An autonomous and controllable aggregation and switching device for power communication

Abstract

This invention discloses an autonomous and controllable aggregation switching device for power communication, applicable to the field of power communication technology. It includes two aggregation switches and multiple redundant paths between them. Each switch includes a service-aware engine, a transmit / receive processing module, a frame-by-frame cancellation controller, an adaptive replication selector, and a path health assessment module. The service-aware engine identifies the power service type. The adaptive replication selector dynamically determines the replication factor k based on service priority and path health. Upon first correctly receiving a frame, the receiver immediately sends a cancellation command in reverse order, and the sender stops redundant transmission of that frame on other paths. Simultaneously, the deduplication window capacity is dynamically adjusted based on the difference between estimated and actual delay. For bounded data blocks, the receiver maintains a bitmap and actively terminates transmission upon completion. When k=2, the two paths transmit in forward and reverse order, respectively. This invention significantly saves bandwidth, improves security, and meets the high reliability requirements of power communication.

Description

Technical Field

[0001] This invention relates to the field of power communication technology, and in particular to an autonomous and controllable aggregation and switching device for power communication. Background Technology

[0002] The power system is a critical national information infrastructure, and the security and reliability of its communication network are directly related to the safe and stable operation of the power grid. In the power communication network, aggregation and switching equipment is located between the backbone network and the access network, undertaking the task of data aggregation and forwarding from a massive number of terminal devices (such as substation protection devices, merging units, new energy station controllers, distribution automation terminals, etc.).

[0003] To ensure zero-packet-loss transmission of critical services (such as relay protection trip signals GOOSE and sampled values ​​SV), the IEEE 802.1CB standard defines a Frame Copy and Cancel (FRER) mechanism: the sending end copies the same data frame into multiple copies and transmits them through multiple independent paths; the receiving end receives the first arriving copy and discards the duplicate copies. This mechanism has been applied in fields such as aviation and industrial control.

[0004] However, existing FRER technology has the following shortcomings in power communication scenarios: Redundant bandwidth is wasted significantly: FRER requires the sender to fully copy all data frames. Even if the receiver has received one copy of a frame, identical copies on other paths will continue to be transmitted, consuming valuable link bandwidth. This waste is particularly pronounced in bounded block data transmission (such as protection setting download, fault waveform upload, configuration file synchronization, etc.).

[0005] The lack of a redundancy strategy tailored to different business needs stems from the vastly different requirements for reliability, latency, and bandwidth across various power services. GOOSE / SV demands ultra-high reliability and extremely low latency, while remote data services such as MMS / IEC 104 have lower redundancy requirements. The existing FRER employs a uniform replication strategy, which cannot be dynamically adjusted based on service type, leading to resource misallocation.

[0006] Single-path data integrity issue: In existing FRERs, each path transmits a complete copy of the data. An attacker only needs to intercept all the data from any one path to reconstruct the original information, thus expanding the attack surface due to multi-path transmission.

[0007] Path delay differences cause difficulties in receiver processing: In dual-path transmission, due to differences in link length and intermediate node processing, the arrival time difference between two copies of the same frame can reach tens of microseconds or even milliseconds. Chinese patent application CN116319534A discloses a seamless redundant transmission method based on an improved FRER (Frame Reset Execution Rate), which alleviates this problem to some extent by using a dynamic history table length. However, this scheme only adjusts based on the basic traffic attributes of the service flow and does not incorporate closed-loop calibration using the estimated path delay difference information provided by the sender. When the delay difference is too large, the receiver's fixed history window is prone to misjudging late-arriving valid copies as duplicate frames and discarding them, causing data corruption.

[0008] Unable to actively terminate after batch transmission is completed: For bounded data block transmission (total number of frames N is fixed), the existing FRER lacks a "transmission completion" judgment mechanism. The sender will continue to send all preset copies, causing unnecessary bandwidth occupation.

[0009] To address the aforementioned issues, this invention proposes an autonomous and controllable aggregation and switching device for power communication and its redundant transmission control method. Through innovative mechanisms such as service awareness, frame-by-frame cancellation, dynamic window adjustment, and forward / reverse transmission, it significantly improves bandwidth utilization and data security. Summary of the Invention

[0010] The core of this invention lies in the frame-by-frame cancellation mechanism and the mechanism of canceling the transmission of all remaining frames after obtaining a complete copy frame. This can significantly reduce the workload of the data transmission path and effectively release bandwidth. At the same time, the frame-by-frame cancellation ensures that the data in each transmission path is incomplete, effectively guaranteeing the security of data during multi-path transmission.

