Network path switching method and device based on multi-cloud network and computer equipment

Abstract

The application relates to a network path switching method, device and computer equipment based on a multi-cloud network. The method comprises the following steps: performing trend prediction processing based on a network performance data set of each network path obtained in advance to obtain change trend data; performing feature extraction processing and adjustment processing on the network performance data set and the change trend data in sequence to obtain a network path switching instruction; and performing a network path switching action based on the network path switching instruction to switch the transmission path of service data from a current network path to a target network path. The method can improve the working efficiency of the cloud network.

Description

Technical Field

[0001] This application relates to the field of cloud computing technology, and in particular to a network path switching method, apparatus and computer equipment based on a multi-cloud network. Background Technology

[0002] In the digital age, the rapid development of information technology has profoundly changed the operating models of enterprises and organizations. Cloud computing, with its powerful computing capabilities, flexible resource allocation, and efficient data storage and processing advantages, has become a core technological support for driving enterprises to enhance competitiveness and promote innovative development. As businesses continue to expand and data volumes grow dramatically, enterprises and organizations are increasingly reliant on cloud computing.

[0003] To mitigate risks, improve service quality, and fully leverage the unique features of different cloud service providers, an increasing number of enterprises are choosing multi-cloud environments as their technology deployment model. This model allows enterprises to flexibly select and combine multiple cloud services according to their specific needs, thereby achieving optimal resource allocation and ensuring the continuous and stable operation of their businesses.

[0004] However, while multi-cloud environments offer enterprises greater flexibility and improved resource utilization efficiency, network connectivity methods still face numerous challenges in practical applications. Existing methods often suffer from a lack of efficient collaboration and information exchange between modules in a multi-cloud network, leading to poor data flow, delayed information sharing, and consequently impacting the network's overall efficiency. Summary of the Invention

[0005] Therefore, it is necessary to provide a network path switching method, apparatus, and computer device based on a multi-cloud network that can improve the working efficiency of multi-cloud networks, addressing the aforementioned technical problems.

[0006] Firstly, this application provides a network path switching method based on a multi-cloud network, including:

[0007] Trend prediction processing is performed based on the pre-acquired network performance data set of each network path to obtain the trend data;

[0008] The network performance dataset and trend data are sequentially processed for feature extraction and adjustment to obtain the network path switching instruction.

[0009] The network path switching action is executed based on the network path switching command, so that the transmission path of business data is switched from the current network path to the target network path.

[0010] In one embodiment, the aforementioned network performance data set includes multiple sub-data sets arranged in chronological order. Trend prediction processing is performed based on the pre-acquired network performance data sets for each network path to obtain trend data, including:

[0011] Iterative calculations are performed on each sub-data set. In the first iteration, the first sub-data set is input into the trend prediction model, and the trend prediction model is used to extract features from the first sub-data set to obtain the trend features of the first iteration.

[0012] In non-first iterations, the previous trend feature obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model. The trend prediction model is then used to perform feature extraction and feature fusion processing on the sub-data set of the current iteration and the previous trend feature in sequence to obtain the trend feature of the current iteration.

[0013] The trend characteristics obtained based on the last sub-data set are identified as trend data.

[0014] In one embodiment, the network performance data set and trend data are sequentially subjected to feature extraction and adjustment processing to obtain a network path switching instruction, including:

[0015] The instruction generation model is used to perform feature extraction and feature fusion on network performance datasets and trend data respectively to obtain fused data features.

[0016] Based on the network optimization objective, the features of the fused data are adjusted to obtain the adjusted fused data features.

[0017] Based on the characteristics of the fused data and the adjusted fused data characteristics, a network path switching instruction is generated.

[0018] In one embodiment, the above-mentioned use of the instruction generation model to perform feature extraction and feature fusion processing on the network performance data set and the trend data respectively, to obtain fused data features, including:

[0019] The network performance dataset is processed by using an instruction generation model to extract performance features.

[0020] The data features are obtained by using an instruction generation model to extract features from trend data.

[0021] The performance characteristics and data characteristics are fused to obtain the fused data characteristics.

[0022] In one embodiment, the method further includes:

[0023] Based on the characteristics of the fused data, the joint loss function of the trend prediction model and the instruction generation model is determined;

[0024] The trend prediction model and instruction generation model are optimized by tuning the parameters based on the joint loss function.

[0025] In one embodiment, the method further includes:

[0026] The execution results corresponding to the network path switching action will be used as the network performance data set for subsequent use.

[0027] Secondly, this application also provides a network path switching device based on a multi-cloud network, comprising:

[0028] The prediction processing module is used to perform trend prediction processing based on the pre-acquired network performance data set of each network path to obtain the trend data.

[0029] The adjustment processing module is used to sequentially perform feature extraction and adjustment processing on the network performance data set and change trend data to obtain network path switching instructions.

[0030] The path switching module is used to execute network path switching actions based on network path switching instructions, so that the transmission path of business data is switched from the current network path to the target network path.

[0031] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0032] Trend prediction processing is performed based on the pre-acquired network performance data set of each network path to obtain the trend data;

[0033] The network performance dataset and trend data are sequentially processed for feature extraction and adjustment to obtain the network path switching instruction.

[0034] The network path switching action is executed based on the network path switching command, so that the transmission path of business data is switched from the current network path to the target network path.

[0035] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0036] Trend prediction processing is performed based on the pre-acquired network performance data set of each network path to obtain the trend data;

[0037] The network performance dataset and trend data are sequentially processed for feature extraction and adjustment to obtain the network path switching instruction.

[0038] The network path switching action is executed based on the network path switching command, so that the transmission path of business data is switched from the current network path to the target network path.

[0039] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0040] Trend prediction processing is performed based on the pre-acquired network performance data set of each network path to obtain the trend data;

[0041] The network performance dataset and trend data are sequentially processed for feature extraction and adjustment to obtain the network path switching instruction.

[0042] The network path switching action is executed based on the network path switching command, so that the transmission path of business data is switched from the current network path to the target network path.

