An industrial internet of things-oriented high-reliability low-latency routing transmission method and system

Abstract

This invention discloses a highly reliable, low-latency routing and transmission method and system for the Industrial Internet of Things (IIoT), relating to the field of IIoT technology. The method includes: obtaining three protocol slices of a data stream and their parsing time sequence and structural deviation sequence based on a non-uniform slicing rule of protocol byte flipping frequency; calculating the coupling fluctuation index of each slice and extracting a structural editing cost index based on protocol field offset changes; obtaining a global accuracy index based on the cost index variance when the evolution judgment condition is met by calculating the covariance between the two indices; if it is greater than a confidence threshold, obtaining the phase compensation time based on the coupling fluctuation index, correcting the target protocol timestamp, and sending the data. This invention has the advantages of adaptive parsing, highly reliable transmission, and accurate spatiotemporal alignment.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial Internet of Things, and in particular to a highly reliable low-latency routing transmission method and system for the industrial Internet of Things. Background Art

[0002] With the advancement of intelligent manufacturing, the access of multi-source heterogeneous devices and the delivery of control instructions have become important contents in the fields of edge computing and routing transmission of the industrial Internet of Things. In industrial sites, the parsing of heterogeneous protocols and routing transmission face the situations of network physical environment interference and device firmware upgrade; the prior art usually relies on static protocol parsing templates or text comparison for data mapping, and this processing method is difficult to adapt to the dynamic load with data byte jump and the non-uniform distribution feature of static frame headers, which has the problem of low protocol extraction efficiency.

[0003] In addition, when parsing delay or abnormal data structure occurs in the industrial site, the existing system is difficult to separate the network fluctuation caused by noise attenuation at the physical layer of the network and link congestion from the protocol version evolution caused by device firmware upgrade and newly added monitoring fields, which easily misjudges network congestion as protocol change to trigger rule reorganization, or misjudges protocol evolution as communication interference to discard messages, resulting in reduced reliability of routing transmission. At the same time, the protocol conversion routing gateway will generate physical residence time when processing heterogeneous data, and the parsing and calculation time consumption will be superimposed on the transmission link of the control instruction, which delays the time when the downstream actuator receives the instruction; due to the lack of a prediction and timing correction mechanism for the dynamic time consumption of protocol parsing, asynchronous multi-unit equipment has space-time misalignment when receiving linkage control instructions. Therefore, there is a need for a routing transmission solution that can decouple network interference and protocol evolution and eliminate gateway residence delay. Summary of the Invention

[0004] Aiming at the technical problems in the background art, the present invention provides a highly reliable low-latency routing transmission method and system for the industrial Internet of Things.

[0005] A highly reliable, low-latency routing and transmission method for the Industrial Internet of Things (IIoT) includes: obtaining a first, second, and third protocol slice of a data stream based on a non-uniform slicing rule of protocol byte flipping frequency; obtaining the parsing time sequence and structural deviation sequence of each protocol slice in the previous parsing cycle; obtaining the coupling fluctuation index corresponding to each protocol slice based on the parsing time sequence and structural deviation sequence; obtaining a structural editing cost index based on the offset change of protocol fields; obtaining the index change characteristics of the coupling fluctuation index and structural editing cost index of each protocol slice; when the index change characteristics meet a preset evolution judgment condition, obtaining a global accuracy index based on the variance of the structural editing cost index; if the global accuracy index is greater than a preset confidence threshold, obtaining a phase compensation time based on the coupling fluctuation index, correcting the timestamp of the target protocol based on the phase compensation time, obtaining target data containing the corrected timestamp, and sending it out.

[0006] Optionally, based on the non-uniform slicing rules of protocol byte flip frequency, the first protocol slice, the second protocol slice, and the third protocol slice of the data stream are obtained, including: obtaining the number of data flips at each byte position in the data stream within the sliding time window, obtaining the protocol byte flip frequency based on the number of data flips; performing non-uniform slicing on the protocol data frame according to the protocol byte flip frequency, dividing the continuous byte segment with a protocol byte flip frequency less than a first frequency threshold into the first protocol slice, dividing the byte segment with a protocol byte flip frequency between the first frequency threshold and the second frequency threshold into the second protocol slice, and dividing the byte segment with a protocol byte flip frequency greater than the second frequency threshold into the third protocol slice.

[0007] Optionally, the parsing time sequence and structural deviation sequence of each protocol slice in the previous parsing cycle are obtained, including: extracting the actual decoding time of each protocol slice at each parsing node and constructing the parsing time sequence; and extracting the offset of each protocol slice in the byte structure relative to the baseline protocol template and constructing the structural deviation sequence.

[0008] Optionally, based on the analysis time sequence and structural deviation sequence of each protocol slice, the coupling volatility index corresponding to each protocol slice is obtained, including: obtaining the time difference between adjacent analysis points from the analysis time sequence of each protocol slice, and obtaining the deviation change between adjacent analysis points from the structural deviation sequence of each protocol slice; obtaining the relative time volatility by dividing the time difference by the average of the analysis time sequence of each protocol slice; obtaining the divisor by dividing the deviation change by the average of the structural deviation sequence of each protocol slice plus one, and performing an exponential mapping operation with the natural constant as the base and the divisor as the exponent to obtain the relative structural volatility; and multiplying the relative time volatility and the relative structural volatility to obtain the coupling volatility index corresponding to each protocol slice.

[0009] Optionally, the structural editing cost index is obtained based on the offset changes of the protocol fields, including: extracting the starting byte offset and field byte length of each semantic field in the base protocol and the protocol to be converted; calculating the absolute value of the offset difference and the absolute value of the length difference of each semantic field between the protocol to be converted and the base protocol; obtaining the sum of the absolute values ​​of the offset differences and the absolute values ​​of the length differences of all semantic fields, dividing the sum by the total byte length of the protocol data frame, and obtaining the structural editing cost index.

[0010] Optionally, when the covariance meets the preset evolution judgment condition, a global accuracy index is obtained based on the variance of the structural editing cost index, including: when the increase in the structural editing cost index of the first protocol slice or the second protocol slice is greater than the preset increase threshold and the coupling fluctuation index is less than the preset fluctuation threshold, it is confirmed that the preset evolution judgment condition is met; the reciprocal is calculated based on the variance of the structural editing cost index, and the reciprocal is normalized to generate a global accuracy index.

[0011] Optionally, the phase compensation time is obtained based on the coupling fluctuation index, the timestamp of the target protocol is corrected based on the phase compensation time, and the target data containing the corrected timestamp is obtained and sent out. This includes: generating a field reorganization mapping matrix of the protocol to be converted based on the mapping template and structure editing cost index of the baseline protocol; obtaining the phase compensation time based on the coupling fluctuation index of the third protocol slice and the average parsing dwell time of a single message; using the phase compensation time to perform a subtraction operation to correct the timestamp field of the target protocol data; reorganizing the data format based on the field reorganization mapping matrix, embedding the corrected timestamp field, obtaining the target data and sending it out.

[0012] Optionally, the phase compensation time amount is obtained based on the coupling fluctuation index of the third protocol slice and the average parsing dwell time of a single message, including: adding a value of one to the coupling fluctuation index of the third protocol slice to obtain a first weight; multiplying the average parsing dwell time by the first weight to obtain the phase compensation time amount; and using the phase compensation time amount to perform a subtraction operation to correct the timestamp field of the target protocol data, including: obtaining the current system absolute time, subtracting the phase compensation time amount from the current system absolute time to obtain the target timestamp and writing it into the timestamp field.

[0013] A highly reliable, low-latency routing and transmission system for the Industrial Internet of Things (IIoT) is also provided. The system includes a management platform, a sensor network platform, and an object platform connected sequentially. The management platform includes: a slicing module, used to obtain the first, second, and third protocol slices of the data stream based on non-uniform slicing rules of protocol byte flipping frequency; obtain the parsing time sequence and structural deviation sequence of each protocol slice in the previous parsing cycle; a feature calculation module, used to obtain the coupling fluctuation index corresponding to each protocol slice based on the parsing time sequence and structural deviation sequence; obtain the structural editing cost index based on the offset change of the protocol fields; a judgment module, used to obtain the index change characteristics of the coupling fluctuation index and the structural editing cost index of each protocol slice; when the index change characteristics meet the preset evolution judgment conditions, obtain the global accuracy index based on the variance of the structural editing cost index; and an output module, used to obtain the phase compensation time amount based on the coupling fluctuation index if the global accuracy index is greater than a preset confidence threshold, correct the timestamp of the target protocol based on the phase compensation time amount, obtain the target data containing the corrected timestamp, and send it out.

