Method and system for locating and delimiting home broadband internet failure

Abstract

The application discloses a home broadband network fault positioning and delimiting method and system. The method comprises the following steps: obtaining a two-level connection duration matrix based on the PPPoE connection duration data of network elements; performing first-order difference operation on the two-level connection duration matrix to obtain a difference operation result based on home broadband service interruption statistical data; calculating the similarity of the curve waveform of the PPPoE connection duration difference operation through a weighted Pearson correlation coefficient; wherein the waveform curve is obtained based on the difference operation result; clustering the curve waveform based on the similarity of the curve waveform through a clustering algorithm, and positioning and delimiting the home broadband fault network element based on the home broadband service interruption statistical data according to the clustering result. The application can utilize an artificial intelligence algorithm, establish a network element unavailability group obstacle identification model based on the PPPoE connection duration field and other state indicators reported by a home gateway soft probe, fit the time difference waveform curve of different network element dimensions after identifying the user service interruption, and effectively delimit the unavailability network element of the home broadband network.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of communication technology, in particular to a home broadband network fault positioning and delimiting method and system. BACKGROUND

[0002] With the development of digital society, telecom operators have built gigabit home broadband (hereinafter referred to as home broadband) to lay the infrastructure of information highway, providing users with wired high-speed Internet environment for developing home gigabit network, Internet TV and other services. The networking structure and service process of home broadband network are as follows Figure 1 .

[0003] In the networking structure, home broadband is composed of four parts: home broadband terminal, home broadband network, content resources and service platform. Among them, the home broadband terminal includes Internet TV, user PC, mobile phone and the like, which accesses the home broadband network to provide users with rich and colorful Internet services; the home broadband network is composed of optical network unit (ONU), optical line terminal (OLT), broadband network gateway (BNG) and transmission network (CMNET), which realizes user authentication, data transmission channel and routing management, and provides high-speed Internet access channel for home broadband users; the content resources include television video content source and service provider, which provide actual access service content for users; the service platform is used for charging service, user account management and the like.

[0004] In the service process, home broadband network access is the process of home broadband users obtaining Internet content resources through wired network, and the specific process includes three steps in turn: authentication process, DNS process and access process. Step 1: authentication process. The user automatically dials through the home ONU, and sends user account, password and other information to the AAA authentication platform through OLT, BNG and CMNET for authentication. After authentication, the user is allocated IP address, DNS address and user access bandwidth, realizing network access. Step 2: DNS process. The user inputs domain name to access website, and sends the required website domain name to the operator DNS server through DNS protocol, and the operator DNS returns the IP address information of the user website; the user uses the IP address to access the server corresponding to the website. Step 3: access process. If the website server is in IDC, the content resources are returned to the user by IDC, which will pass through the DPI analysis system of IDC outlet. If the website server is in CDN distribution, the content resources are returned to the user by CDN. If the website server is in different network, the website content may be accessed through acceleration channel, telecom, unicom or international channel.

[0005] The unavailability of home broadband network elements will directly cause the interruption of batch user services, affecting the satisfaction of home broadband users. Home broadband services need to identify the quality problems and influence range of network performance, service quality, user experience and the like, and according to the identified network performance, service quality, user experience and the like, to carry out delimiting positioning and problem root cause diagnosis.

[0006] Currently, user-level problem localization and analysis methods mainly fall into two categories: those based on user signaling data tracing and those based on user indicator segmentation and horizontal comparison. Figure 2 and Figure 3 As shown, both methods rely on the experience of business experts to identify network anomalies. The difference lies in the approach: segmenting and comparing user metrics online brings this experience online, solidifying specific rules within the IT system. In contrast, tracing based on user signaling data relies on offline analysis by business experts, manually analyzing signaling content one by one within the timeframe of the user issue. Tracing based on user signaling data is primarily used to backtrack signaling data streams for single-user, single-service behavior after a user complaint. It extracts error codes and other information from the signaling data, starting with basic information and analyzing multiple layers including the control plane, business plane, domain name layer, and protocol handshake layer to analyze the complaint. Basic information analysis includes determining if the complainant is infected with a virus, accessed an illegal address, generated detailed bills, used an abnormal terminal, or used high-traffic services to identify the cause of the complaint. Further analysis is then conducted from multiple levels based on signaling error code information. The segmentation and horizontal comparison based on user metrics mainly start with user complaints. For different problem types, on the one hand, business experts sort out the set of metrics that are strongly related to the problem and confirm the network nodes where the metrics converge. Based on experience, they sort out the horizontal and vertical comparison rules of the metrics and output the problem location conclusions on the user side, network side, and business side based on the rules. On the other hand, they analyze the metrics on the user side, network side, and business side based on experience and directly output the conclusion that there is a problem with the node based on the degradation of the metrics.

