Data transmission optimization method comprehensively considering data slicing and multi-path routing selection

Abstract

The invention discloses a data transmission optimization method comprehensively considering data slicing and multi-path routing selection, belongs to the technical field of data transmission, and solves the problems of low security and low transmission efficiency of the existing data transmission mode. The method comprises the steps that a data slice training set is used for training a data slice large model to generate a data slice rule, a data transmission result based on the data slice rule is used for calculating a loss function, and model parameters of the data slice large model are optimized; based on the multi-path routing training set, training a multi-path routing planning large model to obtain a multi-path routing planning large model passing training; and performing data slicing processing on the real-time to-be-transmitted data by using a data slicing rule predicted by the trained data slicing large model, and performing multi-path routing transmission on each data slice of the real-time to-be-transmitted data by using an optimal multi-path routing algorithm predicted by the trained multi-path routing planning large model. And recombining each data slice to complete transmission.

Description

Technical Field

[0001] The present invention relates to the technical field of data transmission, and in particular, to a data transmission optimization method that comprehensively considers data slicing and multi-path routing selection. Background Art

[0002] With the rapid development of the Internet and information technology, the demand and complexity of data transmission have been continuously increasing. In modern network environments, especially in business scenarios involving sensitive information, the security and efficiency of data transmission are particularly important. To address these challenges, network slicing and multi-path routing technologies have gradually become key means to improve data transmission performance.

[0003] Currently, the main drawback of data transmission is that the data slicing process and the path routing process are not closely integrated, making it difficult to cooperate together to achieve the data transmission process. Specifically, in the research process of data slicing, data is divided into multiple independent slices to enhance data security. However, this research does not involve how to route the sliced data to achieve data transmission, that is, the problem of how to efficiently and securely transmit the sliced data is not considered. In fact, the setting of relevant parameters for data transmission may affect the way of data slicing. In the research of multi-path routing planning, only reinforcement learning methods are used to plan multi-path routing, which does not involve data slicing and other preprocessing processes before the data to be transmitted, and does not address different requirements for multi-path routing algorithms under different complex scenarios.

[0004] In summary, although the research in these two directions in the prior art has its own advantages, due to the lack of a comprehensive consideration of the data transmission mechanism, there are obvious deficiencies in its practical application. Therefore, how to design a data transmission optimization method that comprehensively considers data slicing and multi-path routing selection to improve the security, efficiency, and user experience of data transmission is a technical problem that urgently needs to be solved at present. Summary of the Invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a data transmission optimization method that comprehensively considers data slicing and multi-path routing selection to solve the problems of low security and transmission efficiency existing in the existing data transmission methods.

[0006] The present invention discloses a data transmission optimization method that comprehensively considers data slicing and multi-path routing selection, and the method includes: Training a data slicing large model with a data slicing training set to generate data slicing rules, calculating a loss function using the data transmission results based on the data slicing rules, and optimizing the model parameters of the data slicing large model; after multiple iterative trainings, obtaining a data slicing large model that passes the training; Train a multi - path routing planning large model based on a multi - path routing training set to obtain a multi - path routing planning large model that has passed the training; Slice the real - time data to be transmitted using the data slicing rules predicted by the trained data slicing large model, perform multi - path routing transmission on each data slice of the real - time data to be transmitted using the optimal multi - path routing algorithm predicted by the trained multi - path routing planning large model, and the receiving party reorganizes each data slice according to the predicted data slicing rules to complete the transmission.

[0007] Based on the above - mentioned solution, the present invention has also made the following improvements: Furthermore, each data slicing training sample in the data slicing training set includes a data sample to be transmitted, its confidentiality level, and transmission parameters.

[0008] Furthermore, train the data slicing large model by performing the following operations: In each training process, use each data slicing training sample in the data slicing training set as input, and the data slicing large model outputs the matching data slicing rules; Slice the data sample to be transmitted according to the generated data slicing rules to obtain several data slices; perform data transmission on each data slice according to the transmission parameters of the data sample to be transmitted; Calculate the comprehensive loss function according to the transmission performance indicators and confidentiality level during the data transmission process, and optimize the model parameters of the data slicing large model according to the comprehensive loss function; Jump to the next training until the preset training end condition is met, and finally obtain a data slicing large model that has passed the training.

