Multi-node communication dual ring network based on time sensitive network

Abstract

Embodiments of the present disclosure provide a multi-node communication double ring network based on a time-sensitive network, applied to the technical field of time-sensitive networks. The network comprises m time-sensitive network end switching devices connected in cascade, m being a positive integer and m>=3; wherein each time-sensitive network end switching device has both network terminal function and Ethernet switch function; the m time-sensitive network end switching devices are interconnected through respective Ethernet interfaces to build a double ring network, so as to form a physically completely isolated inner ring and outer ring double redundancy structure; each time-sensitive network end switching device comprises n logical ports, n being a positive integer and n>=5. In this way, the spanning tree convergence speed and network bandwidth utilization can be improved, and broadcast storm and infinite loop can be effectively prevented.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and more particularly to the field of time-sensitive networking technology, specifically to a multi-node dual-ring network based on time-sensitive networking. Background Technology

[0002] Ethernet ring topology is a network topology that connects multiple Ethernet switches in a closed ring structure via point-to-point links. The ring architecture has a simple and clear topology; compared to other complex network topologies, each node only needs to establish connections with two adjacent nodes. Data has the characteristic of cyclic transmission within the ring. When a node fails, data transmission is not interrupted but can cleverly find alternative paths to continue its journey through the network. This fault tolerance ensures that the network can maintain basic communication functions even in the face of partial failures, guaranteeing the continuity and stability of network services.

[0003] Frame Replication and Elimination for Reliability (FRER) technology in Time-Sensitive Networking (TSN) is a core mechanism developed to address the insufficient reliability of Ethernet in mission-critical systems such as industrial control, automotive electronics, and power protection. Its technological background can be understood from three dimensions: industry-driven needs, current Ethernet shortcomings, and the standard evolution path. It achieves carrier-grade reliability with minimal cost (only a small increase in header and replica bandwidth) while maintaining the openness and cost-effectiveness of Ethernet. In the TSN architecture, FRER, along with time synchronization and traffic scheduling, forms the "deterministic communication triangle," providing an indispensable communication foundation for fields such as the Industrial Internet, intelligent transportation, and energy digitalization.

[0004] However, current Ethernet is designed based on the Best-Effort model, which has the following fatal flaws: Single path dependency means that data is transmitted along only one path, and a failure of a link or node will cause service interruption. The recovery speed is slow, relying on the Spanning Tree Protocol (STP) or its improved version STP (Rapid STP, RSTP) protocol, with a convergence time of up to seconds (30-50s), which cannot meet the real-time requirements of industrial control, power protection and other scenarios; It lacks redundancy mechanisms and the 1+1 protection capability similar to that of telecommunications SDH / OTN; Low bandwidth utilization; under normal conditions, half of the links are idle (single ring, single blockage point).

[0005] Furthermore, while STP and its improved versions RSTP and MSTP (Multiple STP) are widely used in current Ethernet networks to prevent loops and broadcast storms, their inherent defects are becoming increasingly apparent in modern high-bandwidth, low-latency, and high-reliability network scenarios.

[0006] The main drawbacks of the STP protocol include: slow convergence speed, failing to meet the carrier-grade self-healing requirement of ≤50ms; building a loop-free topology by blocking redundant links, resulting in 50% or more of the physical links being idle; inability to achieve load balancing unless multiple MSTP instances are used, but the configuration is complex; STP is a pure Layer 2 protocol without service differentiation, with all VLANs sharing the same spanning tree; inability to provide differentiated protection for critical services such as control flow and video flow; and fault recovery applies a "one-size-fits-all" approach to all traffic, without supporting priority switching. Summary of the Invention

[0007] This disclosure provides a dual-ring network for multi-node communication based on time-sensitive networking and a method for multi-node communication based on time-sensitive networking.

[0008] According to a first aspect of this disclosure, a multi-node dual-ring network based on time-sensitive networking is provided. The network includes: Cascaded m time-sensitive network end-switching devices, where m is a positive integer and m≥3; Each time-sensitive network-end switch has both network terminal and Ethernet switch functions; m time-sensitive network-end switches are interconnected through their respective Ethernet interfaces to form a dual-ring network, creating a physically completely isolated inner and outer ring dual-redundancy structure; each time-sensitive network-end switch includes n logical ports, where n is a positive integer and n≥5.