[0014] Optionally, the slicing module is also used to: obtain the number of data flips at each byte position in the data stream within the sliding time window, obtain the protocol byte flip frequency based on the number of data flips; perform non-uniform slicing on the protocol data frame according to the protocol byte flip frequency, divide the continuous byte segment with the protocol byte flip frequency less than the first frequency threshold into the first protocol slice, divide the byte segment with the protocol byte flip frequency between the first frequency threshold and the second frequency threshold into the second protocol slice, and divide the byte segment with the protocol byte flip frequency greater than the second frequency threshold into the third protocol slice.

[0015] The beneficial effects of this invention are reflected in: In the entire highly reliable, low-latency routing and transmission method for the Industrial Internet of Things (IIoT), firstly, by using a non-uniform slicing rule based on the protocol byte flip frequency, the dependence on preset templates is reduced. Static frame headers, function codes, and dynamic payloads are stripped based on data jump probability, improving the feature localization efficiency of heterogeneous protocols. Secondly, a coupling fluctuation index and a structural editing cost index are constructed, mapping time delay and spatial structural deviation to a two-dimensional topological phase space. A mean-normalized exponential model is used to quantify the impact of structural jumps on the computational power consumption of parsing, providing data support for fault state assessment. Furthermore, interference decoupling and evolution boundary determination are performed by calculating the covariance of the spatiotemporal dual indices, distinguishing between network physical layer noise attenuation and device firmware upgrades. When a firmware upgrade is confirmed, a reassembly mechanism is initiated, reducing resource consumption and packet dropping, and improving the reliability of routing transmission. Finally, a spatiotemporal phase compensation mechanism for asynchronous multi-unit protocols is introduced. The average parsing dwell time of the edge gateway is calculated using the coupling fluctuation index, and the timestamp of the target protocol data is corrected, offsetting the communication delay generated during protocol conversion. This ensures that the linkage control commands are aligned in time phase, meeting the requirements of collaborative control. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0017] Figure 1 This is a schematic diagram illustrating the steps of the highly reliable, low-latency routing and transmission method for the Industrial Internet of Things according to the present invention. Figure 2 This is a schematic diagram of some steps in S1 of the highly reliable, low-latency routing and transmission method for the Industrial Internet of Things of the present invention. Figure 3 This is a schematic diagram of a portion of step S2 in the highly reliable, low-latency routing and transmission method for the Industrial Internet of Things of the present invention; Figure 4 This is a schematic diagram of another part of step S2 in the highly reliable and low-latency routing transmission method for industrial IoT of the present invention; Figure 5 This is a schematic diagram of some steps in S3 of the highly reliable, low-latency routing and transmission method for the Industrial Internet of Things of the present invention; Figure 6 This is a schematic diagram of part of step S4 in the highly reliable, low-latency routing and transmission method for industrial IoT of the present invention; Figure 7 This is a schematic diagram of the composition of the highly reliable, low-latency routing transmission system for the Industrial Internet of Things according to the present invention; Figure 8 This is a schematic diagram of the optimized industrial Internet of Things (IoT) involved in this invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0020] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] This invention provides a highly reliable, low-latency routing and transmission method for the Industrial Internet of Things (IIoT), such as... Figure 1 As shown, in one embodiment, the method includes: S1. Based on the non-uniform slicing rules of protocol byte flipping frequency, obtain the first protocol slice, the second protocol slice, and the third protocol slice of the data stream; obtain the parsing time sequence and structural deviation sequence of each protocol slice in the previous parsing cycle of the current parsing cycle; S2. Based on the parsing time sequence and structural deviation sequence of each protocol slice, obtain the coupling fluctuation index corresponding to each protocol slice; obtain the structural editing cost index based on the offset change of the protocol field. S3. Obtain the coupling fluctuation index and structural editing cost index of each protocol slice and their index change characteristics; when the index change characteristics meet the preset evolution judgment conditions, obtain the global accuracy index based on the variance of the structural editing cost index. S4. If the global accuracy index is greater than the preset confidence threshold, the phase compensation time is obtained according to the coupling fluctuation index, the timestamp of the target protocol is corrected according to the phase compensation time, the target data containing the corrected timestamp is obtained and sent.

[0022] In this embodiment, it should be noted that in S1, existing solutions typically rely on preset templates to parse the protocol, which is difficult to adapt to industrial environments where data structures change frequently. S1 does not rely on a fixed frame header or payload structure, but instead uses a sliding time window (e.g., 1000 milliseconds) to count the frequency of state flips at each byte position in the protocol data.

[0023] Specifically, the number of flips for each byte within the recorded time window is divided by the total number of samples. When the flip frequency of a byte segment is less than 0.01 (e.g., only flipped twice, frequency 0.002), it is identified as a static frame header and classified as the first protocol slice; when the flip frequency is between 0.01 and 0.5 (e.g., 0.15), it is classified as the second protocol slice, representing the function code; when the flip frequency is greater than 0.5 (e.g., 0.85), it is identified as dynamic payload data and classified as the third protocol slice. Subsequently, the actual decoding time of each slice in each parsing iteration is extracted to construct a parsing time sequence (e.g., 1.2, 1.5, 1.3 milliseconds, etc.), and the structural offset is extracted to construct a structural deviation sequence. This process transforms the protocol structure into a physical jump probability distribution, achieving adaptive localization of heterogeneous protocol slices and improving the flexibility of feature extraction.

[0024] In S2, the focus is on quantifying the impact of protocol structure jumps on parsing computational power consumption and calculating the memory reorganization cost of protocol conversion. To avoid errors caused by simply linearly adding the time difference and structure difference, a coupling volatility index and a structure editing cost index are constructed. For the coupling volatility index, the time difference and structural deviation change between adjacent parsing points are first obtained.

[0025] Subsequently, the relative time volatility (e.g., 0.357) is obtained by dividing the time difference (e.g., the absolute value of the difference between 1.8 and 1.3) by the average value of the time series (e.g., 1.4 milliseconds). Simultaneously, the relative structural volatility (e.g., 1.4918) is obtained by dividing the change in structural deviation (e.g., the absolute value of the difference between 12 and 8) by the average value of the structural deviation series plus one (e.g., 8 plus 1 equals 9 bytes), and then using this as an exponent for mapping operations. Multiplying these two results in the coupled volatility index (e.g., 0.6527), which reflects the nonlinear amplification effect of structural changes on time delay. For the structural editing cost index, the sum of the absolute values ​​of the differences in the offsets and lengths of each semantic field between the protocol to be converted and the baseline protocol is calculated and normalized by dividing by the total protocol length (e.g., 64 bytes) (resulting in 0.09375). These calculations, by introducing mean normalization and an exponential model, objectively reflect the resistance state and conversion overhead of heterogeneous protocol parsing, providing data support for subsequent state assessment.

[0026] In S3, the aim is to decouple network physical layer noise attenuation from protocol changes caused by device firmware upgrades at the underlying logical level. It continuously monitors the covariance and numerical characteristics of the coupling fluctuation index and structural editing cost indicators of each protocol slice.

[0027] When network congestion or link jitter occurs, the coupling fluctuation index will rise while the structure editing cost index will remain close to zero. This indicates a transmission anomaly, and protocol mapping changes will not be performed.

[0028] When a device firmware upgrade results in the addition of new monitoring fields, it will detect an increase in the structural editing cost index of the first or second protocol slice (e.g., 0.09375) exceeding a preset increase threshold (e.g., 0.05). Simultaneously, the coupling fluctuation index will fluctuate briefly before falling back and remaining below a preset fluctuation threshold (e.g., 0.2) (e.g., 0.12). Based on this characteristic, it is determined that the evolution condition is met, confirming a protocol version upgrade. At this point, the variance of the structural editing cost index of each slice is extracted (e.g., 0.001), its reciprocal (e.g., 1000) is taken, and after normalization, a global accuracy index (e.g., 0.92) is generated. This judgment logic distinguishes between network interference and protocol evolution, avoiding resource waste and packet loss caused by misjudgment, and improving the decision reliability of the gateway in complex network environments.