[0009] Furthermore, the comprehensive loss function is expressed as: (1) where 、 、 、 represent the resource utilization rate loss function, data transmission delay loss function, packet loss rate loss function, and confidentiality level loss function respectively; 、 、 、 represent the weight parameters of the resource utilization rate loss function, data transmission delay loss function, packet loss rate loss function, and confidentiality level loss function respectively.

[0010] Furthermore, the transmission performance indicators include the actual transmission volume, actual average delay, and the number of lost packets.

[0011] Furthermore, the resource utilization rate loss function Expressed as: (2) Wherein, and respectively represent the actual transmission volume and the theoretical maximum transmission volume of all data slices of the data sample to be transmitted; Data transmission delay loss function Expressed as: (3) Wherein, and respectively represent the actual average delay and the theoretical average delay of all data slices of the data sample to be transmitted; Packet loss rate loss function Expressed as: (4) Wherein, and respectively represent the number of lost packets and the total number of transmitted packets of the data slices of the data sample to be transmitted.

[0012] Furthermore, the confidentiality level loss function is expressed as as: (5) Wherein, represents the actual number of data slices, represents the minimum number of data slices corresponding to the confidentiality level of the currently transmitted data sample.

[0013] Furthermore, in the multi-path routing training set, the multi-path routing sample includes: the transmission parameters of each routing path, the data volume and confidentiality level of the data sample to be transmitted, and the optimal multi-path routing algorithm label.

[0014] Furthermore, train the multi-path routing planning large model by performing the following operations: During the training process of the multi-path routing planning large model, use the network state of each routing path, the data volume and confidentiality level of the data sample to be transmitted as the input, and the corresponding optimal multi-path routing algorithm label as the output to train the multi-path routing planning large model.

[0015] Furthermore, transmit the real-time data to be transmitted by performing the following operations: The data sender receives and determines whether the real-time data to be transmitted can be sliced. If so, input the real-time data to be transmitted, its confidentiality level, and transmission parameters into the trained data slicing large model to predict and output the data slicing rule; The data sender performs data slicing on the real-time data to be transmitted according to the data slicing rules, obtaining a number of data slices; Input the transmission parameters of each routing path from the data sender to the data receiver, the data volume and security level of each data slice of the data to be transmitted into the trained multi-path routing planning large model, and predict and output the optimal multi-path routing algorithm; Obtain the multi-path routing of the data to be transmitted according to the predicted optimal multi-path routing algorithm, and transmit each data slice of the real-time data to be transmitted to the data receiver according to the multi-path routing; The data receiver recombines each received data slice according to the data slicing rules to complete the data transmission.

[0016] Compared with the prior art, the present invention can at least achieve one of the following beneficial effects: The data transmission optimization method provided by the present invention that comprehensively considers data slicing and multi-path routing selection can effectively solve the problems of low security and transmission efficiency existing in the existing data transmission methods, and has the following technical effects: (1) In terms of security improvement 1) Combination of data slicing and security level: The data slicing training set contains data samples to be transmitted and their security levels. The trained data slicing large model will generate corresponding slicing rules according to the security level of the data. For data with a high security level, more refined and dispersed slicing rules can be generated, so that even if the data is intercepted during transmission, it is difficult for attackers to restore the complete sensitive information, thus enhancing the confidentiality of the data; 2) Multi-path transmission reduces risks: By adopting multi-path routing transmission, different slices of data are sent through multiple paths. Even if some of these paths are attacked or fail, as long as there are still some paths that can normally transmit data slices, the receiver may be able to recombine the complete data according to the slicing rules, reducing the risk of complete data loss or leakage caused by single-path failure or attack, and improving the security and reliability of data transmission.