[0009] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the time-sensitive network terminal switching device is used to control the data frame transmission time of each priority transmission queue according to a preset gating period; based on the credit value of each transmission queue, and according to a preset traffic scheduling algorithm, to schedule the data transmission rate of each priority transmission queue; to copy the data frame to be transmitted into four data frame replicas, and to simultaneously inject them into the dual-ring network in four directions: inner ring clockwise, inner ring counterclockwise, outer ring clockwise, and outer ring counterclockwise. The data in each sending queue is the data that the source host writes to the sending queue cache of the time-sensitive network switching device through a preset interface.

[0010] In addition to the aspects and any possible implementations described above, a further implementation is provided in which each priority sending queue satisfies a preset priority rule; The preset priority rules include: Use a polling method to query and determine whether the command buffer of the sending queue is not empty, whether the gating of the sending queue is enabled, and whether the credit value is greater than or equal to zero; When the transmission command buffer of the sending queue is not empty, the gating of the sending queue is enabled, and the credit value is greater than or equal to zero, the transmission command is read from the storage of the corresponding sending queue, the sending queue is marked as an active queue, and the transmission is scheduled. The transmission commands are tagged and stored in a cache corresponding to the queue priority.

[0011] In addition to the aspects and any possible implementations described above, a further implementation is provided in which time-sensitive network-end switching devices in the dual-ring network are time-synchronized so that the frequency and phase difference between adjacent time-sensitive network-end switching devices are kept within a preset error range.

[0012] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein controlling the data frame transmission time of each priority transmission queue according to a preset gating period includes: Time is divided into multiple cyclic periods, each containing multiple time slots; Within each time slot segment, a gating state is configured for each sending queue. When the gating is enabled, the sending queue of the corresponding priority is allowed to send data.

[0013] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein scheduling the data transmission rate of each priority sending queue based on the credit value of each sending queue and according to a preset traffic scheduling algorithm includes: Each sending queue maintains a credit value. When data transmission is idle, the credit value increases with an upward slope, and during data transmission, the credit value decreases with a downward slope. When the credit value is greater than or equal to zero, the corresponding priority sending queue is allowed to send data.

[0014] In addition to the aspects and any possible implementations described above, a further implementation is provided in which the lifetime value of a data frame is decremented by 1 each time it passes through an intermediate node in the dual-ring network, and the node directly discards the data frame when the lifetime value reaches zero. The Time-to-Live (TTL) value of the data frame is added to the header of the data frame by the source device; the initial value of the TTL value is set to the ring hop count, which is obtained by autonomously probing the number of time-sensitive network switching devices connected in series on the dual ring network using the link layer discovery protocol frame; the destination device removes the TTL field before receiving the data frame.

[0015] In addition to the aspects and any possible implementations described above, a further implementation is provided in which an intermediate node in the dual-ring network receives a data frame with a redundant tag; queries the media access control address table; if a media access control table entry is found and the destination port is the host port of the local connection, the data frame is sent to the host, and the flow table copy port does not participate in forwarding; if no media access control table entry is found, the flow table is queried for flow identification and flow copying, and the data frame is forwarded to the next level adjacent node; In this process, the destination device is configured with a frame elimination agent to remove redundant labels and send the original data frame to the destination host; the Media Access Control List (MAC) has a higher priority than the flow table replication port; the terminal device forwards the packet according to the MAC, while intermediate devices forward it according to the flow table.

[0016] In addition to the aspects and any possible implementations described above, a further implementation is provided in which the logical port types in the dual-ring network include internal ports and external ports, with the internal ports connected to the dual-ring network and the external ports connected to the host machine or external time-sensitive network devices. In this process, the network management processor of the time-sensitive network end switching device reads the configuration file and configures the parameters; when a data frame enters from an external port, it learns the associated table entries of the source medium access control address and port number; if a data frame with a matching source medium access control address but a mismatched port number is found, the current data frame is discarded.

[0017] In addition to the aspects described above and any possible implementation, a further implementation is provided in which the Ethernet data stream operating in the dual-ring network includes terminal communication data, network management data, time synchronization data, and data streams from external time-sensitive networks.