[0007] In terms of boundary determination and localization capabilities, existing methods mainly rely on the accumulation of expert experience, and rules need to be summarized and supplemented manually, lacking standardization, automatic learning, and rapid iteration capabilities. Methods based on signaling backtracking and horizontal comparison of indicators heavily depend on the judgment of business experts, resulting in a lack of standardization and objectivity in existing user problem analysis techniques. Furthermore, since unavailable network elements are only one factor affecting user complaints, expert experience can only pinpoint some network factors, and there are also issues with inaccurate localization. For example, user error code distribution can only roughly determine whether the cause is on the wireless side or the core side. Existing boundary determination and localization technologies are powerless against factors outside the scope of business experts' expertise, and they also cannot quickly obtain data for new service complaints for automatic updates and rapid escalation. Summary of the Invention

[0008] The present invention aims to at least partially solve one of the technical problems in the related art.

[0009] To address this, the present invention proposes a method for locating and delimiting home broadband internet faults. Utilizing artificial intelligence algorithms, based on status indicators such as the PPPoE connection duration field reported by the home gateway's soft probe, a network element unavailability group fault identification model is established. After the user's service is interrupted, the time difference waveform curves of different network element dimensions are fitted to effectively delimit the unavailable network elements for home broadband internet access, which has strong practical value.

[0010] Another objective of this invention is to provide a home broadband internet fault location and demarcation system.

[0011] The third objective of this invention is to provide a computer device.

[0012] A fourth objective of this invention is to provide a non-transitory computer-readable storage medium.

[0013] To achieve the above objectives, this invention provides a method for locating and delimiting home broadband internet faults, comprising:

[0014] A two-level connection duration matrix is ​​obtained based on PPPoE connection duration data of network elements;

[0015] The first-order difference operation is performed on the two-level connection duration matrix to obtain the difference operation result based on the home broadband service interruption statistics.

[0016] The similarity of the waveform curves from the PPPoE connection duration difference operation is calculated using a weighted Pearson correlation coefficient; the waveform curves are obtained based on the difference operation results.

[0017] Clustering algorithms are used to cluster curve waveforms based on their similarity, and the faulty network elements of home broadband based on home broadband service interruption statistics are located and delimited according to the clustering results.

[0018] The home broadband internet fault location and delimitation method of this invention may also have the following additional technical features:

[0019] In one embodiment of the present invention, a two-level connection duration matrix is ​​obtained based on the PPPoE connection duration data of network elements, including:

[0020] Collect PPPoE connection duration data for all network elements;

[0021] The connection duration data is sorted by reportTime to obtain the duration sorting results;

[0022] The two-level connection duration matrix is ​​obtained based on the duration arrangement results; the two-level connection duration matrix includes the duration of the first-level connection and the duration of the second-level connection.

[0023] In one embodiment of the present invention, the first-level connection duration and the second-level connection duration are differentially calculated according to reportTime, so as to obtain the curve waveform of the multi-dimensional vector based on the differential calculation results of the PPPoE connection duration of all periodic granularities.

[0024] In one embodiment of the invention, prior to the weighted Pearson correlation coefficient, the method further includes:

[0025] Obtain the input data used to calculate the Pearson correlation coefficient; wherein, the input data includes the PPPoE time difference vector of different objects under the same dimension;

[0026] The Pearson correlation coefficient is calculated by multiplying the PPPoE time difference vector by a weighting factor.

[0027] In one embodiment of the present invention, the similarity of the curve waveforms is calculated using a weighted Pearson correlation coefficient, including:

[0028] Calculate the correlation coefficient between the PPPoE time difference vectors;

[0029] Set a threshold for the correlation coefficient;

[0030] The similarity of the curve waveforms represented by the PPPoE time difference vector is calculated based on the comparison results of the correlation coefficient and the correlation coefficient threshold.

[0031] In one embodiment of the present invention, clustering curve waveforms based on their similarity using a clustering algorithm includes:

[0032] Clustering algorithms are used to cluster the curve waveforms corresponding to the difference operation results of the first-level connection duration and the second-level connection duration, respectively. The clustering algorithm divides the curve waveform dataset into different classes or clusters according to the similarity calculation method of weighted Pearson correlation coefficient.

[0033] In one embodiment of the present invention, a clustering algorithm is used to cluster the curve waveforms corresponding to the difference operation results of the first-level connection duration and the difference operation results of the second-level connection duration, including:

[0034] A curve waveform dataset is constructed based on the curve waveforms corresponding to the difference operation results of the first-level connection duration and the second-level connection duration.

[0035] Traverse the curve waveform dataset and compare the waveform similarity between a single curve waveform data and the curve waveform data within a cluster. If the waveform similarity is greater than a first preset threshold, then the single curve waveform data is placed within a cluster; otherwise, the single curve waveform data is placed into a new cluster to obtain multiple clusters.