[0043] The aforementioned network path switching method, apparatus, and computer equipment based on multi-cloud networks first perform trend prediction processing based on pre-acquired network performance data sets of each network path to obtain trend data; then, feature extraction and adjustment processing are sequentially performed on the network performance data sets and trend data to obtain network path switching instructions; finally, network path switching actions are executed based on the network path switching instructions to switch the transmission path of service data from the current network path to the target network path. In this method, by generating network path switching instructions based on network performance data sets and trend prediction data, dynamic path optimization can be performed according to the real-time status and future trends of the network, thereby helping to reduce network congestion, optimize resource allocation, and improve the working efficiency of multi-cloud networks. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is an internal structural diagram of a computer device in one embodiment;

[0046] Figure 2 This is a flowchart illustrating a network path switching method based on a multi-cloud network in one embodiment;

[0047] Figure 3 This is a flowchart illustrating a network path switching method based on a multi-cloud network in another embodiment;

[0048] Figure 4 This is a flowchart illustrating a network path switching method based on a multi-cloud network in another embodiment;

[0049] Figure 5 This is a flowchart illustrating a network path switching method based on a multi-cloud network in another embodiment;

[0050] Figure 6 This is a flowchart illustrating a network path switching method based on a multi-cloud network in another embodiment;

[0051] Figure 7 This is a structural block diagram of a network path switching device based on a multi-cloud network in another embodiment. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0053] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 1 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data during network path switching based on a multi-cloud network. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a network path switching method based on a multi-cloud network.

[0054] Those skilled in the art will understand that Figure 1The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0055] In one exemplary embodiment, such as Figure 2 As shown, a network path switching method based on a multi-cloud network is provided, which can be applied to... Figure 1 The following steps, 201 to 203, are used as an example of computer equipment.

[0056] Step 201: Perform trend prediction processing based on the pre-acquired network performance data set of each network path to obtain change trend data.

[0057] In this context, a network path refers to the transmission path of data from the source node to the destination node. Business data is transmitted through different physical or virtual links, forming multiple available network paths. Due to factors such as bandwidth, latency, and packet loss rate, the network performance of each network path varies.

[0058] A network performance dataset refers to the measurement results of actual performance indicators of a network path. These indicators include, but are not limited to, bandwidth (data transmission rate), latency (time delay in data transmission), and packet loss rate (the proportion of data lost). By monitoring network performance datasets, the health status of each network path can be assessed.

[0059] Trend data refers to the predicted future performance trend of a network path based on a set of network performance data. Based on trend data, potential future performance changes of a network path can be predicted, such as decreased bandwidth or increased latency.

[0060] In this embodiment of the application, the computer device can perform trend prediction processing based on the network performance data set of each network path obtained in advance to obtain the change trend data.

[0061] In some embodiments, the computer device first collects network performance data in real time from various network paths. Since the collected data may contain noise or missing values, data preprocessing is required after collection, such as removing outliers, filling in missing data, and smoothing, to ensure the accuracy and completeness of the data.

[0062] Next, data analysis and feature extraction are required. This process involves extracting valuable features from the processed network performance data. These features can include the average, maximum, and minimum bandwidth values, as well as fluctuation ranges, reflecting the overall health of the network path. Furthermore, computer equipment can analyze the time-series characteristics of the data, such as trends and periodic changes, to identify potential patterns of change in network performance.

[0063] After feature extraction, an appropriate prediction model can be selected based on the extracted features to generate the trend of each network path over a future period of time, i.e., trend data.

[0064] In some embodiments, computer devices visualize trend data to facilitate understanding and decision-making. Predictive results can be intuitively displayed through charts, curves, or dashboards, helping users quickly identify potential network bottlenecks or problems. Simultaneously, detailed analysis reports are generated, summarizing future trends in network paths and providing a scientific basis for subsequent network path switching.

[0065] Step 202: Perform feature extraction and adjustment processing on the network performance data set and change trend data in sequence to obtain the network path switching instruction.

[0066] Feature extraction refers to the process of extracting key features from network performance datasets and trend data. Extracted features may include network path stability, peak latency, and transmission volatility.

[0067] Adjustment processing refers to the process of processing and optimizing the extracted feature data. This may include steps such as data standardization and normalization to eliminate the influence of different data units, or adjusting data biases by weighting and selecting the most important features, thereby improving the accuracy of decision-making.

[0068] Network path switching instructions are control commands generated based on the results of feature extraction and adjustment processing. They specify when and how to switch the current network path to another target network path.

[0069] In this embodiment, the computer device sequentially performs feature extraction and adjustment processing on the network performance data set and change trend data to obtain a network path switching instruction.

[0070] In some embodiments, the computer device needs to perform feature extraction on the network performance dataset. During this stage, the computer device extracts key features from the performance data of each network path, such as bandwidth, latency, packet loss rate, and latency jitter. These features help assess the current condition of each network path. For example, by analyzing the average bandwidth and latency fluctuations, the computer device can determine whether the current network path is stable and suitable to continue as a data transmission channel. Furthermore, the computer device needs to extract long-term trends and short-term fluctuation patterns from the trend data. This information helps predict future performance changes in network paths, thereby supporting path switching decisions.

[0071] After features are extracted, the next step is feature adjustment. The goal of this stage is to optimize the feature data to make it more suitable for subsequent decision-making. Common processing methods include standardization or normalization, which eliminates differences in the units of measurement between different features, making them have the same scale. Furthermore, some features may be more important in path switching decisions, so their influence can be highlighted through weighting. For example, latency and packet loss rate are important for real-time communication and therefore may need to be given higher weights. Feature selection techniques can also be used to remove redundant or irrelevant features, thereby simplifying the data and improving decision-making efficiency.

[0072] After feature extraction and adjustment, the computer device can build a decision model based on the processed feature data to determine whether network path switching is necessary. The decision model can make basic path selection decisions using threshold judgments, such as triggering a switch when latency or packet loss rate exceeds a preset tolerance value. In more complex cases, the computer device can use machine learning models (such as decision trees and support vector machines) to analyze historical data, learn network path performance patterns, and thus make more accurate predictions and decisions.

[0073] Finally, based on the output of the decision model, the computer device generates a network path switching instruction. This instruction specifies when and where the path switch should occur to ensure the smooth transmission of business data. During the path switch, the computer device adjusts routing, load balancing, and other configurations according to the instruction, thereby switching the data flow from the current network path to the target network path.

