Network failure prediction method, electronic device, storage medium and computer program product

Abstract

The application discloses a network fault prediction method, an electronic device, a storage medium and a computer program product, and relates to the technical field of network fault prediction. The method comprises the following steps: acquiring first information of a network to be monitored, wherein the first information comprises index data of a first link in a historical time period adjacent to a first moment; determining a first prediction interval of each first link based on the first information; determining a network fault probability of each first link in the corresponding first prediction interval based on the index data of each first link; and the first prediction interval takes the first moment as a starting moment.

Description

Technical Field

[0001] This application relates to the field of big data technology, and in particular to a network fault prediction method, electronic device, storage medium and computer program product. Background Technology

[0002] Network technology has been applied to production activities in various industries. Once a network failure occurs, it will affect business operations. Most existing network failure prediction methods are based on real-time data obtained at the current moment to predict whether a network failure will occur. However, considering factors such as computation time and network latency, by the time the prediction result is obtained, the network failure has already occurred, resulting in poor predictive effectiveness. Summary of the Invention

[0003] To address the related technical issues, embodiments of this application provide a network fault prediction method, an electronic device, a storage medium, and a computer program product.

[0004] The technical solution of this application embodiment is implemented as follows:

[0005] This application provides a network fault prediction method, the method comprising:

[0006] Obtain first information about the network to be monitored, including indicator data of the first link in a historical time period adjacent to the first moment;

[0007] Based on the first information, a first prediction interval is determined for each first link, with the first prediction interval starting from a first time point;

[0008] Based on the metric data of each first link, the network failure probability of each first link within the corresponding first prediction interval is determined.

[0009] In the above scheme, determining the first prediction interval for each first link based on the first information includes:

[0010] Based on one or more indicator data of the second link, one or more second prediction intervals of the second link are determined, and one second prediction interval corresponds to one type of indicator data; the second link represents any link in the first link.

[0011] The first prediction interval of the second link is determined based on one or more second prediction intervals of the second link.

[0012] In the above scheme, determining one or more second prediction intervals for the second link based on one or more indicator data of the second link includes:

[0013] Determine the standard deviation and data growth rate of the first indicator data of the second link; wherein, the first indicator data represents any one of one or more indicator data of the second link, and the data growth rate is determined based on the first indicator data at a first time and the first indicator data at a second time, and the second time represents the nearest time before the first time.

[0014] The second prediction interval for the second link is determined based on the standard deviation and data growth rate of the first indicator data of the second link.

[0015] The method in the above scheme further includes:

[0016] The time period between the first moment and the third moment is defined as the historical time period of the third link, and the third moment is updated to the first moment of the third link; the third link represents the link in the first link where the indicator data changes abruptly, and the third moment represents the moment when the indicator data of the third link changes abruptly; or...

[0017] If the index data of the second link does not change abruptly within the first prediction interval, the first prediction interval of the second link is determined as the historical time period of the second link, and the end time of the first prediction interval of the second link is determined as the first time of the second link; the second link represents any link in the first link.

[0018] The method in the above scheme further includes:

[0019] If the indicator data of the second link is outside the set range within the first prediction interval, it is determined that the indicator data of the second link has undergone a sudden change; wherein,

[0020] The second link represents any link in the first link, and the set range is determined based on the standard deviation of the first indicator data corresponding to the second link at the first time and the first indicator data of the second link; the first indicator data represents any one of one or more indicator data of the second link.

[0021] The method in the above scheme further includes:

[0022] If the network failure probability of any first link is greater than or equal to a set threshold, second information is obtained. The second information includes the links in the network where the indicator data has changed abruptly and the corresponding number of data changes.

[0023] The third information is determined based on the second information, and the third information represents information related to network faults.

[0024] In the above scheme, determining the third information based on the second information includes:

[0025] Based on the second information and the first network topology of the network, a second network topology is determined. The second network topology includes faulty nodes and link-related nodes whose indicator data has changed abruptly. The faulty nodes represent the first link-related nodes.

[0026] The third information is determined based on the shape or characteristics of the second network topology.

[0027] In the above scheme, after obtaining the second information, the method further includes:

[0028] The number of data mutations corresponding to the links in the network where the indicator data has mutated is reset to zero.

[0029] In the above scheme, determining the network failure probability of each first link within the corresponding first prediction interval based on the indicator data of each first link includes:

[0030] The indicator data of the second link is input into the set model to obtain the probability of each indicator data of the second link causing a network failure; the second link represents any link in the first link, and the set model is used to output the probability of the indicator data causing a network failure based on the indicator data;

[0031] Based on the first indicator data of the second link and the probability of the first indicator data causing network failure, the weight coefficient of the first indicator data is determined; the first indicator data represents any one of one or more indicator data of the second link.

[0032] Based on all weight coefficients of the second link and the probability of network failure caused by all index data of the second link, the probability of network failure of the second link in the corresponding first prediction interval is determined.

[0033] This application embodiment also provides a network fault prediction device, the device comprising:

[0034] The first acquisition module is used to acquire first information about the network to be monitored, the first information including indicator data of the first link in a historical time period adjacent to the first moment;

[0035] The first determining module is used to determine the first prediction interval for each first link based on the first information, wherein the first prediction interval starts at the first time point;

[0036] The second determining module is used to determine the network failure probability of each first link within the corresponding first prediction interval based on the indicator data of each first link.

[0037] This application also provides an electronic device, including: a processor and a communication interface; wherein,

[0038] The processor is configured to acquire first information about the network to be monitored, the first information including indicator data of the first link in a historical time period adjacent to the first moment; determine a first prediction interval for each first link based on the first information; determine the network failure probability of each first link within the corresponding first prediction interval based on the indicator data of each first link; the first prediction interval is based on the first moment as the starting moment.

[0039] This application also provides an electronic device, including a processor and a memory for storing a computer program capable of running on the processor.

[0040] When the processor runs the computer program, it executes the steps of any of the above methods.

[0041] This application also provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of any of the above methods.

[0042] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above methods.

