Passive optical network ring network protection switching method and system based on double cascade tree

Abstract

This invention discloses a passive optical network ring protection switching method and system based on a dual-cascaded tree, relating to the field of optical communication network technology. The invention constructs a dual-cascaded tree ring topology by configuring two serial interfaces on a single root node; downlink dual-path synchronous broadcast of superframe data, with nodes relaying hop-by-hop; uplink dynamically switching receiving paths based on node identifiers, uniformly scheduling bandwidth resources; and combining distributed fault detection, automatically switching ports and reactivating disconnected nodes when a link is interrupted. The system includes a root node and multi-cascaded node devices, supporting hybrid optoelectronic cabling. This invention offers low networking cost, low transmission latency, and high reliability, meeting the needs of high-reliability communication scenarios such as industrial control and automotive electronics.

Description

Technical Field

[0001] This invention relates to the field of optical communication network technology, specifically to a passive optical network ring protection switching method and system based on a dual-cascaded tree. Background Technology

[0002] In current industrial automation, automotive electronics, and intelligent robotics fields, the connection between distributed sensors, servo controllers, camera equipment, and the main control unit mostly adopts point-to-point copper cable wiring, combined with Controller Area Network (CAN) bus or Ethernet control automation technology (EtherCAT) to achieve data transmission. With the rapid increase in sensor data bandwidth requirements and the increasingly stringent requirements for low-latency control, traditional copper cable solutions are gradually showing significant technical bottlenecks: First, copper cables have a low bandwidth limit, making it difficult to support the synchronous transmission of multiple high-definition sensor data streams. Signal attenuation becomes severe after transmission distances exceed tens of meters, failing to meet long-distance deployment needs. Second, copper cables have weak electromagnetic interference (EMI) resistance, easily leading to data errors and insufficient reliability in strong electromagnetic environments such as industrial sites and vehicle power compartments. Third, existing solutions typically require the simultaneous deployment of data networks and control buses, resulting in complex system architectures and high material and maintenance costs.

[0003] Passive Optical Network (PON) technology, with its advantages of high bandwidth, low latency, long distance, and resistance to electromagnetic interference, is gradually entering the fields of industrial and automotive communication. Traditional PONs adopt a point-to-multipoint tree topology, connecting optical line terminals (OLTs) and multiple optical network units (ONUs) through optical splitters. However, this topology has the risk of single point of failure—if the backbone fiber breaks, all downstream nodes will lose connection, which cannot meet the availability requirements of high-reliability industrial scenarios.

[0004] Existing optical network solutions with protection capabilities mainly fall into two categories: one is the dual-optical-line terminal ring network solution, which uses two core devices to build a ring topology for fault switching, but the equipment cost doubles, and it requires complex two-end coordination protocols, resulting in high end-to-end transmission latency and making it difficult to adapt to time-sensitive services; the other is the single-end protection tree solution, which only provides redundant protection for the backbone fiber and cannot cover branch link faults, thus limiting the protection scope. In addition, most existing solutions only support fiber optic transmission, which is costly for short-distance deployments and cannot flexibly adapt to the field requirements of mixed fiber optic and copper cabling.

[0005] Therefore, the industry urgently needs a passive optical network solution based on a single core device, supporting cascaded ring topology, having full-link fault protection, and compatible with hybrid optoelectronic cabling, to meet the industrial-grade communication requirements of high reliability and low latency while controlling costs. Summary of the Invention

[0006] To address the aforementioned shortcomings of existing technologies, this invention provides a passive optical network ring protection switching method and system based on a dual-cascaded tree. By configuring two serial deserializer interfaces on a single root node device, a dual-cascaded tree ring network topology is constructed. Combining dynamic bandwidth allocation, dynamic path switching, and distributed fault detection mechanisms, rapid protection switching for full-link faults is achieved under a low-cost architecture with a single core device. At the same time, it supports hybrid deployment of optical fiber and copper cable, taking into account high reliability, low transmission latency, and flexible networking capabilities.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] The passive optical network ring protection switching method based on a two-cascaded tree includes the following steps:

[0009] S1. Configure a media access control module and a dynamic bandwidth allocation module in a single root node device, and set two serial-to-parallel conversion transceiver interfaces as the primary cascade port and the backup cascade port respectively. Cascade multiple node devices to the two ports in sequence to form a bidirectional ring network topology. Define the node connected to the primary cascade path as the primary node and the node connected to the backup cascade path as the backup node.

[0010] S2, in the downlink direction, the root node broadcasts the same superframe data to all nodes on the corresponding cascade path simultaneously through the primary cascade port and the backup cascade port. Each cascade node forwards the received downlink data to its own node for local processing and continues to forward it to the next level node.

[0011] S3, in the uplink direction, the dynamic bandwidth allocation module allocates uplink burst transmission windows to all nodes. The root node dynamically selects the corresponding serial-to-parallel conversion transceiver interface to receive the burst data packets uploaded by the node according to the node type corresponding to the current burst window.

[0012] S4. When a link interruption fault is detected, the nodes on both sides of the fault location switch the uplink and downlink functions and clock source of their two serial-to-parallel conversion transceiver interfaces. The root node reactivates the disconnected node through the uninterrupted cascade path to complete the protection switching.