[0017] (2) In terms of improved transmission efficiency 1) Optimized data slicing rules: By continuously training and optimizing the data slicing large model using the data slicing training set, the generated slicing rules can comprehensively consider factors such as transmission parameters. For example, according to conditions such as network bandwidth and delay, the data is sliced into a size and quantity that are more suitable for the transmission network, making the transmission after data slicing more efficient, avoiding problems such as transmission redundancy or transmission blockage caused by unreasonable slicing, and thus improving the overall transmission efficiency; 2) Intelligent multi-path routing planning: The large multi-path routing planning model trained based on the multi-path routing training set can predict the optimal multi-path routing algorithm according to the data volume, confidentiality level of different data slices, and transmission parameters of each routing path. This algorithm can reasonably allocate the transmission of each data slice on different paths, make full use of network resources, achieve load balancing, avoid the bottleneck of single-path transmission, and effectively improve the speed and efficiency of data transmission.

[0018] In the present invention, the above technical solutions can also be combined with each other to achieve more preferred combination schemes. Other features and advantages of the present invention will be described in the following specification. Moreover, some advantages can be made obvious from the specification, or understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained through the content specifically pointed out in the specification and the drawings. Brief Description of the Drawings

[0019] The drawings are only for the purpose of showing specific embodiments, and are not considered as a limitation to the present invention. Throughout the drawings, the same reference signs denote the same components; Figure 1 It is a flowchart of a data transmission optimization method that comprehensively considers data slicing and multi-path routing selection provided by an embodiment of the present invention; Figure 2 It is a flowchart of a data transmission method provided by an embodiment of the present invention. Detailed Embodiments

[0020] The following will specifically describe the preferred embodiments of the present invention with reference to the drawings. The drawings form a part of this application and are used together with the embodiments of the present invention to explain the principles of the present invention, and are not used to limit the scope of the present invention.

[0021] A specific embodiment of the present invention discloses a data transmission optimization method that comprehensively considers data slicing and multi-path routing selection. The flowchart of this method is as Figure 1 shown.

[0022] Step S1: Use the data slice training set to train the data slice large model to generate data slice rules, calculate the loss function using the data transmission results based on the data slice rules, and optimize the model parameters of the data slice large model; after multiple iterative trainings, obtain a data slice large model that passes the training.

[0023] During the data transmission process, for the data to be transmitted with a relatively high confidentiality level, more data slicing operations are required to enhance the security of its data transmission. At the same time, when the data to be transmitted exceeds the transmission capacity of the network transmission unit, data slicing needs to be performed on the data to be transmitted. In addition, during the data transmission process, it is also affected by other parameters such as the network state. Therefore, to solve this problem, this embodiment designs the following training method for the data slicing large model to achieve intelligent data slicing of the data to be transmitted.

[0024] Step S11: Construct a data slicing training set for training the data slicing large model.

[0025] Preferably, in this embodiment, each data slicing training sample in the data slicing training set includes the data sample to be transmitted (data packet) and its confidentiality level, and transmission parameters (including network state and transmission capacity of the network transmission unit). The specific description is as follows.

[0026] (1) Data packet The data packet consists of two parts, namely the header and the data part. The header contains the key information of the data packet, which is used to guide the transmission and processing of the data packet in the network. This information may include source address, destination address, protocol type, data packet length, checksum, etc. The format and content of the header will vary according to different network protocols (such as IP protocol, TCP protocol, etc.). The data part is the actual payload to be transmitted by the data packet, which contains user data or application data. The composition of each data slice should be processed accordingly according to the identification, flag, and fragment offset in the header of the data packet.

[0027] Identification: All data slices of the same data packet use the same identification.

[0028] Flag: When DF (Don't Fragment) = 1, slicing is prohibited; when DF = 0, slicing is allowed. That is, for the data packet formed by the data slice, if its DF = 0, slicing can still be performed.

[0029] When MF (More Fragment) = 1, there are subsequent slices; when MF = 0, there are no subsequent slices, that is, this slice is the last slice of this packet. Only when DF = 0 does MF make sense.

