An intelligent operation and maintenance and security system for industrial switches based on edge computing
By integrating intelligent operation and maintenance and security functions into industrial switches through edge computing, the problem of collaborative linkage caused by the functional dispersion in existing technologies is solved. It realizes accurate fault location, adaptive early warning, national cryptographic encryption and wired and wireless integrated scheduling, improves the autonomous operation reliability and security of industrial networks, and adapts to the networking requirements of new industrial scenarios.
Patent Information
- Application Number
- CN202611587914.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-14
- Publication Date
- 2026-11-17
AI Technical Summary
The existing split-type and external architecture of industrial switches makes it impossible for operation and maintenance, security and scheduling capabilities to work together. Hidden cable faults are difficult to troubleshoot, high wireless traffic can easily occupy control channels, and operation and maintenance capabilities fail when the network is down. They cannot adapt to the integrated networking requirements of new industrial scenarios such as flexible production lines, machine vision, and digital twins.
The intelligent operation and maintenance and security system for industrial switches based on edge computing is divided into three layers from top to bottom: edge data acquisition layer, local edge computing engine layer, and operation and maintenance security interaction layer. It integrates intelligent operation and maintenance diagnosis, national cryptographic security, and wired and wireless unified scheduling functions on the local switch, realizing real-time multi-dimensional data acquisition, accurate fault location, adaptive early warning, national cryptographic encryption, unified traffic scheduling, and local visualized operation and maintenance.
It achieves multi-functional intrinsic integration, eliminates the performance and reliability shortcomings of external devices, ensures the autonomous operation reliability of the network in disconnection scenarios, takes into account the high real-time performance and security compliance requirements of industrial control, adapts to the wired and wireless integrated collaborative management and control of new industrial scenarios, forms a closed-loop management and control capability of operation and maintenance and security linkage, and reduces the reliance on manual on-site operation and maintenance.
Smart Images

Figure CN123123101A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of industrial internet communication and network security technology, specifically to an intelligent operation and maintenance and security system for industrial switches based on edge computing. Background Technology
[0002] With the deep integration of industrial internet and edge computing technology, functions such as real-time control, intelligent operation and maintenance, and security protection in production sites are continuously being pushed to the network edge. As the core node of edge networking, industrial Ethernet switches have evolved from single data forwarding devices to multifunctional edge intelligent nodes.
[0003] Currently, industrial sites generally adopt a split architecture of "basic industrial switch + external security gateway + cloud operation and maintenance platform": operation and maintenance diagnosis relies heavily on centralized cloud analysis, and the device can only identify obvious faults such as link connectivity on the local device; national cryptographic encryption, access control and other security functions are deployed in series through independent gateways; high-speed wireless access such as Wi-Fi 7 mostly adopts external extension solutions, and wired and wireless traffic scheduling and security management are isolated from each other.
[0004] This split-type, externally-mounted functional overlay mode has core technical defects: the dispersed functions lead to the inability to coordinate and link operation and maintenance, security and scheduling capabilities, and there are prominent problems such as difficulty in diagnosing hidden cable faults, the easy occupancy of control channels by high wireless traffic, and failure of operation and maintenance capabilities when the network is down. It cannot adapt to the integrated networking needs of new industrial scenarios such as flexible production lines, machine vision, and digital twins, and has become the core technical bottleneck restricting the intelligent upgrading of industrial switches. Summary of the Invention
[0005] To address the aforementioned technical issues, this invention provides an intelligent operation and maintenance and security system for industrial switches based on edge computing. This system is mounted on a wide-temperature fanless industrial switch hardware platform and consists of three layers from top to bottom: an edge data acquisition layer, a local edge computing engine layer, and an operation and maintenance security interaction layer.
[0006] The edge data acquisition layer uses various sensors, acquisition circuits, and protocol stacks in the switch hardware to achieve real-time acquisition of multi-dimensional data. It is used to collect network operation data, device status data, cable characteristic data, wireless access data, and security characteristic data in real time, and transmits the collected data to the local edge computing engine layer.
[0007] The local edge computing engine layer, the core processing layer of the system, is implemented based on an industrial-grade multi-core ARM processor and includes three core components: an intelligent operation and maintenance diagnostic engine, a national cryptographic integrated security engine, and a wired and wireless unified scheduling module. The intelligent operation and maintenance diagnostic engine is used to accurately locate cable faults based on collected data, adaptively classify and warn of faults, and coordinate with the industrial ring network protocol to achieve fault self-healing.
[0008] The national cryptographic integrated security engine is inherently integrated into the switch forwarding plane, eliminating the need for an external independent security gateway and avoiding additional forwarding latency. It is used to implement national cryptographic hardware encryption, access control, edge-side abnormal traffic identification, and full-process operation auditing.