[0036] The system iterates through multiple clusters and obtains the number of data within each cluster. Based on the comparison between the number of data within each cluster and a second preset threshold, it determines whether there are similar curve waveforms among the nodes, thereby obtaining the waveform curve clustering results.

[0037] In one embodiment of the present invention, locating and delimiting faulty network elements of broadband services based on broadband service interruption statistics according to clustering results includes:

[0038] To obtain the cluster count result, query the number of clusters in the waveform curve clustering results.

[0039] The similarity of the waveform curves of network elements is determined based on the cluster count query results, and the faulty network elements of home broadband based on the home broadband service interruption statistics are located and delineated according to the similarity judgment results.

[0040] To achieve the above objectives, another aspect of the present invention proposes a home broadband internet fault location and demarcation system, comprising:

[0041] The connection duration acquisition module is used to obtain a two-level connection duration matrix based on the PPPoE connection duration data of network elements;

[0042] The duration difference operation module is used to perform first-order difference operation on the two-level connection duration matrix to obtain the difference operation result based on the home broadband service interruption statistics.

[0043] The weighted waveform judgment module is used to calculate the similarity of the waveform curves obtained from the PPPoE connection duration difference operation by using a weighted Pearson correlation coefficient; wherein, the waveform curve is obtained based on the difference operation result;

[0044] The clustering network element location module is used to cluster curve waveforms based on the similarity of the curve waveforms using a clustering algorithm, and to locate and delimit faulty network elements of home broadband based on the statistics of home broadband service interruption according to the clustering results.

[0045] The home broadband internet fault location and delimitation method and system of this invention establishes a home broadband network connection duration model based on the PPPoE connection duration field status index reported by the optical network unit (ONU), obtains connection duration waveform curves in different geographical dimensions, and then performs waveform similarity clustering through Pearson distance. Combined with the home broadband network resource attributes, it can effectively and automatically delimit and locate the faulty network element of the home broadband, and has strong application value.

[0046] To achieve the above objectives, a third aspect of this application provides a computer device, including a processor and a memory; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, for implementing the home broadband internet fault location and demarcation method as described in the first aspect embodiment.

[0047] To achieve the above objectives, a fourth aspect of this application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the home broadband internet fault location and demarcation method as described in the first aspect embodiment.

[0048] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0049] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0050] Figure 1 This is a diagram of the existing home broadband network architecture and service flow.

[0051] Figure 2 This is a schematic diagram of existing location analysis based on signaling data backtracking;

[0052] Figure 3 This is an existing diagram illustrating segmentation and horizontal comparison based on user metrics;

[0053] Figure 4 This is a flowchart of a method for locating and delimiting broadband internet faults according to an embodiment of the present invention;

[0054] Figure 5 This is a flowchart of the correlation coefficient calculation according to an embodiment of the present invention;

[0055] Figure 6 This is a schematic diagram of waveform similarity in various cities according to an embodiment of the present invention;

[0056] Figure 7 This is a logic diagram for locating and delimiting faulty network elements in a home broadband network according to an embodiment of the present invention;

[0057] Figure 8 This is a schematic diagram illustrating the location of faulty network elements in a home broadband system according to an embodiment of the present invention.

[0058] Figure 9 This is a schematic diagram of the structure of a home broadband internet fault location and demarcation system according to an embodiment of the present invention;

[0059] Figure 10 It is a computer device according to an embodiment of the present invention. Detailed Implementation

[0060] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0061] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0062] The following description, with reference to the accompanying drawings, outlines a method, system, device, and storage medium for locating and delimiting broadband internet faults according to embodiments of the present invention.

[0063] First, the terms that may be used in this invention are explained, as shown in Table 1:

[0064] Table 1

[0065]

[0066]

[0067] Figure 4 This is a flowchart of a home broadband internet fault location and delimitation method according to an embodiment of the present invention, such as... Figure 4 As shown, the method includes, but is not limited to, the following steps:

[0068] S1. A two-level connection duration matrix is ​​obtained based on the PPPoE connection duration data of network elements.

[0069] It is understood that the ONU in this embodiment of the invention has a built-in soft probe that reports the PPPoE protocol connection duration upTime of the ONU at the reportTime time period t, which represents the actual running time of the ONU since restarting. If the ONU restarts, upTime will be counted again from zero.

[0070] In one embodiment of the present invention, PPPoE connection duration data of all network elements are collected; the connection duration data are arranged according to reportTime to obtain a duration arrangement result; a two-level connection duration matrix is ​​obtained based on the duration arrangement result; wherein, the two-level connection duration matrix includes a first-level connection duration and a second-level connection duration.