[0011] To solve the above problems, the present invention adopts the following technical solution.

[0012] An autonomous and controllable aggregation and switching device for power communication includes a first aggregation switch, a second aggregation switch, and multiple redundant transmission paths for data transmission between the two. The first aggregation switch and the second aggregation switch are respectively located at the data sending end and the data receiving end. Both the first aggregation switch and the second aggregation switch include a service awareness engine, a transmission processing module, a reception processing module, and a frame-by-frame cancellation controller. The frame-by-frame cancellation controller is signal-connected to the reception processing module and the transmission processing module, respectively. The autonomous and controllable aggregation and switching equipment also includes an adaptive replication selector and a path health assessment module. The adaptive replication selector is connected to the service awareness engine and the path health assessment module by signal, respectively.

[0013] Furthermore, the service awareness engine is used to identify the power service type to which the Ethernet data stream entering the device belongs. Power service types include at least: GOOSE, SV, PTP, MMS, IEC 104, and bounded block transport services.

[0014] Furthermore, the aggregation switch also includes a bounded block transmission controller, which is signal-connected to a frame-by-frame cancellation controller. When transmitting a data block with a predetermined total number of frames N, the bounded block transmission controller maintains a bitmap of length N. The bitmap is used to record the sequence numbers of correctly received frames. When it is detected that all bits in the bitmap have been recorded and marked, it is determined that the data block has been completely received, and a completion confirmation message is sent to the peer device through the reverse channel. After receiving the completion confirmation message, the transmission processing module immediately stops sending the remaining frames of the data block that have not yet been sent.

[0015] Furthermore, the estimated path delay difference information is extracted from the received duplicate frames and recorded as ΔT1. The actual arrival time difference ΔT2 between two duplicate frames with the same sequence number in the same data stream is recorded. The delay difference drift δ=|ΔT2-ΔT1| is calculated, and the frame elimination history window capacity W=W is dynamically adjusted accordingly. 基础 +α·δ; Among them, W 基础 α represents the base window capacity (the default size of the historical window used by the receiver for frame deduplication when the transmission delays of the two redundant paths are exactly the same (δ=0)). α represents the scaling factor, which is used to convert the delay difference drift δ into an increment of the window capacity, where α is a positive real number.

[0016] Furthermore, the path health assessment module calculates the comprehensive health score H for each path based on at least three dimensions: bit error rate m1, latency jitter m2, port queue depth m3, link utilization m4, and historical failure count m5. First, the bit error rate m1, latency jitter m2, port queue depth m3, link utilization m4, and historical fault count m5 are dimensionless and normalized, and uniformly mapped to the [0,1] interval. Bit error rate m1: The original value range is [0,1], which can be used directly; Time delay jitter m2: Normalized to m2′=min(m2 / M) 2,max ,1), where M 2,max This is the maximum allowable latency jitter threshold for the system. Port queue depth m3: Normalized to m3′=min(m3 / M) 3,max ,1), where M 3,max The maximum capacity of the port queue (determined by the device hardware specifications); Link utilization m4: The original value range is [0,1], which can be used directly; Historical failure count m5: Normalized to m5′=min(m5 / M) 5,ma x,1), where M 5,max The threshold for the maximum number of failures allowed within the statistical period; After normalization, the formula for calculating the overall health score H is: H=W1·(1-m1)+W2·(1-m2′)+W3·(1-m3′)+W4·(1-m4)+W5·(1-m5′); Where W1, W2, W3, W4, and W5 are configurable weights, satisfying ∑ i=1 5 Wi=1.