[0009] The wired and wireless unified scheduling module is designed for mixed networking scenarios of Wi-Fi7 wireless terminals and wired ports. It is used to perform unified queue scheduling of traffic from wired ports and Wi-Fi7 wireless terminals to ensure priority forwarding of industrial control messages.
[0010] The operation and maintenance security interaction layer is used to provide local visual network management, encrypted remote operation and maintenance interfaces, and local data storage services.
[0011] Preferred: The specific content collected by the edge data acquisition layer includes: network operation data, device status data, cable characteristic data, wireless access data, and security characteristic data; Preferred network operation data includes: real-time traffic of 10 Gigabit ports, packet forwarding latency, packet loss rate, VLAN configuration status, ring network topology, and real-time status of each link; Preferred: Device status data includes: main switching chip temperature, CPU core temperature, power input voltage, multi-point temperature measurement data of the casing, and operating status of each hardware module; Preferred cable characteristic data includes: network cable differential impedance value, link attenuation value, and TDR reflection signal time-domain waveform acquired through dedicated hardware acquisition circuit, with sampling accuracy reaching the nanosecond level; Preferred: Wireless access data includes: number of Wi-Fi 7 terminals connected, signal strength of each terminal, real-time uplink and downlink bandwidth, terminal MAC address and authentication information; Preferred security feature data includes: source / destination IP and MAC addresses, access behavior sequence, traffic burst characteristics, and IPsec encrypted tunnel connection status.
[0012] The preferred intelligent operation and maintenance diagnostic engine includes: a cable fault precise location unit, an adaptive fault early warning unit, and a ring network self-healing coordination unit. The cable fault precise location unit injects nanosecond-level high-speed pulse signals into the network cable via hardware acquisition circuitry. When there is a break or impedance mismatch in the cable, the pulse signal is reflected, and the acquisition circuitry records the round-trip time and amplitude of the reflected signal. The adaptive fault early warning unit calculates the deviation of the sampled value from the normal baseline in real time, determines the early warning level, and links the corresponding hardware protection sidechain. The ring network self-healing coordination unit is deeply integrated with the self-developed proprietary ring network protocol. When a ring network link fault is detected, a hardware-level parallel switching mechanism is used to complete the seamless switching of the backup link within ≤5ms, ensuring uninterrupted industrial control services. Simultaneously, information such as the fault link number, switching duration, and affected port range is pushed to the operation and maintenance diagnostic engine, automatically generating fault alarms and achieving a synchronous response of "self-healing + alarm".
[0013] The preferred formula for calculating fault distance is: Where: L is the physical distance from the fault point to the switch port, in meters (m); v is the propagation speed of the electrical signal in the corresponding network cable, in meters per second (m / s); t is the round-trip time of the pulse signal from transmission to reception of the reflected signal, in seconds (s).
[0014] Preferably, fault type identification is achieved by combining impedance deviation rate with reflection characteristics. The formula for impedance deviation rate is: , where: Z 实测 The actual differential impedance value acquired by the hardware, in ohms (Ω); Z 标准 For the standard differential impedance of the corresponding network cable, the standard value for industrial Ethernet twisted-pair cables is 100Ω. Fault diagnosis rules: If the reflected signal amplitude is 100% and the measured impedance is infinite: it is determined to be a broken wire fault; if 10% < ΔZ < 50%: it is determined to be an impedance offset fault, usually due to poor connector contact or localized cable damage; if ΔZ ≤ 10% but the link attenuation exceeds the preset threshold: it is determined to be a link attenuation exceeding the standard fault, usually due to cable aging or moisture.
[0015] Preferred: The adaptive fault early warning unit employs a sliding window dynamic baseline algorithm to avoid false alarms and missed alarms caused by fixed thresholds; specifically, it includes: selecting historical sampling data from the past N hours as a sliding window, calculating the overall mean and overall standard deviation of the data within the window, and constructing a normal baseline for the parameters; and calculating the deviation of the current sampled value in real time. , where μ is the population mean within the sliding window; σ is the population standard deviation; and X is the real-time collected value; then the alarm level is determined according to the graded early warning rules.
[0016] Preferred: Tiered warning rule: When D≥3: Level 1 warning is triggered (notification of attention), the system records the operation log and does not interfere with the operation of the business; when D≥5: Level 2 alarm is triggered (serious anomaly), the system actively pushes alarm information and links the corresponding protection strategy (such as high temperature load reduction, port rate limiting, etc.); when D<3, no warning is triggered.
[0017] Preferably: the deviation of the current sampled value: Where i is the data collection number within a preset time period; I is the total number of data collections within the preset time period, i = 1, 2, ..., I; X i This refers to the data collected, numbered i; μ is the overall mean of the data within the preset time period, and t... i X0 is the time elapsed between the acquisition time of data number i and the current time point; X0 is the unit deviation of the data acquisition type.