[0029] In S4, the impact of physical dwell time on the phase of control commands caused by edge computing gateways processing heterogeneous data is eliminated.

[0030] When the global accuracy metric (e.g., 0.92) exceeds the confidence threshold (e.g., 0.85), the spatiotemporal phase compensation mechanism is activated. First, a field reassembly mapping matrix for the protocol to be converted is generated based on the baseline protocol template and the structural editing cost metric. To calculate the phase compensation time, the average parsing dwell time of a single message is obtained (e.g., 45 milliseconds), and the coupling fluctuation index of the third protocol slice (e.g., 0.58) is extracted. This index is then multiplied by a value of 1.58 to obtain the first weight, which is then multiplied by the average dwell time to calculate the phase compensation time (e.g., 71.1 milliseconds). When encapsulating the target protocol data, the current system relative runtime (e.g., 50000 milliseconds) is obtained, and the aforementioned phase compensation time is subtracted to obtain the corrected target timestamp (e.g., 49928.9 milliseconds). Finally, the target message is formatted using the field reassembly mapping matrix, and this corrected timestamp is written into it.

[0031] This compensation calculation offsets the physical delay caused by the gateway protocol conversion process by deducting the timestamp in advance, so that the linkage control commands are aligned in time phase at the receiving end, thus meeting the requirements of collaborative control of multi-source asynchronous devices.

[0032] In summary, the highly reliable, low-latency routing and transmission method for the Industrial Internet of Things (IIoT) firstly reduces reliance on preset templates by using a non-uniform slicing rule based on protocol byte flipping frequency, and improves the feature localization efficiency of heterogeneous protocols by stripping static frame headers, function codes, and dynamic payloads based on data jump probability. Secondly, it constructs a coupling fluctuation index and a structural editing cost index, mapping time delay and spatial structure deviation to a two-dimensional topological phase space, and uses a mean-normalized exponential model to quantify the impact of structural jumps on the computational power consumption of parsing, providing data support for fault state assessment. Furthermore, it uses the covariance of spatiotemporal dual indices to decouple interference and determine evolution boundaries, distinguishing between network physical layer noise attenuation and device firmware upgrades. When a firmware upgrade is confirmed, a reassembly mechanism is initiated, reducing resource consumption and packet dropping, and improving the reliability of routing transmission. Finally, it introduces a spatiotemporal phase compensation mechanism for asynchronous multi-unit protocols, using the coupling fluctuation index to calculate the average parsing dwell time of the edge gateway, correcting the timestamp of the target protocol data, offsetting the communication delay caused by the protocol conversion process, and aligning the linkage control commands in time phase to meet the requirements of collaborative control.

[0033] like Figure 2 As shown, in one specific implementation, S1 includes: S11, obtaining the number of data flips at each byte position of the data stream within the sliding time window, and obtaining the protocol byte flip frequency based on the number of data flips; S12. The protocol data frame is non-uniformly sliced ​​according to the protocol byte flip frequency. The continuous byte segment with the protocol byte flip frequency less than the first frequency threshold is divided into the first protocol slice, the byte segment with the protocol byte flip frequency between the first frequency threshold and the second frequency threshold is divided into the second protocol slice, and the byte segment with the protocol byte flip frequency greater than the second frequency threshold is divided into the third protocol slice. S13. Extract the actual decoding time of each protocol slice at each parsing node and construct the parsing time sequence; S14. Extract the offset of each protocol slice relative to the baseline protocol template in the byte structure and construct a structural deviation sequence.

[0034] In this embodiment, it should be noted that in S11, the underlying physical flipping pattern is extracted by sliding sampling of the data stream within a fixed time window. It should be noted that the size of the aforementioned sliding time window is typically obtained based on statistical analysis of the sampling period and protocol frame length of historical communication data from the industrial site. Specifically, continuous communication messages from the device under normal operating conditions can be collected, the average transmission interval of complete protocol data frames can be extracted, and the window value can be determined by weighted summation or rounding combined with the system's tolerance for edge gateway parsing latency. For example, if the average transmission interval of a message from an industrial sensor in 100 historical communication records is 800 milliseconds, and the maximum network hysteresis redundancy time allowed by the control system is 200 milliseconds, adding these two values ​​together allows the sliding time window to be reasonably set to 1000 milliseconds.

[0035] Specifically, a time window of 1000 milliseconds is set, and the number of data flips at each byte position within this window is recorded in real time. For example, if a specific byte only undergoes two state flips within this time window, the number of data flips is divided by the total number of samples to calculate the protocol byte flip frequency of that byte as 0.002. This step no longer relies on manual parsing of protocol semantics, but objectively characterizes the activity level of bytes through data transition probabilities. This solves the problem of difficulty in feature extraction caused by the opaque protocol structure of heterogeneous devices, and provides reliable data support for subsequent non-uniform slicing.

[0036] In S12, the protocol data frames are non-uniformly sliced ​​based on the protocol byte flipping frequency. It should be noted that the preset first and second frequency thresholds are typically obtained through unsupervised clustering analysis of byte flipping probabilities in a limited number of historical parsing records for various known benchmark protocols. Specifically, the flipping frequency of each byte position in the historical normal data stream is extracted, and clustering algorithms such as K-Means are used to divide it into three categories: static, semi-static, and dynamic regions. Then, the frequency values ​​of the boundaries of each category are extracted as the corresponding boundary thresholds. For example, analysis of 1000 historical messages reveals that the maximum flipping frequency in the static frame header region rarely exceeds 0.008, the frequency in the function code region is concentrated between 0.05 and 0.45, while the frequency in the dynamic payload region is consistently above 0.55. Based on these boundary distribution characteristics of historical data, the first frequency threshold for distinguishing the static region can be set to 0.01, and the second frequency threshold for distinguishing the function code and dynamic payload can be set to 0.5.

[0037] Furthermore, a first frequency threshold is set to 0.01, and a second frequency threshold is set to 0.5. For the aforementioned consecutive byte segments with a flip frequency of 0.002, since it is less than 0.01, it is classified as a first protocol slice, representing a relatively static frame header. If the measured frequency of a byte segment is 0.15, which is between the two thresholds, it is classified as a second protocol slice, representing a function code. If the measured frequency reaches 0.85, which is greater than 0.5, it is classified as a third protocol slice, representing dynamic payload data. This classification logic dynamically locates the boundaries based on the inherent activity of the data, enabling adaptive separation of the protocol's control and data areas, thus improving the adaptability and slicing accuracy of protocol parsing.

[0038] In S13, after the protocol slices are completed, the actual decoding time of each protocol slice at each parsing node is extracted to construct a parsing time sequence. For example, for the third protocol slice, its actual time consumption in multiple parsing processes is continuously recorded to obtain time data such as 1.2 milliseconds, 1.5 milliseconds, 1.3 milliseconds, and 1.8 milliseconds, and these discrete time points are arranged in chronological order to construct a complete sequence. This step aims to quantify the fluctuation of computing power consumption when processing a specific protocol slice, solving the problem that existing evaluation methods only focus on overall latency and ignore the parsing resistance of local slices. By finely recording the time trajectory of individual slices, it lays the foundation for the subsequent construction of a coupled evaluation model of time and space dimensions.

[0039] In S14, synchronized with the time series, the offset of each protocol slice relative to the baseline protocol template in byte structure is extracted to construct a structural deviation sequence. During actual operation, the difference between the starting byte position and the baseline position at each moment when parsing the slice is recorded. For example, the offset of the third protocol slice in the previous parsing iterations is recorded as 8 bytes. This step transforms the physical changes in the protocol structure into a specific numerical sequence, resolving the structural blind spot problem caused by implicit firmware upgrades or fine-tuning of data frame formats. By continuously tracking the offset trajectory of bytes in memory space, minute spatial structural changes can be accurately identified, thus providing a precise spatial dimension data source for quantifying the physical cost of protocol version evolution.