[0017] An autonomous and controllable aggregation and switching device for power communication, wherein its data transmission control method includes the following steps: Step S1: First, the service awareness engine identifies the power service type of the Ethernet data stream entering the device and obtains the service priority level; Step S2: The path health assessment module monitors the health status of multiple transmission paths in real time and calculates the comprehensive health score for each path; Step S3: The adaptive replication strategy selector dynamically determines the replication factor k of the data block to be transmitted (k is a positive integer, indicating that each data frame will be copied into k copies and sent on k paths respectively) based on the service priority level and the comprehensive health score of each transmission path, and selects k transmission paths. Step S4: The sending processing module divides the data block to be transmitted into N data frames with unique sequence numbers, and copies each data frame into k copies. Each copy carries the same sequence number and estimated path delay difference information. Then, the k copies are sent through the selected k paths respectively. Step S5: After the first aggregation switch at the receiving end correctly receives a frame with a certain frame sequence number for the first time, it immediately sends a copy frame cancellation command with the corresponding frame sequence number to the second aggregation switch at the sending end through the reverse channel. Step S6: Upon receiving the cancellation command, the second aggregation switch at the sending end immediately stops sending all unsent or currently being transmitted copy frames with the frame sequence number on the transmission path; Step S7: The aggregation switch at the data receiving end extracts the estimated path delay difference information from the received duplicate frames, dynamically adjusts the frame to eliminate historical window capacity based on the local actual arrival time difference, and performs deduplication and forwarding of duplicate frames based on the adjusted window.

[0018] Furthermore, the reverse channel is either an independent transmission path or at least one of multiple redundant paths.

[0019] Furthermore, the replication factor k satisfies: 1 ≤ k ≤ 3, and k is an integer. The specific value of k is dynamically determined according to the following rules: A1: When the service type is MMS or IEC 104, k=1; A2: When the service type is bounded block transmission, k=2; A3: When the business type is GOOSE, SV, or PTP: A31: When the overall health score H of all selected transmission paths is greater than or equal to the first threshold, then k=1; A32: When the health values ​​of all transmission paths are below the first threshold, and one or more transmission paths are above the second threshold, then k=2; A33: When H for all paths is less than or equal to the second threshold, then k = 3; Among them, the first threshold is ≥0.9, the second threshold is ≥0.7, and the first threshold is > the second threshold.

[0020] Furthermore, when k=2, the sending processing module sends each data frame in ascending order of sequence number on one transmission path, and sends each data frame in descending order of sequence number on another transmission path; when k=3, the sending processing module sends each data frame in random order of sequence number or in a polling manner on the third transmission path.

[0021] Compared with the prior art, the advantages of this invention are: (1) As long as the aggregation switch at the receiving end receives the copy frame corresponding to the complete sequence number, the data block is considered to have been completely received. At this time, even if the copy frames in the two transmission paths are not completely sent, the transmission of the remaining frames in the two transmission paths can be stopped, which greatly shortens the time to obtain complete data. At the same time, this embodiment terminates redundant transmission immediately after the batch data transmission is completed, which can save 30%-50% of bandwidth occupation and reduce the processing load of the sending end.

[0022] (2) By setting up a frame-by-frame cancellation mechanism, the data transmission of the transmission path ends after all duplicate frames are fully received. After the receiving end receives the first duplicate frame n, it can cancel duplicate frames n on other transmission paths in real time. The entire transmission process is also a process of continuously canceling the transmission of redundant duplicate frames. The sending end does not need to complete the transmission of all redundant copies, terminates invalid tasks in advance, and releases the sending queue, CPU, and memory resources. This enables the device to process more data streams simultaneously and improves the overall throughput.

[0023] (3) By setting up a dynamic window, the correctness of deduplication can be effectively guaranteed: by calculating the delay difference drift δ in real time, the window capacity is dynamically expanded to ensure that the late-arriving valid copy always falls within the window, so that it is correctly identified as a duplicate frame and discarded, avoiding data corruption or duplicate submission; at the same time, memory resources are effectively saved: when the delay difference is small, the window automatically shrinks to the basic capacity W. 基础 This effectively avoids excessive memory usage over a long period, adapts to the limited memory resources of power equipment (especially in scenarios with a large number of concurrent streams), and effectively avoids misjudging valid frames due to insufficient windows, indirectly improving the reliability of data transmission, which is especially crucial for critical services such as GOOSE that do not allow packet loss.

[0024] (4) The copy frames of different transmission paths are sent in forward and reverse order, which increases the hit rate of cancellation instructions from less than 50% to more than 90%, greatly reducing invalid redundant traffic and further improving the complete transmission speed of data blocks. Attached Figure Description

[0025] Figure 1 This is the main system block diagram of the present invention; Figure 2 This is a data transmission flowchart for bounded data services according to the present invention; Figure 3 This is a schematic diagram illustrating the frame-by-frame cancellation method used in the present invention when data is transmitted across multiple paths. Figure 4 This is a schematic diagram of forward and reverse transmission when there are two transmission paths in this invention; Figure 5 This is a schematic diagram illustrating the various transmission methods used when transmitting data through the three paths of this invention. Detailed Implementation

[0026] The technical solutions will now be clearly and completely described with reference to the accompanying drawings in the embodiments of the present invention.