[0018] Preferred: Unit deviation , where max(X) is the maximum value of the collected values, min(X) is the minimum value of the collected values, and η is the step value.
[0019] The preferred integrated security engine comprises: a national cryptographic hardware encryption unit, an access control and firewall unit, an edge abnormal traffic identification unit, and an operation auditing unit; the national cryptographic hardware encryption unit has built-in hardware acceleration modules for the SM1 and SM4 domestic encryption algorithms, and adopts a multi-group parallel processing architecture to achieve end-to-end encrypted data transmission; the access control and firewall unit, based on ACL rules and industrial protocol whitelists, achieves unauthorized device access interception and industrial protocol control; the edge abnormal traffic identification unit, based on a local feature library, identifies network attacks and abnormal traffic in real time, triggers alarms, and automatically blocks abnormal traffic; the operation auditing unit locally stores all configuration operations, access behaviors, and security event logs, supporting offline fault tracing and security auditing.
[0020] Preferred formula for calculating single-packet encryption latency: , where: L 报文 k is the length of a single message; k is the message priority level; w k The scheduling factor is numbered k, with the highest priority message w1>1 having exclusive access to its dedicated encryption unit and not being overtaken by lower priority messages; N k R represents the number of parallel encryption units with priority number k; 加密 The throughput of a single encryption unit.
[0021] Preferred: Total forwarding latency per packet after enabling national cryptographic encryption: T 转发α represents the inherent processing latency of the switch hardware forwarding, α is the pipeline overlap coefficient with a value range of 0 < α < 1, representing the proportion of time overlap between encryption processing and forwarding table lookup; β is the edge multitasking correction coefficient.
[0022] Preferably, the operation and maintenance security interaction layer can include: a local visual network management unit, a remote operation and maintenance interface unit, and a local storage unit. The local visual network management unit supports local access via a web interface, displaying network topology, device status, fault alarms, and security events, and supports batch parameter configuration and firmware upgrades; the remote operation and maintenance interface unit supports remote operation and maintenance access under an encrypted channel, automatically switching to a pure local operation and maintenance mode when the network connection is lost; the local storage unit has built-in industrial-grade storage media, storing no less than 180 days of operation logs, fault records, and security audit data.
[0023] Technical effects and advantages of the present invention: Addressing the architectural shortcomings of existing industrial switches with their split-type and externally mounted functionalities, the present invention adopts an edge-integrated design, possessing multi-dimensional technological advancements and significant creative advantages: First, it eliminates the performance and reliability shortcomings of external devices at the architectural level. Intelligent operation and maintenance, national cryptographic security, and wired and wireless scheduling functions are all internally integrated into the local edge computing unit of the switch, abandoning the traditional model of relying on external security gateways and cloud operation and maintenance. This fundamentally eliminates the additional forwarding latency and single point of failure risks introduced by external devices. Even in network disconnection scenarios, it can still maintain complete operation and maintenance and security protection capabilities, significantly improving the autonomous operation reliability of industrial field networks.
[0024] Second, it overcomes the industry's technical bias that security features and real-time performance are mutually exclusive. The national cryptographic encryption unit is deeply embedded in the forwarding plane. Combined with the overlapping processing mechanism of the forwarding-encryption pipeline and the hierarchical priority scheduling strategy, it can achieve full-link national cryptographic encryption and multi-level access protection without significantly affecting the forwarding latency of industrial control messages, thus balancing the high real-time requirements of industrial control with the requirements of industrial data security compliance.
[0025] Third, it enables integrated wired and wireless collaborative management and control. A unified traffic scheduling, operation and maintenance monitoring, and security management system is constructed to ensure the priority of industrial control message forwarding, while supporting stable concurrent access of high-speed wireless terminals. This addresses the problems of fragmented wired and wireless systems and the easy encroachment of high-bandwidth wireless services on control channels in existing technologies, adapting to new industrial scenarios such as machine vision and flexible production lines.
[0026] Fourth, it establishes a closed-loop management capability that integrates operation and maintenance with security. Fault diagnosis, ring network self-healing, and security protection work in tandem to construct a complete closed loop of "monitoring-early warning-response-source tracing," improving the efficiency of network anomaly response and handling, and reducing reliance on on-site manual maintenance in industrial settings. The solution adopts a modular design, adaptable to multi-port switch hardware platforms, and possesses good engineering versatility and scenario adaptability. Attached Figure Description
[0027] Figure 1 This is a structural block diagram of an intelligent operation and maintenance and security system for industrial switches based on edge computing proposed in this invention.
[0028] Figure 2 This is a schematic diagram illustrating the process of accurately locating cable faults in an intelligent operation and maintenance and security system for industrial switches based on edge computing, as proposed in this invention.