[0074] Step 203: Execute a network path switching action based on the network path switching instruction to switch the transmission path of service data from the current network path to the target network path.

[0075] The network path switching action refers to the actual process of switching the network path according to the network path switching instruction. This may include operations such as adjusting the route, reconfiguring the connection, and switching the load on the network device, with the aim of transferring business data from the current network path to a better target path.

[0076] Business data refers to the data stream that needs to be transmitted over the network. It usually represents the transmission content of a specific application, such as web page access data, file downloads, video streams, etc.

[0077] In this embodiment, the computer device executes a network path switching action based on a network path switching instruction, so that the transmission path of the service data is switched from the current network path to the target network path.

[0078] In some embodiments, after receiving a path switching instruction, the computer device needs to parse and verify the instruction to determine its validity and accuracy. The verification process may include verifying whether the target network path is available and whether the switching conditions have been met. For example, information such as the target path, switching timing, and priority contained in the network path switching instruction will be carefully checked to avoid network interruptions or instability caused by instruction errors.

[0079] Next, the computer equipment will evaluate the availability and stability of the target network path. This process involves checking the current network performance metrics of the target network path, such as bandwidth, latency, and packet loss rate, to ensure that the target path can meet the transmission requirements of the business data.

[0080] Once the target network path is confirmed to be available, the computer equipment will perform necessary resource scheduling and routing configuration. This includes updating routing tables, adjusting routing rules on network devices, and possibly reconfiguring load balancing policies to ensure a smooth transition of data flow from the current path to the target path. Furthermore, other relevant devices in the network (such as firewalls, routers, or load balancers) may also need to adjust their settings to accommodate the new path configuration.

[0081] The path switching operation is formally executed after all configurations are complete. At this point, the computer equipment will forward service data to the target network path according to the new routing rules. After the path switch is completed, the computer equipment will monitor the performance of the new target path in real time to verify the switching effect. This may include checking key performance indicators such as bandwidth, latency, and packet loss rate of the target path to ensure that the switched target network path meets the expected performance standards. If the monitoring results show that the performance of the target network path is lower than expected, the computer equipment should take contingency measures in a timely manner, such as falling back to the original path or selecting another available path.

[0082] In the aforementioned network path switching method based on multi-cloud networks, trend prediction processing is first performed on the pre-acquired network performance data set of each network path to obtain trend data. Then, feature extraction and adjustment processing are performed on the network performance data set and trend data in sequence to obtain network path switching instructions. Finally, the network path switching action is executed based on the network path switching instructions to switch the transmission path of service data from the current network path to the target network path. In this method, by generating network path switching instructions based on network performance data set and trend prediction data, dynamic path optimization can be performed according to the real-time status and future trends of the network, thereby helping to reduce network congestion, optimize resource allocation, and improve the working efficiency of multi-cloud networks.

[0083] In one exemplary embodiment, such as Figure 3 As shown, the aforementioned network performance data set includes multiple sub-data sets arranged in chronological order. Based on this, the step of "performing trend prediction processing based on the pre-acquired network performance data sets of each network path to obtain trend data" includes steps 301 to 303. Wherein:

[0084] Step 301: Perform iterative calculations for each sub-data set. In the first iteration, input the first sub-data set into the trend prediction model and use the trend prediction model to perform feature extraction processing on the first sub-data set to obtain the trend features of the first iteration.

[0085] In this context, a subdata set refers to a subset of the overall data. Typically, these subdata sets are divided by time period, representing network performance data for a specific timeframe. Each subdata set may contain network performance metrics such as bandwidth, latency, and packet loss rate, used to analyze network trends.

[0086] A trend prediction model is a model used to predict the future trend of data. By analyzing each sub-data set, the model can discover patterns and regularities in the data and predict future changes based on these regularities.

[0087] Feature extraction refers to extracting useful information for problem analysis from various sub-data sets. These features are typically representative or key information portions of the data, such as bandwidth fluctuations or periodic changes in latency. In trend forecasting, the goal of feature extraction is to transform complex sub-data sets into concise representations that can be used for modeling and prediction.

[0088] Trend characteristics are features extracted from each sub-dataset that describe the patterns of data change. In network performance analysis, trend characteristics may include trends in bandwidth (increase or decrease), latency (increase or decrease), etc. These characteristics are used to reflect the changing trends of each sub-dataset over time.

[0089] In this embodiment, during the initial iteration, the computer device inputs the first sub-data set into the trend prediction model for processing. The first sub-data set contains preliminary performance data of the network path or other relevant data. The trend prediction model performs feature extraction processing on the first sub-data set to extract key trend features, thus obtaining the trend features for the first iteration.

[0090] In some embodiments, the trend prediction model can use a recurrent neural network (RNN) because RNNs can remember previous information and combine it with the current input, thereby capturing the temporal dependencies in the data. Therefore, they have a unique advantage when dealing with multiple time-series sub-data sets. Assume the input time series is... The hidden state of RNN The update formula is as follows:

[0091]

[0092] Where W is the weight matrix and b is the bias term. It is an activation function. It is the hidden state from the previous moment. This represents the input data of the recurrent neural network at time step [time]. During training, the RNN adjusts the weights and biases using the backpropagation algorithm to minimize the error between the predicted and actual values. The loss function typically uses mean squared error (MSE) or cross-entropy. For long sequences of data, the problems of vanishing or exploding gradients may occur. Using a Long Short-Term Memory (LSTM) network to improve the RNN structure can alleviate gradient explosion. Taking LSTM as an example, it introduces an input gate, a forget gate, and an output gate to control the flow and retention of information. Input gate... This determines how much of the currently entered information is saved:

[0093]

[0094] in, The output of the input gate, The activation function is usually the sigmoid function. This is the weight matrix. This is the bias term for the input gate.

[0095] Forgotten Gate This determines how much of the past is forgotten:

[0096]

[0097] in, For the output of the forget gate, This is the weight matrix. This is the bias term for the forget gate.

[0098] Output gate This determines how much of the past is forgotten:

[0099]

[0100] in, For the output of the output gate, This is the weight matrix. The bias term for the gate output.