[0027] First implementation method: This embodiment mainly focuses on data transmission of bounded data blocks, and the specific details are as follows: like Figures 1-2 An autonomous and controllable aggregation and switching device for power communication includes a first aggregation switch, a second aggregation switch, and multiple redundant transmission paths for data transmission between the two. The first aggregation switch and the second aggregation switch are respectively located at the data sending end and the data receiving end. Both the first aggregation switch and the second aggregation switch include a service awareness engine, a transmission processing module, a reception processing module, and a frame-by-frame cancellation controller. The frame-by-frame cancellation controller is signal-connected to the reception processing module and the transmission processing module, respectively. The service awareness engine is used to identify the power service type of the Ethernet data stream entering the device. When the power service type is identified as bounded block data transmission service, this embodiment is applicable for specific implementation.

[0028] The autonomous and controllable aggregation and switching equipment also includes an adaptive replication selector and a path health assessment module. The adaptive replication selector is connected to the service awareness engine and the path health assessment module respectively. The adaptive replication selector is used to determine the replication factor k based on the service type obtained from the service awareness engine and the health status of the transmission path obtained from the health assessment module. Generally, the value range of the replication factor k is: 1≤k≤3. When the service type is bounded data block transmission, k=2.

[0029] In this embodiment, since the service type is bounded data block transmission, k is fixed at 2, that is, there are two transmission paths between the first aggregation switch and the second aggregation switch. When transmitting data, the sending processing module divides the data block to be transmitted into N data frames with unique sequence numbers, and copies each data frame into two copies. Each copy carries the same sequence number and estimated path delay difference information. Then, the two copies are sent through the two selected paths respectively.

[0030] The aggregation switch also includes a bounded block transmission controller, which is signal-connected to a frame-by-frame cancellation controller. When transmitting a data block with a predetermined total number of frames N, the bounded block transmission controller maintains a bitmap of length N. The bitmap is used to record the sequence numbers of correctly received frames. When it is detected that all bits in the bitmap have been recorded and marked, that is, the receiving processing module on the second aggregation switch has received the copy frames of all sequence numbers, it is determined that the data block has been completely received. At this time, a completion confirmation message can be sent to the peer device through the reverse channel. After receiving the completion confirmation message, the sending processing module immediately stops sending the remaining frames of the data block that have not yet been sent. The reverse channel is an independent transmission path or at least one of multiple redundant paths.

[0031] Compared to traditional FRERs, where data transmission requires all data frames to be sent completely for a specific transmission path before stopping, this embodiment considers the data block to have been completely received as soon as the receiving end's aggregation switch receives the copy frame corresponding to the complete sequence number. At this point, even if the copy frames in both transmission paths have not been completely sent, the transmission of the remaining frames in both transmission paths can be stopped, significantly shortening the time to obtain complete data. In addition, this embodiment terminates redundant transmission immediately after the batch data transmission is completed, which can save 30%-50% of bandwidth usage and reduce the processing load on the sending end.

[0032] Second implementation method: This implementation introduces a more granular frame-by-frame cancellation mechanism, dynamic window adjustment, and adaptive replication factor decision based on path health, making it suitable for scenarios with higher requirements for real-time performance and bandwidth efficiency (such as critical services like GOOSE, SV, and PTP).

[0033] The workflow of this implementation method is as follows: Step 1: The business awareness engine identifies the business type (e.g., GOOSE). Simultaneously, the path health assessment module calculates a comprehensive health score H for each path based on at least three dimensions: bit error rate, latency jitter, port queue depth, link utilization, and historical failure count. Each indicator is first processed to be dimensionless and normalized, uniformly mapped to the [0,1] interval. Bit error rate m1: The original value range is [0,1], which can be used directly; Time delay jitter m2 (unit: μs): normalized to m2′=min(m2 / M) 2,max ,1), where M 2,max The maximum allowable latency jitter threshold for the system (configurable, default value is 1000μs). Port queue depth m3 (unit: number of packets): Normalized to m3′=min(m3 / M) 3,max ,1), where M 3,max The maximum capacity of the port queue (determined by the device hardware specifications); Link utilization m4: The original value range is [0,1], which can be used directly; Historical failure count m5 (unit: times): Normalized to m5′=min(m5 / M) 5,ma x,1), where M 5,max The maximum allowed number of failures per hour for the statistical period (configurable, default value is 10 times / hour).