[0043] The network fault prediction method, electronic device, storage medium, and computer program product provided in this application's embodiments determine a first prediction interval for each first link based on first information of the network to be monitored. The first information includes indicator data of the first link in a historical time period adjacent to the first moment, and the first prediction interval starts from the first moment. Based on the indicator data of each first link, the network fault probability of each first link within the corresponding first prediction interval is determined. This scheme determines the network fault probability of the first prediction interval based on indicator data of a historical time period, i.e., it determines the network fault probability of a future time period based on historical data. This allows for faster acquisition of the network fault probability of a future time period, enabling the determination of the network fault probability before a network fault occurs and improving the effectiveness of network fault prediction. Attached Figure Description

[0044] Figure 1 This is a schematic flowchart of a network fault prediction method according to an embodiment of this application;

[0045] Figure 2 This is a schematic diagram illustrating the application process of the network fault prediction method according to an embodiment of this application;

[0046] Figure 3 This is a schematic diagram of the network fault prediction device according to an embodiment of this application;

[0047] Figure 4 This is a schematic diagram of the electronic device structure according to an embodiment of this application. Detailed Implementation

[0048] Network technology has been applied to production activities in various industries. Once a network failure occurs, it will affect business operations, especially in scenarios such as carrier networks, the power industry, and the financial industry. The losses caused by network failures are usually calculated in seconds, so preventing network failures is particularly important.

[0049] Existing technical solutions primarily involve analyzing equipment operating patterns and fault characteristics, identifying influencing factors, and using combinatorial theory and information fusion techniques to train neural network prediction models for network fault prediction. These solutions typically deploy tools at the network egress point, collect network operational data, train the model using machine learning algorithms, and use real-time data as input to determine whether a fault has occurred based on the output.

[0050] The disadvantages of existing technical solutions are as follows:

[0051] The network operation data that has occurred is collected and input into the network fault prediction model to determine whether a network fault has occurred. However, considering factors such as computation time and network latency, in a rapidly changing network environment, by the time the prediction result is obtained, the network fault has already occurred.

[0052] In a stable operating environment, the real-time acquisition, processing and input of data into the model calculation results in significant resource overhead.

[0053] Setting a fixed time interval to periodically acquire data may result in network failures during the sampling interval.

[0054] For scenarios like weather forecasting, equipment malfunction prediction, and battery life prediction, the likelihood of an event occurring can be determined by considering its influencing factors. This approach aligns well with objective laws and is less affected by external factors. For example, weather forecasting can use cloud formations, atmospheric conditions, ocean currents, temperature, and humidity to predict rain, which is consistent with natural science and has minimal impact from external factors like artificial rainmaking. However, the network environment is not a single device and cannot be judged solely based on its operational status. Furthermore, networks are significantly affected by external factors, such as system crashes or service interruptions caused by hacker attacks overloading the target system, network loops caused by unplugging and plugging in network cables, and network flooding caused by misconfigurations. These situations are sudden, change rapidly, and require timely responses. Existing technical solutions are not designed for rapid responses to such emergencies.

[0055] Based on this, in various embodiments of this application, a first prediction interval for each first link is determined based on first information of the network to be monitored. The first information includes indicator data of the first link in a historical time period adjacent to the first moment, and the first prediction interval starts from the first moment. Based on the indicator data of each first link, the network failure probability of each first link within the corresponding first prediction interval is determined. The above scheme determines the network failure probability of the first prediction interval based on indicator data of a historical time period, that is, it determines the network failure probability of a future time period based on historical data. This allows for faster acquisition of the network failure probability of a future time period, enabling the determination of the network failure probability before a network failure occurs and improving the effectiveness of network failure prediction.

[0056] The present application will now be described in further detail with reference to the accompanying drawings and embodiments.

[0057] This application provides a network fault prediction method applied to electronic devices, which may have a display screen or display function, or may be connected to a display screen. For example... Figure 1 As shown, it includes:

[0058] Step 101: Obtain the first information of the network to be monitored, the first information including the indicator data of the first link in the historical time period adjacent to the first moment.

[0059] Here, the first link can include all links of the network to be monitored; it can also include only a portion of the links of the network to be monitored, the specific portion of the links to be monitored being determined based on actual needs. In practical applications, the first link can include all links of the network to be monitored.

[0060] The metrics data for the first link represent the network operation metrics data of the first link in the network to be monitored. The metrics data for each first link may include one or more of the following. For example, the metrics data for each first link may include one or more of the following: traffic, latency, jitter, and packet loss rate.

[0061] The first moment can be understood as the current moment. It should be understood that the first information includes the indicator data of the first link at the first moment.

[0062] The historical time period can be set based on actual needs. It is worth noting that the historical time period is not fixed and can be adjusted.

[0063] Step 102: Determine the first prediction interval for each first link based on the first information.

[0064] The first prediction interval starts at the first time point. The first prediction interval can be understood as a future time period.

[0065] It is worth noting that the indicator data of the first link in the first information can be distinguished by the link identifier (id, identifier) ​​and / or other link identifiers.

[0066] Here, determining the first prediction interval for each first link based on the first information may include:

[0067] Determine the standard deviation of each indicator data in the second link, determine the data growth rate of the indicator data with the smallest standard deviation, and determine the first prediction interval of the second link based on the smallest standard deviation and the data growth rate; or,

[0068] Determine the standard deviation and data growth rate of any indicator data in the second link, and determine the first prediction interval of the second link based on the standard deviation and data growth rate; or,

[0069] Determine the prediction interval corresponding to each indicator data of the second link, and determine the first prediction interval of the second link from all determined prediction intervals; wherein,

[0070] The second link represents any link in the first link.

[0071] In practical applications, the method of determining the first prediction interval for each first link based on the first information can be to determine the prediction interval corresponding to each indicator data of the second link, and then determine the first prediction interval of the second link from all the determined prediction intervals.

[0072] In one embodiment, determining the first prediction interval for each first link based on the first information includes:

[0073] Based on one or more indicator data of the second link, one or more second prediction intervals of the second link are determined, and one second prediction interval corresponds to one type of indicator data; the second link represents any link in the first link.

[0074] The first prediction interval of the second link is determined based on one or more second prediction intervals of the second link.

[0075] Here, based on one or more indicator data of the second link, one or more second prediction intervals of the second link are determined. This can be to determine the first parameter and data growth rate of the first indicator data of the second link, and to determine the second prediction interval of the second link based on the first parameter and data growth rate of the first indicator data of the second link. The first indicator data represents any one of the one or more indicator data of the second link. The first parameter can be the standard deviation or the variance, and there is no specific limitation.

[0076] The first prediction interval of the second link is determined based on one or more second prediction intervals of the second link. This can be done by determining the minimum prediction interval among one or more second prediction intervals of the second link as the first prediction interval of the second link, or by determining the average value of all second prediction intervals of the second link as the first prediction interval of the second link.

