A dual-redundant 100G switching system with multi-module collaborative management
The dual-redundant 100G switching system with multi-module collaborative management solves the problem of module-level redundancy backup under 100Gbps high-speed data exchange in existing technologies. It realizes the organic integration of 100G switching capability and module-level dual redundancy, constructs a multi-module collaborative redundancy system with a full-function stack, realizes second-level fault switching and seamless service conversion, and improves the system's data throughput and reliability.
Patent Information
- Application Number
- CN202611450858.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-17
- Publication Date
- 2026-11-17
AI Technical Summary
Existing high-reliability communication systems cannot achieve module-level redundancy backup under 100Gbps high-speed data exchange, resulting in problems such as conflict between communication bandwidth and redundancy backup, incomplete redundancy hierarchy, lack of unified management for master-slave switching, and difficulty in balancing high reliability and processing performance.
The dual-redundant 100G switching system, which adopts multi-module collaborative management, includes primary and backup power modules, FMQL modules, 100G switching modules, and high-reliability MCU modules. It achieves module-level dual redundancy through independent power supply, data, and management links. The high-reliability MCU module acts as a unified arbitration unit for health management and master-slave switching.
It achieves the organic integration of 100G high-speed switching capability and module-level dual redundancy, and constructs a multi-module collaborative redundancy system covering the full function stack, realizing second-level fault switching and seamless service conversion, improving the system's data throughput, continuous operation reliability and survivability in harsh environments.
Smart Images

Figure CN123123299A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of high-reliability network communication and fault-tolerant computing technology, specifically to a dual-redundant 100G switching system with multi-module collaborative management. It is particularly suitable for applications with extremely high requirements for communication bandwidth and system reliability, such as military equipment, marine electronics, nuclear power control, and industrial robots. In particular, this invention relates to the deep integration of 100G high-speed network switching technology, FPGA collaborative processing technology, and high-reliability MCU master-slave management technology, achieving single-point fault tolerance and seamless service switching during long-term continuous operation of a high-reliability platform through a module-level dual-redundant architecture. Background Technology
[0002] With the continuous improvement of information technology in military equipment, intelligent ship electronics, digital nuclear power control systems, and the autonomy of industrial robots, the requirements for data throughput, real-time performance, and continuous operational reliability of high-reliability communication platforms are increasing. In application scenarios such as military battlefields, ocean-going vessels, nuclear power plant control, and industrial robots, communication equipment needs to operate unattended for extended periods in harsh environments, facing multiple threats including extreme temperatures, strong electromagnetic interference, vibration and shock, and component aging. A failure can lead to communication link interruption, causing mission failure or safety incidents. Therefore, high-reliability redundancy backup technology has become crucial in the design of these high-reliability communication systems. A dual-redundant 100G switching system with multi-module collaborative management can achieve 100Gbps high-speed data exchange, protocol processing, and system-level redundancy backup, meeting the dual requirements of communication bandwidth and fault tolerance in high-reliability scenarios.
[0003] Currently, high-reliability communication systems mainly adopt the following technical solutions.
[0004] (1) Standalone cold backup solution based on low-speed bus interconnection Traditional high-reliability communication equipment often uses low-speed buses such as 1553B and CAN for inter-module interconnection, with communication rates typically in the hundreds of Mbps range. This cannot meet the high-capacity data transmission requirements of current military equipment, marine electronics, nuclear power control, and industrial robots. In terms of redundancy design, a whole-system cold backup approach is often adopted, where a manual or host computer command switches to the backup device after the primary device fails. The disadvantages of this approach are: 1) low bus bandwidth, unable to support high-speed data exchange at 100Gbps; 2) long cold backup switching time (typically minutes), resulting in complete service interruption during the switch; and 3) low utilization of whole-system backup resources, failing to achieve fine-grained redundancy at the module level.
[0005] (2) Single-module hot backup solution based on commercial switching chips Some communication equipment uses commercial 100G Ethernet switching chips to build high-speed switching networks and achieves hot backup through dual switching chips. However, this solution has the following shortcomings when directly applied to high-reliability scenarios: 1) Commercial switching chips have not undergone high-reliability screening and hardening treatment, and are at high risk of bit error rate and functional abnormality under harsh environments (strong electromagnetic interference, extreme temperature, vibration, etc.); 2) Redundancy is only implemented at the switching layer, and the FPGA coprocessing module is still a single point (non-redundant) design, which does not constitute a complete module-level redundancy system; 3) There is a lack of a unified redundancy management mechanism, the backup strategies of each module are independent of each other, the coordination of fault switching is poor, and it is difficult to ensure seamless service connection.
[0006] (3) Fault-tolerant scheme based on FPGA logic redundancy Existing technologies also employ FPGA triple modular redundancy (TMR) or logic-level redundancy design methods to improve reliability. However, these solutions primarily address fault tolerance of the FPGA's internal logic resources and cannot solve system failures caused by module-level hardware failures (such as chip damage or power supply failure). Furthermore, TMR solutions have high resource overhead (typically requiring more than three times the logic resources), making them difficult to implement in scenarios where a significant amount of FPGA resources are already being used for 100G switching protocol processing.
[0007] (4) Centralized management scheme based on integrated electronics Existing integrated electronic systems mostly adopt a centralized management architecture, with a central management computer uniformly monitoring and controlling each functional module. The disadvantages of this approach are: 1) The central management computer itself is prone to becoming a single point of failure; once the management computer fails, all redundancy switching mechanisms cannot be executed; 2) When the management channels (such as IIC, UART) between the management computer and each functional module fail, the monitoring capability of the module will be lost; 3) There is a lack of independent management units with high reliability levels, and management functions may fail first under harsh environmental conditions.