[0034] After normalizing all indicators, the formula for calculating the overall health score H is: H=W1·(1-m1)+W2·(1-m2′)+W3·(1-m3′)+W4·(1-m4)+W5·(1-m5′) Where W1, W2, W3, W4, and W5 are configurable weights, satisfying ∑ i=1 5 Wi=1, and the default values ​​can be set to W1=0.25, W2=0.25, W3=0.20, W4=0.15, and W5=0.15. Each weight can be adjusted according to actual operation and maintenance experience.

[0035] Step 2: Decision on the adaptive replication factor k. The replication factor k satisfies: 1 ≤ k ≤ 3, and k is an integer. The specific value of k is dynamically determined according to the following rules: A1: When the service type is MMS or IEC 104, k=1; A2: When the service type is bounded block transmission, k=2; A3: When the business type is GOOSE, SV, or PTP: A31: When the overall health score H of all selected transmission paths is greater than or equal to the first threshold, then k=1; A32: When the health values ​​of all transmission paths are below the first threshold, and one or more transmission paths are above the second threshold, then k=2; A33: When H for all paths is less than or equal to the second threshold, then k = 3; Among them, the first threshold is ≥0.9, the second threshold is ≥0.7, and the first threshold is greater than the second threshold.

[0036] The third step, after determining the value of k based on the service type and overall health status, involves sending data and canceling it frame by frame. The specific details are as follows (taking a dual transmission path with k=2 as an example): like Figure 3 The sending end copies each data frame into two copies (a total of two data frames), and sends them through paths A and B. Each frame carries a sequence number n and an estimated path delay difference ΔT1. Assuming path A is faster, the copied frame n arrives at the receiving end first. After receiving the frame for the first time, the receiving end immediately sends a cancellation instruction C(n) through the reverse channel. Upon receiving C(n), the sending end immediately deletes all copies with sequence number n from the path queues and sends a cancellation instruction downstream.

[0037] By implementing a frame-by-frame cancellation mechanism, building upon the first implementation's conclusion that data transmission along a path ends upon complete reception of all duplicate frames, the receiving end can cancel duplicate frames n on other transmission paths in real time after receiving the first duplicate frame n. The entire transmission process is also a continuous process of canceling redundant duplicate frame transmissions. The sending end does not need to complete the transmission of all redundant copies, prematurely terminating invalid tasks and releasing transmission queues, CPU, and memory resources. This allows the device to process more data streams simultaneously, improving overall throughput.

[0038] Furthermore, in traditional FRER, each path transmits a complete copy of the data. An attacker only needs to intercept all the data from any one path to reconstruct the original information, and multiple transmission paths actually increase the attack surface.

[0039] In this scheme, due to the frame-by-frame cancellation, the actual set of frames transmitted on each path at any given time is a proper subset of the entire set (for example, path A actually transmitted frames 1, 3, 5..., and path B actually transmitted frames 2, 4, 6...). Even if an attacker intercepts a single path, they cannot obtain complete data; even if two paths are intercepted simultaneously, due to the real-time nature of the cancellation command, the data on the two paths are dynamically complementary rather than completely overlapping, making it extremely difficult to merge, thus significantly improving data security.

[0040] In the frame-by-frame cancellation mechanism, to ensure transmission reliability during the period when the receiving end sends back the cancellation command, the following processing logic is adopted: (1) After the receiving end correctly receives the data frame with sequence number n for the first time, it immediately adds the sequence number n to the local 'received set' and sends a cancellation instruction C(n) to the sending end through the reverse channel. (2) Before receiving confirmation of the cancellation instruction from the sender, the receiver will directly discard all subsequent duplicate frames with sequence number n and will not process or forward them again. (3) After receiving the cancellation instruction C(n), the sending end immediately stops sending all the copy frames with sequence number n on the path; for the copy frames that are already in the transmission (in-flight), since they cannot be intercepted midway, they will still reach the receiving end and be deduplicated and discarded by the 'received set' mechanism on the receiving end side. (4) Since the receiver only sends the cancellation command after the first correct receipt of a frame, at least one copy has been successfully received and stored, and no data loss will occur even if all other copies are discarded subsequently; (5) The 'received set' and the dynamic frame elimination history window work together: the window is used to identify duplicate frames within the normal latency range, and the 'received set' is used to identify and discard redundant frames during the cancellation instruction delay period and in extreme latency scenarios. The two complement each other and work together to ensure reliable data transmission and efficient deduplication under the frame-by-frame cancellation mechanism.

