CN114244691A - Video service fault positioning method and device and electronic equipment - Google Patents

Abstract

The invention discloses a fault location method, device and electronic equipment for video services, belonging to the technical field of communication. The specific implementation scheme includes: acquiring the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network; based on the pre-trained service quality perception model and the data of the first index parameter, analyzing and obtaining the users of the video service Experience quality information; evaluate the service quality of the video service according to the user experience quality information of the video service; if the service quality of the video service does not meet the preset requirements, based on the pre-trained fault location model set and the second index parameter The fault location model set includes multiple fault location models, and each fault location model corresponds to a network domain. As a result, in the case of video services, end-to-end fault location for the global network can be automatically implemented, and the location efficiency can be improved.

Description

Fault locating method, device and electronic device for video service

technical field

The invention belongs to the technical field of communications, and in particular relates to a fault location method, device and electronic equipment for video services.

Background technique

Currently, during video playback, abnormal situations such as long initial delay, frequent freezes, and black and blurry screens often occur due to network failures. In response to this abnormal situation, the existing fault location method is that maintenance personnel detect and check the possible fault points of the network one by one through probe detection. It can be seen from this that the existing fault location method is inefficient.

SUMMARY OF THE INVENTION

The purpose of the embodiments of the present invention is to provide a fault location method, device and electronic device for video services, so as to solve the problem of low efficiency of existing fault location methods when facing video services.

In order to solve the above-mentioned technical problems, the present invention is achieved in this way:

In a first aspect, an embodiment of the present invention provides a fault location method for a video service, including:

acquiring data of the first index parameter and data of the second index parameter of the video service in the multi-domain network;

Based on the pre-trained service quality perception model and the data of the first index parameter, the user experience quality information of the video service is obtained by analysis; wherein the service quality perception model is used to characterize the data of the first index parameter and user experience The relationship between quality information;

Evaluate the service quality of the video service according to the user experience quality information of the video service;

Under the condition that the service quality of the video service does not meet the preset requirements, the fault location is performed based on the pre-trained fault location model set and the data of the second index parameter;

Wherein, the fault location model set includes multiple fault location models, and each fault location model corresponds to a network domain.

Optionally, the fault location is performed based on the pre-trained fault location model set and the data of the second index parameter, including:

The index parameter data of each network domain in the data of the second index parameter is respectively input into the fault location model corresponding to each network domain to perform fault inference.

Optionally, the first index parameters include: index parameters in the wireless network, index parameters in the core network, and index parameters in the bearer network.

Optionally, the second index parameters include: index parameters in the wireless network, index parameters in the transmission network, index parameters in the core network, index parameters in the bearer network, and index parameters in the user data center.

Optionally, before the acquisition of the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network, the method includes:

For each network domain, perform the following processes respectively to obtain the fault location model set:

obtaining a first sample data set of the network domain; wherein, the first sample data set includes index parameter data of the network domain;

Model training is performed by using the first sample data set to obtain a fault location model corresponding to the network domain.

Optionally, before the acquisition of the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network, the method includes:

Obtaining a second sample data set; wherein, the second sample data set includes the index parameter data of the video service in the multi-domain network and the user experience quality information;

Preprocessing the second sample data set to obtain a target data set;

Model training is performed using the target data set to obtain the service quality perception model.

Optionally, performing preprocessing on the second sample data set to obtain a target data set, including:

Using the maximum information coefficient method, calculate the maximum information coefficient between the data of each type of index parameter in the index parameter data and the associated user experience quality information;

According to the calculated maximum information coefficient, the target data set is selected from the second sample data set; wherein, the target data set includes the data of the first index parameter in the index parameter data and the data of the first index parameter and the first index parameter. The user experience quality information associated with the data of the index parameter, and the maximum information coefficient between the data of the first index parameter and the user experience quality information is greater than a preset threshold.

Optionally, after the fault location is performed, the method further includes:

Send the resulting failure information to the failed network domain.