[0077] In practical applications, determining the first prediction interval of a second link based on one or more second prediction intervals of the second link can be achieved by determining the smallest prediction interval among one or more second prediction intervals of the second link as the first prediction interval of the second link.

[0078] In this embodiment, a second prediction interval corresponds to a type of indicator data. Each type of indicator data in the second link has a corresponding second prediction interval, which is a comprehensive consideration. Based on one or more second prediction intervals of the second link, the first prediction interval of the second link is determined, which can improve the accuracy of the first prediction interval of the second link.

[0079] In one embodiment, determining one or more second prediction intervals for the second link based on one or more indicator data of the second link includes:

[0080] Determine the standard deviation and data growth rate of the first indicator data of the second link; wherein, the first indicator data represents any one of one or more indicator data of the second link, and the data growth rate is determined based on the first indicator data at a first time and the first indicator data at a second time, and the second time represents the nearest time before the first time.

[0081] The second prediction interval for the second link is determined based on the standard deviation and data growth rate of the first indicator data of the second link.

[0082] Here, the data growth rate is determined based on the first indicator data at the first time point and the first indicator data at the second time point. It can be determined by dividing the first value by the second value and taking the absolute value of the quotient. Alternatively, it can be determined by multiplying the absolute value of the quotient of the first value by the second value and the first deviation coefficient. The first deviation coefficient can be set based on actual needs. Here, the first value is the difference between the first indicator data at the first time point and the first indicator data at the second time point, and the second value is the difference between the first time point and the second time point.

[0083] In practical applications, the standard deviation of the first indicator data in the second link can be determined using the following formula:

[0084]

[0085] Where, σ xN represents the standard deviation of the first indicator data in the second link. x M represents the number of first indicator data points in the second link. x x is the mean of the first indicator data of the second link. i This is the i-th first indicator data of the second link.

[0086] The data growth rate of the first indicator data in the second link can be determined using the following formula:

[0087]

[0088] Where, k x x represents the data growth rate of the first indicator data in the second link. t1 The first indicator data for the first moment of the second link, x t2 The first indicator data is the data at the second moment of the second link; T is the difference between the first moment and the second moment, which can be understood as the data collection interval.

[0089] The second prediction interval of the second link is determined based on the standard deviation and data growth rate of the first indicator data of the second link. This can be achieved by dividing the standard deviation of the first indicator data of the second link by the data growth rate of the first indicator data of the second link; or by multiplying the quotient obtained by dividing the standard deviation of the first indicator data of the second link by the data growth rate of the first indicator data of the second link by the second deviation coefficient. The second deviation coefficient can be set based on actual needs.

[0090] In practical applications, the second prediction interval of the second link can be determined using the following formula:

[0091]

[0092] Where, Δt x σ represents the second prediction interval corresponding to the first indicator data of the second link. x k is the standard deviation of the first indicator data in the second link. x The data growth rate of the first indicator data for the second link.

[0093] In this embodiment, the second prediction interval of the second link is determined based on the standard deviation and data growth rate of the first indicator data of the second link, which improves the efficiency of determining the second prediction interval of the second link. The standard deviation can better reflect the data characteristics of the first indicator data of the second link. By determining the second prediction interval of the second link through the standard deviation, the accuracy of the second prediction interval of the second link can be improved.

[0094] In one embodiment, the network fault prediction method further includes:

[0095] If the indicator data of the second link is outside the set range within the first prediction interval, it is determined that the indicator data of the second link has undergone a sudden change; wherein,

[0096] The second link represents any link in the first link, and the set range is determined based on the standard deviation of the first indicator data corresponding to the second link at the first time and the first indicator data of the second link; the first indicator data represents any one of one or more indicator data of the second link.

[0097] It should be understood that a certain indicator data of the second link corresponds to a set range; if any indicator data of the second link is outside the corresponding set range within the first prediction interval, it is determined that the corresponding indicator data of the second link has a sudden change.

[0098] Here, the range is determined based on the standard deviation of the first indicator data of the second link at the first time and the first indicator data of the second link. Alternatively, the range can be determined based on a third and fourth value, for example, a range of [x...]. t1 -σ x x t1 +σ x ], x t1 σ represents the first indicator data corresponding to the second link at the first moment. x The standard deviation of the first indicator data of the second link can be used; alternatively, the set range can be determined by the product of the third value and the third deviation coefficient, and the product of the fourth value and the fourth deviation coefficient. The third and fourth deviation coefficients can be set based on actual needs. The third and fourth deviation coefficients can be the same or different. The third value is the difference between the standard deviations of the first indicator data of the second link at the first time and the first indicator data of the second link, and the fourth value is the sum of the standard deviations of the first indicator data of the second link at the first time and the first indicator data of the second link.

[0099] In this embodiment, the conditions for a sudden change in the monitoring index data are clearly defined. The sudden change in the monitoring index data is used to monitor rapid changes in the network under monitoring, while slow or normal changes in the network under monitoring are ignored, which helps to improve the accuracy of subsequent determination of network failure probability.

[0100] In one embodiment, the network fault prediction method further includes:

[0101] The time period between the first moment and the third moment is defined as the historical time period of the third link, and the third moment is updated to the first moment of the third link; the third link represents the link in the first link where the indicator data changes abruptly, and the third moment represents the moment when the indicator data of the third link changes abruptly; or...

[0102] If the index data of the second link does not change abruptly within the first prediction interval, the first prediction interval of the second link is determined as the historical time period of the second link, and the end time of the first prediction interval of the second link is determined as the first time of the second link; the second link represents any link in the first link.

[0103] It should be understood that the third time is within the first prediction interval; in the case of multiple third links, the time period between the first time of the fourth link and the third time of the fourth link is determined as the historical time period of the fourth link, and the third time of the fourth link is updated to the first time of the fourth link; the fourth link represents any link in the third link.

[0104] In this embodiment, the update method of the historical time period and the first moment is clearly defined in the case of sudden changes in indicator data and no sudden changes in indicator data. It can dynamically adjust the historical time period according to the indicator data of the network to be monitored, thereby changing the first prediction interval and improving the accuracy and practicality of network fault prediction.

[0105] Step 103: Based on the indicator data of each first link, determine the network failure probability of each first link within the corresponding first prediction interval.