[0030] Fragment offset: The relative position of a certain middle slice in the original IP packet.

[0031] A simple example of the data slicing rule is shown in Table 1. It can be seen from Table 1 that: The data packet to be transmitted: 20 bytes for the header and 3800 bytes for the data part; Rule: Each slice does not exceed 1420 bytes (the transmission capacity of the network transmission unit); Identification: 666; Flags: DF = 0, indicating that slicing is allowed; MF = 0, indicating that there are no subsequent slices; Fragment offset: The fragment offset of the first data slice is 0, and the fragment offset of subsequent data slices depends on the length of the data part of the previous data slice.

[0032] Table 1 Example of Data Slice Rules 。

[0033] (2)Confidentiality Level Before actual data transmission, the data sender classifies and grades the data to be transmitted according to the importance of the data to be transmitted (mainly considering the security of data transmission) to determine the confidentiality level of the data to be transmitted. In the specific implementation process, the confidentiality level can be adaptively set according to the specific application scenario, and this embodiment does not impose too many restrictions on this. Considering the security of data transmission, for data with a higher confidentiality level, it is inclined to use more data slices for transmission; while for data with a lower confidentiality level, it is inclined to use fewer data slices for transmission. Therefore, the confidentiality level also has a certain impact on the setting of data slice rules.

[0034] (3)Network Status Network status includes information such as bandwidth, latency, and packet loss rate, which can be obtained through network monitoring tools and also has a direct impact on the formulation of data slice rules.

[0035] In the specific implementation process, considering that the network status changes dynamically over time, the following methods can be used to process the network status during the data sample transmission process: (1) Time series vectorization / matrix representation: Present the time series data of the network status changing over time in the form of a vector or matrix for subsequent mathematical operations and model processing; (2) Sliding window aggregation analysis: Set a fixed time window length (such as every second or every minute) to perform aggregation analysis on the network status within the window. For example, calculate key indicators such as the average bandwidth, latency, and packet loss rate in the past 1 second to more intuitively grasp the dynamic change trend of the network status; (3) Feature value normalization processing: Scale the values of the network status to a fixed range (such as 0 - 1 or - 1 to 1) to avoid adverse effects on the model performance due to overly large differences in feature scales and ensure that the model can more accurately evaluate and predict the network status.

[0036] (4)Transmission Capacity of the Network Transmission Unit The transmission capacity of the network transmission unit determines the maximum data length allowed during the transmission process. That is, the data to be transmitted needs to perform a single data transmission limited by the transmission capacity of the network transmission unit.

[0037] Preferably, in this embodiment, the data slicing rules include: the number of data slices, the size of each data slice, and the offset. The data slicing rules can help the data slicing large model learn the slicing rules in different scenarios.

[0038] It should be noted that the data samples to be transmitted in the data slicing training set are all samples with data slicing rule requirements. Specifically, for some scenarios sensitive to latency, data slicing may introduce additional processing latency, so it is possible to choose not to slice and directly transmit the complete data packet. Or, in the case where slicing is explicitly prohibited in the data packet, data slicing is not required either.

[0039] In addition, it should be noted that each data slice in the data slicing rule of the same data sample to be transmitted has the same identifier to ensure data uniqueness. If the identifier is lost or inconsistent, it may lead to confusion or data errors; at the same time, retaining the identifier helps to maintain the correlation and interpretability between data slices at the model output stage, facilitating subsequent analysis and verification.

[0040] Step S12: Train the data slicing large model using the data slicing training set to obtain a trained data slicing large model.

[0041] Preferably, the data slicing large model can be implemented using a Transform model. Below, the specific implementation process of this step is described in detail as follows.

[0042] Step S121: In each training process, use each data slice training sample in the data slicing training set as input, and the data slicing large model outputs the matching data slicing rules (also known as "data splitting rules").

[0043] This data slicing rule is used to guide the splitting and organization of data during subsequent data transmission processes to adapt to different transmission conditions and requirements.