[0101] Through these gating mechanisms, LSTM can better handle long-term dependencies in long sequence data, thereby more accurately predicting the changing trends of network traffic.

[0102] Step 302: In the non-first iteration process, the previous trend feature obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model. The trend prediction model is used to perform feature extraction and feature fusion processing on the sub-data set of the current iteration and the previous trend feature in sequence to obtain the trend feature of the current iteration.

[0103] The non-first iteration process refers to each round of calculation after the first iteration. Unlike the first iteration, the non-first iteration relies not only on the current sub-data set but also on the trend characteristics obtained from the previous iteration, in order to perform more accurate analysis and prediction of the data.

[0104] Feature fusion refers to the process of merging or integrating data features from different sources or time points. In subsequent iterations, the computer combines the trend features from the previous iteration with the features of the current iteration's sub-dataset. The purpose of feature fusion is to enhance the understanding of data changes and obtain more comprehensive and accurate trend features by combining different features.

[0105] In this embodiment, during the non-first iteration process, the computer device performs further calculations based on the results of the previous iteration in each iteration. During this process, the computer device inputs the trend characteristics of the previous iteration (i.e., the output characteristics of the previous iteration) together with the new sub-data set of the current iteration into the trend prediction model.

[0106] The trend prediction model performs feature extraction on the sub-dataset of the current iteration to extract the main trend features of the current data. Then, the computer system performs feature fusion processing on the trend features from the previous round and the features extracted this time to obtain the trend features of the current iteration.

[0107] Step 303: The trend characteristics obtained based on the last sub-data set are determined as trend data.

[0108] In this embodiment of the application, after several rounds of iteration, the computer device will eventually determine the overall trend data based on the trend characteristics obtained from the last sub-data set.

[0109] In some embodiments, the process of acquiring network performance data sets for each network path may include: using Simple Network Management Protocol (SNMP) to retrieve network performance data sets from the network, including port traffic, error rate, etc. The performance metric values ​​corresponding to a specific object identifier can be obtained through SNMP's Get and GetNext commands. Furthermore, by combining sampled flow (sFlow) technology, in-depth analysis of network traffic can be performed. This analysis can obtain rich information about the source address, destination address, protocol type, and application type of the traffic, which helps to more accurately evaluate network performance. By comprehensively utilizing SNMP and sFlow technologies, network path performance parameters can be obtained comprehensively, in real-time, and accurately, providing strong support for subsequent analysis, decision-making, and optimization.

[0110] In the above embodiments, during the first iteration, feature extraction is performed using the first sub-dataset to generate preliminary trend features. In subsequent iterations, by fusing the trend features obtained from the previous iteration with the current sub-dataset, trend prediction can be continuously adjusted and optimized. This iterative process effectively utilizes historical and new data, gradually improving the accuracy of trend prediction, enabling the trend prediction model to adapt to changes in network conditions and provide more accurate trend predictions.

[0111] In one exemplary embodiment, such as Figure 4 As shown, the above-mentioned "performing feature extraction and adjustment processing on the network performance data set and change trend data sequentially to obtain network path switching instructions" includes steps 401 to 403. Among them:

[0112] Step 401: Use the instruction generation model to perform feature extraction and feature fusion processing on the network performance data set and the trend data respectively to obtain fused data features.

[0113] Among them, the instruction generation model is a data-driven algorithm model. Its main function is to generate network path switching instructions based on the input network performance data and trend data to execute path switching operations.

[0114] Fusion data features refer to the comprehensive features obtained after feature extraction and feature fusion processing. Fusion data features integrate key information from network performance data and trend data, reflecting the current state of the network path and its future trends. Fusion data features provide comprehensive input for subsequent network path switching decisions, helping to determine whether a path switch is necessary or to select the optimal path.

[0115] In this embodiment of the application, the network performance data set and the trend data are first merged into a data set. Then, the model data set is generated by the instruction and the feature extraction process is performed to extract relevant features. Then, multiple relevant features are fused to obtain fused data features.

[0116] Step 402: Adjust the fused data features based on the network optimization objective to obtain the adjusted fused data features.

[0117] Network optimization objectives refer to the desired optimization effects achieved during network path switching or network resource allocation. These objectives can include increasing network throughput, reducing latency, minimizing packet loss, optimizing bandwidth allocation, and reducing energy consumption. Determining network optimization objectives is fundamental to network path switching decisions; different objectives lead to different switching decisions.

[0118] Adjustment processing refers to the process of adjusting the characteristics of fused data based on network optimization objectives. Depending on the specific optimization goal, the fused data characteristics are further optimized to better guide subsequent decision-making. For example, under the optimization objective of increasing throughput, the adjusted fused data characteristics may focus more on the correlation between bandwidth and traffic, while under the objective of reducing latency, the adjusted characteristics may focus more on network latency characteristics.

[0119] In this embodiment, after obtaining the fused data features, the computer device adjusts the fused data features based on network optimization objectives. During the adjustment process, the computer device weights or filters the fused data features according to preset optimization objectives, so that the final fused data features can most effectively support the achievement of network optimization objectives.

[0120] Step 403: Generate network path switching instructions based on the fused data features and the adjusted fused data features.

[0121] In this embodiment of the application, based on the adjusted fused data features, the instruction generation model will generate corresponding network path switching instructions to execute path switching operations at an appropriate time.

[0122] In the above embodiments, by extracting and fusing features from network performance data and trend data, and adjusting the data in conjunction with network optimization objectives, a network path switching instruction is finally generated, thereby realizing automated network optimization and intelligent scheduling. This helps business data dynamically adapt to network changes, enabling the network to achieve optimal performance in different usage scenarios.

[0123] In one exemplary embodiment, such as Figure 5 As shown, the above-mentioned "using the instruction generation model to perform feature extraction and feature fusion processing on the network performance data set and the trend data respectively to obtain fused data features" includes steps 501 to 503. Wherein:

[0124] Step 501: Use the instruction generation model to perform feature extraction processing on the network performance data set to obtain performance features.

[0125] Performance characteristics are key information extracted from network performance datasets, typically including various metrics representing network operating status, such as bandwidth, latency, packet loss rate, and throughput. These characteristics reflect the network's performance status over a specific time period and form the basis for network optimization decisions.