[0106] Here, based on the indicator data of each first link, the network failure probability of each first link within the corresponding first prediction interval is determined. This can be achieved by determining the probability of each indicator data of the second link causing a network failure and the proportion of each indicator data of the second link, and then determining the network failure probability of the second link within the corresponding first prediction interval based on the probability of each indicator data of the second link causing a network failure and the proportion of each indicator data of the second link. Alternatively, it can be achieved by determining the probability of each indicator data of the second link causing a network failure, assigning weights of 0.4, 0.3, 0.2, and 0.1 according to the sum of the probabilities of each indicator data of the second link causing a network failure from high to low, multiplying the sum of the probabilities of each indicator data of the second link causing a network failure by the corresponding weights, and then summing the results to obtain the network failure probability of the second link within the corresponding first prediction interval. Here, the second link represents any link in the first link.

[0107] In this embodiment, based on the index data of each first link, the network failure probability of each first link within the corresponding first prediction interval is determined. The first prediction intervals corresponding to different first links may be different, which can improve the accuracy of determining the network failure probability.

[0108] In one embodiment, determining the network failure probability of each first link within the corresponding first prediction interval based on the indicator data of each first link includes:

[0109] The indicator data of the second link is input into the set model to obtain the probability of each indicator data of the second link causing a network failure; the second link represents any link in the first link, and the set model is used to output the probability of the indicator data causing a network failure based on the indicator data;

[0110] Based on the first indicator data of the second link and the probability of the first indicator data causing network failure, the weight coefficient of the first indicator data is determined; the first indicator data represents any one of one or more indicator data of the second link.

[0111] Based on all weight coefficients of the second link and the probability of network failure caused by all index data of the second link, the probability of network failure of the second link in the corresponding first prediction interval is determined.

[0112] Here, the indicator data of the second link is input into the setting model. This can be done by inputting the indicator data of the second link separately according to different categories, or by inputting all the indicator data of the second link into the setting model. It is worth noting that inputting the indicator data of the second link and its corresponding setting range into the setting model helps to improve the accuracy of the probability of each indicator data of the second link causing network failure. The setting model can be understood as a network failure prediction model. It should be understood that the setting model can be trained in advance and can be continuously trained and updated.

[0113] Based on the first indicator data of the second link and the probability of the first indicator data causing network failure, the weight coefficient of the first indicator data is determined. It can be a second parameter that determines the first indicator data of the second link and the probability of the first indicator data causing network failure. The weight coefficient of the first indicator data is determined based on the second parameter of the first indicator data of the second link and all the second parameters of the second link. The second parameter can be covariance or correlation coefficient, and there is no specific limitation.

[0114] Based on the probability of network failure caused by all weight coefficients and all indicator data of the second link, the network failure probability of the second link within the corresponding first prediction interval is determined. This can be achieved by multiplying the sum of the probabilities of network failure caused by the first indicator data of the second link by the weight coefficients of the first indicator data to obtain the network failure probability of the first indicator data, and then summing the network failure probabilities of all the first indicator data of the second link to obtain the network failure probability of the second link within the corresponding first prediction interval. Alternatively, it can be achieved by multiplying the sum of the probabilities of network failure caused by the first indicator data of the second link by the weight coefficients of the first indicator data to obtain the network failure probability of the first indicator data, and then adding or subtracting the sum of the network failure probabilities of all the first indicator data of the second link from a set deviation value to obtain the network failure probability of the second link within the corresponding first prediction interval. The set deviation value can be set based on actual needs.

[0115] Taking the second link, which includes four types of indicator data, as an example, the network failure probability of the second link within the corresponding first prediction interval can be determined by the following formula:

[0116] P = w a P a +w b P b +w c P c +w d P d ,in,

[0117]

[0118] μ a =cov(A, P) a (A)),

[0119] μ b =cov(B, P) b (B)),

[0120] μ c =cov(C,P) c (C)),

[0121] μ d =cov(D,P) d (D)),

[0122] P is the network failure probability of the second link within the corresponding first prediction interval, w a P represents the weighting coefficient of the first type of indicator data in the second link. a w is the sum of the probabilities of the first type of metric data in the second link causing network failure. b P represents the weighting coefficient for the second type of indicator data in the second link. bw is the sum of the probabilities that the second type of metric data in the second link causes network failure. c P represents the weighting coefficient of the third indicator data in the second link. c w is the sum of the probabilities that the third indicator data of the second link causes network failure. d P represents the weighting coefficient for the fourth indicator data in the second link. d μ is the sum of the probabilities that the fourth indicator data of the second link causes network failure. a The probability P that the first type of indicator data A and the first type of indicator data cause network failure in the second link. a The covariance of (A), μ b The probability P that the second type of indicator data B and the second type of indicator data cause network failure in the second link. b The covariance of (B), μ c The probability P that the third and third indicator data of the second link cause network failure. c The covariance of (C), μ d The probability P that the fourth indicator data D and the fourth indicator data of the second link cause network failure. d (D) covariance.

[0123] In this embodiment, the method of determining the network failure probability of each first link within the corresponding first prediction interval based on the indicator data of each first link can improve the efficiency of determining the network failure probability.

[0124] In one embodiment, the network fault prediction method further includes:

[0125] If the network failure probability of any first link is greater than or equal to a set threshold, second information is obtained. The second information includes the links in the network where the indicator data has changed abruptly and the corresponding number of data changes.

[0126] The third information is determined based on the second information, and the third information represents information related to network faults.

[0127] Here, the threshold can be set based on actual needs. For example, the threshold can be set to 70%.

[0128] In practical applications, the number of data mutations can be understood as the mutation value.

[0129] For example, a mutation value is set for each first link in the network to be monitored, with an initial value of 0. When a mutation is detected in the indicator data, the mutation value of the link in which the indicator data has changed is increased by 1, as shown in Table 1. Table 1 is only an example and does not constitute a limitation.

[0130] Table 1:

[0131] First link ID 1 2 3 4 5 6 H0 time 0 0 0 0 0 0 H1 time 1 0 0 0 0 0 H2 time 2 1 0 0 0 0 H3 time 3 2 1 0 0 0 H4 time 4 3 2 0 0 1 H5 Moment 5 4 3 0 0 1

[0132] For example, if the probability of a network failure in the first link is greater than or equal to a set threshold at time H5, then the second information includes the number of data mutations in the first link 1 being 5, the number of data mutations in the first link 2 being 4, the number of data mutations in the first link 3 being 3, and the number of data mutations in the first link 6 being 1.

[0133] Here, the third piece of information may include one or more of the following: probability of network failure, cause of network failure, and scope of impact of network failure.