[0013] Preferably, the dynamic bandwidth allocation module allocates uplink burst transmission windows using either a group allocation method or an interleaved allocation method. In the group allocation method, burst windows are first allocated to all primary nodes and then to all backup nodes within each allocation cycle. In the interleaved allocation method, burst windows are allocated alternately to primary nodes and backup nodes in groups within each allocation cycle until the entire allocation cycle is filled.

[0014] Preferably, the root node initiates the discovery window process using two different allocation identifiers, which activate the primary node and the backup node respectively. At the same time, it allocates segmented node identifier ranges to the nodes, with the primary node allocated a low-value range identifier and the backup node allocated a high-value range identifier.

[0015] Preferably, the root node determines the node type based on the specified binary bits of the node identifier, thereby switching the uplink data receiving path and receiving clock to match the node transmission path corresponding to the current burst window.

[0016] Preferably, the detection of link interruption faults is divided into two methods: root node side and node side. The root node side judges the link fault by burst window loss indication, and the node side judges the link fault by downlink superframe out-of-synchronization detection or received signal loss detection. After the fault is triggered, the node automatically switches the port function and clock source of the serial-to-parallel conversion transceiver interface, and the root node initiates a fast activation process through the normal link to restore communication of the lost node.

[0017] A passive optical network ring protection switching system based on a dual-cascaded tree includes a root node device and multiple node devices.

[0018] The root node device includes a media access control module, a dynamic bandwidth allocation module, a primary serial-to-parallel conversion transceiver interface, a backup serial-to-parallel conversion transceiver interface, and a path selection control module. The media access control module connects to two cascaded paths through two serial-to-parallel conversion transceiver interfaces, forming a bidirectional ring topology with multiple node devices in sequence. The dynamic bandwidth allocation module is used to allocate uplink burst transmission windows to all nodes. The path selection control module is used to select the uplink data path and receiving clock of the corresponding serial-to-parallel conversion transceiver interface according to the node type corresponding to the burst window.

[0019] The node device includes two serial-to-parallel conversion transceiver interfaces, a cascade control module, a local media access control module, and a fault detection module. The two serial-to-parallel conversion transceiver interfaces are respectively connected to the upstream and downstream devices. The cascade control module is used to realize the forwarding of uplink and downlink data and the access of local data. The fault detection module is used to detect link interruption faults and trigger protection switching. During switching, the uplink and downlink functions and clock sources of the two serial-to-parallel conversion transceiver interfaces are switched.

[0020] Preferably, the root node device further includes a downlink clock adaptation module and a multiplexer. The downlink clock adaptation module is set on the backup downlink data path and works with the multiplexer to realize the clock domain conversion of downlink data, supporting the transmission of two downlink data paths under different clock domains.

[0021] Preferably, the cascaded control module of the node device includes an asynchronous first-in-first-out (FIFO) buffer and a recentering control unit; the asynchronous FIFO buffer is used to compensate for the jitter and rate difference of the transmit and receive clocks on the data forwarding path, and the recentering control unit is used to reset the read and write pointers of the buffer to a preset position at the start, end or during transmission of the local node burst window.

[0022] Preferably, the fault detection of the root node device is achieved through a sudden window loss indication, and the fault detection of the node device is achieved through downlink superframe out-of-sync detection or signal loss detection; after a fault occurs, the root node performs a fast activation process through an uninterrupted cascading path to restore the communication connection of the lost node.

[0023] Preferably, the cascaded links between node devices support optical fiber, copper cable, or a hybrid of both transmission media; node devices can be expanded to include additional node devices through inter-chip interfaces, sharing the same optical transceiver module to access the passive optical network.

[0024] Compared with the prior art, the beneficial effects of the present invention are:

[0025] This invention requires only a single root node device to construct a dual-cascaded tree ring topology, eliminating the need for dual-core device redundancy and significantly reducing system hardware costs. The ring topology provides full-link fault protection, ensuring that any single-point link interruption does not affect service operation, thus significantly improving network availability.

[0026] This invention employs a node-level data relay transparent transmission mechanism, where downlink data is forwarded step by step without complex routing processing, and uplink burst windows are precisely scheduled based on dynamic bandwidth allocation. The end-to-end transmission latency is much lower than that of traditional switched ring networks, meeting the latency requirements of time-sensitive control services.

[0027] This invention supports fiber optic, copper, and hybrid optoelectronic cabling. For long-distance scenarios, fiber optics are used to ensure bandwidth and anti-interference capabilities, while copper cables are used for short-distance scenarios to reduce deployment costs. It can be flexibly adapted to diverse deployment environments such as industrial production lines, vehicles, and robots.

[0028] This invention provides a dynamic path switching mechanism based on node identifiers, enabling unified scheduling of uplink bandwidth for two cascaded paths. It supports both group allocation and interleaved allocation modes, and can optimize scheduling strategies according to service models to improve bandwidth resource utilization.