[0071] Specifically, PPPoE connection duration data for all ONUs were collected and sorted by reportTime. The results of this time-based sorting were used to construct a two-level connection duration matrix, as shown in Table 2. The first-level connection duration represents the connection duration of each ONU. In the first-level ONU connection duration matrix, each cell represents the PPPoE connection duration reported by a specific ONU at the corresponding reportTime. The second-level connection duration represents the ONU connection duration data for each reportTime, aggregated according to geographical location, yielding the connection durations of all ONUs under OLT, BNG, city, region, and province.

[0072] Table 2

[0073]

[0074] Understandably, theoretically, without considering reporting time errors and ONU clock calibration issues, if no network failure occurs, the absolute value of the difference between the upTime reported by each ONU within two adjacent reportTimes is t; if a network failure occurs, causing the ONU connection to be interrupted, the absolute value of the difference between the upTime reported by the ONU within two adjacent reportTimes is less than t; if the difference in upTimes is equal to zero, it means that the ONU connection is completely interrupted within the reportTime.

[0075] S2, perform a first-order difference operation on the two-level connection duration matrix to obtain the difference operation result based on the home broadband service interruption statistics.

[0076] It is understood that, in this embodiment of the invention, a first-order difference operation is performed on the second-level connection duration matrix to obtain the home broadband service interruption statistics for ONU, OLT, BNG, city, and region within each reportTime, as shown in Table 3.

[0077] Table 3

[0078]

[0079] Specifically, in the secondary connection duration matrix, the primary connection duration is differentially calculated based on `reportTime`. This involves taking the PPPoE connection duration reported by the ONU soft probe, subtracting the previous period's duration from each period to obtain the `UpTime` time difference, dividing the result by the `reportTime` period `t`, and subtracting this value from 1. If the result is 0, it indicates the ONU is functioning normally within that period; if the result is greater than 0 and less than 1, it indicates an ONU interruption within that period; if the result is 1, it indicates a complete ONU interruption within that period; if the result is greater than 1, it indicates the ONU soft probe did not report any data, and the value needs to be set to 1 to account for missing data. The specific calculation method is as follows: `upTime`t It is the connection duration reported by reportTime, upTime t-1 It is the connection duration reported at the previous reportTime, and t is the reportTime period.

[0080] ΔupTime=min[1,(upTime t -upTime t-1 ) / t]

[0081] Specifically, in the second-level connection duration matrix, the second-level connection duration is differentially calculated based on reportTime. This involves subtracting the previous period's upTime from the connection duration data for each period. The result is then divided by the reportTime period t, and finally by the number of ONUs. If the result is 1, it indicates that the ONU status is normal within that period; if the result is greater than 0 and less than 1, it indicates that the ONU has been interrupted within that period; if the result is 0, it indicates that the ONU has been completely interrupted within that period; if the result is negative, it indicates that the ONU soft probe has not reported any data, and the negative value needs to be set to 0 to account for missing data. The specific calculation method is as follows: upTime t It is the connection duration reported by reportTime, upTime t-1 It is the connection duration reported at the previous reportTime, t is the reportTime period, and N is the number of downstream ONUs.

[0082] ΔupTime=min[1,(∑upTime t -∑upTime t-1 ) / (N*t)]

[0083] S3 calculates the similarity of the waveform curves from the PPPoE connection duration difference operation by using a weighted Pearson correlation coefficient; where the waveform curves are obtained based on the difference operation results.

[0084] Understandably, data extraction is performed to obtain the similarity curves of each waveform. Differential operations are then performed on the first-level and second-level connection durations according to reportTime, resulting in a multi-dimensional vector waveform based on the differential operation results of the PPPoE connection durations at all periodic granularities.

[0085] Specifically, starting from the province, the differential calculation results of the PPPoE connection duration at all t-cycle granularities for different regions, cities, BNGs, OLTs, and ONUs over 24 hours are extracted to form a waveform of a multi-dimensional vector, corresponding to the columns in Table 3, including ONU first-level duration differential, OLT second-level duration differential, city second-level duration differential, etc.

[0086] In one embodiment of the present invention, the waveform similarity of the degradation degree of different regions, cities, BNGs, OLTs, and ONUs is calculated based on the time difference curve waveform.

[0087] It is understandable that the weighted Pearson correlation coefficient distance of this invention adds a weight factor to each input vector based on the Pearson correlation coefficient, highlighting the influence of the nearest element on the similarity discrimination effect and downplaying the influence of elements that are far away from the current time.

[0088] In one embodiment of the present invention, input data for calculating the Pearson correlation coefficient is obtained; wherein, the input data includes PPPoE time difference vectors of different objects under the same dimension; the PPPoE time difference vectors are multiplied by a weighting factor to calculate the Pearson correlation coefficient.