[0041] Step 4: Dynamic window adjustment. The receiving processing module extracts the estimated path delay difference information from the received duplicate frames and records it as ΔT1. It records the actual arrival time difference ΔT2 between two duplicate frames with the same sequence number in the same data stream. It calculates the delay difference drift δ=|ΔT2-ΔT1| and dynamically adjusts the frame elimination history window capacity W=W based on this. 基础 +α·δ; Among them, W 基础 α represents the base window capacity (the default size of the historical window used by the receiver for frame deduplication when the transmission delays of the two redundant paths are exactly the same (δ=0)). α represents the scaling factor, which is used to convert the delay difference drift δ into an increment of the window capacity, where α is a positive real number.

[0042] The estimated path delay difference information is generated by the sending end and embedded into the copied frame. The specific method for obtaining it is as follows: Sending end maintains path delay statistics: The sending end estimates the average transmission delay (or one-way delay) of each redundant path to the receiving end by periodically sending probe frames (such as the IEEE 1588 PTP delay request-response mechanism) or by measuring the round-trip time (RTT) of historical data frames. For example, the average delay of path A is D. A The average delay of path B is D. B .

[0043] Calculate the estimated delay difference: The sending end calculates the estimated path delay difference based on the currently used redundant path combination. For a dual-path scenario, the estimated delay difference is denoted as ΔT1 = |D|. A -D B | This value reflects the approximate time difference between two paths transmitting the same frame in the absence of sudden congestion.

[0044] Embedded Data Frame: When constructing each replicated frame, the transmission processing module writes the estimated delay difference information (which can be quantized as an integer or floating-point number) into a reserved field in the frame header (such as the reserved bit in the 802.1CB redundancy tag, or a custom Option field). The receiving end extracts this field to obtain ΔT1.

[0045] Dynamic update: The sending end can periodically or after each transmission update the path delay statistics based on the actual delay difference fed back by the receiving end (e.g., carried by cancellation command or ACK), thereby correcting the estimated delay difference of subsequent frames.

[0046] By setting up a dynamic window, the correctness of deduplication can be effectively guaranteed: by calculating the latency drift δ in real time, the window capacity is dynamically expanded to ensure that late-arriving valid copies always fall within the window, thus being correctly identified as duplicate frames and discarded, avoiding data corruption or duplicate submissions; at the same time, memory resources are effectively saved: when the latency difference is small, the window automatically shrinks to the basic capacity W. 基础 This effectively avoids excessive memory usage over a long period, adapts to the limited memory resources of power equipment (especially in scenarios with a large number of concurrent streams), and effectively avoids misjudging valid frames due to insufficient windows, indirectly improving the reliability of data transmission, which is especially crucial for critical services such as GOOSE that do not allow packet loss.

[0047] In addition, it can adapt to network fluctuations. In power communication networks, path delay may vary with load, electromagnetic interference, or equipment aging. The dynamic window does not require manual configuration and can automatically track changes in delay difference, improving system robustness. By adjusting the dynamic window, the redundant transmission deduplication mechanism can adapt to changes in path delay in real time, optimizing resource utilization while ensuring accuracy, and significantly enhancing the equipment's adaptability to different power communication environments.

[0048] Step 5: Deduplication and forwarding: The receiving end checks the subsequent arriving copies for duplicates based on the adjusted window: if the sequence number is already in the window, it is discarded; otherwise, it is forwarded and added to the window.

[0049] The third implementation method: This implementation method further optimizes the sending end's sequencing strategy based on the second implementation method, improves the hit rate of cancellation commands, and extends to three-path redundancy scenarios.