[0018] According to a second aspect of this disclosure, a multi-node communication method based on time-sensitive networking is provided. This method is applied to the aforementioned dual-ring network for multi-node communication based on time-sensitive networking. The method includes: Upon power-up, each time-sensitive network-end switching device in the dual-ring network synchronizes its time. The source device receives data written to the sending queue by the source host and sends the data to the intermediate device based on the sending schedule; Intermediate devices receive data sent by source devices, identify and copy the data stream, and then forward the data to the destination device to complete the intermediate forwarding. The destination device performs frame cancellation and sends the data to the destination host. The Ethernet data streams operating in the dual-ring network include terminal communication data, network management data, time synchronization data, and data streams from external time-sensitive networks. This application provides a multi-node dual-ring network based on Time-Sensitive Networking (TSN), comprising m cascaded TSN end-switches, where m is a positive integer and m≥3. Each TSN end-switches functions as both a network terminal and an Ethernet switch, transforming the TSN end-switches from a simple "switch + network card" combination into a revolutionary architecture of "deep integration of services and network." This makes the service terminal a "native node" of the ring network, eliminating last-hop uncertainty and ensuring deterministic guarantees for local and ring network flows. The m TSN end-switches are interconnected through their respective Ethernet interfaces to form a dual-ring network, creating a physically isolated inner and outer ring dual-redundancy structure. Each TSN end-switches includes n logical ports, where n is a positive integer and n≥5, thereby shortening the communication path. Since the TSN end-switches are part of the ring, their service flows have four-path redundancy, enabling zero-packet-loss communication even when any three paths are interrupted. Based on this, the network can improve spanning tree convergence speed, network bandwidth utilization, and effectively prevent broadcast storms and infinite loops.

[0019] It should be understood that the description in the Summary of the Invention is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0020] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. The drawings are provided for a better understanding of the invention and are not intended to limit the scope of this disclosure. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein: Figure 1 A network topology diagram of a multi-node communication dual-ring network based on time-sensitive networking according to an embodiment of the present disclosure is shown; Figure 2 A schematic diagram of a dual-ring network data flow communication process according to an embodiment of the present disclosure is shown; Figure 3 A flowchart of a multi-node communication method based on time-sensitive networking according to an embodiment of the present disclosure is shown. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0022] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0023] This disclosure provides a multi-node dual-ring network based on time-sensitive networking (TSN), comprising m cascaded TSN end-switches. Each TSN end-switches functions as both a network terminal and an Ethernet switch, enabling a "deep integration of services and network" architecture and ensuring deterministic guarantees for local and ring network flows. The m TSN end-switches are interconnected through their respective Ethernet interfaces to form a dual-ring network, creating a physically isolated inner and outer ring dual-redundancy structure. Each TSN end-switches includes n logical ports, thereby shortening the communication path, improving the spanning tree convergence speed and network bandwidth utilization, and effectively preventing broadcast storms and infinite loops.

[0024] Figure 1 A network topology diagram of a multi-node communication dual-ring network based on time-sensitive networking according to an embodiment of the present disclosure is shown.

[0025] In some embodiments, the above-described time-sensitive network-based multi-node dual-ring communication network includes: Cascaded m time-sensitive network end-switching devices, where m is a positive integer and m≥3; Each time-sensitive network-end switch has both network terminal and Ethernet switch functions; m time-sensitive network-end switches are interconnected through their respective Ethernet interfaces to form a dual-ring network, creating a physically completely isolated inner and outer ring dual-redundancy structure; each time-sensitive network-end switch includes n logical ports, where n is a positive integer and n≥5.

[0026] In some embodiments, the values ​​of m and n can be set according to the user's actual needs. For example, for ease of subsequent explanation, such as Figure 1As shown, m=8 and n=7, that is, the above TSN-based multi-node communication dual-ring network includes 8 TSN end-switching devices with 5 physical ports. Each TSN end-switching device contains 7 logical ports, and the 7 logical ports internally exchange data through a 7×7 switching matrix.

[0027] Ports 2 and 3 connect to the inner ring Ethernet interface, ports 1 and 4 connect to the outer ring Ethernet interface, port 5 is a service port connecting to the Ethernet interface of an external TSN device, port 6 is a management interface connecting to the internal network management processor, and port 7 is a terminal port connecting to the host machine's Peripheral Component Interconnect Express (PCIe) interface. Therefore, to prevent loops between the inner and outer rings, they are logically isolated using port isolation configuration.