[0089] Specifically, for inputs X and Y, they are multiplied by weighting factors, and then the Pearson correlation coefficient is calculated:

[0090]

[0091] In one embodiment of the present invention, it is recommended that the weighting factor of the weighted Pearson correlation coefficient distance be set according to the dimension number. For example, if the inputs X and Y are 5-dimensional vectors, then K is (1,2,3,4,5). If X and Y are 3-dimensional vectors, then K is (1,2,3).

[0092] In one embodiment of the present invention, the similarity between two waveforms is determined using a weighted Pearson correlation coefficient. The output range is -1 to +1, where 0 represents no correlation, negative values ​​represent negative correlation, and positive values ​​represent positive correlation. This includes:

[0093] 1) When the correlation coefficient is 0, the two vectors X and Y are uncorrelated;

[0094] 2) When the value of X increases (decreases), the value of Y decreases (increases), the two vectors X and Y are negatively correlated, and the correlation coefficient is between -1.0 and 0.0;

[0095] 3) When the value of X increases (decreases), the value of Y increases (decreases), the two vectors X and Y are positively correlated, and the correlation coefficient is between 0.0 and +1.0.

[0096] In one embodiment of the present invention, the correlation coefficient between PPPoE time difference vectors is calculated; a correlation coefficient threshold is set; and the similarity of the curve waveforms represented by the PPPoE time difference vectors is calculated based on the comparison result between the correlation coefficient and the correlation coefficient threshold.

[0097] Specifically, the inputs X and Y are the PPPoE time differences of different objects in the same dimension. The correlation coefficient ρ between X and Y is calculated, and a threshold β is set for this correlation coefficient. The comparison between the correlation coefficient and the threshold is used to calculate the similarity of the waveforms represented by the PPPoE time difference vectors. If ρ > β, it indicates that the waveforms represented by the two vectors X and Y are similar; otherwise, the waveforms represented by the two vectors X and Y are not similar. The calculation logic is as follows: Figure 5 As shown.

[0098] It is understandable that the Pearson coefficient, ranging from -1 to 1, is typically used to determine the correlation between variables and thus characterize the similarity between two variables. This invention only considers positively correlated similarity, as shown in Table 4:

[0099] Table 4

[0100] Correlation coefficient absolute value Similarity (0.8.1.0] Very high (0.6-0.8] High (0.4-0.6] General (0.2-0.4] Low [0.-0.2] Very low

[0101] Based on the provided historical similar waveform data, a similarity threshold is set by calculating the similarity of historical similar waveforms. Through comparative analysis, one embodiment of the present invention uses 0.85 as the correlation threshold; if the correlation coefficient corr>=0.85, the waveforms are similar; otherwise, the waveforms are not similar. Figure 6 For waveform similarity in different cities, such as Figure 6 As shown.

[0102] S4 uses a clustering algorithm to cluster the curve waveforms based on their similarity, and locates and delimits the faulty network elements of home broadband based on the statistics of home broadband service interruption according to the clustering results.

[0103] It is understood that the embodiments of the present invention use clustering algorithms to cluster the waveform curves of the difference results of the first-level connection duration and the difference results of the second-level connection duration, respectively.

[0104] As can be understood, clustering algorithms divide a dataset into different classes or clusters based on a weighted Pearson distance similarity algorithm, maximizing the similarity of data objects within the same cluster while maximizing the dissimilarity of data objects in different clusters. Clustering algorithms include:

[0105] 1. Construct a curve waveform dataset based on the curve waveforms corresponding to the difference operation results of the first-level connection duration and the second-level connection duration;

[0106] 2. Traverse the dataset, comparing the similarity of each individual data point with the data within its cluster. If the similarity is greater than a threshold, place the individual data point into the same cluster; otherwise, place it into a new cluster. The final result is as follows: Figure 6 The multiple clusters shown;

[0107] 3. After clustering, different clusters are obtained. The clusters are traversed and the number of data in the cluster is obtained. If there is a cluster with more than 1 data, the nodes have similar waveforms. Otherwise, there are no similar curves.

[0108] In one embodiment of the present invention, the specific clustering operation steps are as follows: Calculate the similarity of PPPoE time difference waveforms in different dimensions. Set a similarity threshold β. Assume there are 5 network elements A, B, C, D, and E.

[0109] 1. Set network element A as cluster C1, where C1 = [A] and the cluster center is A.

[0110] 2. Calculate the waveform similarity between network elements A and B. If the waveforms are similar, add B to cluster C1, where C1 = [A, B]. If the waveforms are dissimilar, make B a new cluster C2, i.e., C2 = [B], with B as the cluster center.

[0111] 3. Calculate the similarity between the C network element and the cluster centers of C1 and C2. If they are similar, they are assigned to either C1 or C2. Otherwise, a new cluster C3 is formed, i.e., C3 = [C].