[0134] Determining the third information based on the second information can be achieved by: determining the number of links in the network under monitoring where indicator data has mutated based on the second information; and determining the cause of the network failure based on the number of links where indicator data has mutated. For example, if the number of links where indicator data has mutated is greater than a set number, then the cause of the network failure is determined to be a broadcast storm or a virus attack. Alternatively, the topology between the links in the network under monitoring where indicator data has mutated can be determined based on the second information; and the cause of the network failure and / or the scope of its impact can be determined based on the topology between the links where indicator data has mutated.

[0135] In this embodiment, when the network failure probability of any first link is greater than or equal to a set threshold, the relevant information of network failure can be determined quickly based on the links in the network to be monitored where the indicator data changes abruptly and the corresponding number of data changes, thus helping to improve network maintenance efficiency.

[0136] In one embodiment, determining the third information based on the second information includes:

[0137] Based on the second information and the first network topology of the network, a second network topology is determined. The second network topology includes faulty nodes and link-related nodes whose indicator data has changed abruptly. The faulty nodes represent the first link-related nodes.

[0138] The third information is determined based on the shape or characteristics of the second network topology.

[0139] Here, based on the second information and the first network topology of the network to be monitored, the second network topology is determined. This can be achieved by combining the first network topology of the network to be monitored and marking the links in the network to be monitored according to the number of data mutations corresponding to the links where the indicator data changes abruptly, from largest to smallest, to obtain the second network topology. Alternatively, the second network topology can be obtained by taking the nodes of the links with a network failure probability greater than or equal to a set threshold as the center, and marking the links according to the first network topology of the network to be monitored and the number of data mutations corresponding to the links where the indicator data changes abruptly, from largest to smallest, ignoring discrete nodes. Discrete nodes are nodes in the network to be monitored that are not connected to either the faulty node or the link where the indicator data changes abruptly.

[0140] The shape or characteristics of the second network topology may include one or more of the following:

[0141] Star-shaped;

[0142] Ring-shaped;

[0143] Dispersion in the specified direction;

[0144] It spreads to each node.

[0145] In practical applications, the third information is determined based on the shape or characteristics of the second network topology, as shown in Table 2.

[0146] Table 2:

[0147]

[0148] It should be understood that Table 2 is merely an example and does not constitute a limitation on the shape or characteristics of the second network topology, nor does it constitute a limitation on the causes of network faults and the scope of impact of network faults corresponding to the shape or characteristics of the second network topology.

[0149] In this embodiment, the second network topology is determined based on the changing trend of the second information, namely the indicator data or the changing sequence of the links. Based on the shape or characteristics of the second network topology, the network fault propagation trend can be obtained, thereby improving the accuracy of the third information and the efficiency of network fault prediction.

[0150] In one embodiment, after obtaining the second information, the network fault prediction method further includes:

[0151] The number of data mutations corresponding to the links in the network where the indicator data has mutated is reset to zero.

[0152] It should be understood that the number of data mutations corresponding to all links in the network under monitoring that experience sudden changes in all indicator data is reset to zero.

[0153] In this embodiment, each time a network fault is predicted to occur, the corresponding number of data mutations is reset to zero after obtaining the second information. This can avoid the data before and after the network fault prediction from affecting each other and improve the accuracy of network fault prediction.

[0154] The following section provides a more detailed description of this application with reference to application examples.

[0155] A schematic diagram illustrating the application process of the network fault prediction method provided in this application embodiment is shown below. Figure 2 As shown, it includes:

[0156] Step 201: Obtain the first information about the network to be monitored.

[0157] The first information includes the historical time period (t1-Δt) adjacent to the first time t1. ls The first link's metric data, t1], includes all links of the network to be monitored.

[0158] The indicator data can be collected through deployed network monitoring tools, with a collection frequency of f and a collection interval of T = 1 / f.

[0159] Step 202: Process the first information to obtain the processed first information; the processed first information includes the data of each indicator for each first link.

[0160] The indicator data is cleaned, and the cleaned indicator data is divided into multiple arrays according to the link ID and the type of indicator data. For example, array A = [a1, a2, a3, ..., an] represents the second link at (t1-Δt) ls The traffic metric data for the second link is [t1, t1]. Array B = [b1, b2, b3, ..., bn] represents the traffic metric data for the second link at (t1-Δt). ls The latency index data is t1], and the array C = [c1, c2, c3, ..., cn] represents the latency index data of the second link at (t1-Δt). ls The jitter index data of the second link at (t1-Δt) is given by array D = [d1, d2, d3, ..., dn]. ls The packet loss rate data of t1] represents any link in the first link.

[0161] Step 203: Based on the processed first information, determine the first prediction interval for each first link.

[0162] For example, based on A, B, C, and D, determine the standard deviation and data growth rate of the first indicator data for the second link; based on the standard deviation and data growth rate of the first indicator data for the second link, determine the second prediction interval for the second link; all second prediction intervals for the second link include Δt. a , △tb , △t c and △t d , where △t a The second prediction interval corresponding to the traffic metric data of the second link, for example, Δt. a For 1 minute, △t b The second prediction interval corresponding to the latency index data of the second link, for example, Δt. b For 3 minutes, △t c The second prediction interval, Δt, corresponds to the jitter index data of the second link. d The second prediction interval corresponds to the packet loss rate index data of the second link; the first prediction interval Δt of the second link is determined based on all the second prediction intervals of the second link. yc The first prediction interval of the second link is Δt yc =min{△t a , △t b , △t c , △t d}

[0163] Step 204: Monitor for sudden changes in indicator data within the first prediction interval and update the first time point and historical time period; set the second link in (t1-Δt) ls If the indicator data does not change drastically within the time period [t1], proceed to step 205; without setting a second link in (t1-Δt) ls If the indicator data changes abruptly within the time period t1, return to step 203 to update the first prediction interval and then execute step 205, and continue to execute step 204.

[0164] Taking traffic flow data as an example, in (t1, t1+Δt) yc Traffic metrics data A' for the second link within the time period ] t1+T a t1+2T a t1+3T , ..., a t1+△tyc All are within the set range [a] t1 -σ a a t1 +σ a Within the specified context, it was determined that the traffic metrics data for the second link did not undergo any sudden changes. t1 σ represents the traffic metrics data for the second link at the first moment. a For the second link at (t1-Δt) ls The standard deviation of traffic indicator data within the time period [t1]; assuming no sudden changes in any indicator data in the second link, the standard deviation of (t1, t1+Δt) is used. yc The time period is determined as the historical time period of the second link, and t1+Δt is used. ycThis is determined to be the first moment of the second link;

[0165] In (t1, t1+Δt) yc A sudden change in the traffic index data of the second link is detected at the third moment within the time period, namely t1+nT. The time period (t1, t1+nT) is determined as the historical time period of the second link, and the t1+nT moment is updated as the first moment of the second link.