[0028] In some embodiments, eight cascaded TSN end-switches are numbered 1-8 sequentially to form a dual-ring network, with the inner and outer rings physically isolated. TSN end-switches 1 and 8 are located at opposite ends of the dual-ring network, connected to host A and host H respectively. The inner ring returns to TSN end-switches 1 from TSN end-switches 1 via 2-3-4-5-6-7-8, and the outer ring returns to TSN end-switches 1 from TSN end-switches 1 via 8-7-6-5-4-3-2. The Ethernet data streams running in the dual-ring network include terminal communication data, network management data, time synchronization data, and data streams from external time-sensitive networks. The processing flow for these different types of data streams is the same in the dual-ring network.

[0029] As can be seen, the aforementioned dual-ring network enables deep integration of services and network. In other words, the TSN-end switching equipment is not a simple superposition of "switch + network card", but an architectural innovation of "deep integration of services and network".

[0030] Specifically, the TSN-end switching equipment combines the functions of a network terminal and a switch, making the service terminal a "native node" of the ring network. This eliminates last-hop uncertainty and achieves deterministic guarantees for local and ring network flows. This deterministic guarantee is subsequently implemented using the IEEE 802.1Qbv time-sensitive traffic scheduling protocol. Locally generated service flows directly complete subsequent IEEE 802.1CB frame copying, cancellation, and scheduling within the TSN-end switching equipment, without needing to go through the complex path between external network cards and switches, significantly reducing latency jitter in end-to-end communication.

[0031] In addition, the aforementioned dual-ring network simplifies the network hierarchy. The ring network consists of multiple TSN end-switching devices that function as both network terminals and switches, replacing the terminals and switches in a typical ring network topology. This eliminates the need for terminal-switch jumpers, reducing interfaces and points of failure, shortening communication paths, and decreasing hop count and latency.

[0032] In some embodiments, in a time-sensitive network-based multi-node dual-ring network, the source-end data transmission and replication mechanism of the TSN-end switching device includes the following steps: The source host writes data packets into the send queue buffer of the TSN switching device through a preset interface, such as the PCIe interface; The transmission scheduling module of the TSN end switching equipment controls the data frame transmission time of each priority transmission queue according to a preset gating period, such as the gating period of the IEEE 802.1Qbv (Time-Aware Scheduling) protocol. TSN switching equipment uses the preset traffic scheduling algorithm of the IEEE 802.1Qav (Traffic Shaping) protocol, such as the Credit-Based Shaper (CBS) algorithm, to schedule the data transmission rate based on the credit value of each sending queue. According to the IEEE 802.1CB (Frame Copying and Deletion) mechanism, the TSN switching equipment copies the data frame to be sent into four data frame copies. The TSN switching equipment simultaneously injects four data frame copies into the dual-ring network in four directions: inner ring clockwise, inner ring counterclockwise, outer ring clockwise, and outer ring counterclockwise.

[0033] In some embodiments, the end-switching device itself is part of the ring, and the service flow it generates naturally has the redundancy capability of four paths. Messages can reach the target through two clockwise paths or two counterclockwise paths. The network can achieve zero packet loss communication in scenarios where any three paths are disconnected.

[0034] In some embodiments, in a time-sensitive network-based multi-node dual-ring communication network, the time synchronization and traffic transmission scheduling mechanism of the dual-ring network includes the following steps: The TSN switching devices in the network synchronize their time using the IEEE 802.1AS (Time Synchronization) protocol, ensuring that the frequency and phase differences between adjacent TSN switching devices remain within a preset error range. For easy verification, the frequency and phase differences within the preset error range can be converted into time differences. The preset time error range can be set according to the user's actual needs, such as within ±50ns. The transmission scheduling module of the TSN end switching equipment divides time into multiple cyclic periods, each period containing multiple time slots, i.e., a preset gating period; Within each time slot segment, a gating state is configured for each sending queue. When the gating is enabled, the corresponding priority queue is allowed to send data. Each sending queue maintains a credit value. When data transmission is idle, the credit value increases with an upward slope, and during data transmission, the credit value decreases with a downward slope. When the credit value is greater than or equal to zero, the corresponding priority sending queue is allowed to send.

[0035] It should be noted that the above priority sending queues meet the preset priority rules, which can be set according to the user's actual needs.