In a second aspect, an embodiment of the present invention provides a fault location device for a video service, including:

a first acquisition module, configured to acquire the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network;

an analysis module, configured to analyze and obtain the user experience quality information of the video service based on the pre-trained service quality perception model and the data of the first index parameter; wherein the service quality perception model is used to represent the first index parameter The relationship between the data and user experience quality information;

An evaluation module, configured to evaluate the service quality of the video service according to the user experience quality information of the video service

a fault location module, configured to perform fault location based on the pre-trained fault location model set and the data of the second index parameter under the condition that the service quality of the video service does not meet the preset requirements;

Wherein, the fault location model set includes multiple fault location models, and each fault location model corresponds to a network domain.

Optionally, the fault location module is specifically used for:

The index parameter data of each network domain in the data of the second index parameter is respectively input into the fault location model corresponding to each network domain to perform fault inference.

Optionally, the first index parameters include: index parameters in the wireless network, index parameters in the core network, and index parameters in the bearer network.

Optionally, the second index parameters include: index parameters in the wireless network, index parameters in the transmission network, index parameters in the core network, index parameters in the bearer network, and index parameters in the user data center.

Optionally, the quality perception device further includes:

The execution module is configured to execute the following processes for each network domain to obtain the fault location model set:

obtaining a first sample data set of the network domain; wherein, the first sample data set includes index parameter data of the network domain;

Model training is performed by using the first sample data set to obtain a fault location model corresponding to the network domain.

Optionally, the quality perception device further includes:

a second obtaining module, configured to obtain a second sample data set; wherein, the second sample data set includes the index parameter data of the video service in the multi-domain network and the user experience quality information;

a preprocessing module for preprocessing the second sample data set to obtain a target data set;

A training module, configured to perform model training by using the target data set to obtain the service quality perception model.

Optionally, the preprocessing module includes:

a calculation unit, configured to use the maximum information coefficient method to calculate the maximum information coefficient between the data of each type of index parameter in the index parameter data and the associated user experience quality information;

A selection unit, configured to select the target data set from the second sample data set according to the calculated maximum information coefficient; wherein, the target data set includes the data of the first index parameter in the index parameter data and The user experience quality information associated with the data of the first index parameter, the maximum information coefficient between the data of the first index parameter and the user experience quality information is greater than a preset threshold.

Optionally, the quality perception device further includes:

The sending module is used for sending the obtained fault information to the network domain where the fault occurs.

In a third aspect, an embodiment of the present invention provides an electronic device, the electronic device includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, the program or instruction being The processor implements the steps of the method according to the first aspect when executed.

In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, where a program or an instruction is stored on the computer-readable storage medium, and when the program or instruction is executed by a processor, the method according to the first aspect is implemented A step of.

In the embodiment of the present invention, the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network can be obtained, and based on the pre-trained service quality perception model and the data of the first index parameter, analysis Obtaining the user experience quality information of the video service, evaluating the service quality of the video service according to the user experience quality information of the video service, and under the condition that the service quality of the video service does not meet the preset requirements. , and perform fault location based on the pre-trained fault location model set and the data of the second index parameter. Therefore, in the case of video services, end-to-end fault location for the global network can be automatically realized, improving the location efficiency, and through the pre-trained service quality perception model, the user experience-oriented network quality can be perceived in time. Users can accurately and comprehensively perceive abnormal situations in advance before a large number of complaints, so as to locate faults in time before a large number of complaints, and improve user experience.

Description of drawings

1 is a flowchart of a method for locating a fault of a video service according to an embodiment of the present invention;

2 is a flowchart of a service processing process in an embodiment of the present invention;

3 is a schematic structural diagram of a video service-oriented end-to-end quality perception and fault location system in an embodiment of the present invention;

4 is a flowchart of a training process of a service quality perception model and a network fault location model in an embodiment of the present invention;

5 is a flowchart of a service quality perception and network fault location reasoning process in an embodiment of the present invention;

6 is a schematic structural diagram of a fault locating device for a video service according to an embodiment of the present invention;

FIG. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

The terms "first", "second" and the like in the description and claims of the present invention are used to distinguish similar objects, and are not used to describe a specific order or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances so that embodiments of the invention can be practiced in sequences other than those illustrated or described herein, and distinguish between "first", "second", etc. The objects are usually of one type, and the number of objects is not limited. For example, the first object may be one or more than one. In addition, "and/or" in the description and claims indicates at least one of the connected objects, and the character "/" generally indicates that the associated objects are in an "or" relationship.