[0008] In summary, existing technologies share the following common problems: 1) The contradiction between communication bandwidth and redundancy backup: Traditional high-reliability redundancy solutions are limited by low-speed interconnect buses and cannot achieve effective module-level redundancy backup under a 100Gbps high-speed switching architecture; 2) Incomplete redundancy hierarchy: Existing solutions mostly implement redundancy for single functional modules (such as only switching redundancy or only computing redundancy), lacking a complete multi-module collaborative redundancy system covering switching, processing, management, and power supply; 3) Lack of unified management for master-slave switching: The master-slave relationship of each module is configured independently, lacking a switching mechanism arbitrated by a high-reliability MCU, resulting in insufficient coordination between fault isolation and service recovery; 4) Difficulty in balancing high reliability and processing performance: High-reliability devices have limited processing capabilities and cannot handle high-speed protocol processing, while high-performance commercial devices lack adaptability to harsh environments, and the two have not yet been effectively integrated at the system architecture level.
[0009] Therefore, the industry urgently needs a highly reliable processing system architecture that can organically integrate 100G high-speed switching capability with multi-module and dual-redundant backup to meet the comprehensive requirements of high bandwidth, high reliability, and rapid fault recovery in scenarios such as military equipment, marine electronics, nuclear power control, and industrial robots, so as to reduce or even overcome the defects of existing technologies and achieve more beneficial technical effects and technological progress.
[0010] The information included in this background section of this application specification, including any references cited herein and any descriptions or discussions thereof, is included for technical reference purposes only and is not intended to limit the scope of this application. Summary of the Invention
[0011] This application is made in view of the foregoing and other further ideas.
[0012] The main objective of this invention is to overcome the shortcomings of the prior art, provide an innovative dual-redundant 100G switching system with multi-module collaborative management, and solve at least the following technical problems.
[0013] (1) The integration of 100G high-speed switching and module-level redundancy Existing high-reliability communication systems are limited by low-speed interconnect buses such as 1553B and CAN, making it impossible to achieve effective module-level redundancy backup under a 100Gbps high-speed data exchange architecture. This invention aims to construct a high-speed data exchange core through 100G switching modules, mainly realizing 100G switching functions, and is equipped with a complete module-level dual redundancy system, enabling the system to have both ultra-high bandwidth and fault tolerance capabilities.
[0014] (2) The problem of constructing a complete multi-module collaborative redundancy system Existing solutions mostly implement redundancy for single functional modules (such as redundancy only on the switching plane or only on the management plane), lacking a full-stack redundancy design covering switching, processing, control, and power supply. This invention aims to construct a complete multi-module collaborative redundancy system including power supply, ZYNQ (FMQL), 100G switching, and a high-reliability MCU, eliminating the potential for single-point-of-failure at the system level.
[0015] (3) Issues of unified redundancy management and rapid failover In existing solutions, the backup strategies for each module are independent, lacking a unified arbitration management mechanism, resulting in poor coordination and long switching times during fault switching. This invention aims to use a highly reliable MCU module as an independent redundancy management arbitration unit, primarily for health management and master-slave switching control of each functional module, achieving rapid isolation of module-level faults and seamless service switching.
[0016] (4) Balancing high performance processing with high reliability High-reliability devices have limited processing capabilities and cannot handle high-speed protocol processing; high-performance devices lack adaptability to harsh environments, and the two have not yet been effectively integrated at the system architecture level. This invention aims to achieve this through a collaborative architecture design, where high-performance devices are responsible for data plane processing (the 100G switching module mainly implements 100G switching functions, and the FMQL module mainly manages, configures, and mirrors data for the switching module), while the high-reliability MCU is responsible for control plane management (mainly performing health management and master-slave switching for each module), thus balancing the requirements of high performance and high reliability.
[0017] To achieve the above-mentioned objectives and others, according to at least one embodiment of the present invention, a dual-redundant 100G switching system with multi-module collaborative management is provided, which may include a primary power supply module, a backup power supply module, a primary FMQL module, a backup FMQL module, a high-reliability MCU module, a primary 100G switching module, and a backup 100G switching module. The dual-redundant 100G switching system with multi-module collaborative management incorporates at least the following concepts and designs.
[0018] (A) Power redundancy architecture The primary and backup power modules employ independent power supply paths, primarily converting 42V DC input to 12V DC output to power the primary and backup functional module groups, respectively. When one power module fails, the other can continuously supply power to the corresponding functional module, preventing a single power supply failure from paralyzing the entire system. Furthermore, telemetry signals output from the power modules are transmitted to external integrated electronic equipment. This equipment uses these signals to determine the normality of the primary and backup power modules. Upon detecting an anomaly, the integrated electronic equipment can control the primary / backup power module shutdown and switching via OC command signals.
[0019] (B) Redundant architecture of FMQL coprocessor layer The primary FMQL module and the backup FMQL module establish a bidirectional data communication link through a 1000 Base-T Ethernet interface to achieve master-slave state synchronization of the coprocessor layer; The primary FMQL module establishes a data communication link with the primary 100G switching module through a PCIe 2.0 X1 high-speed interface; The backup FMQL module establishes a data communication link with the backup 100G switching module through the PCIe 2.0 X1 high-speed interface.
[0020] (C) 100G Switching Layer Redundancy Architecture The configuration synchronization and status mirroring of the primary 100G switching module and the backup 100G switching module are completed via PCIE 2.0 x1 links through their respective FMQL45T900 units.
[0021] Both the primary 100G switching module and the backup 100G switching module support 100Gbps high-speed Ethernet data switching and provide high-speed network interfaces to the outside world.
[0022] Both the primary 100G switching module and the backup 100G switching module support 1000 Base-T Ethernet data switching and provide gigabit network interfaces.