[0040] like Figure 3 As shown, in one specific embodiment, S2 includes: S21, obtaining the time difference between adjacent parsing points from the parsing time sequence of each protocol slice, and obtaining the deviation change of adjacent parsing points from the structural deviation sequence of each protocol slice; S22. The relative time volatility is obtained by dividing the time difference by the average of the parsing time sequence of each protocol slice. S23. The divisor is obtained by dividing the change in deviation by the average value of the structural deviation sequence of each protocol slice plus one. The relative structural volatility is obtained by performing an exponential mapping operation with the natural constant as the base and the divisor as the exponent. S24. Multiply the relative time volatility by the relative structural volatility to obtain the coupled volatility index corresponding to each protocol slice.

[0041] In this embodiment, it should be noted that in S21, during the feature calculation stage, the time difference between adjacent parsing points is obtained from the parsing time sequence of each protocol slice, and the deviation change between adjacent parsing points is obtained from the structural deviation sequence of each protocol slice. Assuming that the parsing time of the third protocol slice is 1.8 milliseconds in the 5th parsing and 1.3 milliseconds in the 4th, the time difference is 0.5 milliseconds; similarly, the structural deviation offset is 8 bytes in the 5th and 8 bytes in the 4th, so the deviation change is 0 bytes. This step filters the absolute basic values ​​of the data by calculating the difference, focusing on the instantaneous rate of change of the operating state, solving the problem that static data cannot reflect dynamic resistance, and enabling subsequent models to extract transient feature changes caused by network congestion or structural abrupt changes in a timely manner.

[0042] In step S22, after obtaining the difference, the relative time volatility is obtained by dividing the time difference by the average parsing time sequence of each protocol slice. Continuing with the aforementioned data, the average parsing time within the current window is calculated to be 1.4 milliseconds. Dividing the absolute value of the time difference (0.5 milliseconds) by 1.4 milliseconds yields a relative time volatility of approximately 0.357. This step, through mean normalization, eliminates the dimensional influence caused by differences in basic parsing time between different devices or protocols. It solves the technical problem of not being able to directly compare the time volatility amplitude between heterogeneous devices, ensuring that the time-dimensional volatility data has a universal evaluation standard, enabling subsequent calculations to objectively reflect the resistance changes in the parsing process.

[0043] In S23, to quantify spatial variation, the change in deviation is divided by the average of the structural deviation sequences of each protocol slice plus one, yielding a divisor. This divisor is then used as the base for an exponential mapping operation to obtain the relative structural volatility. Assuming the average structural deviation within the current window is 8 bytes, adding one results in 9 bytes; for the aforementioned deviation change of 0 bytes, the divisor is 0, and the natural exponent yields a relative structural volatility of 1. When the structural deviation abruptly changes to 12 bytes in another period, the change is 4 bytes, the new average is 9 bytes, adding one results in 10 bytes, and the exponential calculation yields a relative structural volatility of approximately 1.4918. This step introduces a smoothing constant to avoid the problem of a zero denominator, solving the difficulty in quantifying structural abrupt changes.

[0044] In S24, the relative time volatility and relative structural volatility are multiplied to obtain the coupling volatility index corresponding to each protocol slice. In the aforementioned fifth analysis, the time volatility is 0.357, the structural volatility is 1, and the product is 0.357; however, in the sixth analysis with the structural abrupt change, the time volatility is 0.4375, the structural volatility is 1.4918, and the product suddenly increases to 0.6527. Averaging the accumulated sequences yields the coupling volatility index of the slice, as shown in the example, 0.58. This step nonlinearly couples time delay with spatial structural deviation, solving the problem that existing linear superposition cannot accurately reflect the exponential amplification of computational power consumption caused by structural jumps, and effectively characterizing the comprehensive resistance state during heterogeneous protocol analysis.

[0045] Specifically, the calculation logic for S21 to S24 can be expressed as follows:

[0046] In the above formula, Represents the coupled volatility index; This represents the total number of messages contained within the parsing period; A discrete-time series index representing the message; and Representing the first Next and first The actual time consumed in parsing this protocol slice, in milliseconds; This represents the average parsing time sequence within the current window, in milliseconds. and Representing the first Next and first The structural offset of the protocol slice during the second parsing, in bytes; This represents the average value of the structural deviation sequence, in bytes. To prevent constants with a denominator of zero; This represents an exponential function with the natural constant as its base. This calculation quantifies the non-linear amplification effect of protocol structure changes on parsing time latency. It should be noted that the total number of packets included in the parsing period is typically calculated based on the edge gateway's memory processing bottleneck and historical average throughput stress test data for different protocols. Specifically, continuous stress testing can be used to statistically determine the extreme range of consecutive cached packets that allow the gateway to maintain stable memory usage without causing computational blockage under different network loads. For example, historical stress tests show that when the cache sequence length exceeds 150 packets, the computational cost of the gateway extracting two-dimensional topology features increases exponentially, leading to data backlog. Conversely, when the length is below 50 packets, the sample size is too small and susceptible to interference from occasional network jumps. Considering both the computational accuracy of state assessment and the hardware overhead of edge computing, the total number of packets in this parsing period can be reasonably set to 100.

[0047] The expression nonlinearly maps the parsing delay in the time dimension to the structural changes in the spatial dimension, quantifying the specific impact of protocol structure jumps on the computing power consumption of the underlying edge gateway. In real-world industrial communication environments, a single parsing delay may be caused by network congestion rather than a change in protocol structure. Therefore, directly adding the time difference and structural deviation linearly would lead to dimensional conflicts and error cancellation.

[0048] Furthermore, this expression is achieved through... The relative time fluctuation rate is calculated by dividing the absolute value of the difference in actual time consumption between two adjacent resolutions by the average resolution time within the current sliding time window, thus converting the absolute time difference into a dimensionless relative ratio, thereby eliminating the impact of differences in the basic computing power of different hardware devices.

[0049] Combining application scenario data, in the first The actual time consumed during each parsing session for Milliseconds, the previous time elapsed for Milliseconds, average parsing time within a window for Milliseconds, the relative time volatility is calculated. For changes in structural dimensions, the expression adopts... Calculate the ratio of the change in deviation to the average structural deviation plus one, and use it as an exponential function with the natural constant as the base. The input is used to implement the exponential amplification mechanism. The increment operation here is a smoothing constant to prevent the denominator from being zero when the protocol structure is in a stable state. The specific value of this smoothing constant is usually derived from historical test data analysis of the static byte drift tolerance of normal messages in the baseline protocol. In this embodiment, the vast majority of non-structural deviations caused by padding bits are distributed between 0 and 0.8 bytes. To prevent such non-substantial changes from being infinitely amplified during exponential mapping and to ensure the continuity of the underlying mathematical calculations, the smallest positive integer greater than the upper limit of this distribution range is taken, thus the smoothing constant can be reasonably set to the value 1. In the... During the second analysis, the structural deviation offset for Byte, previous structural offset for Byte, average structural deviation for Byte, at this time the exponentiation part is The final product is This indicates that when the structure remains unchanged, the computation degenerates into pure time fluctuations.

[0050] In the case of a structural mutation caused by a firmware upgrade The actual time consumed during each parsing session for Milliseconds, average time updated to for Milliseconds, structural deviation offset Mutation Byte, average structural deviation updated to for Bytes; at this time, the relative time volatility is The exponentiation part becomes The product of the two reaches After processing the sequence of length , After summing and averaging all terms, the coupling fluctuation index of the third protocol slice is obtained as follows: .

[0051] This exponential amplification calculation process based on mean normalization objectively restores the memory addressing resistance caused by unknown byte offsets to the parser, solving the problem that static monitoring cannot characterize the dynamic parsing resistance state of heterogeneous protocols.

[0052] like Figure 4 As shown, in one specific embodiment, S2 further includes: S25, extracting the starting byte offset and field byte length of each semantic field in the base protocol and the protocol to be converted; S26. Calculate the absolute value of the offset difference between each semantic field and the base protocol, as well as the absolute value of the length difference. S27. Obtain the sum of the absolute values ​​of the offset differences and length differences of all semantic fields, divide the sum by the total byte length of the protocol data frame, and obtain the structure editing cost index.