[0112] 4. Calculate the cluster similarity between the D network elements and the existing cluster centers, and divide them according to the rules in step 2.

[0113] 5. Repeat step 4 to calculate all cities in the loop.

[0114] Preferably, the clustering pseudocode is implemented as follows:

[0115]

[0116]

[0117] Furthermore, in this embodiment of the invention, following the steps described above, the waveform curves of the difference results of the first-level connection duration and the difference results of the second-level connection duration are clustered, and the number of clusters is queried.

[0118] This invention embodiment queries the number of clusters in the waveform curve clustering results to obtain the cluster count query result; based on the cluster count query result, it determines the similarity of the network element waveform curves, and uses the similarity judgment result to locate and delimit the faulty network elements of home broadband based on the home broadband service interruption statistics. The steps are as follows:

[0119] 6.1 If the number of clusters in the differential results of the second-level connection duration at the city level is 1, it means that the connection duration waveforms of all ONUs are exactly the same, and there is no faulty network element in the home broadband network.

[0120] 6.2 If the number of clusters in the difference results of the secondary connection duration at the city level is greater than 1, it indicates that the waveforms of these ONU network elements are not completely similar, and further investigation is required by following the steps below.

[0121] 6.3 If the number of network elements in a cluster is equal to 1, it indicates that the network element is degraded and needs to be addressed according to... Figure 7 According to the flowchart, cluster the primary connection duration to identify the truly degraded ONUs.

[0122] 6.4 If the number of network elements in a cluster is greater than 1, then it needs to be done according to... Figure 7 The overall process continues to investigate the causes by clustering the duration of secondary connections at a higher level.

[0123] 6.5 Check the number of clusters in step 6.1. If the number of clusters is equal to 1, it means that all network element waveforms are completely similar, and the algorithm ends.

[0124] According to the embodiments of the present invention Figure 7 The overall flowchart shows that each node performs clustering and similarity calculations according to the second step of the process until the algorithm ends, which can identify the causes of degradation in each layer. Figure 8 This is a schematic diagram illustrating the location of faulty network elements in the home broadband system according to the present invention. The final output reasons are:

[0125] Anomalies in multiple cities: Waveform clusters are similar across multiple cities;

[0126] Anomalies in multiple districts and counties within a single city: multiple districts and counties within a single city exhibit similar clustering patterns;

[0127] Multiple BNG anomalies in a single county: The waveforms of multiple BNGs within a single county are similar;

[0128] Single BRAS with multiple OLT anomalies: Multiple OLT waveforms under a single BRAS exhibit similar clustering.

[0129] Single OLT with multiple ONU anomaly: Multiple ONUs with highly similar waveforms cluster similarly at the same OLT node;

[0130] Single PON port (first-level splitting segment) anomaly: waveforms are highly similar, and second-level splitting clusters are similar at the same node;

[0131] Secondary beam splitting anomaly: Users with highly similar waveforms cluster similarly at the secondary beam splitting nodes;

[0132] Users with short-term online / offline cycles: Users whose secondary spectrometers fail to fit but experience frequent interruptions.

[0133] Therefore, this invention uses a clustering algorithm to cluster curve waveforms based on the similarity of the curve waveforms, and locates and delimits faulty network elements of home broadband based on the home broadband service interruption statistics according to the clustering results.

[0134] In summary, this invention employs a waveform fitting similarity algorithm and Pearson correlation coefficient to intelligently discover anomalies in network elements, determining whether a network element is abnormally unavailable based on these anomalies. This invention does not rely on manual judgment or traditional rule-based judgment through indicator aggregation. It eliminates the need for manually maintaining a set of indicators strongly correlated with the unavailability of certain types of network elements, significantly reducing the workload of manual sorting and analysis. It avoids potential problems such as misconfiguration, omissions, and location errors that may occur with manual sorting and analysis, achieving a much higher accuracy rate than user-rule-based methods. It can more accurately delineate unavailable network elements for home broadband access, applicable to handling home broadband faults. This provides intelligent analysis tools for actual broadband users and can also be used by back-end network processing personnel, significantly improving the efficiency of complaint handling and problem resolution, preventing mass service interruptions caused by unavailable home broadband network elements, and effectively improving the satisfaction of a large number of users. This solution has a very broad market prospect and high commercial value.

[0135] The home broadband internet fault location and delimitation method according to embodiments of the present invention changes the original collusion method of roughly determining the problematic network element by aggregating the number of interrupted users based on user affiliation. Instead, it proposes a method based on PPPoE time difference waveform fitting to establish a network element outage group fault identification model. This method can more accurately delimit unavailable network elements for home broadband internet access, take targeted solutions, avoid batch user service interruptions caused by unavailable home broadband internet access network elements, and effectively improve the satisfaction of batch users.