[0166] Step 205: Based on the established model and the indicator data of each first link, determine the network failure probability of each first link within the corresponding first prediction interval.

[0167] For example, when setting the second link at (t1-Δt) ls If the indicator data does not change drastically within the time period [t1], the second link will be in (t1-Δt) ls Input the indicator data within the time period [t1] into the set model to obtain the probability of each indicator data of the second link causing network failure, for example, P. b (100) = 0.7 represents the probability that a network failure will occur within 100 milliseconds (ms) of latency. Calculate the probability of network failure caused by the second link in (t1-Δt) ls The covariance of each indicator data and the probability of each indicator data causing network failure within the time period [t1] is used to determine the weight coefficient of each indicator data based on the covariance of each indicator data in the second link; the weight coefficients are multiplied by the sum of the corresponding types of indicator data to obtain the network failure probability of each indicator data, and then summed to obtain the probability of network failure of the second link in Δt. yc The probability of network failure within the network;

[0168] Without setting a second link in (t1-Δt) ls Whether the indicator data undergoes a sudden change within the time period [t1, t1+Δt], and whether the second link is in the range (t1, t1+Δt) yc Assuming no sudden changes in the indicator data within the time period, based on the second link at (t1, t1+Δt) yc The indicator data within the time period determines the first prediction interval △t after the update. yc ';The second link will be in (t1, t1+Δt) yc Input the indicator data within the specified time period into the model to obtain the probability that each indicator data of the second link will cause a network failure; calculate the probability of the second link in (t1, t1+Δt) yc The covariance of each indicator data point and the probability of each indicator data point causing network failure within the time period is used to determine the weight coefficient of each indicator data point based on the covariance of each indicator data point in the second link. The weight coefficients are then multiplied by the sum of the corresponding types of indicator data to obtain the network failure probability of each indicator data point, and these are summed to obtain the network failure probability of the second link in (t1+Δt).yc ,t1+Δt yc +Δt yc The probability of network failure within a given time period;

[0169] Without setting a second link in (t1-Δt) ls If the indicator data changes abruptly within the time interval [t1], and the indicator data of the second link changes abruptly at time t1+nT, then the updated first prediction interval Δt is determined based on the indicator data of the second link within the time interval (t1, t1+nT]. yc "; Input the indicator data of the second link within the time period (t1, t1+nT] into the set model to obtain the probability of each indicator data of the second link causing network failure; calculate the covariance of each indicator data and the probability of each indicator data causing network failure within the time period (t1, t1+nT] of the second link, and determine the weight coefficient of each indicator data based on the covariance of each indicator data of the second link; multiply the weight coefficient by the sum of the corresponding types of indicator data to obtain the network failure probability of each indicator data, and then add them together to obtain the probability of the second link causing network failure within the time period (t1+nT, t1+nT+Δt] of the second link." yc The probability of network failure within a given time period.

[0170] It should be understood that step 205 can be executed in the configuration model, and the configuration model outputs the network failure probability of each first link.

[0171] Step 206: Based on the network failure probability of the first link of the network to be monitored, determine the third information and execute step 207; obtain the actual network failure of the network to be monitored, and return to step 205 to train and update the set model based on the actual network failure of the network to be monitored.

[0172] If the probability of network failure on any first link is greater than or equal to a set threshold, second information is obtained. Based on the second information and the first network topology of the network to be monitored, the second network topology is determined. Based on the shape or characteristics of the second network topology, third information is determined. The third information includes the probability of network failure, the cause of network failure, and the scope of impact of network failure.

[0173] Step 207: Network Fault Warning.

[0174] Push network fault warnings and third-party information; and reset the number of data mutations for all links in the network under monitoring that have experienced sudden changes in indicator data to zero.

[0175] The network fault prediction method provided in this application focuses on rapidly changing indicator data. Based on the trend of indicator data changes, it predicts whether a network fault will occur in the first prediction interval according to this trend. The first prediction interval is dynamically adjusted based on the indicator data change trend, ensuring timely response while effectively saving resource consumption. The prediction interval is determined by the standard deviation and data growth rate of indicator data over a historical period, enabling prediction of the probability of network faults in the first prediction interval, avoiding resource waste caused by real-time processing. Sudden change detection is performed on the indicator data within the prediction interval, enabling timely perception of network changes during sudden network mutations, improving the timeliness and accuracy of network fault prediction. Furthermore, by combining second information with the network topology, the network fault propagation trend is derived, ensuring... This application can accurately predict the cause and scope of network failures, enabling faster fault location and improved network maintenance efficiency. It predicts the probability of network failures at the first prediction interval based on historical time-period indicator data, providing forward-looking insights and valuable time for fault handling. Updating the first prediction interval reduces the likelihood of network failures occurring within a fixed sampling period in a rapidly changing network environment, making it suitable for unexpected situations such as hacker attacks. For fault monitoring of each link in the network under monitoring, it can identify small-scale network failures, such as equipment malfunctions and network loops. This application can accurately predict network failures and determine their causes and scope of impact, while also possessing the ability to adaptively adjust and quickly respond to sudden changes in indicator data, enhancing network stability and security.

[0176] To implement the network fault prediction method provided in this application embodiment, this application embodiment also provides a network fault prediction device, such as... Figure 3 As shown, the device includes:

[0177] The first acquisition module 301 is used to acquire first information of the network to be monitored, the first information including indicator data of the first link in a historical time period adjacent to the first moment;

[0178] The first determining module 302 is used to determine a first prediction interval for each first link based on the first information, wherein the first prediction interval starts at a first time.

[0179] The second determining module 303 is used to determine the network failure probability of each first link within the corresponding first prediction interval based on the indicator data of each first link.

[0180] In one embodiment, the first determining module 302 is specifically used to determine one or more second prediction intervals for the second link based on one or more indicator data of the second link, wherein one second prediction interval corresponds to one type of indicator data; the second link represents any link in the first link.

[0181] The first prediction interval of the second link is determined based on one or more second prediction intervals of the second link.

[0182] In one embodiment, the first determining module 302 is specifically used to determine the standard deviation and data growth rate of the first indicator data of the second link; wherein, the first indicator data represents any one of one or more indicator data of the second link, and the data growth rate is determined based on the first indicator data at a first time and the first indicator data at a second time, and the second time represents the nearest time before the first time.