[0023] (D) High-reliability MCU unified management architecture The high-reliability MCU module, as an independent redundancy management arbitration unit of the system, is mainly responsible for health management and master-slave switching of each module. It establishes management and control links with the primary FMQL module, backup FMQL module, primary 100G switching module and backup 100G switching module through GPIO, IIC and UART low-speed management interfaces, respectively. It is used to monitor the operating status of each module in real time, execute master-slave arbitration decisions, and issue switching commands when a module failure is detected.
[0024] (E) Master-slave switching control mechanism The high-reliability MCU module periodically sends heartbeat detection signals to the corresponding functional modules through each management and control link and receives response signals from each module. When a module fails to receive a response signal for several consecutive times, or detects abnormal status information reported by the module, it is determined that the module has failed. The preset master-slave switching process is then initiated to control the business traffic to switch from the failed master module to the corresponding backup module to continue running.
[0025] More specifically, according to one aspect of the present invention, a dual-redundant 100G switching system with multi-module collaborative management is provided, comprising: a primary power supply module and a backup power supply module, both employing independent power supply paths to convert 42V DC input to 12V DC output, supplying power to the primary functional module group and the backup functional module group respectively; a primary FMQL module and a backup FMQL module, which establish a bidirectional data communication link through a 1000 Base-T Ethernet interface; a primary 100G switching module and a backup 100G switching module, both supporting 100Gbps high-speed Ethernet data exchange and 1000 Base-T Ethernet data exchange; and a high-reliability MCU module, which serves as an independent redundancy management arbitration unit, connected to the primary FMQL module, the backup FMQL module, the primary 100G switching module, and the backup 100G switching module respectively through a management control link, for real-time monitoring of the operating status of each module, execution of master-slave arbitration decisions, and fault switching control.
[0026] According to one or more embodiments of the present invention, the primary FMQL module is connected to the primary 100G switching module via a PCIe 2.0×1 high-speed interface, and the backup FMQL module is connected to the backup 100G switching module via a PCIe 2.0×1 high-speed interface. The configuration synchronization and status mirroring of the primary 100G switching module and the backup 100G switching module are completed through their respective FMQL modules via the PCIe 2.0 × 1 interconnect link; The high-reliability MCU module establishes management and control links with the primary FMQL module, backup FMQL module, primary 100G switching module, and backup 100G switching module through GPIO, IIC, and UART low-speed management interfaces, respectively.
[0027] According to one or more embodiments of the present invention, the primary power module and the backup power module each include an OC control circuit, a soft-start circuit, an EMI filter circuit, an isolated DC / DC converter, and an output filter circuit. Among them, the OC control circuit receives the OC command signal from the external integrated electronic equipment and controls the enable / disable of the isolated DC / DC converter; The dual-redundant 100G switching system with multi-module collaborative management is configured to shut down the power supply channel of the abnormal power module by means of an OC command signal when the telemetry signals output by the primary power module and the backup power module are transmitted to the external integrated electronic equipment and the external integrated electronic equipment detects that one of the power modules is abnormal.
[0028] According to one or more embodiments of the present invention, both the primary 100G switching module and the backup 100G switching module are implemented using the Shengke CTC8120 switching chip, providing a 100G Base-SR4 fiber optic interface and a 1000 Base-T electrical port to the outside; and the 100G Base-SR4 fiber optic interface is connected to an external 100G fiber optic network, while the 1000 Base-T electrical port is connected downwards to the corresponding FMQL module.
[0029] According to one or more embodiments of the present invention, both the primary FMQL module and the backup FMQL module are implemented using the Fudan Microelectronics FMQL45T900 chip. The programmable logic portion of the Fudan Microelectronics FMQL45T900 chip implements a PCIe 2.0×1 interface controller, a Gigabit Ethernet MAC controller, and an LVDS high-speed serial interface controller. The ARM processor portion of the Fudan Microelectronics FMQL45T900 chip runs an embedded PetaLinux system, which is responsible for switching chip management, network protocol stack processing, and communication with the MCU module. Furthermore, a redundant synchronization link is established between the primary FMQL module and the backup FMQL module through a 1000 Base-T Ethernet interface, a GPIO interface, and an LVDS interface.
[0030] According to one or more embodiments of the present invention, the high-reliability MCU module is implemented using the Guoke Anxin AS32S601 microcontroller. It connects the reset pin, status indicator pin and interrupt pin of each module through the GPIO interface, connects the sensors and EEPROM of each module through the IIC bus, communicates with each module for heartbeat detection through the UART serial port, and connects to external integrated electronic equipment through the RS422 interface.
[0031] According to one or more embodiments of the present invention, a high-reliability MCU module is configured to perform the following health management and master-slave switchover steps: Perform heartbeat detection according to the preset monitoring cycle: send heartbeat request frames to each functional module through the UART serial port, and receive response frames containing voltage, temperature and current status; Perform voltage and temperature monitoring: Read the key power rail voltage of each module through AD sampling, and read the temperature sensor data of each module through the IIC bus; Anomaly detection: When a module fails to return a heartbeat response for 3 consecutive times, the core temperature exceeds +105℃ or the case temperature exceeds +85℃, or the critical power rail voltage deviates from the nominal value by more than ±10%, the module is considered to be faulty. Master-Slave Switching: After confirming the fault, a fault alarm is sent to the external integrated electronic equipment, the primary 100G switching module port is shut down and the backup 100G switching module port is enabled, an activation command is sent to the backup FMQL module to switch to its operating mode, the power supply to the faulty module is shut down to achieve fault isolation, and the status after the switch is reported to the external integrated electronic equipment.