[0036] For example, preset priority rules include: Use a polling method to query and determine whether the command buffer of the sending queue is not empty, whether the gating of the sending queue is enabled, and whether the credit value is greater than or equal to zero; When the transmission command buffer of the sending queue is not empty, the gating of the sending queue is enabled, and the credit value is greater than or equal to zero, the transmission command is read from the storage of the corresponding sending queue, the sending queue is marked as an active queue, and the transmission is scheduled. The transmission commands are tagged and stored in a cache corresponding to the queue priority.

[0037] As can be seen, the above method achieves deterministic scheduling, that is, the TSN end switching equipment implements time-based scheduling function and credit value-based scheduling function at the terminal interface.

[0038] Specifically, the time-based scheduling function enables local streams to be scheduled strictly according to a predetermined time, and to accurately schedule network traffic of different priorities or types to meet the critical business requirements such as deterministic communication, low jitter, and hard real-time. The credit-based scheduling function enables specific message queues to reserve bandwidth. When faced with sudden sending at the application layer, the logic layer will output smoothly according to the credit rhythm, reducing the instantaneous bandwidth of the output port, which is used to ensure the maximum bandwidth limit and low jitter transmission of soft real-time business streams (such as audio and video, status monitoring) on ​​the shared link.

[0039] At the same time, the above method also achieves integrated time synchronization and scheduling. For example, the terminal scheduling engine, TSN switching engine, and MAC layer of the end-switching device share the same hardware clock, which supports unified time scheduling of local data streams and forwarded data streams, avoiding cross-device time synchronization errors between network cards and switches in typical ring network topologies.

[0040] In some embodiments, the loop protection and storm suppression mechanism of a dual-ring network includes the following steps: The source device adds a Time To Live (TTL) field to the header of the data frame, and the initial value of TTL is set to the number of ring hops h. Each time a data frame passes through an intermediate node, the TTL value is decremented by 1. When the TTL value reaches zero, the node discards the data frame directly. Before receiving the data frame, the destination device removes the TTL field from the data frame to ensure compatibility with upper-layer protocols.

[0041] In some embodiments, port isolation configurations logically isolate the inner and outer rings from each other, preventing network loops from occurring between the inner and outer rings.

[0042] In some embodiments, the inner loop ports (port 1 and port 2) and the outer loop ports (port 3 and port 4) are isolated from each other to prevent loop storms caused by IEEE 802.1CB frame duplication; ports 1 and 2 are configured as hybrid ports, allowing passage through VLAN lists 1-6; ports 3 and 4 are configured as hybrid ports, allowing passage through VLAN lists 1-6; port 5 is configured as a service port, supporting VLAN tag pass-through.

[0043] In some embodiments, in a time-sensitive network-based multi-node dual-ring network, the data frame forwarding mechanism of the TSN-end switching device includes the following steps: The intermediate node receives a data frame with a redundancy tag (RTAG) from the ring network interface; The Media Access Control (MAC) address table is queried. If a MAC entry is found and the destination port is the host port of the local connection, the data frame is sent to the host. The port of the flow table copy does not participate in the forwarding. If no MAC entry is found, the flow table is queried for flow identification and flow replication, and the data frame is forwarded to the next-level adjacent TSN switching device. The destination device configures a frame cancellation agent to remove the RTAG tag and send the raw data frame to the destination host.

[0044] In some embodiments, the MAC table has a higher priority than the flow table replication port; the destination device forwards the packet according to the MAC table, while intermediate devices forward it according to the flow table.

[0045] In some embodiments, in a time-sensitive network-based multi-node dual-ring communication network, the storm suppression frame filtering mechanism includes the following steps: Upon power-up, the network management processor of the TSN-side switching device reads the configuration file and configures parameters such as IEEE 802.1AS, 802.1CB, 802.1Qav, and 802.1Qbv. The number of TSN end switching devices connected in series on the dual-ring network is autonomously detected using the Link Layer Discovery Protocol (LLDP) frames, which are used to exchange system information between network devices, to obtain the ring network hop count h. The logical ports of the TSN end switching devices in the dual-ring network are divided into internal ports and external ports. The internal ports are connected to the dual-ring network, and the external ports are connected to the host machine or external TSN devices. When a data frame enters from an external port, the association table entries for the source MAC address and port number are learned; If a data frame with a matching source MAC address but a mismatched port number is found, it means that this type of data frame has already been received on another external port, and the current data frame should be discarded.