[0044] Step S122: Perform data slicing processing on the data samples to be transmitted according to the generated data slicing rules to obtain several data slices; perform data transmission on each data slice according to the transmission parameters of the data samples to be transmitted.

[0045] Step S123: Calculate the comprehensive loss function according to the transmission performance index and security level during the data transmission process, and optimize the model parameters of the data slicing large model according to the comprehensive loss function.

[0046] Step S124: Jump to the next training until the preset training end condition is met, and finally obtain a data slice large model that passes the training.

[0047] That is, after obtaining the optimized model parameters in step S123, the optimized model parameters are reapplied to the generation of data slicing rules and the data transmission process, continue to calculate the loss function and optimize the model parameters until the preset training end condition is met (such as the loss function converges to a specified threshold, the change of model parameters tends to be stable, reaching the maximum number of iterations, etc.), and finally obtain a data slice large model that passes the training.

[0048] In step S123, the transmission performance metrics in the data transmission process include the actual transmission volume, the actual average delay, and the number of packet losses. Preferably, in the process of calculating the loss function according to the transmission performance metrics in the data transmission process, in order to comprehensively measure the quality of the data slicing rules, multiple loss functions can be defined to evaluate different performance objectives, such as the resource utilization loss function (measuring the bandwidth usage efficiency), the data transmission delay loss function, and the packet loss rate loss function.

[0049] 1) Resource utilization loss function , (1) Among them, 、 respectively represent the actual transmission volume and the theoretical maximum transmission volume of all data slices of the data sample to be transmitted. The actual transmission volume of all data slices specifically refers to the amount of data that can actually be transmitted to the receiving end and can be obtained by monitoring network devices; the theoretical maximum transmission volume is the amount of data of all data slices and can be measured by a network performance testing tool (such as iPerf).

[0050] Calculate the resource utilization loss function , aiming to maximize the broadband usage efficiency. The smaller (close to 0), the higher the bandwidth usage efficiency.

[0051] 2) Data transmission delay loss function , (2) Among them, 、 respectively represent the actual average delay and the theoretical average delay of all data slices of the data sample to be transmitted. The actual average delay of all data slices is the average of the transmission delays experienced by all data slices in actual transmission. The theoretical average delay of all data slices is the average expected transmission delay under ideal conditions (no congestion, queuing delay, etc.).

[0052] Assume that the sending time and receiving time of the data slice are respectively represented as and . Then, the actual delay of the data slice is represented as: (3) The actual average delay of all data slices is represented as: (4) represents the total number of data slices.

[0053] In addition, network simulation tools (such as NS2, OMNeT++) can be used to simulate the network environment and measure the theoretical average delay.

[0054] Calculate the data transmission delay loss function aiming to minimize the data transmission time. The actual average delay is always greater than or equal to the theoretical average delay because there are various additional delay factors in the actual network, while the theoretical delay only considers the basic transmission time under ideal conditions. The smaller (closer to 1), the closer the transmission time is to the ideal state.

[0055] 3) Packet loss rate loss function , (5) where and respectively represent the number of lost packets and the total number of transmitted packets of the data slices of the data samples to be transmitted.

[0056] Calculate the packet loss rate loss function aiming to reduce the probability of data loss. The smaller (closer to 0), the more reliable the data transmission.

[0057] Preferably, in this embodiment, when the data slice large model is at a high security level, it tends to output a larger number of data slices, so as to ensure that in the case of malicious attacks or data interception, even if the attacker or interceptor obtains a part of the data slices, they can only get a small part of the data slices and it is difficult to restore the complete data content, thus achieving the effect of ensuring data security. Accordingly, the following security level loss function is set.

[0058] 4) Security level loss function , (6) Among them, represents the actual number of data slices, represents the minimum number of data slices corresponding to the confidentiality level of the currently transmitted data sample.

[0059] In the specific implementation process, different minimum numbers of data slices can be set according to the actual network environment corresponding to different confidentiality levels. For example: for a low confidentiality level, the minimum number of data slices is 2; for a medium confidentiality level, the minimum number of data slices is 5; for a high confidentiality level, the minimum number of data slices is 15. Therefore, when the actual number of data slices is less than the corresponding minimum number of data slices required, the loss value of the confidentiality level loss function increases proportionally; when the minimum number of data slices required is met or exceeded, the loss value is 0.