[0050] like Figure 4 Specifically, when k=2, the sending processing module sends each data frame in ascending order of sequence number on one transmission path, and sends each data frame in descending order of sequence number on the other transmission path. For example, for a data block of 100 frames: the sending order of path A is 1, 2, 3, ..., 100; the sending order of path B is 100, 99, 98, ..., 1.

[0051] Copy frame 1 is at the head of the queue in path A (sent earliest) and at the tail of the queue in path B (sent latest). The receiver will likely receive copy frame 1 from path A first, while copy frame 1 from path B is still waiting to be sent at the tail of the queue; a cancellation command can easily remove it. Similarly, copy frame 100 is sent earliest in path B. When the receiver receives copy frame 100 from path B first, copy frame 100 from path A has not yet been sent, resulting in extremely high cancellation efficiency. For intermediate frames, due to the staggered forward and reverse order, the cancellation hit rate is also much higher than for frames sent in the same order.

[0052] By sending data in forward and reverse order, the hit rate of cancellation commands was increased from less than 50% to over 90%, greatly reducing invalid and redundant traffic, and further improving the speed of complete data block transmission.

[0053] like Figure 5 Three-path transmission strategy (k=3): The transmission strategies for the first two transmission paths are the same as when k=2, that is, on one transmission path, each data frame is sent in ascending order of sequence number, and on another transmission path, each data frame is sent in descending order of sequence number; on the third transmission path, each data frame is sent in random order of sequence number or in a round-robin manner.

[0054] When path health is poor or service reliability requirements are extremely high, the adaptive replication strategy selector can enable three-path redundancy (k=3). The two transmission methods for the third transmission path are as follows: Random order: The third path sends frames in a randomly arranged order (ensuring each copied frame is sent once). The random order is not strongly correlated with forward or reverse order, further diversifying the risk.

[0055] Polling method: Divide the sequence numbers into several groups (e.g., 4 groups), and send one frame from each group in turn. Specifically, define the period as 4, and send sequence numbers 1, 26, 51, 76, 27, 52, 77... (adjusted according to N) in sequence. The polling method is simple to implement and can ensure that the transmission timing of each frame on the third path is evenly distributed.

[0056] In practice, the data transmission method of the third transmission path can be selected and set according to actual needs.

[0057] The above description is merely a preferred embodiment of the present invention; it encompasses all the protection scope of the present invention. Any equivalent substitutions or modifications made by those skilled in the art within the technical scope disclosed in the present invention, based on the technical solutions and improved concepts of the present invention, should be covered within the protection scope of the present invention.

Claims

1. An autonomous and controllable aggregation and switching device for power communication, characterized in that: It includes a first aggregation switch, a second aggregation switch, and multiple redundant transmission paths for data transmission between the two. The first aggregation switch and the second aggregation switch are respectively located at the data sending end and the data receiving end. Both the first aggregation switch and the second aggregation switch include a service awareness engine, a transmission processing module, a reception processing module, and a frame-by-frame cancellation controller. The frame-by-frame cancellation controller is signal-connected to the reception processing module and the transmission processing module, respectively. The autonomous and controllable aggregation and switching equipment also includes an adaptive replication selector and a path health assessment module. The adaptive replication selector is signal-connected to the service awareness engine and the path health assessment module, respectively. The aggregation switch also includes a bounded block transmission controller, which is signal-connected to a frame cancellation controller. When transmitting a data block with a predetermined total number of frames N, the bounded block transmission controller maintains a bitmap of length N, which is used to record the sequence numbers of correctly received frames. When it is detected that all bits in the bitmap have been recorded and marked, it is determined that the data block has been completely received, and a completion confirmation message is sent to the peer device through the reverse channel. After receiving the completion confirmation message, the sending processing module immediately stops sending the remaining frames of the data block that have not yet been sent. The receiving processing module extracts the estimated path delay difference information from the received duplicate frames and records it as ΔT1, records the actual arrival time difference ΔT2 between two duplicate frames with the same sequence number in the same data stream, calculates the delay difference drift δ=|ΔT2-ΔT1|, and dynamically adjusts the frame cancellation history window capacity W=W based on this. 基础 +α·δ; Among them, W 基础 α represents the base window capacity, and α represents the scaling factor, which is used to convert the delay drift δ into an increment of the window capacity, where α is a positive real number.