[0032] According to one or more embodiments of the present invention, a dual-redundant 100G switching system with multi-module collaborative management is configured to perform management and configuration steps of the FMQL module for the primary 100G switching module and the backup 100G switching module: Chip initialization: After the system is powered on, a reset release signal is sent to the switching chip through the PCIe 2.0×1 interface to configure the port rate mode, VLAN division rules and flow control policies; Switching entry distribution: Based on the forwarding entries generated by the network protocol stack, the MAC address table, IP routing table and ACL rules are written into the TCAM memory inside the switching chip through the PCIe interface; Status monitoring and dynamic adjustment: Periodically read the port statistics register of the switching chip, dynamically adjust the port load balancing strategy according to the number of bytes sent and received and the error frame rate, and issue a port isolation command when the error frame rate exceeds the threshold; Configure mirror synchronization: After the primary FMQL module completes the configuration, it sends the configuration data packet to the backup FMQL module through the 1000 Base-T interface, and the backup FMQL module performs the same configuration operation on the backup switching chip.
[0033] According to one or more embodiments of the present invention, a dual-redundant 100G switching system with multi-module collaborative management is configured to perform the following system power-on and power supply management steps: Initial power-on: Receives 28V command power and 42V bus power. The main and backup power modules output 12V voltage, which is converted to 3.3V to power the MCU module. Functional module power-on: After the MCU completes its self-test, it sequentially enables the primary FMQL module, the primary 100G switching module, the backup FMQL module, and the backup 100G switching module. Status Synchronization: After the primary and backup FMQL modules complete the mirror synchronization of the switching chip configuration via the 1000 Base-T interface, the system enters a dual-redundancy operating state; and Fault power supply isolation: When an abnormality is detected in the primary and backup functional modules, the power supply to the primary FMQL module and the primary and backup 100G switching modules is shut off to ensure uninterrupted power supply to the backup functional modules.
[0034] According to one or more embodiments of the present invention, the configuration of the dual-redundant 100G switching system with multi-module collaborative management adopts a homogeneous redundancy architecture, including two completely symmetrical functional chains on the primary side and the backup side, and state synchronization and collaborative switching are achieved between the functional chains on the primary side and the backup side through high-speed data links and low-speed management links.
[0035] In summary, compared with the prior art, the present invention and one or more embodiments thereof can achieve at least the following beneficial effects.
[0036] (1) It realizes the organic integration of 100G high-speed switching capability and module-level dual redundancy.
[0037] According to the present invention, a high-speed data switching core can be constructed using a 100G switching module, which mainly realizes the 100G switching function. The main backup modules are configured in each functional layer of the system (FMQL layer, 100G switching layer), which solves the problem that traditional high-reliability systems are limited by low-speed buses and cannot carry large-capacity data transmission. At the same time, it ensures that services are not interrupted when any switching module fails, which significantly improves the data throughput and continuous operation reliability of the high-reliability communication platform.
[0038] (2) A complete multi-module collaborative redundancy system covering the full function stack was constructed.
[0039] According to this invention, a complete multi-module collaborative redundancy system covering power supply, FMQL (ZYNQ), 100G switching, and a high-reliability MCU can be constructed. Each functional unit within the system is configured with independent primary and backup copies, and a three-dimensional redundancy topology is formed between each level through high-speed data links and low-speed management links. The failure of any module will not cause the entire system to fail, completely eliminating the risk of single-point failure. Specifically, the FMQL module mainly manages and configures the 100G switching module and performs data mirroring, while the MCU module mainly performs health management and master-slave failover for each module.
[0040] (3) A second-level fault switching was achieved by unified arbitration by a highly reliable MCU.
[0041] According to the present invention, a highly reliable MCU module can be introduced as an independent redundancy management arbitration unit, which mainly performs health management and master-slave switching for each module. Through the GPIO / IIC / UART management and control link, unified heartbeat detection, status monitoring and master-slave switching control are performed on each functional module. The switching decision is centrally executed by a single arbitration node, which avoids the poor coordination problem caused by independent switching of each module in the existing solution. The fault switching time is shortened from minutes to seconds in the existing solution, realizing truly seamless service switching.
[0042] (4) At the system architecture level, high performance and high reliability are achieved.
[0043] According to the present invention, high-performance processing devices (ZYNQ (FPGA), 100G switching chip) and high-reliability management devices can be organically integrated at the system architecture level: the 100G switching module mainly implements 100G switching functions, the FMQL module mainly manages and configures the switching module and performs data mirroring, and the high-reliability MCU is responsible for system redundancy management and fault arbitration on the control plane (mainly performing health management and master-slave switching for each module). The two perform their respective duties and work together, fundamentally solving the contradiction between the insufficient processing power of high-reliability devices and the insufficient reliability of general devices.
[0044] (5) The power redundancy design further enhances the system’s on-orbit survivability.
[0045] According to the present invention, an independent dual-power redundant design can be adopted. The primary power supply module and the backup power supply module mainly convert the 42V input to 12V to independently power each module. When one power supply fails, the other power supply can continue to provide power, and the highly reliable MCU can synchronously start the corresponding functional module to switch to the other side, forming dual redundancy protection of power supply and function, further improving the system's survivability and task continuity in harsh environments.
[0046] Further embodiments of this application can achieve other advantageous technical effects not listed one by one. These other technical effects may be partially described below and can be expected and understood by those skilled in the art after reading this application. Attached Figure Description
[0047] The above-described features and advantages of these embodiments, as well as other features and advantages, and the ways in which they are implemented, will become more apparent from the following description in conjunction with the accompanying drawings, and embodiments of this application can be better understood. A brief description of the drawings is as follows.
[0048] Figure 1 This is a block diagram of the module composition of a dual-redundant 100G switching system with multi-module collaborative management according to an embodiment of the present invention.
[0049] Figure 2 This is a diagram illustrating the overall architecture of a dual-redundant 100G switching system with multi-module collaborative management according to an embodiment of the present invention.