[0136] To achieve the above embodiments, such as Figure 9 As shown, this embodiment also provides a home broadband internet fault location and delimitation system 10, which includes a connection duration acquisition module 100, a duration difference calculation module 200, a weighted waveform judgment module 300, and a clustering network element location module 400.

[0137] The connection duration acquisition module 100 is used to obtain a two-level connection duration matrix based on the PPPoE connection duration data of network elements;

[0138] The duration difference operation module 200 is used to perform first-order difference operation on the two-level connection duration matrix to obtain the difference operation result based on the home broadband service interruption statistics.

[0139] The weighted waveform judgment module 300 is used to calculate the similarity of the waveform curves of the PPPoE connection duration difference operation by using a weighted Pearson correlation coefficient; wherein, the waveform curve is obtained based on the difference operation result;

[0140] The clustering network element location module 400 is used to cluster curve waveforms based on the similarity of the curve waveforms using a clustering algorithm, and to locate and delimit the faulty network elements of home broadband based on the statistics of home broadband service interruption based on the clustering results.

[0141] Furthermore, the connection duration acquisition module 100 described above is also used for:

[0142] Collect PPPoE connection duration data for all network elements;

[0143] The connection duration data is sorted by reportTime to obtain the duration sorting results;

[0144] The two-level connection duration matrix is ​​obtained based on the duration arrangement results; the two-level connection duration matrix includes the duration of the first-level connection and the duration of the second-level connection.

[0145] Furthermore, the aforementioned duration difference calculation module 200 is also used to perform difference calculations on the first-level connection duration and the second-level connection duration according to reportTime, so as to obtain the curve waveform of the multi-dimensional vector based on the difference calculation results of the PPPoE connection duration at all periodic granularities.

[0146] Furthermore, prior to the aforementioned weighted waveform judgment module 300, a correlation coefficient calculation module is also included, used for:

[0147] Obtain the input data used to calculate the Pearson correlation coefficient; wherein, the input data includes the PPPoE time difference vector of different objects under the same dimension;

[0148] The Pearson correlation coefficient is calculated by multiplying the PPPoE time difference vector by a weighting factor.

[0149] Furthermore, the aforementioned weighted waveform judgment module 300 is also used for:

[0150] Calculate the correlation coefficient between the PPPoE time difference vectors;

[0151] Set a threshold for the correlation coefficient;

[0152] The similarity of the curve waveforms represented by the PPPoE time difference vector is calculated based on the comparison results of the correlation coefficient and the correlation coefficient threshold.

[0153] Furthermore, the aforementioned clustering element localization module 400 is also used for:

[0154] Clustering algorithms are used to cluster the curve waveforms corresponding to the difference operation results of the first-level connection duration and the second-level connection duration, respectively. The clustering algorithm divides the curve waveform dataset into different classes or clusters according to the similarity calculation method of weighted Pearson correlation coefficient.

[0155] Furthermore, the aforementioned clustering element localization module 400 is also used for:

[0156] A curve waveform dataset is constructed based on the curve waveforms corresponding to the difference operation results of the first-level connection duration and the second-level connection duration.

[0157] Traverse the curve waveform dataset and compare the waveform similarity between a single curve waveform data and the curve waveform data within a cluster. If the waveform similarity is greater than a first preset threshold, then the single curve waveform data is placed within a cluster; otherwise, the single curve waveform data is placed into a new cluster to obtain multiple clusters.

[0158] The system iterates through multiple clusters and obtains the number of data within each cluster. Based on the comparison between the number of data within each cluster and a second preset threshold, it determines whether there are similar curve waveforms among the nodes, thereby obtaining the waveform curve clustering results.

[0159] Furthermore, the aforementioned clustering element localization module is also used for:

[0160] To obtain the cluster count result, query the number of clusters in the waveform curve clustering results.

[0161] The similarity of the waveform curves of network elements is determined based on the cluster count query results, and the faulty network elements of home broadband based on the home broadband service interruption statistics are located and delineated according to the similarity judgment results.

[0162] According to the home broadband internet fault location and delimitation system of the present invention, a method based on PPPoE time difference waveform fitting to establish a network element outage group fault identification model is proposed. This method can more accurately delimit unavailable network elements for home broadband internet access, take targeted solutions, avoid batch user service interruptions caused by unavailable home broadband internet access network elements, and effectively improve the satisfaction of batch users.

[0163] To implement the methods of the above embodiments, the present invention also provides a computer device, such as... Figure 10 As shown, the computer device 600 includes a memory 601 and a processor 602; wherein, the processor 602 reads the executable program code stored in the memory 601 to run a program corresponding to the executable program code, so as to implement the various steps of the above method.

[0164] To implement the above embodiments, this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method as described in the foregoing embodiments.