[0029] This invention employs a distributed fault detection mechanism, where the root node and nodes independently detect link status. After a fault is triggered, the node automatically switches its port function, and the root node completes node reactivation through a fast discovery window. The failover time is short, and the service interruption duration can be controlled within milliseconds, ensuring the continuity of industrial control services. Attached Figure Description

[0030] Figure 1 This is a flowchart illustrating the passive optical network ring protection switching method based on a dual-cascaded tree as described in this invention.

[0031] Figure 2 This is a schematic diagram of the architecture of the passive optical network ring protection switching system based on a double-cascaded tree as described in this invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0033] Example 1

[0034] like Figure 1 As shown in the figure, this embodiment elaborates on the complete implementation process of the passive optical network ring protection switching method based on a dual-cascaded tree. The method is based on a time-sensitive passive optical network architecture, forming a closed ring topology by cascading a single root node device with multiple node devices, thereby achieving bidirectional data transmission and link fault protection switching. The entire method can be divided into four core steps: topology construction and node configuration, downlink data broadcast transmission, uplink bandwidth dynamic scheduling and data transmission, and fault detection and protection switching. These steps will be explained in detail below.

[0035] S1, Topology Construction and Node Configuration:

[0036] This step is the system initialization and topology setup phase. The core is to connect multiple node devices in series to form a closed ring network through the two-way serializer / deserializer (SerDes) interface of a single root node, and complete the node role definition, identifier allocation and basic parameter configuration, so as to provide the topology foundation for subsequent data transmission and protection switching.

[0037] The root node device corresponds to the optical line terminal (OLT) in a traditional passive optical network (PON). It is the core control unit of the entire network, integrating a single Media Access Control (MAC) module and a single Dynamic Bandwidth Allocation (DBA) module. Externally, it provides two independent SerDes interfaces, defined as the primary cascading port and the backup cascading port, respectively. Both ports support optical module access or direct copper cable connection, allowing selection of the transmission medium based on the deployment scenario. Multiple node devices are connected in series, with the first and last ends connected to the primary and backup cascading ports of the root node, respectively, ultimately forming a closed ring topology of "root node-node-node-…-root node".

[0038] At the logical level, all nodes cascaded from the primary cascade port along the first direction are defined as primary nodes, forming the primary cascade tree; all nodes cascaded from the backup cascade port along the opposite direction are defined as backup nodes, forming the backup cascade tree. Under normal operating conditions, both cascade trees are in operation simultaneously, jointly carrying network traffic to achieve load balancing and path redundancy.

[0039] To achieve node identification and dynamic path switching, the system adopts a segmented node identifier allocation mechanism: all node identifiers are divided into two consecutive intervals, with the lower value interval (0 to 63) allocated to primary nodes and the higher value interval (64 to 127) allocated to backup nodes. The 6th bit of the node identifier is used as a node type determination bit; when this bit is 0, it corresponds to a primary node, and when it is 1, it corresponds to a backup node. The root node can quickly determine the path to which a node belongs by extracting this bit, without the need for complex table lookup operations, significantly improving the path switching speed.

[0040] Correspondingly, the system sets two different activation allocation identifiers for node discovery and activation on the two paths: when the allocation identifier is FFFF, it corresponds to the discovery window of the primary path, and only the primary node responds to the discovery window and completes registration and activation; when the allocation identifier is FFFE, it corresponds to the discovery window of the backup path, and only the backup node responds to the discovery window and completes registration and activation. During the initial activation phase, the root node sequentially issues discovery windows with the two allocation identifiers, completing the registration, ranging, and identifier allocation for all nodes on both paths, verifying the connectivity and synchronization of the two paths.

[0041] Each node device is configured with two SerDes interfaces by default. One is defined as an uplink port, connecting to the upstream device facing the root node; the other is defined as a downlink port, connecting to the downstream device away from the root node. By default, the node uses the recovery clock received from the uplink port as the system master clock, and also uses this clock as the transmission clock for the downlink port. This enables hop-by-hop clock transmission along the cascaded path, ensuring clock synchronization throughout the entire link and reducing clock jitter across nodes.

[0042] The dual-cascaded tree ring topology constructed in this step, compared to traditional optical splitter-type passive optical networks, does not require the deployment of optical splitters. Nodes can be deployed point by point along the link, making it very suitable for chain-distributed equipment scenarios such as industrial production lines, robot joints, and vehicle bodies. At the same time, the dual paths naturally have redundancy capabilities. If any link fails, communication can be guaranteed through the other path, without the need to add additional core equipment, and the system material cost is significantly reduced.

[0043] S2, downlink data broadcast transmission:

[0044] This step describes the data transmission process in the downlink direction. The root node uses a point-to-multipoint broadcast mechanism to send the same superframe data to all nodes simultaneously through two cascaded paths. Nodes relay and forward the data level by level and extract local service data to achieve synchronous distribution of downlink data across the entire network.

[0045] The basic unit of downlink transmission is a superframe, which consists of two parts: a control field and a data field. The control field carries dynamic bandwidth allocation authorization information, synchronization timing information, discovery window identifier, and system configuration instructions. The data field carries service data packets in Gigabit Encapsulation Method (GEM) format, and can simultaneously carry multiple different types of service traffic. The MAC module of the root node generates unified superframe data and sends it to both the primary and backup downlink data paths.