[0029] Figure 3 This is a schematic diagram of the workflow of an intelligent operation and maintenance and security system for industrial switches based on edge computing proposed in this invention. Detailed Implementation
[0030] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the invention, and should not be construed as limiting the invention. Rather, embodiments of the invention include all variations, modifications, and equivalents falling within the spirit and scope of the appended claims.
[0031] Example 1 refer to Figure 1 In this embodiment, an intelligent operation and maintenance and security system for industrial switches based on edge computing is proposed. It is mounted on a wide-temperature fanless industrial switch hardware platform and is divided into three layers from top to bottom: edge data acquisition layer, local edge computing engine layer, and operation and maintenance security interaction layer.
[0032] The edge data acquisition layer utilizes various sensors, acquisition circuits, and protocol stacks within the switch hardware to achieve real-time acquisition of multi-dimensional data. This includes collecting network operation data, device status data, cable characteristic data, wireless access data, and security characteristic data in real time, and transmitting the acquired data to the local edge computing engine layer. Specifically, the acquired content includes: network operation data, device status data, cable characteristic data, wireless access data, and security characteristic data. Network operation data includes: real-time traffic of 10 Gigabit ports, packet forwarding latency, packet loss rate, VLAN configuration status, ring network topology, and real-time status of each link; device status data includes: main switch chip temperature, CPU core temperature, power input voltage, multi-point temperature measurement data of the casing, and operating status of each hardware module; cable characteristic data includes: differential impedance value of network cable, link attenuation value, and time-domain waveform of TDR reflection signal obtained through dedicated hardware acquisition circuit, with sampling accuracy reaching the nanosecond level; wireless access data includes: number of Wi-Fi 7 terminals connected, signal strength of each terminal, real-time uplink and downlink bandwidth, terminal MAC address and authentication information; security characteristic data includes: packet source / destination IP and MAC address, access behavior sequence, traffic burst characteristics, and IPsec encrypted tunnel connection status. All collected data is transmitted to the local edge computing engine layer through the internal bus. The sampling interval is set to 1ms~1min according to the data type, with high real-time parameters such as ring network status and control message latency using a 1ms sampling interval, and parameters such as device temperature and historical traffic using a 1min sampling interval.
[0033] The local edge computing engine layer, the core processing layer of the system, is implemented based on an industrial-grade multi-core ARM processor and includes three core components: an intelligent operation and maintenance diagnostic engine, a national cryptographic integrated security engine, and a wired and wireless unified scheduling module. The intelligent operation and maintenance diagnostic engine is used to achieve precise cable fault location, adaptive fault classification and early warning based on collected data, and to achieve fault self-healing collaboration in conjunction with the industrial ring network protocol. The intelligent operation and maintenance diagnostic engine can include: a cable fault precise location unit, an adaptive fault early warning unit, and a ring network self-healing collaboration unit. The workflow of the cable fault precise location unit is as follows: [reference needed]. Figure 2 The hardware can acquire the high-speed, nanosecond-level pulse signals injected into the network cable by the acquisition circuit to calculate the fault distance and impedance deviation rate. When there are breaks or impedance mismatches in the cable, the pulse signal will be reflected, and the acquisition circuit records the round-trip time and amplitude of the reflected signal. The fault distance calculation formula is: Where: L is the physical distance from the fault point to the switch port, in meters (m); v is the propagation speed of the electrical signal in the corresponding network cable, in meters per second (m / s); for CAT6 industrial twisted-pair cables, the speed factor is approximately 0.67, i.e., v ≈ 2.0 × 10⁻⁶. 8m / s; In actual deployment, v can be calibrated on-site using standard cables of known length to eliminate errors caused by batch differences in cables; t is the round-trip time of the pulse signal from transmission to reception of the reflected signal, in seconds (s). Fault type identification is achieved by combining impedance deviation rate with reflection characteristics. The formula for impedance deviation rate is: , where: Z 实测 The actual differential impedance value acquired by the hardware, in ohms (Ω); Z 标准 The standard differential impedance for industrial Ethernet twisted-pair cables is 100Ω. Fault diagnosis rules: If the reflected signal amplitude is 100% and the measured impedance is infinite: a broken wire fault is identified; if 10% < ΔZ < 50%: an impedance offset fault is identified, usually due to poor connector contact or localized cable damage; if ΔZ ≤ 10% but link attenuation exceeds a preset threshold: a link attenuation exceeding the limit fault is identified, usually due to cable aging or moisture. For example: In an industrial site, a CAT6 network cable is connected to a switch port. The hardware acquires a reflected signal round-trip time t = 350ns, and the measured differential impedance Z... 