2. The autonomous and controllable aggregation and switching equipment for power communication according to claim 1, characterized in that: The service awareness engine is used to identify the power service type to which the Ethernet data stream entering the device belongs. The power service type includes at least: GOOSE, SV, PTP, MMS, IEC 104, and bounded block transport services.

3. The autonomous and controllable aggregation and switching equipment for power communication according to claim 1, characterized in that: The path health assessment module calculates the comprehensive health score H for each path based on at least three dimensions: bit error rate m1, latency jitter m2, port queue depth m3, link utilization m4, and historical failure count m5. First, the bit error rate m1, latency jitter m2, port queue depth m3, link utilization m4, and historical fault count m5 are dimensionless and normalized, and uniformly mapped to the [0,1] interval. Bit error rate m1: The original value range is [0,1], which can be used directly; Time delay jitter m2: Normalized to m2′=min(m2 / M) 2,max ,1), where M 2,max This is the maximum allowable latency jitter threshold for the system. Port queue depth m3: Normalized to m3′=min(m3 / M) 3,max ,1), where M 3,max The maximum capacity of the port queue (determined by the device hardware specifications); Link utilization m4: The original value range is [0,1], which can be used directly; Historical failure count m5: Normalized to m5′=min(m5 / M) 5,ma x,1), where M 5,max The threshold for the maximum number of failures allowed within the statistical period; After normalization, the formula for calculating the overall health score H is: H=W1·(1-m1)+W2·(1-m2′)+W3·(1-m3′)+W4·(1-m4)+W5·(1-m5′); Where W1, W2, W3, W4, and W5 are configurable weights, satisfying ∑ i=1 5 Wi=1.

4. The autonomous and controllable aggregation and switching equipment for power communication according to claim 3, characterized in that: Its data transmission control method includes the following steps: Step S1: First, the service awareness engine identifies the power service type of the Ethernet data stream entering the device and obtains the service priority level; Step S2: The path health assessment module monitors the health status of multiple transmission paths in real time and calculates the comprehensive health score for each path; Step S3: The adaptive replication strategy selector dynamically determines the replication factor k of the data block to be transmitted based on the service priority level and the comprehensive health score of each transmission path, and selects k transmission paths. Step S4: The sending processing module divides the data block to be transmitted into N data frames with unique sequence numbers, and copies each data frame into k copies. Each copy carries the same sequence number and estimated path delay difference information. Then, the k copies are sent through the selected k paths respectively. Step S5: After the first aggregation switch at the receiving end correctly receives a frame with a certain frame sequence number for the first time, it immediately sends a copy frame cancellation command with the corresponding frame sequence number to the second aggregation switch at the sending end through the reverse channel. Step S6: Upon receiving the cancellation command, the second aggregation switch at the sending end immediately stops sending all unsent or currently being transmitted copy frames with the frame sequence number on the transmission path; Step S7: The aggregation switch at the data receiving end extracts the estimated path delay difference information from the received duplicate frames, dynamically adjusts the frame to eliminate historical window capacity based on the local actual arrival time difference, and performs deduplication and forwarding of duplicate frames based on the adjusted window.

5. The autonomous and controllable aggregation and switching equipment for power communication according to claim 4, characterized in that: The reverse channel is an independent transmission path or at least one of multiple redundant paths.

6. The autonomous and controllable aggregation and switching equipment for power communication according to claim 5, characterized in that: The replication factor k satisfies: 1 ≤ k ≤ 3, and k is an integer. The specific value of k is dynamically determined according to the following rules: A1: When the service type is MMS or IEC 104, k=1; A2: When the service type is bounded block transmission, k=2; A3: When the business type is GOOSE, SV, or PTP: A31: When the overall health score H of all selected transmission paths is greater than or equal to the first threshold, then k=1; A32: When the health values ​​of all transmission paths are below the first threshold, and one or more transmission paths are above the second threshold, then k=2; A33: When H for all paths is less than or equal to the second threshold, then k = 3; Among them, the first threshold is ≥0.9, the second threshold is ≥0.7, and the first threshold is greater than the second threshold.

7. The autonomous and controllable aggregation and switching equipment for power communication according to claim 6, characterized in that: When k=2, the sending processing module sends each data frame in ascending order of sequence number on one transmission path, and sends each data frame in descending order of sequence number on another transmission path; when k=3, the sending processing module sends each data frame in random order of sequence number or in a polling manner on the third transmission path.

Images