[0046] Since the two SerDes interfaces belong to different clock domains, a downlink clock adaptation module (i.e., a downlink gearbox) is installed on the backup downlink data path to ensure synchronous transmission of superframes on both paths. This module performs data conversion across clock domains. Downlink data from the primary path is directly sent to the primary SerDes interface, while downlink data from the backup path undergoes bit width and clock domain conversion via the clock adaptation module before being sent to the backup SerDes interface. This ensures that the transmission timing of the two superframes is aligned, preventing excessive time differences between the two paths on the node side.

[0047] The downlink end-to-end transmission delay is composed of the link transmission delay and the node forwarding and processing delay, and its delay calculation formula is as follows:

[0048]

[0049] in, Let be the number of cascade levels from the root node to the target node. The transmission delay of a single-level link is determined by the length of the transmission medium and the speed of signal propagation; in optical fiber, it is approximately 5 microseconds per kilometer. For downlink forwarding processing latency of a single node, since the data adopts a transparent relay mechanism, it only needs to pass through SerDes and the buffer, and the processing latency can be controlled in the hundreds of nanoseconds.

[0050] The complete process for a node device to receive downlink data is as follows: The SerDes on the uplink port recovers the data and clock from the received optical or electrical signals, converts the serial data into parallel data, and sends it to the cascaded control module. The cascaded control module divides the data into two paths: one path is sent to the local MAC module, where the local MAC parses the superframe control domain information, extracts the GEM service data packets belonging to the current node, and delivers them to the upper-layer application; the other path is sent to the data relay path, where, after asynchronous buffering and clock synchronization, the SerDes on the downlink port converts it into a serial signal and forwards it to the next-level node device.

[0051] Because of the clock hop-by-hop transmission mechanism, the downlink transmission clock of each node comes from the uplink port's receive recovery clock. The clocks of the entire cascaded path remain from the same source, avoiding the cumulative jitter caused by multiple clock domains. The transmission jitter of downlink data can be controlled at the nanosecond level, fully meeting the requirements of high-precision synchronization services.

[0052] Under normal operating conditions, the primary and backup paths broadcast the same downlink data simultaneously. Nodes can receive downlink signals from both directions at the same time, selecting the path with better signal quality as the primary receiving path and the other as a backup. When the signal deteriorates in one path, the node can seamlessly switch to the other path without waiting for the root node to switch over, further improving the reliability of downlink transmission.

[0053] S3, Uplink bandwidth dynamic scheduling and data transmission:

[0054] This step is the data transmission process in the uplink direction. It adopts a point-to-point burst transmission mechanism. The dynamic bandwidth allocation module of the root node uniformly schedules the uplink transmission time of all nodes. Each node only sends data within the allocated burst window to avoid uplink signal conflicts between multiple nodes. The root node dynamically switches the uplink receiving path according to the node type corresponding to the burst window to achieve unified reception of uplink data from both paths.

[0055] The dynamic bandwidth allocation module operates at a fixed cycle, each cycle being called a DBA cycle, with a typical cycle length of 125 microseconds, adaptable to the low-latency scheduling requirements of industrial control applications. The DBA module supports two window allocation modes, which can be switched via register configuration according to the business scenario:

[0056] The first method is the group allocation mode: within each DBA cycle, uplink burst windows are first allocated continuously to all nodes of the primary path, and then uplink burst windows are allocated continuously to all nodes of the backup path. In this mode, the root node has fewer receive path switching times and lower path switching overhead, making it suitable for scenarios with high single-path service traffic and uneven traffic between paths.

[0057] The second method is the staggered allocation mode: within each DBA cycle, burst windows for primary and standby nodes are allocated alternately in groups. That is, a group of primary node windows is allocated first, followed by a group of standby node windows, and this process is repeated until the entire DBA cycle is filled. In this mode, the uplink scheduling of nodes is more even, and the maximum waiting latency of a single node is lower, making it suitable for latency-sensitive scenarios with a large number of nodes.

[0058] Uplink bandwidth utilization is determined by the proportion of effective data within a burst window, and its calculation formula is as follows:

[0059]

[0060] in, This represents the total number of burst windows within a single DBA cycle. For the first The effective data transmission duration for each window This represents the total length of the DBA cycle. Each burst window includes a preamble, control field, valid data field, and check field. The preamble and control field are overhead components; the larger the window, the lower the overhead percentage and the higher the bandwidth utilization.

[0061] The uplink reception on the root node side employs a dynamic path switching mechanism: the path selection control module obtains the node identifier corresponding to the current burst window from the DBA module, extracts the 6th bit of the node identifier to determine the node type; when it is determined to be a primary node, the primary SerDes interface's receive data path and receive clock are selected, causing the primary path's receive clock data recovery module to lock the current burst signal; when it is determined to be a backup node, the receiver switches to the backup SerDes interface's receive data path and receive clock. The entire switching process is automatically completed by hardware logic without software intervention, and the switching latency can be controlled at the nanosecond level, perfectly matching the rapid switching requirements of burst windows.