实测 =132Ω. Calculate the fault distance: L = (2.0 × 10 8 ×350×10 -9 ) / 2=35m; Calculate impedance deviation rate: ΔZ=(|132-100| / 100)×100%=32%. Judgment result: Impedance offset fault, the fault point is located 35 meters away from the port. The system automatically marks the fault location and type on the operation and maintenance interface, and operation and maintenance personnel can go directly to troubleshoot. The adaptive fault early warning unit is used to calculate the deviation of the sampled value from the normal baseline in real time, determine the early warning level, and link the corresponding hardware protection sidechain. A sliding window dynamic baseline algorithm can be used to avoid false alarms and missed alarms caused by fixed thresholds. Specifically, it includes: selecting historical sampling data from the past N hours as a sliding window, calculating the overall mean and overall standard deviation of the data within the window, and constructing the parameter normal baseline. Real-time calculation of the deviation of the current sampled value: Where μ is the overall mean within the sliding window; σ is the overall standard deviation; and X is the real-time collected value. The hierarchical warning rules are as follows: When D ≥ 3: a Level 1 warning is triggered (notification of concern), the system records the operation log, and does not interfere with business operations; when D ≥ 5: a Level 2 alarm is triggered (serious anomaly), the system actively pushes alarm information and links corresponding protection strategies (such as high-temperature load reduction, port rate limiting, etc.); D < 3 does not trigger a warning. For example, taking the temperature warning of the switch's main chip as an example, the sliding window is set to the past 24 hours, with a sampling interval of 5 minutes, for a total of 288 sampling points. The calculated average temperature within the window is μ = 46℃, and the overall standard deviation is σ = 2.8℃. The current real-time sampled temperature X = 61℃. The deviation is calculated as: D = |61 - 46| / 2.8 ≈ 5.36; D ≈ 5.36 ≥ 5, triggering a Level 2 high-temperature alarm. The system immediately pushes the alarm information to the operation and maintenance platform, and automatically starts the heat dissipation optimization strategy, adjusting the power consumption of non-core modules to prevent equipment from crashing due to high temperature. The method for calculating the deviation of the current sampled value can also be... Where i is the data collection number within a preset time period, which can be determined according to the actual situation. This time period can be longer than the sliding window duration to improve data stability, or shorter than the sliding window duration, ideally around 1 minute, to improve data sensitivity and avoid misjudgments caused by occasional data fluctuations. The specific time period depends on the actual situation and will not be elaborated upon here. I is the total number of data collections within the preset time period, i = 1, 2, ..., I; X i It is the collected data numbered i; μ is the overall mean of the data within the preset time period, t i This is the time elapsed between the acquisition time of data point i and the current time point, in seconds (s). Other possible values are also possible, but will not be detailed here. X0 represents the unit deviation for the data acquisition type, which can be set for each data acquisition type or calculated. Where max(X) is the maximum value of the collected data, and min(X) is the minimum value of the collected data. Alternatively, two boundary values can be calibrated instead. η is a step value, which can be 1-10, and other value settings are not excluded; details are omitted here. The judgment level threshold can be set based on experience; details are omitted here. This method can take into account the fluctuation characteristics of historical data, applying weighted attenuation to long-period sampled data to weaken its impact. It can quickly identify and separate abnormal data with large fluctuation amplitudes, effectively improving the computational sensitivity and response efficiency of edge fault judgment. The ring network self-healing collaborative unit is deeply integrated with the self-developed private ring network protocol. When a ring network link fault is detected, a hardware-level parallel switching mechanism is adopted to complete the seamless switching of the backup link within ≤5ms, ensuring uninterrupted industrial control services. Simultaneously, information such as the fault link number, switching time, and affected port range is pushed to the operation and maintenance diagnostic engine, automatically generating fault alarms and achieving synchronous response of "self-healing + alarm".
[0034] The national cryptographic integrated security engine is natively integrated into the switch's forwarding plane, eliminating the need for an external, independent security gateway and avoiding additional forwarding latency. It is used to implement national cryptographic hardware encryption, access control, edge-side abnormal traffic identification, and full-process operation auditing. The national cryptographic integrated security engine includes: a national cryptographic hardware encryption unit, an access control and firewall unit, an edge abnormal traffic identification unit, and an operation auditing unit. The national cryptographic hardware encryption unit integrates hardware acceleration modules for the SM1 and SM4 domestic encryption algorithms and adopts a packet-level parallel processing architecture.