[0060] The above loss function reflects the deviation degree between the data slicing rule and the ideal transmission effect from different dimensions. Preferably, a comprehensive loss function can be constructed, and the relationship between the above different loss functions can be balanced through weights. The comprehensive loss function is expressed as: (7) Among them, , , , respectively represent the weight parameters of the resource utilization loss function , the data transmission delay loss function , the packet loss rate loss function , and the confidentiality level loss function . In the specific implementation process, the weight parameters can be preset according to the actual situation or expert experience.

[0061] Step S2: Train the multi-path routing planning large model based on the multi-path routing training set to obtain a multi-path routing planning large model that has passed the training.

[0062] In the process of route discovery, the multi-path routing protocol will form multiple paths to the destination node. Multi-path routing has the following advantages: (1) Load balancing: Distribute the traffic to multiple paths to avoid overloading a certain path, thereby improving the overall bandwidth utilization rate; (2) Fault tolerance: When a certain path fails, the data can continue to be transmitted through other paths, enhancing the reliability of the system; (3) Fast transmission: By having multiple paths to parallelly wait for the data to be transmitted, the total delay of data transmission can be reduced.

[0063] Preferably, in the multi-path routing training set provided in this embodiment, the multi-path routing samples include: the transmission parameters of each routing path (such as bandwidth, latency, packet loss rate, etc.), the data volume and confidentiality level of the sample data to be transmitted, and the optimal multi-path routing algorithm label. During the label annotation process, when the network state is good and the traffic is stable, a suitable algorithm is annotated; when the network is congested and the data is highly important, an algorithm that can provide high availability and fast response is annotated, etc.

[0064] During the training process of the multi-path routing planning large model, the network state of each routing path, the data volume and confidentiality level of the sample data to be transmitted are grouped as the input, and the corresponding optimal multi-path routing algorithm label is used as the output to train the multi-path routing planning large model. By optimizing the loss function (such as cross-entropy loss, mean square error, etc.), the model learns the pattern of selecting a suitable routing algorithm under different network states.

[0065] Common multi-path routing algorithm labels include: Equal-Cost Multi-Path Routing (ECMP), Weighted Multi-Path Routing (WCMP), Global-Aware Network Traffic Allocation Optimization Algorithm Based on Multi-Path Routing, Multi-Path Transmission Control Protocol (MPTCP), Greedy Algorithm, and Genetic-Algorithm-Based Multi-Path Routing. They are respectively adapted to the following situations: (1) Equal-Cost Multi-Path Routing (ECMP): Scenarios where there are multiple paths with the same cost in the network and simple load balancing is required, and there are no complex requirements for traffic allocation; (2) Weighted Multi-Path Routing (WCMP): Networks with unbalanced resources, scenarios that require high availability and stability, multi-link environments, and dynamic network environments; (3) Global-Aware Network Traffic Allocation Optimization Algorithm Based on Multi-Path Routing: Large-scale data processing, real-time data analysis, and microservice architectures, scenarios that require high availability and stability. In network slicing technology, independent virtual networks can be created for different types of services, and the global-awareness algorithm can optimize the traffic allocation of each slice to ensure efficient use of resources; (4) Multi-Path Transmission Control Protocol (MPTCP): Mobile devices that need to transmit data in parallel on multiple network interfaces (such as WiFi and mobile networks), applications with high requirements for bandwidth and fault tolerance, and scenarios suitable for high availability and fast response; (5) Greedy Algorithm: Occasions where the network structure is simple, path selection is not complex, and the requirements for computing resources are relatively low, suitable for preliminary path selection strategies or small networks; (6) Genetic-Algorithm-Based Multi-Path Routing: In complex network environments, the need to find near-global optimal solutions, suitable for occasions where the optimization problem is relatively complex, the network topology changes frequently, and flexible adaptation is required.