[0053] In this embodiment, it should be noted that in S25, the starting byte offset and field byte length of each semantic field in the base protocol and the protocol to be converted are extracted. For example, the protocol contains field 1 and field 2. For field 1, the base offset is 4 bytes, and the target offset is 6 bytes; the base length is 2 bytes, and the target length is 4 bytes. For field 2, the offset remains unchanged at 8 bytes, but the length changes from 4 bytes to 2 bytes. This step, by directly reading the layout parameters in the protocol physical frame, concretizes the protocol reconstruction problem as changes in memory address and data block length, solving the problem of difficulty in quantifying mapping costs during cross-protocol conversion, and providing the necessary basic data for calculating the physical overhead of memory reorganization.

[0054] In S26, the absolute values ​​of the offset differences and length differences for each semantic field between the protocol to be converted and the base protocol are calculated. For field 1, the absolute value of the offset difference is 6 minus 4, which equals 2 bytes, and the absolute value of the length difference is 4 minus 2, which equals 2 bytes. For field 2, the absolute value of the offset difference is 0 bytes, and the absolute value of the length difference is 4 minus 2, which equals 2 bytes. This step, through absolute value calculation, eliminates interference from the direction of forward or backward movement of fields in memory, considers the changes in data block movement and truncation, and solves the problem of difficulty in uniformly evaluating operations during protocol conversion, so that each reassembly operation can be quantified into specific byte processing costs.

[0055] In S27, the absolute values ​​of the offset differences and length differences of all semantic fields are summed. This sum is then divided by the total byte length of the protocol data frame to obtain the structural editing cost index. Assuming the total byte length of the protocol data frame is 64 bytes, for field 1, the sum of the two absolute values ​​is 4 bytes, which, divided by 64, yields 0.0625; for field 2, the sum of the absolute values ​​is 2 bytes, which, divided by 64, yields 0.03125. Summing all the values, the final structural editing cost index is 0.09375. This step, by calculating the byte displacement and reassembly overhead in memory and normalizing the total length, solves the technical problem of not being able to uniformly measure the conversion cost for protocols of different lengths, reflecting the physical cost required for cross-protocol conversion.

[0056] Specifically, the calculation logic for S25 to S27 can be expressed as follows:

[0057] In the above formula, Represents the cost of structural editing; The total number of semantic fields contained in the protocol data frame; Indexes representing semantic fields; and Representing the first The starting byte offset of each semantic field in the protocol to be converted and the base protocol, in bytes; and Representing the first The byte length of each semantic field in the protocol to be converted and the base protocol, in bytes; This represents the total length of the protocol data frame in bytes. This metric quantifies the physical cost of cross-protocol translation by calculating the displacement and reassembly overhead of bytes in memory.

[0058] In the process of adapting communication protocols for industrial IoT devices, due to the different definitions of message semantic fields by various manufacturers, the underlying physical process of converting one protocol format to another essentially involves re-fragmenting and relocating byte blocks in the gateway's memory. Existing methods often rely on complex text rule comparisons or semantic similarity algorithms, which not only consume significant edge computing resources but also struggle to provide a quantifiable conversion cost. This expression abandons highly complex graph matching logic and directly calculates the layout parameters of the underlying physical data frame.

[0059] Furthermore, in the expression Part used to calculate the first The absolute value of the difference in starting byte offset between the protocol to be converted and the base protocol for each semantic field represents the distance the memory pointer needs to move to find the starting point of the new data payload; while Part of it is used to calculate the absolute value of the difference in field byte lengths. This item represents the size of the memory data block that needs to be reallocated or reduced when reorganizing the new protocol format.

[0060] Based on the data in the application scenario, let's assume the total byte length of the currently processed protocol data frame is... for bytes, which contain a total of semantic fields, i.e., the total number of semantic fields. for For fields Its starting byte offset in the base protocol for Bytes, the target starting byte offset in the protocol to be converted. for Bytes; Base field byte length for Bytes, target field byte length for Bytes. The single calculated value produced by this field is... For fields Assuming its offset remains unchanged, but the length in the protocol to be converted is changed by... bytes shortened to If the value is in bytes, then the individual calculated value of this field is... The overall structural editing cost index is obtained by summing up the individual calculated values ​​of all fields. The total byte length of the protocol data frame is used. The expression uses the sum of the differences in each term as the denominator for division and normalization, transforming the specific number of bytes into a unified dimensionless cost indicator.

[0061] This computational logic solves the problem of not being able to establish a unified evaluation standard for the conversion cost between long and short messages. It enables gateways to directly quantify the memory reorganization overhead caused by device firmware upgrades through low-complexity arithmetic addition and division operations, providing a structural dimension basis for distinguishing network jitter from actual protocol version evolution.

[0062] like Figure 5 As shown, in one specific embodiment, S3 further includes: S31, when it is determined that the increase in the structural editing cost index of the first protocol slice or the second protocol slice is greater than a preset increase threshold, and the coupling fluctuation index is less than a preset fluctuation threshold, it is confirmed that the preset evolution determination condition is met. S32. Calculate the reciprocal of the variance based on the structural editing cost index, and generate a global accuracy index after normalizing the reciprocal.

[0063] In practice, this step aims to distinguish between transmission anomalies at the network physical layer and protocol version upgrades of the device itself. If the structure editing cost index is detected to be approaching zero but the coupling fluctuation index is increasing, it is determined to be a delay caused by network noise, and no protocol conversion mapping is processed. If the structure editing cost index increases abruptly, while the coupling fluctuation index remains stable at a low level after a brief change, it is determined to be a protocol change caused by a device firmware upgrade. At this time, it is confirmed that the preset evolution judgment condition is met, and a global accuracy index representing the evolution confidence is generated based on the consistency of the changes in the structure editing cost of each slice (i.e., the smaller the variance, the higher the consistency, and the higher the global accuracy index).

[0064] In this embodiment, it should be noted that in S31, during the state determination phase, it is determined whether the increase in the structural editing cost index and the coupling fluctuation index meet the preset evolution determination conditions. It should be noted that the preset increase threshold and preset fluctuation threshold are typically obtained through supervised training and statistical boundary delineation based on a sample dataset of characteristic indicators of the edge computing gateway under historical real network congestion and device firmware upgrade environments. Specifically, the peak values ​​of the coupling fluctuation index under multiple network packet loss or link jitter states, and the increase in structural editing cost under multiple real firmware upgrade states are collected. The critical threshold is established by calculating the upper and lower limits of the confidence interval of the corresponding feature distribution. For example, historical tests show that the peak value of the fluctuation index caused by physical layer noise mostly remains above 0.2, while the known increase in single structural editing cost caused by real firmware upgrades is greater than 0.04. Based on this, the preset fluctuation threshold can be set to 0.2 to filter out pure network communication interference, while the preset increase threshold can be set to 0.05 to accurately capture the substantial evolutionary step of the protocol structure characteristics.

[0065] A preset increase threshold of 0.05 and a preset fluctuation threshold of 0.2 are set. When the structural editing cost index of the first or second protocol slice experiences a step increase to 0.09375, with an increase greater than 0.05, and the coupling fluctuation index falls back to 0.12 (less than 0.2), the preset evolution judgment condition is confirmed to be met. This step utilizes the combination of abrupt changes in structural cost and stable time resistance to eliminate interference caused by the attenuation of network physical layer noise, solving the problem of misjudging network congestion as protocol changes in industrial settings and ensuring the accuracy of protocol version evolution judgment.

[0066] In step S32, after the evolution conditions are met, the reciprocal of the variance of the structural editing cost index is calculated. This reciprocal is then normalized to generate a global accuracy index. The variance of the structural editing cost index for each slice is extracted. Assuming the calculated result is 0.001, a smaller variance indicates a consistent internal pattern in the changes across slices. The reciprocal of 0.001 is taken to obtain 1000, which is then mapped and normalized to generate a global accuracy index of 0.92. This step assesses the confidence level of evolution by measuring the consistency of changes across multiple slice structures, resolving the problem of a single index being affected by local anomalies. By generating a quantified accuracy index, a reliable decision-making basis is provided for subsequent decisions on whether to activate the phase compensation mechanism.