[0165] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0166] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for locating and delimiting faults in home broadband internet access, characterized in that, The method includes the following steps: A two-level connection duration matrix is ​​obtained based on PPPoE connection duration data of network elements; Perform a first-order difference operation on the two-level connection duration matrix to obtain the difference operation result based on the home broadband service interruption statistics; The similarity of the waveform curves obtained from the PPPoE connection duration difference operation is calculated using a weighted Pearson correlation coefficient; wherein the waveform curves are obtained based on the difference operation results. Clustering algorithms are used to cluster the curve waveforms based on their similarity, and the faulty network elements of the home broadband service interruption statistics are located and delimited based on the clustering results.

2. The method according to claim 1, characterized in that, A two-level connection duration matrix is ​​obtained based on PPPoE connection duration data from network elements, including: Collect PPPoE connection duration data for all network elements; The connection duration data is sorted according to reportTime to obtain the duration sorting result; The two-level connection duration matrix is ​​obtained based on the duration arrangement result; wherein, the two-level connection duration matrix includes the first-level connection duration and the second-level connection duration.

3. The method according to claim 2, characterized in that, The first-level connection duration and the second-level connection duration are differentially calculated according to reportTime, so as to obtain the curve waveform of the multidimensional vector based on the differential calculation results of the PPPoE connection duration of all periodic granularities.

4. The method according to claim 3, characterized in that, Prior to the weighted Pearson correlation coefficient, the method further includes: Obtain the input data used to calculate the Pearson correlation coefficient; wherein, the input data includes the PPPoE time difference vector of different objects under the same dimension; The Pearson correlation coefficient is calculated by multiplying the PPPoE time difference vector by a weighting factor.

5. The method according to claim 4, characterized in that, The similarity of the curve waveforms is calculated by weighting the Pearson correlation coefficient, including: Calculate the correlation coefficient between the PPPoE time difference vectors; Set the correlation coefficient threshold for the aforementioned correlation coefficient; The similarity of the curve waveform represented by the PPPoE time difference vector is calculated based on the comparison result between the correlation coefficient and the correlation coefficient threshold.

6. The method according to claim 5, characterized in that, Clustering the curve waveforms using a clustering algorithm based on their similarity includes: Clustering algorithms are used to cluster the curve waveforms corresponding to the difference operation results of the first-level connection duration and the second-level connection duration, respectively; wherein, the clustering algorithm divides the curve waveform dataset into different classes or clusters according to the similarity calculation method of weighted Pearson correlation coefficient.

7. The method according to claim 6, characterized in that, Clustering algorithms are used to cluster the curve waveforms corresponding to the difference operation results of the first-level connection duration and the second-level connection duration, respectively, including: A curve waveform dataset is constructed based on the curve waveforms corresponding to the differential operation results of the first-level connection duration and the second-level connection duration. The curve waveform dataset is traversed, and the waveform similarity between a single curve waveform data and the curve waveform data within a cluster is compared. If the waveform similarity is greater than a first preset threshold, the single curve waveform data is placed within a cluster; otherwise, the single curve waveform data is placed into a new cluster, so as to obtain multiple clusters. The system iterates through the multiple clusters and obtains the number of data within each cluster. Based on the comparison between the number of data within each cluster and a second preset threshold, it determines whether there are similar curve waveforms among the nodes, thereby obtaining the waveform curve clustering result.

8. The method according to claim 7, characterized in that, Based on the clustering results, faulty network elements of the home broadband service interruption statistics are located and delimited, including: Query the number of clusters in the waveform curve clustering results to obtain the cluster count query result; The similarity of the network element waveform curves is determined based on the cluster count query results, and the faulty network elements of the home broadband service interruption statistics are located and delineated based on the similarity judgment results.

9. A home broadband internet fault location and demarcation system, characterized in that, include: The connection duration acquisition module is used to obtain a two-level connection duration matrix based on the PPPoE connection duration data of network elements; The duration difference operation module is used to perform first-order difference operation on the two-level connection duration matrix to obtain the difference operation result based on the home broadband service interruption statistics. The weighted waveform judgment module is used to calculate the similarity of the waveform curves obtained from the PPPoE connection duration difference operation by using a weighted Pearson correlation coefficient; wherein the waveform curves are obtained based on the difference operation results. The clustering network element location module is used to cluster the curve waveform based on the similarity of the curve waveform using a clustering algorithm, and to locate and delimit the faulty network elements of the home broadband service interruption statistics based on the clustering results.

10. The system according to claim 9, characterized in that, The connection duration acquisition module is also used for: Collect PPPoE connection duration data for all network elements; The connection duration data is sorted according to reportTime to obtain the duration sorting result; The two-level connection duration matrix is ​​obtained based on the duration arrangement result; wherein, the two-level connection duration matrix includes the first-level connection duration and the second-level connection duration.

Images