[0126] In this embodiment, the instruction generation model extracts performance characteristics from a network performance dataset, such as network bandwidth, latency, packet loss rate, and throughput. These performance characteristics reflect the current operating status of the network and provide a basis for subsequent analysis.

[0127] Step 502: Use the instruction generation model to perform feature extraction processing on the trend data to obtain data features.

[0128] Data features are key information extracted from trending data, typically including the trend, magnitude, and periodicity of network performance changes. These features can reflect potential future changes in the network, such as predicting latency increases or decreases and bandwidth fluctuations, thus providing a basis for decisions such as network path switching.

[0129] In this embodiment, the instruction generation model performs feature extraction processing on the trend data to extract data features such as bandwidth growth trends, latency change trends, and traffic fluctuations. These data features can reflect future network change patterns.

[0130] Step 503: Perform feature fusion processing on the performance features and data features to obtain fused data features.

[0131] Among them, fused data features are a comprehensive set of features obtained through feature extraction and fusion processing. Fusion data features contain key information about the network's current performance (such as bandwidth, latency, throughput, etc.) and its changing trends (such as future performance changes and fluctuations).

[0132] In this embodiment, performance characteristics and data characteristics are fused to obtain fused data characteristics. Through feature fusion, the computer device can integrate the current state of the network and future trends to obtain more comprehensive network performance characteristics.

[0133] In the above embodiments, by extracting features from network performance data and trend data respectively, the key information in these two types of data can be fully explored. By fusing the features of these two types of data, the current network performance and future trends can be comprehensively considered, thereby providing a more accurate network path prediction.

[0134] In one exemplary embodiment, such as Figure 6 As shown, the above method may further include steps 601 to 602. Wherein:

[0135] Step 601: Based on the characteristics of the fused data, determine the joint loss function of the trend prediction model and the instruction generation model.

[0136] The joint loss function is used to simultaneously optimize the trend prediction model and the instruction generation model. It combines the losses of the trend prediction model and the instruction generation model into a single overall loss. By minimizing the joint loss function, the trend prediction model and the instruction generation model can jointly improve their performance.

[0137] In this embodiment of the application, the joint loss function of the trend prediction model and the instruction generation model is determined based on the characteristics of the fused data.

[0138] In some embodiments, a loss function needs to be defined that can simultaneously evaluate the performance of the trend prediction model and the instruction generation model. This loss function combines the objectives of the trend prediction model and the instruction generation model: the trend prediction model aims to accurately predict the changing trend of network performance, while the instruction generation model aims to generate optimized network path switching instructions based on these predictions. The joint loss function weighted and merges the errors of these two objectives to form an overall loss metric. The joint loss function includes prediction errors (such as trend prediction errors) and instruction generation errors (such as path switching errors) to ensure that the trend prediction model and the instruction generation model can be jointly optimized during training to achieve the best network optimization results.

[0139] In some embodiments, the loss value corresponding to the joint loss function can be obtained by first determining the evaluation rules corresponding to the trend prediction model and the instruction generation model, then obtaining the loss value corresponding to the evaluation rule based on the evaluation rule, and finally weighting the various loss values ​​to obtain the final loss value. The evaluation rule may include:

[0140] (1) Network latency assessment rules:

[0141] Define the maximum average network latency threshold as 50 milliseconds, calculate the average network latency over a period of time (e.g., 1 minute), and if the average network latency exceeds 50 milliseconds, then the loss value = (actual average latency - 50) × 10.

[0142] (2) Bandwidth utilization evaluation rules:

[0143] Set a minimum bandwidth utilization target of 70%, and calculate the actual bandwidth utilization. If the actual bandwidth utilization is lower than 70%, the loss value is calculated as (70 - actual utilization) × 5.

[0144] (3) Network topology adaptability assessment rules:

[0145] When the network topology changes, a timer is started, requiring the analysis and decision-making module to respond and adjust the network connection strategy within 10 seconds. If no response is received within 10 seconds, the penalty value is 20.

[0146] (4) Traffic burst handling assessment rules:

[0147] Monitor sudden increases in traffic, set a threshold of 50% increase in a short period of time (e.g., 30 seconds), and calculate the packet loss rate during the traffic surge. If the packet loss rate exceeds 5%, the loss value is calculated as (packet loss rate - 5) × 100.

[0148] (5) Safety assessment rules:

[0149] Check the data encryption status. If it is not encrypted, the loss value is 50. Monitor the number of attacks received. For each attack exceeding the preset security attack threshold, the loss value is 10.

[0150] (6) Fault recovery time assessment rules:

[0151] When a network failure occurs, a timer is started, requiring the failure recovery time to be no more than 5 minutes. If it exceeds 5 minutes, the loss value = (actual recovery time - 5) × 5.

[0152] (7) Quality of Service (QoS) Assessment Rules

[0153] Compare the availability guarantee metrics with the QoS agreement signed with the cloud service provider, such as a requirement of 99.9% availability, and calculate the actual availability. If the actual availability is lower than the agreement requirement, the loss value = (agreement availability - actual availability) × 1000.

[0154] Suppose the loss formula obtained from the results generated based on the above rules is as follows:

[0155]

[0156] The loss value corresponding to the network latency assessment rule is... The loss value corresponding to the bandwidth utilization assessment rule is The loss value corresponding to the network topology adaptability evaluation rule is The loss value corresponding to the traffic burst handling assessment rule is The loss value corresponding to the security assessment rule is The loss value corresponding to the fault recovery time assessment rule is The loss value corresponding to the Quality of Service (QoS) assessment rules is And set the weight parameters respectively. to Finally, the values ​​are added together to obtain the loss value. This is a regularization term used to prevent overfitting.

[0157] Step 602: Perform parameter tuning on the trend prediction model and instruction generation model based on the joint loss function to optimize the trend prediction model and instruction generation model.

[0158] Parameter tuning refers to optimizing the performance of the trend prediction model and instruction generation model during training by adjusting the model's hyperparameters (such as learning rate and regularization parameters). In this process, the trend prediction model and instruction generation model adjust based on feedback from the joint loss function to improve their performance on the task. Parameter tuning typically uses methods such as gradient descent, grid search, and random search to find the optimal combination of hyperparameters to achieve the best prediction accuracy and instruction generation performance of the trend prediction model and instruction generation model.