[0067] like Figure 6 As shown, in one specific embodiment, S4 further includes: S41, generating a field reorganization mapping matrix of the protocol to be converted based on the mapping template of the baseline protocol and the structural editing cost index; S42. Add a value of one to the coupling fluctuation index of the third protocol slice to obtain the first weight; S43. Multiply the average parsing dwell time of a single message by the first weight to obtain the phase compensation time. S44. Obtain the current absolute time, subtract the phase compensation time from the current absolute time to get the target timestamp and write it into the timestamp field; S45. Reorganize the data format according to the field reorganization mapping matrix, embed the corrected timestamp field, obtain the target data and send it out.

[0068] After calculating the compensation time, the absolute time when the gateway issues the command is subtracted from the phase compensation time, which is equivalent to making an advance correction on the timestamp. This offsets the dwell delay caused by the protocol parsing and conversion process at the device node, so that the receiving device is aligned in physical time phase, ensuring the reliability and low latency of the routing transmission.

[0069] In this embodiment, it should be noted that in S41, the preset confidence threshold is usually obtained by comprehensively evaluating the system's fault tolerance rate for the erroneously triggered protocol reassembly mechanism and the distribution characteristics of the global accuracy index in historical misjudgment cases. Specifically, by organizing the network logs of the protocol version that were accurately identified and incorrectly judged in the historical operation data, the maximum value of the global accuracy index under misjudgment cases is extracted as a reference baseline, and a certain safety margin is added on this basis to determine the confidence threshold. For example, by statistically analyzing the past 50 cases where the cost of structural editing changed due to extremely severe network interference (i.e., the evolution of misjudgment), it was found that the highest calculated global accuracy index was only 0.81. In order to ensure that the gateway only performs the expensive protocol reassembly mapping under extremely high confidence, a safety margin of 0.04 is artificially added, and the preset confidence threshold can be reasonably set to 0.85.

[0070] Furthermore, during the conversion compensation phase, when the accuracy index of 0.92 is greater than the confidence threshold of 0.85, S41 generates a field reorganization mapping matrix for the protocol to be converted based on the mapping template of the baseline protocol and the previously obtained structural editing cost index of 0.09375. Based on the calculated offsets and length changes, the addressing rules and data extraction logic of the memory blocks are reorganized to generate a new mapping matrix file. This step, by concretizing the abstract change index into actual conversion rules, solves the technical obstacle of data not being parsed correctly after firmware upgrades of heterogeneous devices, enabling the edge gateway to dynamically adapt to protocol version evolution and providing a structural conversion baseline for subsequent message format reorganization and data processing.

[0071] In step S42, the coupling fluctuation index of the third protocol slice is increased by a value of one to obtain the first weight. In the previous step, the coupling fluctuation index of the third protocol slice was calculated to be 0.58. Adding a constant one to this index yields the first weight of 1.58. The design logic of this step is that the coupling fluctuation index reflects the proportion of additional time consumption due to resolution difficulties. Adding one to it converts it into a multiplicative weighting factor, transforming the dynamic resolution resistance into an amplification factor of physical delay. This operation solves the technical problem of the difficulty in quantifying and extracting dynamic fluctuation delay during protocol conversion, providing mathematical support for subsequent calculations of the spatiotemporal phase delay caused by gateway congestion.

[0072] In S43, the average parsing dwell time of a single packet is multiplied by the first weight to obtain the phase compensation time. The average parsing dwell time of a single packet processed by the current edge computing gateway, obtained through underlying monitoring, is 45 milliseconds. Multiplying 45 milliseconds by the first weight of 1.58 yields a phase compensation time of 71.1 milliseconds. This step, by combining the base dwell time with dynamically fluctuating weight factors, restores the physical time lost by the packet within the gateway due to protocol parsing and format reassembly. This calculation logic solves the problem of neglecting the internal processing latency of the gateway in existing routing transmissions, enabling dynamic perception of the timing impact of its own load on the communication link.

[0073] Specifically, the calculation logic for S42 to S43 can be expressed as follows:

[0074] In the above formula, This represents the phase compensation time, in milliseconds. This represents the average parsing dwell time for processing a single message, in milliseconds. The coupling fluctuation index, representing the extracted third protocol slice, is a dimensionless parameter.

[0075] In industrial collaborative control scenarios, the linkage of multi-source asynchronous devices places strict requirements on the consistency of time phase. However, when edge computing gateways perform dynamic parsing, field reorganization, and format conversion of heterogeneous protocols, they inevitably incur a certain amount of physical processing time, which fluctuates dynamically with changes in network congestion and the complexity of the protocol structure. If this dwell time is ignored and data is sent directly, it will cause a spatiotemporal misalignment between the time when downstream execution devices receive instructions and their actual acquisition and generation time. This expression aims to dynamically and adaptively estimate the current overall network and computing power latency based on environmental fluctuation data calculated in the preceding steps.

[0076] Furthermore, the average parsing dwell time in the expression This characterizes the basic physical time consumption for processing a single message under current stable load conditions; since the parsing process in actual communication is often affected by the resistance interference of protocol heterogeneity, the coupling fluctuation index of the third protocol slice is used. A computational model is introduced. By adding a numerical value of one to the coupled volatility index, a dynamic weighting factor consisting of a base unit of one and a volatility increment is constructed.

[0077] Based on actual test data from application scenarios, let's assume we obtain the average parsing and dwell time for the current gateway to process a single packet. for Milliseconds, while simultaneously extracting the coupling fluctuation index of the third protocol slice representing dynamic payload data. for Based on the above calculation logic, the volatility index is first increased by one to obtain the weighting multiple. Subsequently, the basic average analytical dwell time was calculated. Milliseconds multiplied by the weight factor The actual phase compensation time was calculated. for Milliseconds. When the gateway is preparing to send the assembled target protocol data, if the current system's relative running time is 50,000 milliseconds, it will use this physically elapsed time to subtract the calculated phase compensation time of 71.1 milliseconds, thereby correcting the target protocol's timestamp field to 49,928.9 milliseconds. This calculation mechanism, which calculates and subtracts timestamps in advance based on its own state in real time, effectively aligns and restores the internal computing power delay caused by protocol conversion at the data level. It solves the control phase misalignment problem caused by gateway dwell delay during cross-protocol routing transmission, ensuring the overall performance of high-reliability, low-latency routing transmission in the Industrial Internet of Things.

[0078] In S44, after confirming the compensation amount, the current absolute time is obtained. The phase compensation time is subtracted from the current absolute time to obtain the target timestamp, which is then written into the timestamp field. Assuming the current gateway's system relative running time is 50,000 milliseconds, a subtraction operation is performed, subtracting 71.1 milliseconds, resulting in a target timestamp of 49,928.9 milliseconds. This value is then written into the timestamp field of the target protocol. This step, by performing a forward subtraction correction on the timestamp, shifts the data's time backward, offsetting the dwell delay. This operation solves the time deviation problem superimposed on the transmission link during the routing conversion process, ensuring that the time point received downstream reflects the true generation time of the data.

[0079] In S45, the data format is restructured based on the field reconstruction mapping matrix, and the corrected timestamp field is embedded to obtain the target data and send it down. Based on the mapping matrix generated in S41, the data payload of the original message is assembled into the target protocol message according to the new offset and length rules. Simultaneously, the target timestamp of 49928.9 milliseconds, aligned with the physical phase, is embedded into the message. After packet assembly, it is sent to the cloud or downstream control equipment. This step completes the reshaping of the data structure and restoration of the time phase, solving the spatiotemporal misalignment problem in collaborative control of multi-source heterogeneous devices, and improving the end-to-end reliability and low-latency collaborative control capabilities of industrial IoT routing transmission.