[0046] To facilitate understanding, the following description uses the example of TSN switching device 1 sending data to TSN switching device 5 to illustrate the process of source-end transmission scheduling, frame duplication, intermediate forwarding, frame elimination, and destination reception. The data frame is sent from host A (the source host), copied into four copies by TSN switching device 1, and transmitted to TSN switching device 5 via the inner and outer rings respectively. Finally, the frame elimination agent of TSN switching device 5 eliminates the redundant copies and sends the original data to host B (the host connected to TSN switching device 5, i.e., the destination host).

[0047] Step 1: Ring network time synchronization.

[0048] like Figure 2 As shown, upon initial power-on, the TSN-end switching devices in the dual-ring network will perform time synchronization to keep the time difference after frequency and phase difference conversion between adjacent TSN-end switching devices within an error range of ±50ns.

[0049] Step 2: Host machine data transmission.

[0050] After time synchronization is complete, the application layer software on host A enters the working state. The application layer software calls the driver to queue and write data packets into the transmit queue of the logical layer of TSN-side switching device 1 through the PCIe interface.

[0051] Step 3: Send the schedule.

[0052] The transmit scheduling module of the TSN end-switching device 1 logic layer integrates the IEEE 802.1Qav and IEEE 802.1Qbv protocols in the TSN protocol, and implements two transmit scheduling functions on the terminal interface: time-based transmit scheduling function and credit value-based transmit scheduling function.

[0053] Among them, the time-based transmission scheduling function—an IEEE 802.1Qbv protocol module—ensures local flows are strictly scheduled according to predetermined times. At the terminal interface, local service flows are precisely scheduled based on time triggers, prioritizing network traffic of different priorities or types to meet critical service requirements such as deterministic communication, low jitter, and hard real-time performance. The time scheduling function ensures that data frames are sent within a pre-planned time window, avoiding competition for bandwidth resources with other service flows on the ring network and achieving end-to-end deterministic communication.

[0054] Credit-based send scheduling—an IEEE 802.1Qav protocol module—resers bandwidth reservation for specific message queues. When faced with sudden bursts of data transmission at the application layer, the logic layer's send scheduling module smoothly outputs data according to a credit-based rhythm, reducing the instantaneous bandwidth of the output port. Credit-based scheduling maintains a credit value for each queue, initially zero. During idle data transmission, the credit value increases with an increasing slope; during data transmission, it decreases with a decreasing slope. When the credit value is greater than or equal to zero, the corresponding priority queue is allowed to send. This credit-based scheduling mechanism ensures that in bursty traffic scenarios, the output port will not experience packet loss or latency jitter due to instantaneous overload.

[0055] It should be noted that the sending queue scheduling module generates transmission commands that include queue priority and tracking tags. First, the transmission commands are stored in the buffer corresponding to the queue priority. Then, a polling method is used to check whether the queue transmission command buffer is not empty, whether the queue's gating is enabled, and whether the credit value is greater than or equal to zero. When all three conditions are met, the transmission command is read from the corresponding queue's storage, and that queue is marked as an active queue and transmission is scheduled.

[0056] The transmission engine requests and obtains descriptors through Direct Memory Access (DMA). The descriptor contains a memory address and a data length. Data of the corresponding length is read from the host memory address via DMA and sent to terminal port 7 of the switching matrix.

[0057] Step 4: Frame copying.

[0058] After receiving this data frame at the ingress port 7 of TSN-side switching device 1, it first performs flow identification and flow replication. The frame replication port for this data frame is ring network ports 1-4. After passing through the switching matrix, it is sent to the egress port of ring network ports 1-4, where a specific RTAG tag is inserted to provide a unique identifier and sequence information for this data frame. The data frame then enters TSN-side switching devices 2 and 8 through the ring network link.

[0059] Step 5: Intermediate forwarding.

[0060] After receiving a data frame with an RTAG tag, ring network ports 3 and 4 of TSN-side switching device 2 first query the MAC table. Since the destination host of the data frame is not this device, no MAC table entry is found. Then, the flow table is queried for flow identification and flow replication. The frame replication port for this data is ring network port 1-2, and after passing through the switching matrix, it is sent to the exit of ring network port 1-2.

[0061] The data frame transmission process of TSN end switching devices 3, 4, 6 and 7 is the same as that of TSN end switching device 2. The data frames eventually enter TSN end switching device 5 connected to host B.

[0062] Step Six: Frame Cancellation and Destination Reception.