[0035] Formula for calculating single-packet encryption latency: , where: L 报文 The length of a single message is expressed in bits; k represents the message priority level, with industrial control messages being the highest priority (k=1), Wi-Fi 7 visual data being the second highest priority (k=2), and ordinary data being the lowest priority (k=3); w k The scheduling factor is numbered k, with the highest priority message w1>1 having exclusive access to its dedicated encryption unit and not being overtaken by lower priority messages; N k R represents the number of parallel encryption units with priority number k; 加密 This represents the throughput of a single encryption unit, expressed in bits per second (bps). This calculation method corresponds to the unique mechanism of prioritizing industrial messages and unified wired / wireless scheduling in this invention. It quantifies the scheduling strategy into a latency formula, distinguishing it from general indiscriminate encryption. It addresses the pain point of high-bandwidth wireless services crowding out control message encryption resources, ensuring that encrypted industrial control messages still meet sub-microsecond real-time requirements. Total forwarding latency per packet after enabling national cryptographic encryption: T 转发The inherent processing latency of switch hardware forwarding is represented by α, which is the pipeline overlap coefficient, ranging from 0 to 1, indicating the percentage overlap between encryption processing and forwarding table lookup time. In this solution, because the national cryptographic unit is embedded in the forwarding channel, the typical overlap coefficient for 64-byte industrial packets is 0.4 to 0.6. β is the edge multi-task correction coefficient. In this solution, the hardware acceleration unit offloads maintenance and encryption algorithms, resulting in a typical β value of 1.05 to 1.1. This means that multi-tasks only bring a 5% to 10% micro-hour latency increase, far lower than the latency superposition of external gateways and independent maintenance platforms. This coefficient accurately quantifies the additional overhead of multi-task parallelism under the hardware offloading architecture, breaking the industry's inherent perception that "feature stacking inevitably compromises real-time performance." It directly demonstrates that the intrinsic integration of multiple functions does not significantly increase forwarding latency, balancing functional integration with industrial-grade real-time requirements. The access control and firewall unit, based on ACL access control lists and industrial protocol whitelists, implements port-level and device-level access permission management; it supports filtering rules based on MAC address, IP address, and industrial protocol type to block unauthorized device access and unauthorized protocol packets. The edge abnormal traffic identification unit has a built-in industrial network attack signature database, which identifies network threats such as port scans, DoS attacks, and abnormal protocol packets in real time on the local edge side without uploading to the cloud for analysis. It triggers alarms in real time when an attack is detected and can automatically configure blocking rules to isolate abnormal traffic. The operation audit unit records all device configuration operations, user login behavior, and security event logs locally, supporting retrieval and tracing by time, user, and event type. Log data is encrypted and stored locally for at least 180 days, meeting the requirements for compliance auditing.
[0036] The wired and wireless unified scheduling module is designed for mixed Wi-Fi 7 wireless terminals and wired ports in networking scenarios. It performs unified queue scheduling for traffic from both wired ports and Wi-Fi 7 wireless terminals, ensuring priority forwarding of industrial control messages. It employs a hybrid scheduling mechanism of weighted fair queues and strict priority queues: industrial control messages such as EtherCAT and Profinet are placed in the strict priority queue for priority forwarding, ensuring stable control command latency; high-bandwidth services such as machine vision and data acquisition from Wi-Fi 7 terminals are placed in the weighted fair queue, with bandwidth dynamically allocated based on the number of terminals and service requirements, preventing a single terminal from monopolizing all resources. Wired ports and wireless terminals share a unified operation and maintenance view and security policies, allowing administrators to monitor the status of all access devices, distribute configurations, and manage security on a single interface.
[0037] The operation and maintenance security interaction layer provides local visual network management, encrypted remote operation and maintenance interfaces, and local data storage services. Specifically, it can include: a local visual network management unit, a remote operation and maintenance interface unit, and a local storage unit. The local visual network management unit provides a web management interface that can be accessed simply by connecting to the switch's local network without installing a client. The interface displays network topology, device operating parameters, fault alarm lists, and security event statistics, and supports operations such as batch port configuration, VLAN partitioning, and remote firmware upgrades. The remote operation and maintenance interface unit supports remote operation and maintenance via an encrypted VPN tunnel. When the on-site network is disconnected from the external network, it automatically switches to a pure local operation and maintenance mode, ensuring that core operation and maintenance functions are unaffected. The local storage unit has a built-in industrial-grade eMMC storage chip that encrypts and stores operation logs, fault records, and security audit data. Data is retained even after power failure and supports offline export and traceability analysis.
[0038] Example 2 refer to Figure 3 The system operation steps are as follows: S1. Data Acquisition: Real-time acquisition of data across five dimensions: network operation, device status, cable characteristics, wireless access, and security characteristics; S2, Edge Intelligent Processing: Performs operation and maintenance diagnosis, security protection, and traffic scheduling processing in parallel locally on the switch. S21. Operation and maintenance diagnosis: Based on the collected data, cable fault location and type identification are performed, fault classification and early warning are performed through dynamic baseline algorithm, and link network protocol is linked to realize link fault self-healing; S22. Security Protection: It inherently implements national cryptographic hardware encryption, access control, abnormal traffic identification and operation auditing, and security processing is embedded in the forwarding process; S23. Traffic scheduling: Perform unified queue scheduling on wired and wireless traffic to ensure priority forwarding of industrial control messages; S3, Interaction and Storage: Displays operation and security status through a local visual interface, supports remote encrypted operation and maintenance, and stores all operation and audit data locally.