[0080] like Figure 7 As shown, a highly reliable, low-latency routing transmission system for the Industrial Internet of Things (IIoT) is also provided. The system includes a management platform, a sensor network platform, and an object platform that are sequentially connected in communication. The management platform includes: The slicing module is used to obtain the first, second, and third protocol slices of the data stream based on the non-uniform slicing rules of protocol byte flipping frequency; and to obtain the parsing time sequence and structural deviation sequence of each protocol slice in the previous parsing cycle of the current parsing cycle. The feature calculation module is used to obtain the coupling fluctuation index corresponding to each protocol slice based on the parsing time sequence and structural deviation sequence of each protocol slice; and to obtain the structural editing cost index based on the offset change of the protocol field. The judgment module is used to obtain the index change characteristics of the coupling fluctuation index and the structural editing cost index of each protocol slice; when the index change characteristics meet the preset evolution judgment conditions, the global accuracy index is obtained based on the variance of the structural editing cost index. The output module is used to obtain the phase compensation time based on the coupling fluctuation index if the global accuracy index is greater than the preset confidence threshold, correct the timestamp of the target protocol based on the phase compensation time, obtain the target data containing the corrected timestamp, and send it out.

[0081] In one specific implementation, the slicing module is further configured to: obtain the number of data flips at each byte position in the data stream within the sliding time window, obtain the protocol byte flip frequency based on the number of data flips; perform non-uniform slicing on the protocol data frame according to the protocol byte flip frequency, divide the continuous byte segment with the protocol byte flip frequency less than the first frequency threshold into the first protocol slice, divide the byte segment with the protocol byte flip frequency between the first frequency threshold and the second frequency threshold into the second protocol slice, and divide the byte segment with the protocol byte flip frequency greater than the second frequency threshold into the third protocol slice.

[0082] To enable those skilled in the art to fully understand and implement the technical solutions described in this specification, the following section, in conjunction with a specific application scenario, provides a detailed deduction and data analysis of the entire implementation principle of a highly reliable, low-latency routing transmission system for the Industrial Internet of Things.

[0083] Suppose a smart manufacturing workshop deploys an edge computing gateway to collect real-time data from temperature sensors (using Modbus RTU protocol, base version) and frequency converters (using Profinet protocol, version to be converted), and uniformly converts the heterogeneous protocols to MQTT protocol for reporting to the cloud. During gateway operation, the system first enters S1 to perform sliding sampling of the industrial IoT data stream within a fixed time window (set to 1000 milliseconds). The system detects that a certain continuous byte segment exhibits only two byte flips within 1000 samples. The protocol byte flip frequency is then calculated as follows. The first frequency threshold is set to 0.01, and the second frequency threshold is set to 0.5. Because... The system classifies the byte segment into the first protocol slice, representing the static frame header. Simultaneously, it detects another byte segment with a flip frequency of 0.15, falling between the two, and classifies it into the second protocol slice, representing the function code. Finally, it detects that the temperature load data byte segment has a flip frequency of 0.85, greater than 0.5, and classifies it into the third protocol slice. Subsequently, the system extracts the actual decoding time of these three slices in 10 consecutive parsing iterations, constructing a parsing time sequence (e.g., the time sequence for the third protocol slice is 1.2, 1.5, 1.3, 1.8… milliseconds), and extracts its structural offset relative to the baseline protocol, constructing a structural deviation sequence.

[0084] Upon entering S2, the system calculates the coupling fluctuation index based on the aforementioned sequence. Taking the third protocol slice as an example, assuming that during the 5th parsing ( Its parsing time Milliseconds, the time taken for the previous parsing. Milliseconds, average parsing time within the current window Milliseconds; at this time, the structural deviation offset Byte, previous structural offset Byte, average structural deviation Bytes. According to the coupling fluctuation index formula ,for This item has a relative time volatility of ;because The relative structural volatility is At this point, the product of these terms is... If in the 6th parsing ( Due to a sudden structural deviation caused by a device firmware upgrade, Bytes, parsing time increased to Milliseconds, average time updated to Milliseconds, mean deviation updated to Bytes; then the relative time volatility is The relative structural volatility is At this point, the product becomes The system analyzes the entire sequence (e.g., ...) The coupling fluctuation index of the third protocol slice is obtained by summing and averaging the results. (Assuming the calculated result is 0.58). Simultaneously, the system calculates the cost index for structural editing. Assuming the total length of the protocol data frame... Bytes, containing 2 semantic fields ( Baseline offset of field 1. Bytes, target offset Byte; Baseline Length bytes, target length Bytes. The offset of field 2 remains unchanged; only the length changes from 4 bytes to 2 bytes. According to the formula... The calculated item for field 1 is The calculated item for field 2 is... Final structural editing cost indicators .

[0085] Next, in S3, the system performs interference decoupling and evolution boundary determination. The system continuously monitors the structural editing cost indicators of the first and second protocol slices. A step increase occurred, abruptly jumping from near 0 to 0.09375, and the increase of 0.09375 was greater than the preset increase threshold (set to 0.05). Simultaneously, the coupling fluctuation index of these slices was monitored. After a brief fluctuation, the value dropped back to 0.12, which is less than the preset fluctuation threshold (set to 0.2). The system thus confirms that the preset evolution judgment condition is met, determining that this is due to a genuine protocol structure change caused by the inverter firmware upgrade, rather than a transmission anomaly caused by network physical layer noise attenuation. After confirming the evolution, the system extracts the variance of the editing cost index for each slice structure (assuming the variance is extremely small, calculated to be 0.001), takes its reciprocal to obtain 1000, and then normalizes it (e.g., mapping to...). After determining the interval, a global accuracy index is generated, assuming the index value is 0.92.

[0086] Finally, in S4, the system determines that the global accuracy index of 0.92 is greater than the preset confidence threshold (e.g., 0.85), and then initiates the phase compensation mechanism. Based on the mapping template of the baseline protocol and the calculated structure editing cost index (0.09375), the system generates a field reassembly mapping matrix for the protocol to be converted. Subsequently, the system obtains the average parsing dwell time for a single packet processed by the current edge computing gateway. Assuming statistics show Milliseconds. The third protocol slice coupling fluctuation index calculated based on S2. The system adds a value of one to it to obtain the first weight. Using formulas The phase compensation time was calculated. millisecond.

[0087] After calculating the compensation amount, the system begins to correct the timestamps of the target protocol data (the messages to be encapsulated as MQTT protocols). Assume the current gateway's system relative runtime... The timeout is 50,000 milliseconds (the corresponding Unix timestamp), which the system corrects by subtraction using the formula. Calculate the target timestamp Milliseconds (can be rounded up in practical applications).

[0088] Finally, the system reassembles the original Modbus / Profinet message data into the target MQTT message format based on the field reconstruction mapping matrix, and writes this corrected target timestamp (49928.9 milliseconds) into the timestamp field of the MQTT message. Due to the 71.1 millisecond delay caused by gateway protocol parsing and structural changes, this proactive deduction correction method allows the cloud or downstream control system to accurately reconstruct the actual physical phase of the data in the industrial field when it receives the MQTT message, effectively eliminating system latency errors during routing and ensuring high reliability and low latency across the entire link.

[0089] It should also be noted that the entire highly reliable, low-latency routing and transmission system for the Industrial Internet of Things (IIoT) can be applied to the optimized IIoT. Figure 7 This is a schematic diagram of the composition of the highly reliable, low-latency routing transmission system for the Industrial Internet of Things (IIoT) of the present invention. Figure 8 This is a schematic diagram illustrating the optimized industrial Internet of Things (IIoT) involved in this invention. (See diagram below.) Figure 7 and Figure 8 As shown, the optimized Industrial Internet of Things (IIoT) includes a user platform, a service platform, a management platform, a sensor network platform, and an object platform that establish communication in sequence. The user platform is configured to provide front-end services to users; users obtain the necessary perception service information through the user platform, process the perception service information, and transform it into user perception information; users analyze the user perception information and make corresponding decisions based on their own wishes, and transform the user perception information into user control information through the corresponding information system and send it to the service platform, thereby demonstrating the user's corresponding service needs and wishes.

[0090] The physical entities of the user platform include various user terminals, such as mobile phones, computers, and dedicated terminals, which provide user services through integration with user information system software.

[0091] The service platform is configured as an API server or other server used to establish communication between the management platform and the user platform to achieve corresponding functions; the physical entity of the service platform includes various servers.

[0092] The management platform is configured to perform at least one of the following: device operation status monitoring and management, data monitoring and management, device parameter management, and lifecycle management; the management platform is the overall operation platform for the Internet of Things, which may include various management sub-platforms, with different management sub-platforms performing different management tasks; the physical entities of the management platform include various servers.