[0062] The uplink data transmission and relay on the node side follow these rules: When the local node obtains uplink burst window authorization, the cascading control module switches the uplink data path and sends the burst data packets generated by the local MAC module into the uplink SerDes interface to send to the next higher level node; during idle periods when the local node does not obtain authorization, the cascading control module releases the uplink data path and forwards the uplink burst data received from the downstream node on the downlink port directly to the next higher level node through the relay path, realizing transparent relay of uplink data from the downstream node.

[0063] To compensate for clock jitter and frequency offset in the relay path, an asynchronous First-In-First-Out (FIFO) buffer is set in the relay data path of the node. The minimum depth of the asynchronous FIFO is determined by the maximum clock frequency offset and the longest burst window duration, calculated as follows:

[0064]

[0065] in, The maximum relative frequency offset between the transmit and receive clocks. The longest emergency window duration. The bit rate of data transmission. This indicates a rounding up operation. By configuring an asynchronous FIFO of appropriate depth, it is possible to ensure that no data overflow or empty read errors occur under the maximum frequency offset and longest window, thus guaranteeing the reliability of relay data.

[0066] S4, Fault Detection and Protection Switching:

[0067] This step is the protection and recovery process after a link failure. It adopts a distributed fault detection mechanism, in which the root node and nodes independently monitor the link status. After a fault is triggered, port switching and node reactivation are automatically executed to achieve rapid service recovery and minimize service interruption time.

[0068] The system supports link fault types including fiber optic cable breakage, cable plugging / unplugging, and signal loss due to severe link degradation. Different detection mechanisms are used on the root node side and the node side.

[0069] The root node uses burst window loss detection to determine faults: it continuously monitors the reception status of each node's corresponding burst window. If it fails to receive uplink burst data from the corresponding node for several consecutive cycles, and after ruling out node-specific faults, it determines that the cascaded link in the corresponding direction has been interrupted. This detection method requires no additional detection packets, fully utilizes the burst window status of normal services, and does not consume additional bandwidth.

[0070] The node side achieves fault judgment through downlink superframe synchronization detection and signal loss detection: The node's fault detection module continuously monitors the received signal status and superframe synchronization status of the downlink port. When the received signal power is lower than the threshold or multiple consecutive superframes fail to synchronize, it determines that the downlink in the corresponding direction has been interrupted and triggers a hardware interrupt to the node's control unit.

[0071] When a link in the ring topology is interrupted, the two nodes on either side of the interruption point will detect the link failure and automatically perform a protection switching operation: the node switches its original downlink port to an uplink port, and its original uplink port to a downlink port. Simultaneously, it switches the source of the system master clock, changing from the original uplink port's recovery clock to the new uplink port's recovery clock. The cascading control module synchronously switches the direction of all data paths, switching the original uplink relay path to a downlink relay path, and the original downlink relay path to an uplink relay path. After the switching is completed, the two interrupted nodes communicate with the root node through the normal link on their respective sides. The entire ring topology becomes two independent cascading chains, with all nodes maintaining connection to the root node, and no node losing connection.

[0072] After the root node detects the loss of burst windows for some nodes, it determines that the corresponding link is faulty and immediately initiates a rapid activation process: It sends a shortened rapid discovery window through a normally functioning cascaded port, re-registers and re-ranges the disconnected nodes on the original faulty path side, and updates the node's path attributes and bandwidth allocation table. After reactivation, the DBA module re-plans the uplink burst windows for all nodes, and all nodes resume normal uplink communication.

[0073] The total duration of the entire protection switching process consists of three parts, calculated as follows:

[0074]

[0075] in, The fault detection duration depends on the configuration of the detection threshold and is typically several DBA cycles. The switching time between node ports and paths is implemented by hardware logic and is typically in the microsecond range. This reduces the time required for rapid activation and bandwidth reallocation of root nodes. By optimizing the rapid discovery window length and activation process, the total failover time can be controlled within milliseconds, fully meeting the business continuity requirements of industrial control scenarios.

[0076] After troubleshooting and link restoration, the system supports automatic or manual switchback modes: In automatic switchback mode, the node continuously monitors the link signal on the faulty side. When the signal recovers and stabilizes to the threshold, it automatically switches back to the original port role, and the root node synchronously restores the normal dual-path working mode; In manual switchback mode, maintenance personnel need to trigger the switchback operation through management commands to avoid frequent switching caused by link instability.

[0077] Example 2

[0078] like Figure 2 As shown, this embodiment details a passive optical network ring protection switching system based on a dual-cascaded tree that implements the above method. The system consists of one root node device and multiple node devices, connected by cascaded links to form a closed ring topology. It supports fiber optic, copper cable, or hybrid optoelectronic transmission media and can be widely used in industrial sensor networks, vehicle control networks, robot internal communication, and other scenarios. The hardware architecture, module functions, and working principles of the root node device and node devices are described in detail below.

[0079] Root node device:

[0080] The root node device is the core control and scheduling center of the entire network, corresponding to the optical line terminal in a traditional passive optical network. It is responsible for core functions such as downlink data encapsulation and broadcasting, uplink bandwidth scheduling and reception, node activation management, fault detection and protection control, etc. The root node device adopts a single MAC dual SerDes hardware architecture, realizing the management of dual cascaded trees under a single core processing unit, which significantly reduces chip area and hardware cost compared to the dual MAC solution.