[0183] The second prediction interval for the second link is determined based on the standard deviation and data growth rate of the first indicator data of the second link.

[0184] In one embodiment, the device further includes:

[0185] The third determining module is used to determine the time period between the first time point and the third time point as the historical time period of the third link, and update the third time point to the first time point of the third link; the third link represents the link in the first link where the indicator data changes abruptly, and the third time point represents the moment when the indicator data of the third link changes abruptly; or...

[0186] If the index data of the second link does not change abruptly within the first prediction interval, the first prediction interval of the second link is determined as the historical time period of the second link, and the end time of the first prediction interval of the second link is determined as the first time of the second link; the second link represents any link in the first link.

[0187] In one embodiment, the device further includes:

[0188] The fourth determining module is used to determine if a sudden change has occurred in the indicator data of the second link when the indicator data of the second link is outside a set range within the first prediction interval; wherein,

[0189] The second link represents any link in the first link, and the set range is determined based on the standard deviation of the first indicator data corresponding to the second link at the first time and the first indicator data of the second link; the first indicator data represents any one of one or more indicator data of the second link.

[0190] In one embodiment, the device further includes:

[0191] The second acquisition module is used to acquire second information when the network failure probability of any first link is greater than or equal to a set threshold. The second information includes the links in the network where the indicator data has changed abruptly and the corresponding number of data changes.

[0192] The fifth determining module is used to determine the third information based on the second information, wherein the third information represents information related to network faults.

[0193] In one embodiment, the fifth determining module is specifically used to determine a second network topology based on the second information and the first network topology of the network. The second network topology includes faulty nodes and link-related nodes whose indicator data has changed abruptly. The faulty nodes represent the first link-related nodes.

[0194] The third information is determined based on the shape or characteristics of the second network topology.

[0195] In one embodiment, after acquiring the second information, the device further includes:

[0196] The zeroing module is used to reset the number of data mutations corresponding to links in the network where indicator data has mutated.

[0197] In one embodiment, the second determining module 303 is specifically used to input the indicator data of the second link into a setting model to obtain the probability of each indicator data of the second link causing a network failure; the second link represents any link in the first link, and the setting model is used to output the probability of the indicator data causing a network failure based on the indicator data;

[0198] Based on the first indicator data of the second link and the probability of the first indicator data causing network failure, the weight coefficient of the first indicator data is determined; the first indicator data represents any one of one or more indicator data of the second link.

[0199] Based on all weight coefficients of the second link and the probability of network failure caused by all index data of the second link, the probability of network failure of the second link in the corresponding first prediction interval is determined.

[0200] In practical applications, the first acquisition module 301, the first determination module 302, the second determination module 303, the third determination module, the fourth determination module, the second acquisition module, the fifth determination module, and the clearing module can be implemented by the processor in the network fault prediction device.

[0201] It should be noted that the network fault prediction device provided in the above embodiments is only illustrated by the division of the above-described program modules when performing network fault prediction. In practical applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the processing described above. In addition, the network fault prediction device and the network fault prediction method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0202] Based on the hardware implementation of the above program modules, and in order to implement the network fault prediction method provided in this application embodiment, this application embodiment also provides an electronic device, such as... Figure 4 As shown, the electronic device 400 includes:

[0203] Communication interface 401 enables information exchange with other network nodes;

[0204] The processor 402 is connected to the communication interface 401 to enable information interaction with other network nodes and to execute the methods provided by one or more of the above-mentioned technical solutions when running a computer program. The computer program is stored in the memory 403.

[0205] Specifically, the processor 402 is used to acquire first information of the network to be monitored, the first information including indicator data of the first link in a historical time period adjacent to the first time; determine a first prediction interval for each first link based on the first information; determine the network failure probability of each first link in the corresponding first prediction interval based on the indicator data of each first link; the first prediction interval starts from the first time.

[0206] In one embodiment, the processor 402 is specifically configured to determine one or more second prediction intervals for the second link based on one or more indicator data of the second link, wherein one second prediction interval corresponds to one type of indicator data; the second link represents any link in the first link.

[0207] The first prediction interval of the second link is determined based on one or more second prediction intervals of the second link.

[0208] In one embodiment, the processor 402 is specifically configured to determine the standard deviation and data growth rate of the first indicator data of the second link; wherein, the first indicator data represents any one of one or more indicator data of the second link, and the data growth rate is determined based on the first indicator data at a first time and the first indicator data at a second time, and the second time represents the nearest time before the first time.

[0209] The second prediction interval for the second link is determined based on the standard deviation and data growth rate of the first indicator data of the second link.

[0210] In one embodiment, the processor 402 is further configured to determine the time period between the first time point and the third time point as the historical time period of the third link, and update the third time point to the first time point of the third link; the third link represents the link in the first link where the indicator data changes abruptly, and the third time point represents the moment when the indicator data of the third link changes abruptly; or,

[0211] If the index data of the second link does not change abruptly within the first prediction interval, the first prediction interval of the second link is determined as the historical time period of the second link, and the end time of the first prediction interval of the second link is determined as the first time of the second link; the second link represents any link in the first link.

[0212] In one embodiment, the processor 402 is further configured to determine that a sudden change has occurred in the indicator data of the second link if the indicator data of the second link is outside a set range within a first prediction interval; wherein,

[0213] The second link represents any link in the first link, and the set range is determined based on the standard deviation of the first indicator data corresponding to the second link at the first time and the first indicator data of the second link; the first indicator data represents any one of one or more indicator data of the second link.

[0214] In one embodiment, the processor 402 is further configured to acquire second information when the network failure probability of any first link is greater than or equal to a set threshold, the second information including the links in the network where the indicator data has changed abruptly and the corresponding number of data changes;

[0215] The third information is determined based on the second information, and the third information represents information related to network faults.

[0216] In one embodiment, the processor 402 is specifically configured to determine a second network topology based on the second information and the first network topology of the network, the second network topology including faulty nodes and link-related nodes whose indicator data has changed abruptly, the faulty nodes representing the first link-related nodes;

[0217] The third information is determined based on the shape or characteristics of the second network topology.

[0218] In one embodiment, after acquiring the second information, the processor 402 is further configured to clear the number of data mutations corresponding to the links in the network where the indicator data has mutated.