[0050] Figure 3 This is a data flow diagram of a dual-redundant 100G switching system with multi-module collaborative management according to an embodiment of the present invention.
[0051] Figure 4 This is a flowchart of MCU health management and master-slave switching according to an embodiment of the present invention. Detailed Implementation
[0052] The details of one or more embodiments of this application will be set forth in the following description of the accompanying drawings and specific embodiments. Other features, objects, and advantages of this application will become clear from these descriptions, drawings, and claims.
[0053] It should be understood that the illustrated and described embodiments are not limited in application to the details of the construction and arrangement of the components set forth in the following description or illustrated in the drawings. The illustrated embodiments may be other embodiments and can be implemented or performed in various ways. In fact, it will be apparent to those skilled in the art that various modifications and variations can be made to the various embodiments / examples of this application without departing from the scope or spirit of this disclosure. For example, features illustrated or described as part of one embodiment may be used with another embodiment to still produce another embodiment. Therefore, this disclosure covers such modifications and variations that fall within the scope of the appended claims and their equivalents.
[0054] Similarly, it is understood that the phrases and terms used in this document are for descriptive purposes and should not be considered restrictive. The use of “including,” “contains,” or “has,” and their variations, in this document is intended to include, in an open-ended manner, the items listed thereafter, their equivalents, and any additional items.
[0055] The present invention will now be described in more detail with reference to several specific embodiments and accompanying drawings.
[0056] It will be understood by those skilled in the art that the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.
[0057] 1. System overall architecture.
[0058] like Figure 1 As shown, a dual-redundant 100G switching system with multi-module collaborative management according to an embodiment of the present invention is provided, which may include a primary power supply module, a backup power supply module, a primary FMQL module, a backup FMQL module, a high-reliability MCU module, a primary 100G switching module, and a backup 100G switching module. The dual-redundant 100G switching system with multi-module collaborative management incorporates at least the following concepts and designs.
[0059] Reference Figure 2 This paper illustrates the overall architecture of a dual-redundant 100G switching system with multi-module collaborative management according to an embodiment of the present invention. In this embodiment, the dual-redundant 100G switching system with multi-module collaborative management can adopt a homogeneous redundancy architecture, including two completely symmetrical functional chains: a primary side and a backup side. State synchronization and collaborative switching are achieved between the primary and backup sides through a high-speed data link and a low-speed management link. The dual-redundant 100G switching system with multi-module collaborative management includes, from top to bottom, a 100G switching layer, an FPGA coprocessing layer, an MCU management layer, and a power supply layer. Specifically, the 100G switching module mainly implements 100G switching functions; the FMQL module mainly manages and configures the 100G switching module and performs data mirroring; the MCU module mainly performs health management and master-slave switching for each module; and the power supply module mainly converts the 42V input to 12V to power each module.
[0060] 2. Specific exemplary configurations of each functional module of the system.
[0061] 2.1 100G switching layer.
[0062] The 100G switching layer may include a primary 100G switching module and a backup 100G switching module. In this embodiment, both the primary 100G switching module and the backup 100G switching module can be implemented using the Shengke CTC8120 switching chip.
[0063] The CTC8120 chip is a switching chip that supports 100Gbps line-speed forwarding capability and can provide multiple 100GBase-SR4 fiber optic interfaces and multiple 1000Base-T electrical ports.
[0064] The CTC8120 module can connect to an external 100G fiber optic network via a 100G Base-SR4 fiber optic interface, enabling high-speed data transmission and reception at 100Gbps. Configuration synchronization and status mirroring of these two switching chips / modules can be achieved via their respective FMQL45T900 units through a PCIe 2.0 x1 interconnect link. Simultaneously, they can be connected downlinked to the corresponding FMQL45T900 units via a 1000 Base-T electrical port, facilitating the exchange of control plane data and low-speed service data.
[0065] 2.2 FPGA coprocessing layer.
[0066] The FPGA coprocessing layer may include a primary FMQL module and a backup FMQL module. In this embodiment, both the primary and backup FMQL modules can be implemented using the Fudan Microelectronics FMQL45T900 chip.
[0067] In this embodiment, the programmable logic section of the FMQL45T900 can be used to implement a PCIe 2.0 X1 interface controller, a Gigabit Ethernet MAC controller, and an LVDS high-speed serial interface controller; the ARM processor section runs an embedded PetaLinux system, which is responsible for performing switch chip management, network protocol stack processing, and communication tasks with the MCU module.
[0068] The primary FMQL45T900 unit and the backup FMQL45T900 unit can establish a redundant synchronization link through the following three interfaces: 1000Base-T Ethernet interface: used for transmitting service data and coprocessor layer status synchronization information; GPIO interface: used to transmit hardware-level heartbeat signals and fast interrupt signals; LVDS interface: used for transmitting high-speed serial status data and large amounts of mirror information.
[0069] 2.3 MCU Management Layer.
[0070] The MCU management layer may include an independent high-reliability MCU module, which is mainly responsible for health management and master-slave switching of each module. In this embodiment, the AS32S601 microcontroller from Guoke Anxin can be used to implement this.
[0071] In this embodiment, the MCU management layer (AS32S601 MCU) can connect to each functional module of the system through the following management and control links: via GPIO interface to the reset pin, status indicator pin, and interrupt pin of each module for outputting hardware reset signals and receiving module status flags; via IIC bus to the IIC slave interface of each module for reading the module's temperature sensor, voltage monitoring chip, and EEPROM configuration information; via UART serial port to the debug serial port of each module for transmitting heartbeat detection data packets and status reporting frames; and via RS422 interface to the integrated electronic equipment or host computer for receiving integrated electronic injection switching commands or uploading system telemetry data.
[0072] 2.4 Power Supply Layer.