[0039] This embodiment uses a flexible production line for automotive parts as an application scenario. The production line is equipped with 16 industrial switches described in this invention forming a ring network, which connects to PLC controllers, servo drivers, 20 Wi-Fi 7 machine vision cameras, temperature sensors, and other devices. The requirements are that the ring network self-healing ≤5ms, data encryption using national cryptographic standards, and early warning of equipment failure.
[0040] System initialization phase: After the switch is powered on, the edge data acquisition layer initializes each acquisition channel and starts full-dimensional data acquisition according to the preset sampling interval; the local edge computing engine initializes the sliding window baseline, loads the national cryptographic certificate and access control rules, and completes the configuration of the Wi-Fi7 wireless module and queue scheduling initialization.
[0041] Normal production operation status: When the production line is in normal operation, the wired and wireless unified scheduling module allocates EtherCAT control messages sent by the PLC to the highest priority queue and image data from the machine vision camera to the second highest priority queue, dynamically allocating bandwidth resources to ensure that control commands are delayed and jitter-free; the national cryptographic integrated security engine encrypts all cross-device data transmissions with SM4, verifies device identity, blocks unauthorized access terminals, and stores all operation records in the local audit log; the intelligent operation and maintenance diagnostic engine continuously calculates baselines for various parameters, monitors equipment temperature, cable status, and ring network links in real time, and only records operating data when there are no abnormalities.
[0042] Latent cable fault handling process: Due to frequent dragging by a robotic arm, a connector on a section of network cable on the production line became loose, causing an impedance shift. After the hardware acquisition unit detected the impedance change, the cable fault location algorithm calculated that the fault point was located at 28 meters, with an impedance deviation rate of 27%. The system triggered a level one alert, marking the fault location and type on the maintenance interface. Maintenance personnel could then accurately troubleshoot and handle the issue during shift changes, preventing the fault from escalating and causing a sudden network outage and production stoppage.
[0043] Ring network link breakage handling process: If a section of the ring network fiber is accidentally broken by a material vehicle, the ring network self-healing coordination unit will complete the backup link switch within 3ms, and the production line control business will not be interrupted. At the same time, the self-healing module will synchronize the fault information to the operation and maintenance engine. The system will immediately trigger a level 2 alarm and push information such as the fault link, switching time, and scope of impact to the operation and maintenance personnel, so as to realize the synchronous completion of fault self-healing and operation and maintenance notification.
[0044] High Temperature Anomaly Handling Procedure: During summer, the ambient temperature in the workshop rises, and the temperature of the main chip of a certain switch continues to rise. The adaptive early warning model calculates that the temperature deviation reaches 4.2, triggering a Level 1 warning. The system automatically optimizes port traffic scheduling, reduces the power consumption of non-core modules, and simultaneously pushes a high temperature alert. If the temperature continues to rise to a deviation of 5 or more, a Level 2 alarm is triggered, and the upstream production system is linked to reduce its load to ensure that the equipment does not fail due to high temperature.
[0045] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.
[0046] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.
Claims
1. An intelligent operation and maintenance and security system for industrial switches based on edge computing, mounted on a wide-temperature fanless industrial switch hardware platform, characterized in that, The intelligent operation and maintenance and security system for industrial switches based on edge computing is divided into three layers from top to bottom: edge data acquisition layer, local edge computing engine layer, and operation and maintenance security interaction layer. The edge data acquisition layer is used to collect network operation data, device status data, cable characteristic data, wireless access data and security characteristic data in real time, and transmit the collected data to the local edge computing engine layer; The local edge computing engine layer includes an intelligent operation and maintenance diagnosis engine, a national cryptographic integrated security engine, and a wired and wireless unified scheduling module; The intelligent operation and maintenance diagnostic engine is used to accurately locate cable faults and provide adaptive fault classification and early warning based on collected data, and to achieve fault self-healing collaboration in conjunction with the industrial ring network protocol; the national cryptographic integrated security engine is used to realize national cryptographic hardware encryption, access control, edge-side abnormal traffic identification and full-process operation auditing. The wired and wireless unified scheduling module is used to perform unified queue scheduling of traffic from wired ports and Wi-Fi7 wireless terminals to ensure priority forwarding of industrial control messages. The operation and maintenance security interaction layer is used to provide local visual network management, encrypted remote operation and maintenance interfaces, and local data storage services.
2. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 1, characterized in that, The edge data acquisition layer specifically collects the following data: network operation data, device status data, cable characteristic data, wireless access data, and security characteristic data. Network operation data includes: real-time traffic of 10 Gigabit ports, packet forwarding latency, packet loss rate, VLAN configuration status, ring network topology, and real-time status of each link. Device status data includes: main switching chip temperature, CPU core temperature, power input voltage, multi-point temperature measurement data of the casing, and operating status of each hardware module. Cable characteristic data includes: obtaining network cable differential impedance values, link attenuation values, and TDR reflection signal time-domain waveforms through dedicated hardware acquisition circuits, with sampling accuracy reaching the nanosecond level. Wireless access data includes: the number of Wi-Fi 7 terminals connected, signal strength of each terminal, real-time uplink and downlink bandwidth, and terminal MAC address and authentication information. Security characteristic data includes: packet source / destination IP and MAC addresses, access behavior sequences, traffic burst characteristics, and IPsec encrypted tunnel connection status.
3. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 1, characterized in that, The intelligent operation and maintenance diagnostic engine includes: a cable fault precise location unit, an adaptive fault early warning unit, and a ring network self-healing collaborative unit. The cable fault precise location unit injects nanosecond-level high-speed pulse signals into the network cable via hardware acquisition circuitry. When there are breaks or impedance mismatches in the cable, the pulse signal is reflected, and the acquisition circuitry records the round-trip time and amplitude of the reflected signal. The adaptive fault early warning unit calculates the deviation of the sampled value from the normal baseline in real time, determines the early warning level, and links the corresponding hardware protection sidechain. The ring network self-healing collaborative unit is deeply integrated with the self-developed proprietary ring network protocol. When a ring network link fault is detected, a hardware-level parallel switching mechanism is used to seamlessly switch to a backup link within ≤5ms, ensuring uninterrupted industrial control services. Simultaneously, information such as the fault link number, switching duration, and affected port range is pushed to the operation and maintenance diagnostic engine, automatically generating fault alarms and achieving a synchronous response of "self-healing + alarm".
4. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 3, characterized in that, The formula for calculating fault distance is: Where: L is the physical distance from the fault point to the switch port; v is the propagation speed of the electrical signal in the corresponding network cable; t is the round-trip time of the pulse signal from transmission to reception of the reflected signal.
5. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 3, characterized in that, Fault type identification is achieved by combining impedance deviation rate with reflection characteristics. The formula for impedance deviation rate is: , where: Z 实测 Z represents the actual differential impedance value acquired by the hardware. 标准 This is the standard differential impedance of the corresponding network cable.
6. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 5, characterized in that, Fault determination rules: If the reflected signal amplitude is 100% and the measured impedance is infinite: it is determined to be a broken wire fault; if 10% < ΔZ < 50%: it is determined to be an impedance offset fault, usually due to poor contact of the connector or local damage to the cable; if ΔZ ≤ 10% but the link attenuation exceeds the preset threshold: it is determined to be a link attenuation exceeding the standard fault, usually due to cable aging or moisture.
7. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 3, characterized in that, The adaptive fault warning unit uses a sliding window dynamic baseline algorithm to avoid false alarms and missed alarms caused by fixed thresholds.
8. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 7, characterized in that, The alarm level is obtained by: calculating the deviation D of the current sampled value in real time; and then determining the alarm level according to the graded early warning rules.
9. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 1, characterized in that, The national cryptographic integrated security engine includes: a national cryptographic hardware encryption unit, an access control and firewall unit, an edge abnormal traffic identification unit, and an operation auditing unit. The national cryptographic hardware encryption unit incorporates hardware acceleration modules for the SM1 and SM4 domestic encryption algorithms, employing a multi-group parallel processing architecture to achieve end-to-end encrypted data transmission. The access control and firewall unit, based on ACL rules and industrial protocol whitelists, intercepts unauthorized device access and manages industrial protocols. The edge abnormal traffic identification unit identifies network attacks and abnormal traffic in real time based on a local feature library, triggering alarms and automatically blocking abnormal traffic. The operation auditing unit locally stores all configuration operations, access behaviors, and security event logs, supporting offline fault tracing and security auditing.
10. The intelligent operation and maintenance and security system for industrial switches based on edge computing according to claim 1, characterized in that, The operation and maintenance security interaction layer specifically includes: a local visual network management unit, a remote operation and maintenance interface unit, and a local storage unit; the local visual network management unit supports local access via the web, displays network topology, device status, fault alarms and security events, and supports batch parameter configuration and firmware upgrades; the remote operation and maintenance interface unit supports remote operation and maintenance access under an encrypted channel, and automatically switches to a pure local operation and maintenance mode when the network connection is lost; the local storage unit has built-in industrial-grade storage media to store no less than 180 days of operation logs, fault records and security audit data.