[0081] The root node device mainly includes the following functional modules:

[0082] Media access control module:

[0083] The MAC module is the core data processing unit of the root node, responsible for the generation, encapsulation, and transmission of downlink superframes, and the parsing, verification, and forwarding of uplink GEM data packets. It also handles protocol processing functions such as node activation and registration, ranging calibration, and protection switching control. Internally, the MAC module integrates a transmit gearbox and a receive gearbox, used for bit width conversion and clock domain adaptation of downlink transmitted data and uplink received data, respectively, matching the serial transmission bit width and internal parallel data bit width of the SerDes interface.

[0084] Dynamic bandwidth allocation module:

[0085] The DBA module is the core of uplink bandwidth scheduling. It operates at a fixed period and supports both group allocation and interleaved allocation scheduling modes, which can be flexibly configured through control registers. The DBA module collects bandwidth request information from each node in real time, and, in conjunction with a preset service level policy, allocates an uplink burst window of the corresponding size to each node, generates authorization information, and sends it to all nodes via downlink superframes. Simultaneously, the DBA module outputs the node identifier and path type corresponding to each window to the path selection control module, serving as the basis for uplink path switching control.

[0086] Primary SerDes interface and backup SerDes interface:

[0087] Two independent SerDes interfaces correspond to the primary and backup cascade paths, respectively, serving as the physical interface between the root node and external links. Each SerDes interface supports high-speed serial data transmission and features a built-in Clock Data Recovery (CDR) module to recover the received clock and data from the received serial data stream. The transmitter supports parallel-to-serial conversion, converting internal parallel data into a high-speed serial signal output. Both SerDes interfaces support both optical module and copper cable interfaces, adaptable to different transmission media.

[0088] Path selection control module:

[0089] The path selection control module is the core of the uplink receive path control. It receives burst window node identifier information from the DBA module, extracts the 6th bit of the node identifier to generate a path selection signal, and controls the multiplexer group to complete the synchronous switching of the uplink data path, receive clock, and CDR enable. This module is implemented with pure hardware logic, resulting in extremely low switching latency. It can meet the high-speed switching requirements of burst windows, ensuring that the receive path has been switched and stabilized when each burst window arrives.

[0090] Downlink clock adapter module:

[0091] The downlink clock adaptation module, also known as the downlink gearbox, is located on the backup downlink data path and is used to convert data from the main clock domain to the backup SerDes clock domain. Since the reference clocks of the two SerDes paths may differ, the clock adaptation module performs buffered conversion between the data bit width and the clock domain, ensuring the timing synchronization of the two downlink superframes and preventing excessive time deviations between the two paths. This module supports bypass mode, and the integration of clock adaptation logic can be selected via register configuration.

[0092] Multiplexer group:

[0093] The root node integrates multiple multiplexers, including a downlink data selector, an uplink data selector, a transmit clock selector, and a receive clock selector. The downlink data selector selects whether the backup downlink path passes through the clock adaptation module; the uplink data selector selects whether the received data from the primary or backup SerDes is sent to the receive gearbox; the transmit and receive clock selectors select the corresponding operating clocks for the transmit and receive gearboxes, respectively. All selectors are uniformly controlled by the path selection control module or configuration register.

[0094] Fault detection module:

[0095] The fault detection module monitors the uplink's operational status, determining link connectivity by counting the number of successful and lost data receptions within each node's burst window. When uplink data is not received from the corresponding node for several consecutive cycles, the fault detection module triggers a fault interruption, notifying the control unit to initiate a protection switching process. This module also supports bit error rate statistics, monitoring link signal quality and providing early warnings of degradation trends before a complete link outage.

[0096] Control status register module:

[0097] The CSR module provides software configuration and status reading interfaces. The operating mode, parameter configuration, and running status of all modules can be implemented by reading and writing the CSR register. Configurable parameters include SerDes enable and reset, DBA scheduling mode, fault detection threshold, gearbox reset, path selection mode, protection switching enable, etc., giving the system high flexibility and customizability to adapt to different application scenarios.

[0098] The clock architecture of the root node device adopts a reference clock plus phase-locked loop scheme: an external high-precision reference clock is input, and after being multiplied by the internal phase-locked loop, the system operating clock and the SerDes reference clock are generated; the primary and backup SerDes share the same reference clock source to ensure the homogeneity of the two clocks and reduce clock deviation across paths.

[0099] Node devices:

[0100] Node devices are distributed access units in a network, corresponding to optical network units in traditional passive optical networks. They are responsible for the access, encapsulation, and uploading of local service data, and also serve as data relays, forwarding uplink and downlink data from upstream and downstream nodes. Each node device is equipped with dual SerDes interfaces and complete cascading control logic, allowing it to function as either an end node for service access or an intermediate node for relay functions, offering flexible networking options.

[0101] The node device mainly includes the following functional modules:

[0102] Dual SerDes interfaces:

[0103] Each node device is equipped with two identical SerDes interfaces, which are used as uplink and downlink ports by default and can be interchanged during fault switching. Each SerDes interface supports burst mode reception and transmission, adapting to the burst uplink transmission characteristics of passive optical networks; it has a built-in CDR module, which can recover the clock and data from the received serial data; it supports optical signal or electrical signal input and output, and can be equipped with optical modules or directly connected to copper cables depending on the transmission medium.