The method for locating the fault of the video service provided by the embodiment of the present invention will be described in detail below with reference to the accompanying drawings through specific embodiments and application scenarios thereof.

Please refer to FIG. 1. FIG. 1 is a flowchart of a method for locating a fault of a video service provided by an embodiment of the present invention. The method is applied to an electronic device. As shown in FIG. 1, the method includes the following steps:

Step 101: Acquire the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network.

In this embodiment, the first index parameter may include index parameters in the wireless network, index parameters in the core network, index parameters in the bearer network, and the like. The data of the first index parameter is, for example, deep packet inspection (Deep Packet Inspection, DPI) data in the corresponding network.

Optionally, for the first index parameter, the index parameters in the wireless network include but are not limited to air interface delay, signal strength, uplink and downlink round-trip delay (Round-Trip Time, RTT), etc.; the index parameters of the core network include but are not limited to. Limited to Transmission Control Protocol (TCP) chain establishment time, number of packet retransmissions, number of out-of-order packets, uplink and downlink RTT delay, etc.; bearer network indicators include but are not limited to TCP link establishment success rate, bearer network bidirectional Delay, bearer network packet loss rate, etc.

In this embodiment, the second index parameters may include index parameters in the wireless network, index parameters in the transmission network, index parameters in the core network, index parameters in the bearer network, and index parameters in the user data center. The data of the second index parameter is, for example, network management data in the corresponding network.

Optionally, for the second index parameter, the index parameter in the wireless network includes but is not limited to: physical resource block (physical resource block, PRB) utilization rate, radio resource control (Radio Resource Control, RRC) number of users, weak coverage type Indicators such as Reference Signal Receiving Power (RSRP), etc., interference indicators such as Signal to Interference plus Noise Ratio (SINR), etc., switching indicators, etc.; indicators in the transmission network include but Not limited to indicators such as bandwidth utilization, CPU utilization, delay, and packet loss rate; indicators in the core network include but are not limited to indicators such as interface bandwidth utilization, CPU utilization, and routing quality; indicators in the bearer network include However, it is not limited to indicators such as interface bandwidth utilization, CPU utilization, disk input/output (I/O) capability, and routing performance; the indicator parameters at the user data center include but are not limited to server performance (such as CPU capability, I/O bus capability, scheduling capability, etc.), server bandwidth, Content Delivery Network (CDN) hit rate, Domain Name System (Domain Name System, DNS) node load and other indicators.

Step 102: Based on the pre-trained service quality perception model and the data of the first index parameter, analyze and obtain the user experience quality information of the video service.

In this embodiment, the service quality perception model is used to represent the correlation between the data of the first index parameter and the user experience quality information.

In an embodiment, the user experience quality information can be selected as QoE data on the terminal side. The QoE data is, for example, a video quality Mean Opinion Score for Video (vMOS).

In another implementation manner, the user experience quality information can be selected as a user experience quality level. For example, the user experience quality level can be divided into five levels, namely level 1, level 2, level 3, level 4 and level 5.

Step 103: Evaluate the service quality of the video service according to the user experience quality information of the video service.

Optionally, in order to evaluate the service quality of the video service, a corresponding relationship between the user experience quality information and the service quality may be established in advance, so that after analyzing and obtaining the user experience quality information, the corresponding service quality is evaluated. For example, the user experience quality level is divided into five levels, respectively corresponding to very good, good, good, poor and very poor service quality.

Step 104: In the case that the service quality of the video service does not meet the preset requirements, the fault location is performed based on the pre-trained fault location model set and the data of the second index parameter.

