarm, and performs real-time arbitration and fusion on the network service policies and preconfigured policies to output a unified policy instruction set; a unified operation situation visualization module, which integrally displays the unified dynamic model, real-time operation state, comprehensive alarm, policy distribution and execution effect, and provides a policy adjustment, resource isolation or service recovery operation interface based on the unified dynamic model and alarm.

2. The system of claim 1, wherein, The specific implementation mode of the flight as a core correlation dimension comprises the following steps: a dynamic data source confidence weight table is maintained for each flight; the real-time packet loss rate, historical accuracy rate and relevance to the current flight phase of the data reported by each data source are used to dynamically calculate and adjust the weight of each data source in data fusion; when the data of different data sources for the same correlation object are inconsistent, weighted arbitration is performed based on the data source confidence weight table to determine the fusion result.

3. The system of claim 1, wherein, The unified dynamic model specifically comprises: a business state graph with a time stamp attribute in a storage layer; in the business state graph, edge weights are dynamically calculated according to real-time topology structure and traffic data, and the propagation strength of the influence of business between nodes is represented; when the node state in the unified dynamic model is updated, associated nodes are traversed along the business state graph, and the quantitative influence value of the state change on associated business indexes is calculated.

4. The system of claim 1, wherein, The satellite-airborne-cabin end-to-end network service chain is specifically represented as a state correlation topology comprising multiple logical segments in the unified dynamic model, the logical segments at least comprising a satellite communication segment, an airborne network equipment segment, a cabin wireless access point segment and a passenger terminal access session segment, and the state parameters of the segments are correlated, stored and updated in the unified dynamic model.

5. The system of claim 1, wherein, The flight class, device class and passenger IP class traffic, connection stability and service quality indicators are calculated and monitored, and the specific implementation is: a hierarchical time window comparison mechanism is used to maintain a second-level sliding window, a minute-level aggregation window and a flight class reference window for each level indicator; the indicator value of the current second-level sliding window is compared with the statistical baseline of the corresponding minute-level aggregation window; the historical mean and fluctuation range of the flight class reference window are combined to dynamically correct the judgment threshold; if the corrected indicator value exceeds the threshold range, it is determined to be abnormal.

6. The system of claim 1, wherein, The passenger real-name authentication compliance is verified, specifically including: real-time verification of the identity information provided by the passenger when accessing with the real-name database, and marking the devices with failed verification, abnormal information or unauthenticated, generating corresponding security alarm events, which will trigger the adaptive dynamic strategy coordination module to perform network access interception actions.

7. The system of claim 1, wherein, The specific problem link is traced according to the unified dynamic model, specifically including: when an alarm is generated for a monitored abnormality, the alarm object is mapped in the unified dynamic model, and the reverse tracing is performed along the state association topology to locate the passenger session, cabin wireless access point, uplink airborne network device and satellite communication link associated with the alarm object in sequence.

8. The system of claim 1, wherein, The pre-defined protection rules specifically include: the core expression form is to ensure that the service target meets the constraint condition when the trigger condition is met; the adaptive dynamic strategy coordination module combines the declarative protection rules with the real-time running state data in the unified dynamic model to compile and generate configuration instruction sequences that can be executed on specific airborne network devices or ground control devices.

9. The system of claim 1, wherein, The dynamic generation of network service strategy specifically includes: after receiving the comprehensive alarm, the adaptive dynamic strategy coordination module matches and instantiates the corresponding intervention strategy instruction from the pre-set emergency response strategy library according to the alarm type and root cause information, which includes but is not limited to speed limiting or blocking instructions for specific IPs, service restart suggestion instructions for specific devices, or global traffic shaping instructions for link quality.

10. The system of claim 1, wherein, The real-time arbitration and fusion of network service strategy and pre-configuration strategy specifically include: the core is to solve the multi-objective optimization problem with constraints; wherein, the optimization objectives include maximizing passenger service satisfaction, minimizing the number of strategy conflicts, and balancing network resource utilization; the constraint conditions are the upper limit of the current available physical resources and the inviolable safety strategy.

11. The system of claim 1, wherein, The specific implementation of the unified policy instruction set is: a hierarchical effect and monitoring rollback strategy is used, for non-critical policy changes, the unified policy instruction set is split into multiple batches, and is sequentially issued in a pre-set order; after each batch of instructions is issued, a pre-set monitoring period is waited, and after confirming that the service chain running state is stable or the abnormal indicators are improved, the next batch of instructions is triggered for issuance; if the service chain state is deteriorated, the subsequent batch of instructions is automatically stopped, and the rollback operation of the issued instructions is performed.

12. The system of claim 1, wherein, The unified running situation visualization module and the unified dynamic model realize bidirectional state synchronization and operation mapping, specifically including: Forward mapping: the user's selection, filtering, drilling operation in the visualization interface is converted into a graph query instruction or attribute filtering condition of the unified dynamic model; Reverse synchronization: the state update, alarm generation, policy taking effect event in the unified dynamic model drives the style, position or value update of the corresponding element in the visualization interface in real time.

Images