[0159] In this embodiment, the computer device, based on a joint loss function, jointly tunes the parameters of the trend prediction model and the instruction generation model. The parameter tuning process includes optimizing model hyperparameters, such as the learning rate, regularization coefficient, and network structure, to improve the model's accuracy and robustness. During this process, the trend prediction model and the instruction generation model adjust their parameters based on feedback from the joint loss function to minimize errors and improve overall performance. For example, the trend prediction model might adjust its prediction accuracy to make predictions of future network states more accurate; simultaneously, the instruction generation model will generate more efficient path switching instructions based on these more accurate predictions.

[0160] The optimized trend prediction model and instruction generation model will be able to better predict the changing trends of network performance and generate more accurate network path switching instructions based on these predictions.

[0161] In the above embodiments, by optimizing the joint loss function, the performance of the trend prediction model and the instruction generation model will be improved simultaneously, thereby making network path switching more intelligent and efficient, which helps to improve the overall stability and performance of the network. This not only can cope with the current state of the network, but also make forward-looking adjustments based on the predicted changing trends, further improving the utilization efficiency and responsiveness of network resources.

[0162] In an exemplary embodiment, the above-mentioned "execute network path switching action based on network path switching instruction to switch the transmission path of service data from the current network path to the target network path" may include:

[0163] Labels are assigned to service data from a pre-configured label forwarding table based on network path switching instructions, so that service data can switch from the current network path to the target network path based on the label.

[0164] The label forwarding table is a mapping table used to store network path switching information, containing the mapping relationship between labels and network paths. During network path switching, labels are used to identify different network paths and corresponding traffic. A pre-configured label forwarding table stores this predefined label information so that when a path switching is needed, the appropriate label can be quickly found and applied for data flow forwarding.

[0165] Labels are identifiers used to identify specific data flows or network paths during network path switching. By assigning labels to service data, the transmission path is switched from the current network path to the target network path, and correct routing and processing are performed based on the labels.

[0166] In this embodiment, during path switching, the computer device needs to obtain the label corresponding to the target network path by looking up a pre-configured label forwarding table. Then, based on the target network path label found in the label forwarding table, the computer device assigns a new label to the currently transmitted service data. This label indicates the new transmission path for the service data. After the service data is assigned a new label, the computer device will perform data flow forwarding operations according to the label, switching the service data from the current network path to the target network path.

[0167] In the above embodiments, traditional network path switching requires a considerable amount of time to recalculate the optimal path or involves more complex signal switching and protocol operations. However, path switching based on a tag forwarding table allows the computer device to quickly determine the data transmission path after tag assignment, thereby significantly reducing path switching latency.

[0168] In one exemplary embodiment, the method further includes:

[0169] The execution results corresponding to the network path switching action will be used as the network performance data set for subsequent use.

[0170] In this embodiment, after the path switching operation is completed, the computer device needs to monitor and record the network performance data on the new path in real time. The monitored network performance data is stored and recorded to form a new network performance data set. The purpose of recording the network performance data set is to provide a basis for subsequent decision-making and network adjustments. In a dynamic environment, network load and performance may change continuously, so it is necessary to continuously update this data to reflect the current network status.

[0171] Computer devices can evaluate the effectiveness of current network paths based on new sets of network performance data and adjust or optimize network configurations as needed. For example, if a path is performing poorly, the computer device may generate a new network path switching instruction to reroute traffic to a better-performing path.

[0172] In the above embodiments, by continuously monitoring and recording network performance after path switching, the computer device can continuously adjust network strategies based on real-time data, thereby improving the overall stability and performance of the network.

[0173] According to some embodiments of this application, a network path switching method based on a multi-cloud network is provided. Taking the application of this method to a computer device as an example, it may include the following steps:

[0174] Step 1: Perform iterative calculations for each sub-data set. In the first iteration, input the first sub-data set into the trend prediction model and use the trend prediction model to perform feature extraction processing on the first sub-data set to obtain the trend features of the first iteration.

[0175] Step 2: In the non-first iteration process, the previous trend feature obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model. The trend prediction model is used to perform feature extraction and feature fusion processing on the sub-data set of the current iteration and the previous trend feature in sequence to obtain the trend feature of the current iteration.

[0176] Step 3: Determine the trend characteristics obtained based on the last sub-data set as the trend data.

[0177] Step 4: Use the instruction generation model to perform feature extraction on the network performance data set to obtain performance features.

[0178] Step 5: Use the instruction generation model to extract features from the trend data to obtain data features.

[0179] Step 6: Perform feature fusion processing on the performance features and data features to obtain fused data features.

[0180] Step 7: Adjust the features of the fused data based on the network optimization objective to obtain the adjusted fused data features.

[0181] Step 8: Generate network path switching instructions based on the fused data features and the adjusted fused data features.

[0182] In some embodiments, the trend prediction model and the instruction generation model described above can be jointly trained, and the training process may include:

[0183] Based on the characteristics of the fused data, a joint loss function for the trend prediction model and the instruction generation model is determined. The joint loss function is then used to tune the parameters of both models to optimize them.

[0184] Step 9: Assign a label to the service data from the pre-configured label forwarding table based on the network path switching instruction, so that the service data can switch from the current network path to the target network path based on the label.

[0185] Step 10: Use the execution results corresponding to the network path switching action as the network performance data set for subsequent use.

[0186] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0187] Based on the same inventive concept, this application also provides a network path switching device for implementing the network path switching method based on a multi-cloud network as described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the network path switching device based on a multi-cloud network provided below can be found in the limitations of the network path switching method based on a multi-cloud network described above, and will not be repeated here.

[0188] In one exemplary embodiment, such as Figure 7 As shown, a network path switching device based on a multi-cloud network is provided, including: a prediction processing module 701, an adjustment processing module 702, and a path switching module 703, wherein:

[0189] The prediction processing module 701 is used to perform trend prediction processing based on the pre-acquired network performance data set of each network path to obtain the change trend data;

[0190] The adjustment processing module 702 is used to sequentially perform feature extraction and adjustment processing on the network performance data set and change trend data to obtain network path switching instructions.