Wherein, the fault location model set includes multiple fault location models, and each fault location model corresponds to a network domain. For example, the wireless network, transmission network, core network, bearer network, and user data center may respectively correspond to a fault location model, and all fault location models constitute a fault location model set.

Optionally, the above step 104 may include: inputting the index parameter data of each network domain in the data of the second index parameter into the fault location model corresponding to each network domain, respectively, to perform fault inference. In this way, end-to-end real-time accurate fault location can be realized for the global network.

In the embodiment of the present invention, the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network can be obtained, and based on the pre-trained service quality perception model and the data of the first index parameter, analysis Obtaining the user experience quality information of the video service, evaluating the service quality of the video service according to the user experience quality information of the video service, and under the condition that the service quality of the video service does not meet the preset requirements. , and perform fault location based on the pre-trained fault location model set and the data of the second index parameter. Therefore, in the case of video services, end-to-end fault location for the global network can be automatically realized, improving the location efficiency, and through the pre-trained service quality perception model, the user experience-oriented network quality can be perceived in time. Users can accurately and comprehensively perceive abnormal situations in advance before a large number of complaints, so as to locate faults in time before a large number of complaints, and improve user experience.

进行表示，

表示最后的分类器模型，f_k表示决策树，每个决策树根据输入样本X_i＝{x₁，x₂，…x_m}产生预测结果，最终通过投票确定最后的分类类别。比如结果是按照vMOS值分为1,2,3,4,5五个等级，分别代表业务质量很好、较好、好、差、很差。In the embodiment of the present invention, the above-mentioned service quality perception model can be selected from, but is not limited to, a deep learning model, a neural network model, a decision tree model, and the like. The above service quality perception model can be obtained by model training using the XGBoost algorithm. The XGBoost algorithm is an ensemble algorithm that uses

to express,

represents the final classifier model, f _k represents the decision tree, each decision tree produces prediction results according to the input samples X _i ={x ₁ , x ₂ ,...x _m }, and finally determines the final classification category by voting. For example, the result is divided into five grades: 1, 2, 3, 4, and 5 according to the vMOS value, which represent the service quality is very good, good, good, poor, and very poor.

Optionally, before the above step 101, a service quality perception model may be obtained by pre-training. The training process of the service quality perception model may include:

Obtain a second sample data set; wherein, the second sample data set includes index parameter data of the video service in the multi-domain network and user experience quality information associated with the index parameter data; the index parameter data is, for example, a wireless network , DPI data in the core network and bearer network;

Preprocessing the second sample data set to obtain a target data set;

Model training is performed using the target data set to obtain the service quality perception model.

Further, the above-mentioned preprocessing process for the second sample data set may include: first, cleaning, deduplicating, and normalizing the data in the second sample data set, and then processing by quintuple, start and end time, etc. Data association is performed between the indicator parameter data and the user experience quality information to obtain the target data set. The five-tuple is, for example, a set consisting of a source IP address, a source port, a destination IP address, a destination port and a transport layer protocol.

In addition, in order to improve the speed of model training, when preprocessing the second sample data set, it is also possible to select the index parameter data that is closely related to the user experience quality information from the second sample data set, and based on the selected training data set Perform model training. For example, when selecting index parameter data, the maximum information coefficient method can be used to select by the obtained maximum information coefficient (Maximal information coefficient, MIC).

Optionally, the above process of preprocessing the sample data set may include: first, using the maximum information coefficient method to calculate the relationship between the data of each type of index parameter in the index parameter data and the associated user experience quality information. maximum information coefficient MIC, and then select the target data set from the second sample data set according to the calculated maximum information coefficient. Wherein, the target data set includes data of the first indicator parameter in the indicator parameter data and user experience quality information associated with the data of the first indicator parameter, the data of the first indicator parameter and the The maximum information coefficient between user experience quality information is greater than the preset threshold. The preset threshold can be preset based on actual requirements. In this way, the most relevant index parameter data and user experience quality information can be selected by using the maximum information coefficient method, so as to improve the speed of model training on the premise of ensuring the accuracy of the model.