[0066] Step S3: Use the data slicing large model and the multi-path routing planning large model that have passed the training to transmit the real-time data to be transmitted.

[0067] The specific implementation process of Step S3 is described as follows.

[0068] Step S31: The data sender receives and determines whether the real-time data to be transmitted can be sliced. If it can, the real-time data to be transmitted, its confidentiality level, and transmission parameters are input into the data slicing large model that has passed the training, and the data slicing rule is predicted and output.

[0069] Step S32: The data sender slices the real-time data to be transmitted according to the data slicing rule to obtain a number of data slices.

[0070] Step S33: Input the transmission parameters of each routing path from the data sender to the data receiver, the data volume and confidentiality level of each data slice of the data to be transmitted into the multi-path routing planning large model that has passed the training, and predict and output the optimal multi-path routing algorithm.

[0071] Step S34: Obtain the multi-path routing of the real-time data to be transmitted according to the predicted optimal multi-path routing algorithm, and transmit each data slice of the data to be transmitted to the data receiver according to the multi-path routing.

[0072] Step S35: The data receiver reorganizes each received data slice according to the data slicing rule to complete the data transmission.

[0073] In the data transmission process of this embodiment, the large model for selecting the multi-path routing algorithm selects an appropriate transmission path according to conditions such as the transmission environment and data level. The data slices after the data packet is split (including the IP information of the access slice and information such as the slice and identification, etc.) are sent simultaneously through multiple paths. The data sender and the receiver obtain the data slicing rule through the consensus algorithm. Finally, the data receiver reorganizes the data slices according to the corresponding rule to complete the data transmission. The consensus algorithm for the data slicing rule means that the data sender obtains the data splitting rule through the consensus algorithm, and the data receiver uses the data splitting rule to reorganize the data. The consensus algorithm can adopt Federated Byzantine Agreement (FBA). The FBA consensus algorithm can reach a consensus in a short time and achieve efficient and secure information confirmation in a large-scale network.

[0074] In summary, the method provided in this embodiment has the following beneficial effects: A large model obtained through training generates appropriate slicing rules, and these rules are used to slice the data, forming different parts of the sliced data to be transmitted that are isolated from each other and do not affect each other, so as to improve the confidentiality and transmission efficiency of the data and enhance the user's business experience. The sliced data is transmitted using multiple paths. Facing various complex network environments, a large model is used to dynamically select appropriate multi-path routing algorithms in various complex environments to enhance bandwidth capabilities, fault tolerance capabilities, achieve load balancing, and reduce latency.

[0075] Based on slicing the data to improve data confidentiality, a large model is used to intelligently generate data splitting rules, enabling the splitting rules to better adapt to data of different sizes and different confidentiality levels. Instead, a large model is used to plan multi-path routing. Through training on a large-scale dataset, the large model can learn from rich knowledge, reducing the dependence on a large amount of training data in a specific environment; the large model has better generalization ability and can adapt to different network environments and traffic patterns.

[0076] Those skilled in the art can understand that all or part of the processes of implementing the method of the above embodiment can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a disk, an optical disc, a read-only memory, or a random access memory, etc.

[0077] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. An optimized data transmission method that comprehensively considers data slicing and multi-path routing selection, characterized in that The method includes: Training a data slice large model using a data slice training set to generate data slice rules, calculating a loss function using the data transmission results based on the data slice rules, and optimizing the model parameters of the data slice large model; obtaining a data slice large model that passes the training after multiple iterations of training; Training a multi-path routing planning large model based on a multi-path routing training set to obtain a multi-path routing planning large model that passes the training; Performing data slicing on the real-time data to be transmitted using the data slice rules predicted by the data slice large model that passes the training, performing multi-path routing transmission on each data slice of the real-time data to be transmitted using the optimal multi-path routing algorithm predicted by the multi-path routing planning large model that passes the training, and the receiving party recombines each data slice according to the predicted data slice rules to complete the transmission.