[0063] When TSN-side switching device 5 receives data frames with RTAG tags at ring network ports 1, 2, 3, and 4, it queries the MAC table and finds the MAC entry, indicating that the destination port is terminal port 7. After passing through the switching matrix, the data is sent to the egress of ring network port 7. Ring network port 7 is configured with a frame removal agent, which will remove the RTAG tags and send the original data frame to host B.

[0064] Furthermore, the following example, a dual-ring network consisting of eight TSN-end switching devices, will be used to illustrate the process of storm suppression.

[0065] Assuming a ring hop count of 8, upon power-up, the host machines are not yet operational. The network management processor inside the TSN-end switching device begins reading configuration files to configure parameters such as IEEE 802.1AS, 802.1CB, and IEEE 802.1Qav / bv. Simultaneously, it autonomously performs device probing using LLDP frames to identify the number of TSN-end switching devices connected in series on the ring network, i.e., the ring hop count is 8.

[0066] In a dual-ring network, the logical ports of all TSN-end switching devices are divided into two types: the four ports connecting the ring network (ports 1-4) are internal ports, and the other three ports (ports 5-7) are external ports.

[0067] The host machine sends data frames using the terminal port (external port), learning the source MAC address and port number of the data frame at the entry point of the external port. Simultaneously, a Time-to-Live (TTL) field is added to the data frame header, with an initial TTL of 8 ring hops. The data frame is then sent from the internal port onto the ring network.

[0068] The adjacent TSN end switching device receives the data frame from the internal port of the ring network. At the ingress, it uses the source MAC address and port number to perform a table lookup: if the lookup result is a match between the source MAC address and the port number, it means that the current type of data frame has been received on other external ports, and the current frame is discarded; if the lookup result is a mismatch between the MAC address and the port number, it means that the current type of data frame has not been received on other external ports, and then the time-to-live field in the data frame is checked.

[0069] If the time to live is zero, discard the current frame; otherwise, store and forward the current data frame. If the egress type is an internal port, decrement the time to live field of the data frame before sending it; if the egress type is an external port, remove the time to live field before sending it.

[0070] According to the above process, when a data frame is transmitted in the ring network, the TTL value is decremented by 1 for each node it passes through. When the TTL value is reduced to zero, the node directly discards the data frame, effectively preventing broadcast storms and infinite loops.

[0071] The above is an introduction to the network device embodiments. The following method embodiments will further illustrate the solution described in this disclosure.

[0072] Figure 3 A block flowchart of a multi-node communication method 300 based on time-sensitive networking according to an embodiment of the present disclosure is shown. The method 300 is applied to a dual-ring network for multi-node communication based on time-sensitive networking. Figure 3 As shown, the method 300 includes: S310. Upon power-up, each time-sensitive network-end switching device in the dual-ring network performs time synchronization. The S320 source device receives data written to the transmission queue by the source host and transmits the data to the intermediate device based on the transmission schedule. The S330 intermediate device receives data sent by the source device, performs data stream identification and copying, and sends the data to the destination device to complete the intermediate forwarding. The S340 destination device receives and forwards data, performs frame cancellation, and sends the data to the destination host. The Ethernet data streams operating in the dual-ring network include terminal communication data, network management data, time synchronization data, and data streams from external time-sensitive networks. It should be noted that those skilled in the art should understand that the embodiments described in this specification are optional embodiments, and the content involved is not necessarily essential to this disclosure.

[0073] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0074] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A multi-node dual-ring network for communication based on time-sensitive networking, characterized in that, include: Cascaded m time-sensitive network end-switching devices, where m is a positive integer and m≥3; Each time-sensitive network terminal switching device has both network terminal and Ethernet switch functions. m time-sensitive network-end switches are interconnected through their respective Ethernet interfaces to form a dual-ring network, creating a physically isolated inner and outer ring dual-redundancy structure; each time-sensitive network-end switch includes n logical ports, where n is a positive integer and n≥5.

2. The multi-node dual-ring network based on time-sensitive networking according to claim 1, characterized in that, The time-sensitive network terminal switching device is used to control the data frame transmission time of each priority transmission queue according to a preset gating period; based on the credit value of each transmission queue, it schedules the data transmission rate of each priority transmission queue according to a preset traffic scheduling algorithm; it copies the data frame to be transmitted into four data frame replicas, and injects them into the dual-ring network simultaneously in four directions: inner ring clockwise, inner ring counterclockwise, outer ring clockwise, and outer ring counterclockwise. The data in each sending queue is the data that the source host writes to the sending queue cache of the time-sensitive network switching device through a preset interface.