For example, assuming that the maximum information coefficient method is used to select the target data set from the sample data set D, the following formulas (1) and (2) can be used to sequentially calculate the data X and the label data of each type of index parameter of the data set D, that is, the user The MIC between the quality of experience information Y, if the MIC[X, Y] is greater than a certain threshold (which can be set according to the needs of the scene), the corresponding X and Y are selected until the cycle is completed, and the selected data forms a new dataset for model training.

In the above formulas (1) and (2), P(X) represents the probability distribution of X; P(Y) represents the probability distribution of Y; P(X, Y) represents the joint probability distribution of X and Y; min(|x |,|y|) means the smaller of the number of grids divided in the X direction and the number of grids divided in the Y direction.

Optionally, for the training of the service quality perception model, the selected new feature data with a large correlation coefficient may be used as input data, and user experience quality information such as vMOS may be used as output data, and the XGBoost algorithm may be used for model training. Specific processes may include:

S1, given a dataset D={(X _i , y _i )} (|D|=n, X _i ∈ R ^m , y _i ∈ R), the number of samples in the dataset is n, and the amount of feature data is m, Initialize an ensemble decision tree from the dataset:

Among them, w is the output leaf vector, and T is the number of leaf nodes.

S2, determine the loss function, the algorithm is trained in an additive manner, and the objective function in the t-th round of training is expressed as:

为正则项，γ和λ为常量，l()为交叉熵损失函数。定义节点分裂损失降低函数为：in,

is the regular term, γ and λ are constants, and l() is the cross-entropy loss function. Define the node splitting loss reduction function as:

Among them, _gi is the first-order gradient of the loss function, and _hi is the second-order gradient of the loss function. I is the sample instance set, G and _H are the first-order gradient and the second-order gradient sum corresponding to the instance I, and GL , _HL , _GR , and _HR are the corresponding left node instance _IL and right node instance IR respectively _. The sum of the first-order gradient and the second-order gradient.

S3, perform a decision tree node split search to generate an integrated decision tree. The specific steps of splitting decision tree nodes are as follows: First, according to the current node instance set I, initialize the first-order gradient G=∑ _i∈I g _i , the second-order gradient H=∑ _i∈I h _i , and the left-split node first-order gradient G _L =0, the first-order gradient of the right split node G _R =G, the second-order gradient of the left split node H _L =0, the second-order gradient of the right split node H _R =H; update _GL , _HL , _GR , H _R , the update follows G _L ←G _L +g _i , H _L ←H _L +h _j , G _R ←GG _L , H _R ←HH _L , and the maximum split loss reduction value is L _split , according to the split The loss reduction value is the largest node split case for node splitting. This process is repeated until the entire decision tree is constructed.

S4, use the test sample to generate a prediction result for the generated decision tree, and the category corresponding to the leaf node with the largest probability is the final classification result. The trained XGBoost model is a service quality perception model, which can accurately infer abnormal user experience through the DPI data of the wireless network, core network, and bearer network.

In the embodiment of the present invention, the training process of the above fault location model set may include:

For each network domain, perform the following processes respectively to obtain the fault location model set:

obtaining a first sample data set of the network domain; wherein, the first sample data set includes index parameter data of the network domain;

Model training is performed by using the first sample data set to obtain a fault location model corresponding to the network domain.

That is to say, for each network domain such as a wireless network, a transmission network, a core network, a bearer network, or a user data center, a corresponding fault location model can be obtained by training separately. Then, the fault localization models obtained by training are integrated to obtain a fault localization model set. For example, when training a fault location model, the model can be trained through unsupervised learning such as the isolated forest iForest algorithm.