[0073] The power supply layer can include a primary power supply module and a backup power supply module, adopting a completely independent redundant power supply architecture, mainly converting the 42V input to 12V to power each module.
[0074] In this embodiment, the internal structure of the primary and backup power supply modules may include: an OC control circuit, a soft-start circuit, an EMI filter circuit, an isolated DC / DC converter, and an output filter circuit.
[0075] The power supply module receives two external power supplies on its input side: (1) 28V command power supply: after passing through the OC control circuit, the enable / disable of the isolated DC / DC converter is controlled by the control signal (2 main and 2 backup OC control signals) to realize command-level control of the main or backup power supply channels; (2) 42V bus power supply: after filtering out high-frequency interference by the EMI filter circuit, it is input to the isolated DC / DC converter for voltage conversion.
[0076] The isolated DC / DC converter converts the 42V bus voltage into the 12V DC voltage required by the system, and then outputs it to the main functional module group or the backup functional module group after being filtered by the output filter circuit.
[0077] The OC control circuit receives OC command signals from external integrated electronic equipment to control the primary / standby switching of the power supply module. Telemetry signals (including output voltage, temperature, and other parameters) output by the power supply module are transmitted to the external integrated electronic equipment, which uses these signals to determine whether the primary and standby power supply modules are functioning correctly. When the integrated electronic equipment detects an anomaly in a single power supply module, it actively shuts off the power supply channel on the faulty side via OC command signals, thus achieving primary / standby switching and preventing damage to functional modules from a faulty power supply.
[0078] 3. Inter-module connections and data flow.
[0079] 3.1 Normal data flow path.
[0080] Reference Figure 3 This document illustrates a data flow diagram of a dual-redundant 100G switching system with multi-module collaborative management according to an embodiment of the present invention. An exemplary software in this embodiment may include the Shengke 8120 module SDK, FMQL module software, and MCU health management software. The Shengke 8120 module SDK primarily implements data plane control functions such as initialization configuration of the 100G switching chip, forwarding table entry management, and VLAN / QoS policy configuration. It is responsible for the register-level driver and protocol stack adaptation of the switching chip, forming the software foundation for the normal operation of the 100G switching function. The FMQL module software performs configuration management and operational status monitoring of the 100G switching modules, achieving data synchronization and status mirroring between the primary / backup switching modules, ensuring seamless service continuity during primary / backup switching. The MCU health management software periodically collects parameters such as temperature and voltage of each module via GPIO / IIC / UART, performs unified health status assessment, executes arbitration decisions upon detecting module faults, issues switching commands, and achieves fault isolation and service migration.
[0081] 3.2 FPGA management and configuration of the switching chip.
[0082] In this embodiment, the FMQL module, such as the FMQL45T900 unit, can manage and configure the CTC8120 switching chip through the PCIe 2.0 X1 interface, specifically including the following steps.
[0083] Step S1: Chip Initialization After the system powers on, the ARM processor of the FMQL45T900 unit first sends a reset release signal to the CTC8120 through the PCIe 2.0 X1 interface, waiting for the chip to complete its bootstrapping. After the chip bootstrapping is complete, the FMQL45T900 performs read and write operations on the PCIe configuration space to configure the port rate mode, VLAN partitioning rules, and flow control policies of the CTC8120.
[0084] Step S2: Issue swap table entries The FMQL45T900 unit writes the MAC address table, IP routing table, and ACL rules into the TCAM memory inside the CTC8120 through the PCIe 2.0 X1 interface, based on the forwarding table entries generated by the upper-layer network protocol stack.
[0085] Step S3: Status Monitoring and Dynamic Adjustment The FMQL45T900 unit can periodically read the port statistics registers of the CTC8120 (including the number of bytes sent and received, the number of error frames, the number of CRC errors, the number of flow control frames, etc.) through the PCIe interface, and dynamically adjust the port load balancing strategy based on the statistical results. When the error frame rate of a port exceeds a preset threshold, the FMQL45T900 unit issues a port isolation command through the PCIe interface to switch the traffic of that port to a redundant port.
[0086] Step S4: Configure mirror synchronization After completing the CTC8120 configuration, the primary FMQL45T900 unit sends the configuration data packet to the backup FMQL45T900 unit through the 1000 Base-T interface. After parsing the data packet, the backup unit performs the same configuration operation on the backup CTC8120 module through its own PCIe 2.0 X1 interface to ensure that the configuration of the switching chips on both the primary and backup sides is completely consistent.
[0087] 4. MCU health management and master-slave switching process.
[0088] like Figure 4 The diagram illustrates an exemplary flowchart of MCU health management and master-slave switching according to an embodiment of the present invention.
[0089] 4.1 Health monitoring mechanism.
[0090] The AS32S601 MCU module performs the following health management operations according to a preset monitoring cycle (set to 500ms in this embodiment).
[0091] Step H1: Heartbeat Detection The MCU sends heartbeat request frames to the primary FMQL45T900 unit, the backup FMQL45T900 unit, the primary CTC8120 module, and the backup CTC8120 module via the UART serial port.
[0092] Upon receiving a heartbeat request frame, each functional module should return a heartbeat response frame within a preset timeout window (set to 100ms in this embodiment). The response frame contains the module's current operating status, including voltage, temperature, and current.
[0093] Step H2: Voltage and temperature inspection The MCU reads the voltage and current status of each module board sequentially through AD sampling, with the voltage monitoring range covering key power rails such as 3.3V, 1.8V, and 1.0V; it reads the data of the 18B20 temperature sensor of each module through the IIC bus, with the temperature monitoring range covering -40℃ to +85℃.