[0191] The path switching module 703 is used to execute network path switching actions based on network path switching instructions, so as to switch the transmission path of service data from the current network path to the target network path.

[0192] In an exemplary embodiment, the prediction processing module 701 is specifically used to perform iterative calculations for each sub-data set. In the first iteration, the first sub-data set is input into the trend prediction model, and the trend prediction model is used to perform feature extraction processing on the first sub-data set to obtain the trend features of the first iteration. In subsequent iterations, the previous trend features obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model, and the trend prediction model is used to perform feature extraction processing and feature fusion processing on the sub-data set of the current iteration and the previous trend features in sequence to obtain the trend features of the current iteration. The trend features obtained based on the last sub-data set are determined as the trend data.

[0193] In an exemplary embodiment, the adjustment processing module 702 is specifically used to perform feature extraction and feature fusion processing on the network performance data set and the trend data using the instruction generation model to obtain fused data features; to adjust the fused data features based on the network optimization objective to obtain adjusted fused data features; and to generate a network path switching instruction based on the fused data features and the adjusted fused data features.

[0194] In an exemplary embodiment, the adjustment processing module 702 is specifically used to perform feature extraction processing on the network performance data set using the instruction generation model to obtain performance features; to perform feature extraction processing on the trend data using the instruction generation model to obtain data features; and to perform feature fusion processing on the performance features and data features to obtain fused data features.

[0195] In one exemplary embodiment, the above-described apparatus may further include:

[0196] The loss function determination module 704 is used to determine the joint loss function of the trend prediction model and the instruction generation model based on the features of the fused data.

[0197] The parameter tuning module 705 is used to perform parameter tuning on the trend prediction model and the instruction generation model based on the joint loss function, so as to optimize the trend prediction model and the instruction generation model.

[0198] In an exemplary embodiment, the path switching module 703 is specifically used to assign a label to service data from a pre-configured label forwarding table based on a network path switching instruction, so that the service data can be switched from the current network path to the target network path based on the label.

[0199] In one exemplary embodiment, the above-described apparatus may further include:

[0200] Feedback module 705 is used to use the execution result corresponding to the network path switching action as a set of network performance data for subsequent use.

[0201] The modules in the aforementioned multi-cloud network path switching device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0202] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0203] Trend prediction processing is performed based on the pre-acquired network performance data set of each network path to obtain the trend data;

[0204] The network performance dataset and trend data are sequentially processed for feature extraction and adjustment to obtain the network path switching instruction.

[0205] The network path switching action is executed based on the network path switching command, so that the transmission path of business data is switched from the current network path to the target network path.

[0206] In one embodiment, the network performance data set includes multiple sub-data sets arranged in chronological order, and the processor, when executing the computer program, further performs the following steps:

[0207] Iterative calculations are performed on each sub-data set. In the first iteration, the first sub-data set is input into the trend prediction model, and the trend prediction model is used to extract features from the first sub-data set to obtain the trend features of the first iteration.

[0208] In non-first iterations, the previous trend feature obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model. The trend prediction model is then used to perform feature extraction and feature fusion processing on the sub-data set of the current iteration and the previous trend feature in sequence to obtain the trend feature of the current iteration.

[0209] The trend characteristics obtained based on the last sub-data set are identified as trend data.

[0210] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0211] The instruction generation model is used to perform feature extraction and feature fusion on network performance datasets and trend data respectively to obtain fused data features.

[0212] Based on the network optimization objective, the features of the fused data are adjusted to obtain the adjusted fused data features.

[0213] Based on the characteristics of the fused data and the adjusted fused data characteristics, a network path switching instruction is generated.

[0214] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0215] The network performance dataset is processed by using an instruction generation model to extract performance features.

[0216] The data features are obtained by using an instruction generation model to extract features from trend data.

[0217] The performance characteristics and data characteristics are fused to obtain the fused data characteristics.

[0218] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0219] Based on the characteristics of the fused data, the joint loss function of the trend prediction model and the instruction generation model is determined;

[0220] The trend prediction model and instruction generation model are optimized by tuning the parameters based on the joint loss function.

[0221] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0222] Labels are assigned to service data from a pre-configured label forwarding table based on network path switching instructions, so that service data can switch from the current network path to the target network path based on the label.

[0223] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0224] The execution results corresponding to the network path switching action will be used as the network performance data set for subsequent use.

[0225] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0226] Trend prediction processing is performed based on the pre-acquired network performance data set of each network path to obtain the trend data;

[0227] The network performance dataset and trend data are sequentially processed for feature extraction and adjustment to obtain the network path switching instruction.

[0228] The network path switching action is executed based on the network path switching command, so that the transmission path of business data is switched from the current network path to the target network path.

[0229] In one embodiment, the network performance dataset includes multiple sub-data sets arranged in chronological order, and the computer program, when executed by a processor, further performs the following steps:

[0230] Iterative calculations are performed on each sub-data set. In the first iteration, the first sub-data set is input into the trend prediction model, and the trend prediction model is used to extract features from the first sub-data set to obtain the trend features of the first iteration.

[0231] In non-first iterations, the previous trend feature obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model. The trend prediction model is then used to perform feature extraction and feature fusion processing on the sub-data set of the current iteration and the previous trend feature in sequence to obtain the trend feature of the current iteration.

[0232] The trend characteristics obtained based on the last sub-data set are identified as trend data.

[0233] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0234] The instruction generation model is used to perform feature extraction and feature fusion on network performance datasets and trend data respectively to obtain fused data features.

[0235] Based on the network optimization objective, the features of the fused data are adjusted to obtain the adjusted fused data features.

[0236] Based on the characteristics of the fused data and the adjusted fused data characteristics, a network path switching instruction is generated.

[0237] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0238] The network performance dataset is processed by using an instruction generation model to extract performance features.

[0239] The data features are obtained by using an instruction generation model to extract features from trend data.

[0240] The performance characteristics and data characteristics are fused to obtain the fused data characteristics.

[0241] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0242] Based on the characteristics of the fused data, the joint loss function of the trend prediction model and the instruction generation model is determined;

[0243] The trend prediction model and instruction generation model are optimized by tuning the parameters based on the joint loss function.