[0104] Cascaded control module:

[0105] The cascaded control module is the core of node data forwarding, responsible for switching uplink and downlink data paths, local data access, and relay data forwarding control. Internally, the module includes a data multiplexer, a clock multiplexer, and a data relay path, which can dynamically switch data flow directions according to the operating status: Under normal operation, uplink data is received from the downlink port, sent to the uplink port via the relay path, and downlink data is received from the uplink port, sent to the downlink port via the relay path; during fault switching, all data paths synchronously reverse direction to adapt to the port role switch.

[0106] Local MAC module:

[0107] The local MAC module is responsible for node protocol processing and local service access. In the downlink direction, it parses control information and GEM data packets in superframes, extracts local service data, and delivers it to upper-layer applications. In the uplink direction, it encapsulates local service data into GEM format and performs burst transmission according to the burst window time authorized by the DBA. The local MAC module also handles protocol interactions such as node activation, registration, ranging, and bandwidth requests, enabling nodes to join the network and operate normally.

[0108] Fault detection module:

[0109] The fault detection module is responsible for monitoring the link status on both sides of the node. By detecting indicators such as downlink superframe synchronization status, received signal strength, and bit error rate, it determines whether a link fault has occurred. When a link signal loss or continuous superframe synchronization failure is detected, the module immediately triggers a hardware interrupt, notifying the node control unit to initiate a protection switching process. This module also supports signal quality monitoring, which can detect link degradation trends in advance and report them to the root node for early warning.

[0110] Asynchronous FIFO buffer group:

[0111] The node internally sets up two sets of asynchronous FIFO buffers, corresponding to the uplink and downlink relay paths respectively. These buffers are used to achieve cross-clock domain conversion of relay data, compensate for jitter and frequency offset between the transmitting and receiving clocks, and prevent data transmission errors. The depth of the FIFO can be configured according to the application scenario. For long-distance, high-frequency-offset scenarios, a deeper FIFO can be configured to improve reliability, while for short-distance, low-latency scenarios, a shallower FIFO can be configured to reduce forwarding latency.

[0112] Re-centering control unit:

[0113] The recentering control unit is used to eliminate read / write pointer drift caused by long-term operation of the asynchronous FIFO, ensuring long-term reliability of data transmission. This unit can reset the asynchronous FIFO's read / write pointers to a preset intermediate position at the start and end of the local node's burst window or during window transmission, offsetting the accumulated pointer offset caused by clock frequency deviation. The recentering operation is completed during data gaps, does not affect normal data transmission, and can be executed periodically.

[0114] Clock management module:

[0115] The clock management module is responsible for managing and switching all clock domains within the node. By default, it selects the uplink SerDes CDR recovery clock as the system master clock, which also serves as the downlink SerDes transmission clock, enabling hop-by-hop clock propagation. When protection switching or port role swapping occurs, the clock management module automatically switches the clock source, selecting the new uplink port recovery clock as the system master clock to ensure continuous and stable clock operation during the switching process and prevent service data loss.

[0116] Laser control module:

[0117] For node devices using fiber optic media, a laser control module is configured to control the on / off state of the optical transmitter's laser, coordinating with uplink burst transmission. The module incorporates a programmable delay logic unit, which can precisely adjust the timing of the laser activation signal to match the transmission delay of the data path, ensuring accurate alignment between the laser activation and data output times and preventing optical signal deviations. Each node is equipped with laser activation input and output signals; when cascaded, the laser control timing can be passed step-by-step, adapting to the burst transmission synchronization requirements of multi-level repeaters.

[0118] The node devices also support expansion capabilities: additional child node devices can be added via inter-chip interfaces. Multiple child nodes can share the same master node's optical transceiver module to access the network, eliminating the need for each node to be configured with an optical module, significantly reducing the cost of densely deployed multi-node scenarios. The extended nodes communicate with the master node via an internal high-speed interface, using a unified GEM data format, which is completely transparent to the root node. The root node does not need to be aware of the extended nodes' existence, and the scheduling and management methods are consistent with ordinary nodes.

[0119] The entire system supports hybrid deployment of fiber optic and copper cables: fiber optic links are used in long-distance, high-interference environments to ensure transmission bandwidth and anti-interference capabilities; direct copper cable connections are used in short-distance, low-cost scenarios, eliminating the cost of optical modules. The system can flexibly combine the two media according to the site environment to achieve the optimal balance between performance and cost, making it highly suitable for complex deployment scenarios such as industrial production lines, automotive electronics, and intelligent robots.

[0120] The embodiments described above are merely examples of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application.