[0094] Step H3: Anomaly Detection The MCU compares the collected monitoring data with preset thresholds. In this embodiment, an example of the set judgment rule is as follows: If a module fails to return a heartbeat response three times in a row, the module is deemed to have failed to communicate. If the core temperature of a module exceeds +105℃ or the module case temperature read exceeds +85℃, the module is determined to be overheated. If the critical power rail voltage of a module deviates from the nominal value by more than ±10%, the power supply of that module is deemed abnormal.
[0095] 4.2 Master-slave switchover process.
[0096] When the MCU determines that either module on the master or slave side has failed, the following master-slave switching process can be executed.
[0097] Step T1: Fault Confirmation and Alarm After confirming that the fault state has stabilized within three consecutive monitoring cycles, the MCU sends a fault alarm frame to the integrated electronic equipment via the RS422 interface. The alarm frame includes the fault module ID, fault type code, and fault occurrence timestamp. Simultaneously, the MCU illuminates the fault indicator light on the system panel via the GPIO interface.
[0098] Step T2: Business Traffic Switching The MCU sends a port shutdown signal to the primary CTC8120 module via the GPIO interface, and simultaneously sends a port enable signal to the backup CTC8120 module. Since the backup CTC8120 module has already completed configuration mirroring synchronization through the FMQL45T900 unit, it can immediately take over the 100G high-speed data exchange service after the port is enabled.
[0099] Step T3: Control plane switching The MCU sends an activation command to the backup FMQL45T900 unit via the UART serial port. After receiving the command, the ARM processor of the unit switches its operating mode from "backup monitoring" to "primary activation", loads the running context from the operating system image, and resumes business processing.
[0100] Step T4: Fault Isolation The MCU sends a shutdown command to the power chip of the faulty module via GPIO, cutting off the power supply path of the faulty module, achieving complete isolation of the power supply on the faulty side, and preventing the fault from spreading.
[0101] Step T5: Status reporting and recovery waiting The MCU periodically reports the system status after the switch (including the list of currently active modules, the list of faulty modules, system runtime, etc.) to the integrated electronic equipment via the RS422 interface. The MCU continuously monitors the status of the faulty modules. If the faulty module recovers normal heartbeat response within a preset recovery window (e.g., 30 minutes), the MCU can reinstate it into the backup candidate queue.
[0102] 5. Power supply management process.
[0103] In this embodiment, the system power-on and power supply management process is as follows.
[0104] Step P1: Initial Power-On After receiving the 28V command power supply and the 42V bus power supply, the OC control circuit is in the default on state, the isolated DC / DC converter is working, and the main power module and the backup power module each output 12V voltage, which is converted into 3.3V to power the MCU.
[0105] Step P2: Power supply isolation between the primary power module and the backup power module The primary power module and the backup power module each output 12V voltage to the external integrated electronic equipment via telemetry signals. When the external integrated electronic equipment detects an abnormal voltage, it triggers a power management command, which shuts down the abnormal power module via an OC command signal, while keeping the normal power module on.
[0106] Step P3: Power on the primary and backup function modules. After the MCU module powers on and completes its self-test, the MCU sequentially enables the primary FMQL45T900 module, the primary CTC8120 module, the backup FMQL45T900 module, and the backup CTC8120 module via GPIO port control signals.
[0107] Step P4: State synchronization completed After the primary and backup modules are powered on, each backup module enters "backup monitoring" mode, waiting for status synchronization. The primary FMQL45T900 unit and the backup FMQL45T900 unit can then complete the switching chip configuration mirror synchronization via the 1000 Base-T interface. After synchronization is complete, the system enters normal dual-redundancy operation.
[0108] Step P5: Power supply fault isolation When the MCU detects an abnormality in any of the main and backup functional modules and triggers a master-slave switching instruction, the MCU sequentially shuts down the main FMQL45T900 module and the main CTC8120 module through GPIO control signals, while ensuring that the backup FMQL45T900 module and the backup CTC8120 module continue to output normally, thus ensuring uninterrupted power supply to the backup functional modules.
[0109] The foregoing description of several embodiments of this application has been provided for illustrative purposes. This foregoing description is not intended to be exhaustive, nor to limit the application to the precise steps and / or forms disclosed; obviously, many modifications and variations can be made in light of the teachings above. The scope of this application and all its equivalents are intended to be defined by the appended claims.
Claims
1. A dual-redundant 100G switching system with multi-module collaborative management, characterized in that, include: The primary power module and the backup power module use independent power supply paths to convert 42V DC input to 12V DC output, which power the primary functional module group and the backup functional module group respectively. The primary FMQL module and the backup FMQL module establish a bidirectional data communication link between them via a 1000 Base-T Ethernet interface; The primary 100G switching module and the backup 100G switching module both support 100Gbps high-speed Ethernet data switching and 1000 Base-T Ethernet data switching. The high-reliability MCU module, as an independent redundant management arbitration unit, is connected to the primary FMQL module, backup FMQL module, primary 100G switching module and backup 100G switching module respectively through management and control links. It is used to monitor the operating status of each module in real time, execute master-slave arbitration decisions and fault switching control.
2. The dual-redundant 100G switching system with multi-module collaborative management according to claim 1, characterized in that, The primary FMQL module is connected to the primary 100G switching module via a PCIe 2.0×1 high-speed interface, and the backup FMQL module is connected to the backup 100G switching module via a PCIe 2.0×1 high-speed interface. The configuration synchronization and status mirroring of the primary 100G switching module and the backup 100G switching module are completed through their respective FMQL modules via a PCIe 2.0 × 1 interconnect link. The high-reliability MCU module establishes management and control links with the primary FMQL module, backup FMQL module, primary 100G switching module, and backup 100G switching module through GPIO, IIC, and UART low-speed management interfaces, respectively.