[0219] In one embodiment, the processor 402 is specifically used to input the indicator data of the second link into a set model to obtain the probability of each indicator data of the second link causing a network failure; the second link represents any link in the first link, and the set model is used to output the probability of the indicator data causing a network failure based on the indicator data;

[0220] Based on the first indicator data of the second link and the probability of the first indicator data causing network failure, the weight coefficient of the first indicator data is determined; the first indicator data represents any one of one or more indicator data of the second link.

[0221] Based on all weight coefficients of the second link and the probability of network failure caused by all index data of the second link, the probability of network failure of the second link in the corresponding first prediction interval is determined.

[0222] It should be noted that the specific processing procedure of processor 402 can be understood by referring to the above method.

[0223] Of course, in practical applications, the various components in electronic device 400 are coupled together through bus system 404. It can be understood that bus system 404 is used to realize the connection and communication between these components. In addition to a data bus, bus system 404 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, in... Figure 4 The general designated all buses as Bus System 404.

[0224] The memory 403 in this embodiment is used to store various types of data to support the operation of the electronic device 400. Examples of such data include any computer program used to operate on the electronic device 400.

[0225] The methods disclosed in the embodiments of this application can be applied to the processor 402, or implemented by the processor 402. The processor 402 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 402 or by instructions in the form of software. The processor 402 may be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 402 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly reflected as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium, which is located in the memory 403. The processor 402 reads the information in the memory 403 and completes the steps of the aforementioned method in combination with its hardware.

[0226] In an exemplary embodiment, the electronic device 400 may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers (MCUs), microprocessors, or other electronic components to perform the aforementioned method.

[0227] It is understood that the memory 403 in this embodiment can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), ferromagnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); magnetic surface memory can be disk storage or magnetic tape storage. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM).The memory 403 described in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0228] In an exemplary embodiment, this application also provides an electronic device, including a processor and a memory for storing a computer program capable of running on the processor, wherein the processor, when running the computer program, performs the steps of any of the methods described above.

[0229] This application embodiment also provides a storage medium, namely a computer storage medium, specifically a computer-readable storage medium, such as a memory 403 that stores a computer program. The computer program can be executed by the processor 402 of the electronic device 400 to complete the steps described in the preceding method. The computer-readable storage medium can be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM.

[0230] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above methods.

[0231] It should be noted that terms such as "first" and "second" are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. The term "and / or" in this article merely describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, the term "one or more" in this article refers to any combination of at least two of any one or more items from a set of A, B, and C. For example, "one or more of A, B, and C" can represent any one or at least two or more elements selected from the set of A, B, and C.

[0232] Furthermore, the technical solutions described in the embodiments of this application can be combined arbitrarily without conflict.

[0233] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application.

Claims

1. A network fault prediction method, characterized in that, The method includes: Obtain first information about the network to be monitored, including indicator data of the first link in a historical time period adjacent to the first moment; Based on the first information, a first prediction interval is determined for each first link, with the first prediction interval starting from a first time point; Based on the metric data of each first link, the network failure probability of each first link within the corresponding first prediction interval is determined.

2. The method according to claim 1, characterized in that, The step of determining the first prediction interval for each first link based on the first information includes: Based on one or more indicator data of the second link, one or more second prediction intervals of the second link are determined, and one second prediction interval corresponds to one type of indicator data; the second link represents any link in the first link. The first prediction interval of the second link is determined based on one or more second prediction intervals of the second link.

3. The method according to claim 2, characterized in that, Determining one or more second prediction intervals for the second link based on one or more indicator data of the second link includes: Determine the standard deviation and data growth rate of the first indicator data of the second link; wherein, the first indicator data represents any one of one or more indicator data of the second link, and the data growth rate is determined based on the first indicator data at a first time and the first indicator data at a second time, and the second time represents the nearest time before the first time. The second prediction interval for the second link is determined based on the standard deviation and data growth rate of the first indicator data of the second link.

4. The method according to any one of claims 1 to 3, characterized in that, The method further includes: The time period between the first moment and the third moment is defined as the historical time period of the third link, and the third moment is updated to the first moment of the third link; the third link represents the link in the first link where the indicator data changes abruptly, and the third moment represents the moment when the indicator data of the third link changes abruptly; or... If the index data of the second link does not change abruptly within the first prediction interval, the first prediction interval of the second link is determined as the historical time period of the second link, and the end time of the first prediction interval of the second link is determined as the first time of the second link; the second link represents any link in the first link.

5. The method according to any one of claims 1 to 3, characterized in that, The method further includes: If the indicator data of the second link is outside the set range within the first prediction interval, it is determined that the indicator data of the second link has undergone a sudden change; wherein, The second link represents any link in the first link, and the set range is determined based on the standard deviation of the first indicator data corresponding to the second link at the first time and the first indicator data of the second link; the first indicator data represents any one of one or more indicator data of the second link.

6. The method according to any one of claims 1 to 3, characterized in that, The method further includes: If the network failure probability of any first link is greater than or equal to a set threshold, second information is obtained. The second information includes the links in the network where the indicator data has changed abruptly and the corresponding number of data changes. The third information is determined based on the second information, and the third information represents information related to network faults.

7. The method according to claim 6, characterized in that, The determination of the third information based on the second information includes: Based on the second information and the first network topology of the network, a second network topology is determined. The second network topology includes faulty nodes and link-related nodes whose indicator data has changed abruptly. The faulty nodes represent the first link-related nodes. The third information is determined based on the shape or characteristics of the second network topology.

8. The method according to claim 6, characterized in that, After obtaining the second information, the method further includes: The number of data mutations corresponding to the links in the network where the indicator data has mutated is reset to zero.

9. The method according to any one of claims 1 to 3, 7 to 8, characterized in that, The determination of the network failure probability of each first link within the corresponding first prediction interval based on the indicator data of each first link includes: The indicator data of the second link is input into the set model to obtain the probability of each indicator data of the second link causing a network failure; the second link represents any link in the first link, and the set model is used to output the probability of the indicator data causing a network failure based on the indicator data; Based on the first indicator data of the second link and the probability of the first indicator data causing network failure, the weight coefficient of the first indicator data is determined; the first indicator data represents any one of one or more indicator data of the second link. Based on all weight coefficients of the second link and the probability of network failure caused by all index data of the second link, the probability of network failure of the second link in the corresponding first prediction interval is determined.

10. An electronic device, characterized in that, This includes a processor and memory for storing computer programs that can run on the processor. When the processor is used to run the computer program, it performs the steps of the method according to any one of claims 1 to 9.

11. A 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 9.

12. A computer program product, comprising a computer program, 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 9.

Images