Claims

1. A passive optical network ring protection switching method based on a dual-cascaded tree, characterized in that, Includes the following steps: S1. Configure a media access control module and a dynamic bandwidth allocation module in a single root node device, and set two serial-to-parallel conversion transceiver interfaces as the primary cascade port and the backup cascade port respectively. Cascade multiple node devices to the two ports in sequence to form a bidirectional ring network topology. Define the node connected to the primary cascade path as the primary node and the node connected to the backup cascade path as the backup node. S2, in the downlink direction, the root node broadcasts the same superframe data to all nodes on the corresponding cascade path simultaneously through the primary cascade port and the backup cascade port. Each cascade node forwards the received downlink data to its own node for local processing and continues to forward it to the next level node. S3, in the uplink direction, the dynamic bandwidth allocation module allocates uplink burst transmission windows to all nodes. The root node dynamically selects the corresponding serial-to-parallel conversion transceiver interface to receive the burst data packets uploaded by the node according to the node type corresponding to the current burst window. S4. When a link interruption fault is detected, the nodes on both sides of the fault location switch the uplink and downlink functions and clock source of their two serial-to-parallel conversion transceiver interfaces. The root node reactivates the disconnected node through the uninterrupted cascade path to complete the protection switching.

2. The passive optical network ring protection switching method based on a dual-cascaded tree according to claim 1, characterized in that, The dynamic bandwidth allocation module allocates uplink burst transmission windows using either a group allocation method or an interleaved allocation method. In the group allocation method, burst windows are first allocated to all primary nodes and then to all backup nodes within each allocation cycle. In the interleaved allocation method, burst windows are alternately allocated to primary nodes and backup nodes in groups within each allocation cycle until the entire allocation cycle is filled.

3. The passive optical network ring protection switching method based on a dual-cascaded tree according to claim 1, characterized in that, The root node initiates the discovery window process using two different allocation identifiers, which activate the primary node and the standby node respectively. At the same time, it allocates segmented node identifier ranges to the nodes, with the primary node allocated a low value range identifier and the standby node allocated a high value range identifier.

4. The passive optical network ring protection switching method based on a dual-cascaded tree according to claim 3, characterized in that, The root node determines the node type based on the specified binary bits of the node identifier, and switches the uplink data receiving path and receiving clock accordingly to match the node transmission path corresponding to the current burst window.

5. The passive optical network ring protection switching method based on a dual-cascaded tree according to claim 1, characterized in that, Link interruption fault detection is divided into two methods: root node side and node side. The root node side judges the link fault by burst window loss indication, while the node side judges the link fault by downlink superframe out-of-synchronization detection or received signal loss detection. After the fault is triggered, the node automatically switches the port function and clock source of the serial-to-parallel conversion transceiver interface, and the root node initiates a fast activation process through the normal link to restore communication with the lost node.

6. A passive optical network ring protection switching system based on a dual-cascaded tree, characterized in that, Includes the root node device and multiple node devices; The root node device includes a media access control module, a dynamic bandwidth allocation module, a primary serial-to-parallel conversion transceiver interface, a backup serial-to-parallel conversion transceiver interface, and a path selection control module. The media access control module connects to two cascaded paths through two serial-to-parallel conversion transceiver interfaces, forming a bidirectional ring topology with multiple node devices in sequence. The dynamic bandwidth allocation module is used to allocate uplink burst transmission windows to all nodes. The path selection control module is used to select the uplink data path and receiving clock of the corresponding serial-to-parallel conversion transceiver interface according to the node type corresponding to the burst window. The node device includes two serial-to-parallel conversion transceiver interfaces, a cascade control module, a local media access control module, and a fault detection module. The two serial-to-parallel conversion transceiver interfaces are respectively connected to the upstream and downstream devices. The cascade control module is used to realize the forwarding of uplink and downlink data and the access of local data. The fault detection module is used to detect link interruption faults and trigger protection switching. During switching, the uplink and downlink functions and clock sources of the two serial-to-parallel conversion transceiver interfaces are switched.

7. The passive optical network ring protection switching system based on a dual-cascaded tree according to claim 6, characterized in that, The root node device also includes a downlink clock adaptation module and a multiplexer. The downlink clock adaptation module is set on the backup downlink data path and works with the multiplexer to realize the clock domain conversion of downlink data, supporting the transmission of two downlink data paths under different clock domains.

8. The passive optical network ring protection switching system based on a dual-cascaded tree according to claim 6, characterized in that, The cascaded control module of the node device includes an asynchronous first-in-first-out buffer and a recentering control unit; the asynchronous first-in-first-out buffer is used to compensate for the jitter and rate difference of the transmit and receive clocks on the data forwarding path, and the recentering control unit is used to reset the read and write pointers of the buffer to a preset position at the start, end or transmission of the local node burst window.

9. The passive optical network ring protection switching system based on a dual-cascaded tree according to claim 6, characterized in that, The fault detection of the root node device is achieved through a sudden window loss indication, and the fault detection of the node device is achieved through downlink superframe out-of-synchronization detection or signal loss detection. After a failure occurs, the root node performs a fast activation process through an uninterrupted cascading path to restore the communication connection of the lost nodes.

10. The passive optical network ring protection switching system based on a dual-cascaded tree according to claim 6, characterized in that, The cascaded links between node devices support transmission media such as optical fiber, copper cable, or a combination of both; node devices can be expanded to include additional node devices through inter-chip interfaces, sharing the same optical transceiver module to access the passive optical network.

Images