3. The multi-node dual-ring network based on time-sensitive networking according to claim 1, characterized in that, Each priority sending queue satisfies a preset priority rule; The preset priority rules include: Use a polling method to query and determine whether the command buffer of the sending queue is not empty, whether the gating of the sending queue is enabled, and whether the credit value is greater than or equal to zero; When the transmission command buffer of the sending queue is not empty, the gating of the sending queue is enabled, and the credit value is greater than or equal to zero, the transmission command is read from the storage of the corresponding sending queue, the sending queue is marked as an active queue, and the transmission is scheduled. The transmission commands are tagged and stored in a cache corresponding to the queue priority.

4. The multi-node dual-ring network based on time-sensitive networking according to claim 3, characterized in that, Time synchronization is performed on each time-sensitive network terminal switching device in the dual-ring network to keep the frequency and phase difference between adjacent time-sensitive network terminal switching devices within a preset error range.

5. The multi-node dual-ring network based on time-sensitive networking according to claim 3, characterized in that, The step of controlling the data frame transmission time of each priority transmission queue according to a preset gating period includes: Time is divided into multiple cyclic periods, each containing multiple time slots; Within each time slot segment, a gating state is configured for each sending queue. When the gating is enabled, the sending queue of the corresponding priority is allowed to send data.

6. The multi-node dual-ring network based on time-sensitive networking according to claim 3, characterized in that, The process of scheduling the data transmission rate of each priority sending queue based on its credit value and a preset traffic scheduling algorithm includes: Each sending queue maintains a credit value. When data transmission is idle, the credit value increases with an upward slope, and during data transmission, the credit value decreases with a downward slope. When the credit value is greater than or equal to zero, the corresponding priority sending queue is allowed to send data.

7. The multi-node dual-ring network based on time-sensitive networking according to claim 1, characterized in that, In the dual-ring network, the lifetime value of a data frame is decremented by 1 each time it passes through an intermediate node. When the lifetime value reaches zero, the node discards the data frame directly. The Time-to-Live (TTL) value of the data frame is added to the header of the data frame by the source device; the initial value of the TTL value is set to the ring hop count, which is obtained by autonomously probing the number of time-sensitive network switching devices connected in series on the dual ring network using the link layer discovery protocol frame; the destination device removes the TTL field before receiving the data frame.

8. The multi-node dual-ring network based on time-sensitive networking according to claim 1, characterized in that, The intermediate node in the dual-ring network receives a data frame with a redundant tag; it queries the media access control address table; if a media access control entry is found and the destination port is the host port of the local connection, the data frame is sent to the host, and the flow table copy port does not participate in forwarding; if no media access control entry is found, the flow table is queried for flow identification and flow copying, and the data frame is forwarded to the next level adjacent node. In this process, the destination device is configured with a frame elimination agent to remove redundant labels and send the original data frame to the destination host; the Media Access Control List (MAC) has a higher priority than the flow table replication port; the terminal device forwards the packet according to the MAC, while intermediate devices forward it according to the flow table.

9. The multi-node dual-ring network based on time-sensitive networking according to claim 1, characterized in that, The logical port types in the dual-ring network include internal ports and external ports. Internal ports connect to the dual-ring network, while external ports connect to the host machine or external time-sensitive network devices. In this process, the network management processor of the time-sensitive network end switching device reads the configuration file and configures the parameters; when a data frame enters from an external port, it learns the associated table entries of the source medium access control address and port number; if a data frame with a matching source medium access control address but a mismatched port number is found, the current data frame is discarded.

10. A multi-node communication method based on time-sensitive networking, applied to a dual-ring network for multi-node communication based on time-sensitive networking as described in any one of claims 1 to 9, characterized in that, The method includes: Upon power-up, each time-sensitive network-end switching device in the dual-ring network synchronizes its time. The source device receives data written to the sending queue by the source host and sends the data to the intermediate device based on the sending schedule; Intermediate devices receive data sent by source devices, identify and copy the data stream, and then forward the data to the destination device to complete the intermediate forwarding. The destination device performs frame cancellation and sends the data to the destination host. The Ethernet data streams operating in the dual-ring network include terminal communication data, network management data, time synchronization data, and data streams from external time-sensitive networks.

Images