For example, taking the fault location model corresponding to the wireless network as an example, the sample data set on the wireless network side is Z, and the isolated forest iForest algorithm is used for model training. The iForest consists of t isolation trees (Isolation Tree, iTree), each iTree is A binary tree structure. The training process of the fault location model for the wireless network may include:

a) randomly select Ψ sample points as subsamples from the sample data set Z, and put them into the root node of the tree;

b) Randomly specify a dimension, randomly generate a cut point p in the current node data, and the cut point is generated between the maximum value and the minimum value of the specified dimension in the current node data;

c) Generate a hyperplane at this cutting point, then divide the current node data space into 2 subspaces, place the data with the specified dimension less than p on the left side of the current node, and place the data greater than or equal to p on the right side of the current node ;

d) Recursively b) and c) in the child nodes, and continuously construct new child nodes, until there is only one data in the child node (can no longer continue to cut) or the child node has reached the limit height. After getting t iTrees, the training of the iForest model ends.

Afterwards, the trained fault location model can be deployed on the cloud server for subsequent fault inference. The reasoning process is as follows: first, input the data x in the relevant time period of the abnormal user experience into the fault location model, make it traverse each iTree, and then calculate which layer of each iTree the data x finally falls on (x is at the height of the tree) , and finally calculate the average height of data x in each iTree. If the average height is lower than the preset threshold, it is determined to be a fault, otherwise there is no fault. The preset threshold can be preset based on actual requirements.

Optionally, after the fault location is performed, the obtained fault information may be sent to the network domain where the fault occurs, so as to issue an alarm for the faulty network to realize automatic operation and maintenance.

The present application will be described below with reference to specific embodiments and accompanying drawings.

As shown in FIG. 2 , the service processing process involved in this embodiment of the present application may include:

Step 21: Collect data. Wherein, the data collected in this step includes three types: 1) DPI data of wireless network, core network, and bearer network; 2) network management data of wireless network, transmission network, core network, bearer network, and user data center; 3) Video QoE data of the terminal.

Step 22: Preprocess the data collected in Step 21, such as cleaning, deduplication, normalization, etc., and then convert the wireless network, core network, bearer network DPI data and video through quintuple, start and end timestamps, etc. QoE data for data association.

Step 23: Determine whether the trained service quality perception model and network fault location model have been deployed. If it has been deployed, model inference can be performed directly. If model initialization training has not been performed, model training will be started.

Step 24: Train the service quality perception model and the network fault location model, and deploy them to the cloud server. For the training process of the service quality perception model and the network fault location model, reference may be made to the above content, and details are not repeated here.

Step 25: Use the trained service quality perception module to monitor the service.

Step 26: Determine whether there is abnormal user experience. If it is determined that there is an abnormal user experience, step 27 is performed; otherwise, step 29 is performed.

Step 27: Use the network fault location model set to locate the fault.

Step 28: Send an alarm notification for the faulty network.

Step 29: End the relevant process or start the next monitoring cycle.

Referring to FIG. 3 , an embodiment of the present invention further provides an end-to-end quality perception and fault location system for video services, which mainly includes a network physical facility layer, a data collection and processing module deployed on a cloud server, and a data collection and processing module deployed on the cloud server. The model training module and the model inference module deployed on the cloud server. The data collection and processing module includes a data collection sub-module and a data processing sub-module. The data collection sub-module can collect data from physical layers such as wireless network, transmission network, core network, bearer network, and user data center network. The data can be preprocessed by the data processing submodule. The model training module includes a service quality perception model and a network fault location model set. The service quality perception model is obtained by training a supervised learning model based on video vMOS tags and service-related data. The network fault location model set is based on each domain data for unsupervised learning. Learn to get. The model inference module includes a service quality perception module, a network fault location module and an alarm notification module. The service quality perception module is used for abnormal user identification, the network fault location module is used for fault location, and the alarm notification module is used for alarm notification. It should be pointed out that the data collection and processing module and the model inference module can be set on one device, or can be set on different physical devices respectively.

With reference to FIG. 3 and FIG. 4 , the training process of the service quality perception model and the network fault location model in this embodiment may include:

S1: The data collection sub-module in the data collection and processing module collects data, including: S1.1, collect the video terminal to collect vMOS value as a label; S1.2, collect the DPI data of the wireless network; S1.3, collect the DPI of the core network Data; S1.3, collect the DPI data of the bearer network;

S2: The data processing sub-module cleans, deduplicates, and normalizes the data collected in S1, and then uses key fields such as quintuple, start and end timestamps to associate.