[0244] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0245] Labels are assigned to service data from a pre-configured label forwarding table based on network path switching instructions, so that service data can switch from the current network path to the target network path based on the label.

[0246] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0247] The execution results corresponding to the network path switching action will be used as the network performance data set for subsequent use.

[0248] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:

[0249] Trend prediction processing is performed based on the pre-acquired network performance data set of each network path to obtain the trend data;

[0250] The network performance dataset and trend data are sequentially processed for feature extraction and adjustment to obtain the network path switching instruction.

[0251] The network path switching action is executed based on the network path switching command, so that the transmission path of business data is switched from the current network path to the target network path.

[0252] In one embodiment, the network performance dataset includes multiple sub-data sets arranged in chronological order, and the computer program, when executed by a processor, further performs the following steps:

[0253] Iterative calculations are performed on each sub-data set. In the first iteration, the first sub-data set is input into the trend prediction model, and the trend prediction model is used to extract features from the first sub-data set to obtain the trend features of the first iteration.

[0254] In non-first iterations, the previous trend feature obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model. The trend prediction model is then used to perform feature extraction and feature fusion processing on the sub-data set of the current iteration and the previous trend feature in sequence to obtain the trend feature of the current iteration.

[0255] The trend characteristics obtained based on the last sub-data set are identified as trend data.

[0256] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0257] The instruction generation model is used to perform feature extraction and feature fusion on network performance datasets and trend data respectively to obtain fused data features.

[0258] Based on the network optimization objective, the features of the fused data are adjusted to obtain the adjusted fused data features.

[0259] Based on the characteristics of the fused data and the adjusted fused data characteristics, a network path switching instruction is generated.

[0260] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0261] The network performance dataset is processed by using an instruction generation model to extract performance features.

[0262] The data features are obtained by using an instruction generation model to extract features from trend data.

[0263] The performance characteristics and data characteristics are fused to obtain the fused data characteristics.

[0264] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0265] Based on the characteristics of the fused data, the joint loss function of the trend prediction model and the instruction generation model is determined;

[0266] The trend prediction model and instruction generation model are optimized by tuning the parameters based on the joint loss function.

[0267] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0268] Labels are assigned to service data from a pre-configured label forwarding table based on network path switching instructions, so that service data can switch from the current network path to the target network path based on the label.

[0269] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0270] The execution results corresponding to the network path switching action will be used as the network performance data set for subsequent use.

[0271] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0272] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0273] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A network path switching method based on multi-cloud networks, characterized in that, The method includes: Trend prediction processing is performed based on the pre-acquired network performance data set of each network path to obtain the trend data; The network performance data set and the trend data are sequentially subjected to feature extraction and adjustment processes to obtain network path switching instructions. Based on the network path switching instruction, a network path switching action is executed to switch the transmission path of service data from the current network path to the target network path. The network performance data set includes multiple sub-data sets arranged in chronological order. The trend prediction processing based on the pre-acquired network performance data sets for each network path yields trend data, including: Iterative calculations are performed on each of the aforementioned sub-data sets. In the first iteration, the first sub-data set is input into the trend prediction model, and the trend prediction model is used to perform feature extraction processing on the first sub-data set to obtain the trend features of the first iteration. In non-first iterations, the previous trend feature obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model. The trend prediction model is then used to perform feature extraction and feature fusion processing on the sub-data set of the current iteration and the previous trend feature in sequence to obtain the trend feature of the current iteration. The trend characteristics obtained based on the last sub-data set are determined as the trend data.

2. The method according to claim 1, characterized in that, The process of sequentially performing feature extraction and adjustment on the network performance data set and the trend data to obtain network path switching instructions includes: The network performance dataset and the trend data are subjected to feature extraction and feature fusion processing respectively using the instruction generation model to obtain fused data features. The fused data features are adjusted based on the network optimization objective to obtain the adjusted fused data features. Based on the fused data features and the adjusted fused data features, the network path switching instruction is generated.

3. The method according to claim 2, characterized in that, The method of using an instruction generation model to perform feature extraction and feature fusion processing on the network performance data set and the trend data respectively, to obtain fused data features, including: The network performance data set is processed by the instruction generation model to obtain performance features; The aforementioned instruction-generated model is used to perform feature extraction processing on the trend data to obtain data features; The performance features and the data features are subjected to feature fusion processing to obtain the fused data features.

4. The method according to claim 2, characterized in that, The method further includes: Based on the characteristics of the fused data, the joint loss function of the trend prediction model and the instruction generation model is determined; The trend prediction model and the instruction generation model are tuned based on the joint loss function to optimize them.

5. The method according to claim 1, characterized in that, The step of executing a network path switching action based on the network path switching instruction to switch the transmission path of service data from the current network path to the target network path includes: Based on the network path switching instruction, a label is assigned to the service data from a pre-configured label forwarding table, so that the service data is switched from the current network path to the target network path based on the label.

6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: The execution results corresponding to the network path switching action will be used as a set of network performance data for subsequent use.

7. A network path switching device based on a multi-cloud network, characterized in that, The device includes: The prediction processing module is used to perform trend prediction processing based on the pre-acquired network performance data set of each network path to obtain the trend data. The adjustment processing module is used to sequentially perform feature extraction and adjustment processing on the network performance data set and the change trend data to obtain network path switching instructions; The path switching module is used to execute a network path switching action based on the network path switching instruction, so that the transmission path of service data is switched from the current network path to the target network path. The network performance data set includes multiple sub-data sets arranged in chronological order. The prediction processing module is specifically used to perform iterative calculations on each of the sub-data sets. In the first iteration, the first sub-data set is input into the trend prediction model, and the trend prediction model is used to perform feature extraction processing on the first sub-data set to obtain the trend features of the first iteration. In subsequent iterations, the previous trend features obtained from the trend prediction model and the sub-data set of the current iteration are input into the trend prediction model, and the trend prediction model is used to perform feature extraction processing and feature fusion processing on the sub-data set of the current iteration and the previous trend features in sequence to obtain the trend features of the current iteration. The trend features obtained based on the last sub-data set are determined as the trend data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images