2. The data transmission optimization method considering data slicing and multi-path routing selection according to claim 1, characterized in that Each data slice training sample in the data slice training set includes a data sample to be transmitted, its confidentiality level, and transmission parameters.

3. The data transmission optimization method considering data slicing and multi-path routing selection according to claim 2, wherein Training the data slice large model by performing the following operations: In each training process, using each data slice training sample in the data slice training set as input, and the data slice large model outputs the matching data slice rules; Performing data slicing on the data sample to be transmitted according to the generated data slice rules to obtain a number of data slices; performing data transmission on each data slice according to the transmission parameters of the data sample to be transmitted; Calculating a comprehensive loss function based on the transmission performance metrics and confidentiality level during the data transmission process, and optimizing the model parameters of the data slice large model according to the comprehensive loss function; Jumping to the next training until the preset training end condition is met, and finally obtaining a data slice large model that passes the training.

4. The data transmission optimization method considering data slicing and multi-path routing selection according to claim 3, characterized in that The comprehensive loss function is expressed as: (1) Among them, , , , respectively represent the resource utilization loss function, the data transmission delay loss function, the packet loss rate loss function, and the security level loss function; , , , respectively represent the weight parameters of the resource utilization loss function, the data transmission delay loss function, the packet loss rate loss function, and the security level loss function.

5. The data transmission optimization method that comprehensively considers data slicing and multi-path routing selection according to claim 4, wherein The transmission performance metrics include the actual transmission volume, the actual average delay, and the number of lost packets.

6. The data transmission optimization method considering data slicing and multi-path routing selection according to claim 5, characterized in that Resource utilization loss function Expressed as: (2) Among them, and respectively represent the actual transmission volume and the theoretical maximum transmission volume of all data slices of the data sample to be transmitted; Data transmission delay loss function It is expressed as: (3) Among them, and respectively represent the actual average delay and the theoretical average delay of all data slices of the data sample to be transmitted; Packet loss rate loss function Expressed as: (4) Among them, and respectively represent the number of lost packets and the total number of transmitted packets of the data slices of the data samples to be transmitted.

7. The data transmission optimization method considering data slicing and multi-path routing selection according to claim 6, characterized in that Confidentiality level loss function representation is as follows: (5) Among them, represents the actual number of data slices, represents the minimum number of data slices corresponding to the confidentiality level of the currently transmitted data sample.

8. The data transmission optimization method considering data slicing and multi-path routing selection according to any one of claims 3-7, characterized in that In the multi-path routing training set, the multi-path routing sample includes: the transmission parameters of each routing path, the data volume and confidentiality level of the data sample to be transmitted, and the optimal multi-path routing algorithm label.

9. The data transmission optimization method considering data slicing and multi-path routing selection according to claim 8, characterized in that Training the multi-path routing planning large model by performing the following operations: During the training of the multi-path routing planning large model, using the network state of each routing path, the data volume and confidentiality level of the data sample to be transmitted as input, and the corresponding optimal multi-path routing algorithm label as output to train the multi-path routing planning large model.

10. The data transmission optimization method considering data slicing and multi-path routing selection according to claim 9, characterized in that Performing transmission on the real-time data to be transmitted by performing the following operations: The data sender receives and determines whether the real-time data to be transmitted can be sliced. If it can, the real-time data to be transmitted, its confidentiality level, and transmission parameters are input into the data slice large model that passes the training, and the predicted output is the data slice rules; The data sender performs data slicing on the real-time data to be transmitted according to the data slice rules to obtain a number of data slices; Inputting the transmission parameters of each routing path from the data sender to the data receiver, the data volume and confidentiality level of each data slice of the data to be transmitted into the multi-path routing planning large model that passes the training, and the predicted output is the optimal multi-path routing algorithm; Obtaining the multi-path routing of the data to be transmitted according to the predicted optimal multi-path routing algorithm, and transmitting each data slice of the real-time data to be transmitted to the data receiver according to the multi-path routing; The data receiver reorganizes the received data slices according to the data slicing rules to complete the data transmission.

Images