[0093] The sensor network platform is configured to perform at least one of the following functions: network management, command management, device status management, data protocol management, data parsing, data classification, data transmission monitoring, and data transmission security management. The sensor network platform provides functions such as data communication, transmission, parsing, identification, and classification, avoiding the direct aggregation of data from various object platforms on the management platform, which would otherwise result in data redundancy and low data processing efficiency. The physical entities of the object platforms include various gateways, edge computing devices, etc.

[0094] The object platform is configured to perform specific production control, detection, measurement and other production tasks; the physical entities of the production objects include various production equipment, sensors and so on.

[0095] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0096] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0097] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A highly reliable, low-latency routing and transmission method for the Industrial Internet of Things (IIoT), characterized in that, The methods include: Based on the non-uniform slicing rules of protocol byte flipping frequency, the first protocol slice, the second protocol slice, and the third protocol slice of the data stream are obtained; Obtain the parsing time sequence and structural deviation sequence of each protocol slice in the previous parsing cycle of the current parsing cycle; Based on the parsing time sequence and structural deviation sequence of each protocol slice, the coupling fluctuation index corresponding to each protocol slice is obtained; the structural editing cost index is obtained based on the offset change of the protocol field. Obtain the characteristics of the coupling fluctuation index and structural editing cost index of each protocol slice; When the characteristics of indicator changes meet the preset evolution judgment conditions, the global accuracy indicator is obtained based on the variance of the structure editing cost indicator. If the global accuracy index is greater than the preset confidence threshold, the phase compensation time is obtained based on the coupling fluctuation index, the timestamp of the target protocol is corrected based on the phase compensation time, the target data containing the corrected timestamp is obtained and sent out.

2. The highly reliable, low-latency routing and transmission method for the Industrial Internet of Things according to claim 1, characterized in that, The non-uniform slicing rule based on protocol byte flipping frequency obtains the first protocol slice, the second protocol slice, and the third protocol slice of the data stream, including: Get the number of data flips at each byte position in the data stream within the sliding time window, and obtain the protocol byte flip frequency based on the number of data flips; The protocol data frame is non-uniformly sliced ​​according to the protocol byte flip frequency. The continuous byte segment with the protocol byte flip frequency less than the first frequency threshold is divided into the first protocol slice, the byte segment with the protocol byte flip frequency between the first frequency threshold and the second frequency threshold is divided into the second protocol slice, and the byte segment with the protocol byte flip frequency greater than the second frequency threshold is divided into the third protocol slice.

3. The highly reliable, low-latency routing and transmission method for the Industrial Internet of Things according to claim 2, characterized in that, The step of obtaining the parsing time sequence and structural deviation sequence of each protocol slice in the previous parsing cycle includes: Extract the actual decoding time of each protocol slice at each parsing node, and construct a parsing time sequence; Extract the offset of each protocol slice relative to the baseline protocol template in the byte structure, and construct a structural deviation sequence.

4. The highly reliable, low-latency routing and transmission method for the Industrial Internet of Things according to claim 1, characterized in that, The step of obtaining the coupling fluctuation index corresponding to each protocol slice based on the parsing time sequence and structural deviation sequence of each protocol slice includes: The time difference between adjacent parsing points is obtained from the parsing time sequence of each protocol slice, and the deviation change between adjacent parsing points is obtained from the structural deviation sequence of each protocol slice. The relative time volatility is obtained by dividing the time difference by the average of the parsed time-consuming sequences of each protocol slice. The divisor is obtained by dividing the change in deviation by the average value of the structural deviation sequence of each protocol slice by one. The relative structural volatility is obtained by performing an exponential mapping operation with the natural constant as the base and the divisor as the exponent. The relative time volatility is multiplied by the relative structural volatility to obtain the coupled volatility index corresponding to each protocol slice.

5. The highly reliable, low-latency routing and transmission method for the Industrial Internet of Things according to claim 1, characterized in that, The step of obtaining the structural editing cost index based on the offset change of the protocol field includes: Extract the starting byte offset and field byte length of each semantic field in the base protocol and the protocol to be converted; Calculate the absolute value of the offset difference between the protocol to be converted and the base protocol for each semantic field, as well as the absolute value of the length difference. The sum of the absolute values ​​of the offset differences and length differences of all semantic fields is obtained, and then divided by the total byte length of the protocol data frame to obtain the structural editing cost index.

6. The highly reliable, low-latency routing and transmission method for the Industrial Internet of Things according to claim 1, characterized in that, When the covariance meets the preset evolution judgment condition, the global accuracy index is obtained based on the variance of the structural editing cost index, including: When the increase in the structural editing cost index of the first protocol slice or the second protocol slice is greater than the preset increase threshold, and the coupling fluctuation index is less than the preset fluctuation threshold, the preset evolution judgment condition is confirmed to be met. The reciprocal of the variance of the structural editing cost index is calculated, and the reciprocal is normalized to generate a global accuracy index.

7. The highly reliable, low-latency routing and transmission method for the Industrial Internet of Things according to claim 1, characterized in that, The process of obtaining the phase compensation time based on the coupling fluctuation index, correcting the target protocol's timestamp based on the phase compensation time, obtaining target data containing the corrected timestamp, and sending it out includes: Based on the mapping template of the baseline protocol and the structural editing cost index, generate a field reorganization mapping matrix for the protocol to be converted; The phase compensation time is obtained based on the coupling fluctuation index of the third protocol slice and the average parsing dwell time of a single message; The timestamp field of the target protocol data is corrected by subtraction using the phase compensation time amount; The data format is reorganized based on the field reconstructing mapping matrix, and the corrected timestamp field is embedded to obtain the target data and send it out.

8. The highly reliable, low-latency routing and transmission method for the Industrial Internet of Things according to claim 7, characterized in that, The step of obtaining the phase compensation time based on the coupling fluctuation index of the third protocol slice and the average parsing dwell time of a single message includes: The first weight is obtained by adding a value of one to the coupling fluctuation index of the third protocol slice. Multiply the average resolution dwell time by the first weight to obtain the phase compensation time. The timestamp field of the target protocol data is corrected by subtraction using phase compensation time, including: Obtain the current system absolute time, subtract the phase compensation time from the current system absolute time to get the target timestamp, and write it into the timestamp field.

9. A highly reliable, low-latency routing and transmission system for the Industrial Internet of Things, characterized in that, The system includes a management platform, a sensor network platform, and an object platform that are connected in sequence. The management platform includes: The slicing module is used to obtain the first, second, and third protocol slices of the data stream based on the non-uniform slicing rules of protocol byte flipping frequency; and to obtain the parsing time sequence and structural deviation sequence of each protocol slice in the previous parsing cycle of the current parsing cycle. The feature calculation module is used to obtain the coupling fluctuation index corresponding to each protocol slice based on the parsing time sequence and structural deviation sequence of each protocol slice; and to obtain the structural editing cost index based on the offset change of the protocol field. The judgment module is used to obtain the index change characteristics of the coupling fluctuation index and the structural editing cost index of each protocol slice; when the index change characteristics meet the preset evolution judgment conditions, the global accuracy index is obtained based on the variance of the structural editing cost index. The output module is used to obtain the phase compensation time based on the coupling fluctuation index if the global accuracy index is greater than the preset confidence threshold, correct the timestamp of the target protocol based on the phase compensation time, obtain the target data containing the corrected timestamp, and send it out.

10. The highly reliable, low-latency routing and transmission system for the Industrial Internet of Things according to claim 9, characterized in that, The slicing module is also used for: Get the number of data flips at each byte position in the data stream within the sliding time window, and obtain the protocol byte flip frequency based on the number of data flips; The protocol data frame is non-uniformly sliced ​​according to the protocol byte flip frequency. The continuous byte segment with the protocol byte flip frequency less than the first frequency threshold is divided into the first protocol slice, the byte segment with the protocol byte flip frequency between the first frequency threshold and the second frequency threshold is divided into the second protocol slice, and the byte segment with the protocol byte flip frequency greater than the second frequency threshold is divided into the third protocol slice.

Images