S3: The model training module uses the associated data of each domain to train the supervised learning model, that is, the associated data is used as the model input data, and the vMOS value is used as the label, and the supervised learning model is used for training to obtain a service quality perception model, so as to detect abnormal users with poor experience. to identify.

S4: The data collection sub-module in the data collection and processing module collects data, including: S4.1, collect network management data of wireless network; S4.2, collect network management data of transmission network; S4.3, collect network management data of core network; S4.4, collecting the network management data of the bearer network; S4.5, collecting the network management data of the user data center.

S5: The data processing sub-module performs cleaning, deduplication, normalization and other processing on the data collected in S4.

S6: The model training module uses the processed data of each domain to perform model training using the unsupervised learning method, and obtains an end-to-end fault composed of five fault location models including wireless network, transmission network, core network, bearer network and user data center. Locate the model set.

With reference to FIG. 3 and FIG. 5 , the service quality perception and network fault location reasoning process in this embodiment may include:

S1: The data collection sub-module in the data collection and processing module collects data, including: S1.1, collect the video terminal to collect vMOS value as a label; S1.2, collect the DPI data of the wireless network; S1.3, collect the DPI of the core network Data; S1.3, collect the DPI data of the bearer network;

S2: The data processing sub-module cleans, deduplicates, and normalizes the data collected in S1, and then uses key fields such as quintuple, start and end timestamps to associate.

S3: Input the associated data into the service quality perception module to identify abnormal user experience, that is, input the associated data to the service quality perception model for inference.

S4: if the service quality perception module determines that there is an abnormal user experience, the service quality perception module sends the service quality alarm to the network fault location model;

S5: The network fault location module sends a fault location-related data request to the data collection module;

S6: The data collection sub-module in the data collection and processing module collects data, including: S6.1, collects network management data of the wireless network; S6.2, collects the network management data of the transmission network; S6.3, collects the network management data of the core network; S6.4, collecting the network management data of the bearer network; S6.5, collecting the network management data of the user data center.

S7: The data processing sub-module performs cleaning, deduplication, normalization and other processing on the data collected in S7.

S8: Input the processed data into each fault location model in the corresponding network fault location model set to identify the fault, and output the result to the alarm notification module;

S9.1-S9.5: The alarm notification sub-module transmits the fault-related information to the network domain where the fault occurs, such as the wireless network, transmission network, core network, bearer network, and user data center.

To sum up, this solution can directly correlate multi-domain network indicators with user QoE, model and build a service quality perception model, so as to realize timely perception of network quality for user experience. Further, according to the abnormal network quality perception, it is proposed to collect and aggregate multi-domain network management data for an overall unified analysis, define an intelligent fault location model set based on unsupervised learning, and automatically realize end-to-end real-time accurate fault location for the global network.

The service quality perception model in this solution can assist operators to evaluate the QoE of video users through network-side data, perceive the quality of network services provided to users, detect user anomalies in time, prevent a large number of complaints from users, and extract the time when abnormal users occur. Abnormal information related to location, etc., assists further fault location.

The network fault location model set in this solution can automatically and accurately locate faults in the entire network domain in real time, and send alarm notifications, which greatly saves human operation and maintenance costs and time costs.

Please refer to FIG. 6. FIG. 6 is a schematic structural diagram of a fault locating apparatus for a video service provided by an embodiment of the present invention. The apparatus is applied to electronic equipment. As shown in FIG. 6, the fault locating apparatus 60 for the video service may include:

The first obtaining module 61 is used to obtain the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network;

An analysis module 62, configured to analyze and obtain the user experience quality information of the video service based on the pre-trained service quality perception model and the data of the first index parameter; wherein the service quality perception model is used to represent the first index The relationship between parameter data and user experience quality information;

An evaluation module 63, configured to evaluate the service quality of the video service according to the user experience quality information of the video service

A fault location module 64, configured to perform fault location based on the pre-trained fault location model set and the data of the second index parameter under the condition that the service quality of the video service does not meet the preset requirement;

Wherein, the fault location model set includes multiple fault location models, and each fault location model corresponds to a network domain.