3. The dual-redundant 100G switching system with multi-module collaborative management according to claim 1 or 2, characterized in that, The primary power module and the backup power module each include an OC control circuit, a soft-start circuit, an EMI filter circuit, an isolated DC / DC converter, and an output filter circuit. The OC control circuit receives OC command signals from external integrated electronic equipment and controls the enable / disable of the isolated DC / DC converter. The dual-redundant 100G switching system with multi-module collaborative management is configured to shut down the power supply channel of the abnormal power module by means of the OC command signal when the telemetry signals output by the primary power module and the backup power module are transmitted to the external integrated electronic device and the external integrated electronic device detects that one of the power modules is abnormal.
4. The dual-redundant 100G switching system with multi-module collaborative management according to any one of the preceding claims, characterized in that, Both the primary 100G switching module and the backup 100G switching module are implemented using the Shengke CTC8120 switching chip, providing a 100G Base-SR4 fiber optic interface and a 1000 Base-T electrical port to the outside. Furthermore, the 100G Base-SR4 fiber optic interface is connected to an external 100G fiber optic network, and the 1000 Base-T electrical port is connected downwards to the corresponding FMQL module.
5. A dual-redundant 100G switching system with multi-module collaborative management according to any one of the preceding claims, characterized in that, Both the primary and backup FMQL modules are implemented using the Fudan Microelectronics FMQL45T900 chip. The programmable logic portion of the Fudan Microelectronics FMQL45T900 chip implements a PCIe 2.0×1 interface controller, a Gigabit Ethernet MAC controller, and an LVDS high-speed serial interface controller. The ARM processor portion of the Fudan Microelectronics FMQL45T900 chip runs an embedded PetaLinux system, responsible for switch chip management, network protocol stack processing, and communication with the MCU module. Furthermore, the primary and backup FMQL modules establish a redundant synchronization link through a 1000 Base-T Ethernet interface, a GPIO interface, and an LVDS interface.
6. The dual-redundant 100G switching system with multi-module collaborative management according to any one of the preceding claims, characterized in that, The high-reliability MCU module is implemented using the Guoke Anxin AS32S601 microcontroller. It connects to the reset pin, status indicator pin, and interrupt pin of each module through the GPIO interface, connects to the sensors and EEPROM of each module through the IIC bus, communicates with each module for heartbeat detection through the UART serial port, and connects to external integrated electronic equipment through the RS422 interface.
7. The dual-redundant 100G switching system with multi-module collaborative management according to any one of the preceding claims, characterized in that, The high-reliability MCU module is configured to perform the following health management and master-slave switchover steps: Perform heartbeat detection according to the preset monitoring cycle: send heartbeat request frames to each functional module through the UART serial port, and receive response frames containing voltage, temperature and current status; Perform voltage and temperature monitoring: Read the key power rail voltage of each module through AD sampling, and read the temperature sensor data of each module through the IIC bus; Anomaly detection: When a module fails to return a heartbeat response for 3 consecutive times, the core temperature exceeds +105℃ or the case temperature exceeds +85℃, or the critical power rail voltage deviates from the nominal value by more than ±10%, the module is considered to be faulty. Master-Slave Switching: After confirming the fault, a fault alarm is sent to the external integrated electronic equipment, the primary 100G switching module port is shut down and the backup 100G switching module port is enabled, an activation command is sent to the backup FMQL module to switch to its operating mode, the power supply to the faulty module is shut down to achieve fault isolation, and the status after the switch is reported to the external integrated electronic equipment.
8. The dual-redundant 100G switching system with multi-module collaborative management according to claim 5, characterized in that, The dual-redundant 100G switching system with multi-module collaborative management is configured to execute the management and configuration steps of the FMQL module for the primary 100G switching module and the backup 100G switching module: Chip initialization: After the system is powered on, a reset release signal is sent to the switching chip through the PCIe 2.0×1 interface to configure the port rate mode, VLAN division rules and flow control policies; Switching entry distribution: Based on the forwarding entries generated by the network protocol stack, the MAC address table, IP routing table and ACL rules are written into the TCAM memory inside the switching chip through the PCIe interface; Status monitoring and dynamic adjustment: Periodically read the port statistics register of the switching chip, dynamically adjust the port load balancing strategy according to the number of bytes sent and received and the error frame rate, and issue a port isolation command when the error frame rate exceeds the threshold; Configure mirror synchronization: After the primary FMQL module completes the configuration, it sends the configuration data packet to the backup FMQL module through the 1000 Base-T interface, and the backup FMQL module performs the same configuration operation on the backup switching chip.
9. The dual-redundant 100G switching system with multi-module collaborative management according to any one of the preceding claims, characterized in that, The dual-redundant 100G switching system with multi-module collaborative management is configured to perform the following system power-on and power supply management steps: Initial power-on: Receives 28V command power and 42V bus power. The main and backup power modules output 12V voltage, which is converted to 3.3V to power the MCU module. Functional module power-on: After the MCU completes its self-test, it sequentially enables the primary FMQL module, the primary 100G switching module, the backup FMQL module, and the backup 100G switching module. Status synchronization: After the primary and backup FMQL modules complete the mirror synchronization of the switching chip configuration through the 1000 Base-T interface, the system enters a dual-redundancy operation state; Fault power supply isolation: When an abnormality is detected in the primary and backup functional modules, the power supply to the primary FMQL module and the primary and backup 100G switching modules is shut off to ensure uninterrupted power supply to the backup functional modules.
10. A dual-redundant 100G switching system with multi-module collaborative management according to any one of the preceding claims, characterized in that, The configuration of the dual-redundant 100G switching system with multi-module collaborative management adopts a homogeneous redundancy architecture, including two completely symmetrical functional chains on the primary side and the backup side. The functional chains on the primary side and the backup side achieve state synchronization and collaborative switching through high-speed data links and low-speed management links.