Optionally, the fault location device 60 further includes:

Optionally, the fault location module 64 is specifically used for:

The index parameter data of each network domain in the data of the second index parameter is respectively input into the fault location model corresponding to each network domain to perform fault inference.

Optionally, the first index parameters include: index parameters in the wireless network, index parameters in the core network, and index parameters in the bearer network.

Optionally, the second index parameters include: index parameters in the wireless network, index parameters in the transmission network, index parameters in the core network, index parameters in the bearer network, and index parameters in the user data center.

Optionally, the quality perception device 60 further includes:

The execution module is configured to execute the following processes for each network domain to obtain the fault location model set:

obtaining a first sample data set of the network domain; wherein, the first sample data set includes index parameter data of the network domain;

Model training is performed by using the first sample data set to obtain a fault location model corresponding to the network domain.

Optionally, the quality perception device 60 further includes:

a second obtaining module, configured to obtain a second sample data set; wherein, the second sample data set includes the index parameter data of the video service in the multi-domain network and the user experience quality information;

a preprocessing module for preprocessing the second sample data set to obtain a target data set;

A training module, configured to perform model training by using the target data set to obtain the service quality perception model.

Optionally, the preprocessing module includes:

a calculation unit, configured to use the maximum information coefficient method to calculate the maximum information coefficient between the data of each type of index parameter in the index parameter data and the associated user experience quality information;

A selection unit, configured to select the target data set from the second sample data set according to the calculated maximum information coefficient; wherein, the target data set includes the data of the first index parameter in the index parameter data and The user experience quality information associated with the data of the first index parameter, the maximum information coefficient between the data of the first index parameter and the user experience quality information is greater than a preset threshold.

Optionally, the quality perception device 60 further includes:

The sending module is used for sending the obtained fault information to the network domain where the fault occurs.

It is understandable that the device 60 for locating the fault of the video service according to the embodiment of the present invention can implement each process of the method embodiment shown in FIG. 1 above, and can achieve the same technical effect.

In addition, an embodiment of the present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program is executed by the processor During execution, each process of the method embodiment shown in FIG. 1 can be implemented and the same technical effect can be achieved. To avoid repetition, details are not described here.

Referring to FIG. 7 , an embodiment of the present invention further provides an electronic device 70 , including a bus 71 , a transceiver 72 , an antenna 73 , a bus interface 74 , a processor 75 and a memory 76 .

In the embodiment of the present invention, the electronic device 70 further includes: a computer program stored on the memory 76 and executable on the processor 75 . It is understandable that when the computer program is executed by the processor 75, each process of the method embodiment shown in FIG. 1 can be implemented, and the same technical effect can be achieved. In order to avoid repetition, details are not repeated here.

In FIG. 7, the bus architecture (represented by bus 71), which may include any number of interconnected buses and bridges, will include one or more processors, represented by processors 75, and memory, represented by memory 76. The various circuits are linked together. The bus 71 may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, etc., which are well known in the art and therefore will not be described further herein. Bus interface 74 provides an interface between bus 71 and transceiver 72 . Transceiver 72 may be a single element or multiple elements, such as multiple receivers and transmitters, providing a means for communicating with various other devices over a transmission medium. The data processed by the processor 75 is transmitted on the wireless medium through the antenna 73 , and further, the antenna 73 also receives the data and transmits the data to the processor 75 .

The processor 75 is responsible for managing the bus 71 and general processing, and may also provide various functions including timing, peripheral interface, voltage regulation, power management, and other control functions. Instead, memory 76 may be used to store data used by processor 75 in performing operations.

Optionally, the processor 75 may be a CPU, ASIC, FPGA or CPLD.

Embodiments of the present invention further provide a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, each process of the method embodiment shown in FIG. 1 can be implemented and the same technology can be achieved. The effect, in order to avoid repetition, is not repeated here.

Computer-readable media includes both permanent and non-permanent, removable and non-removable media, and can be implemented by any method or technology for storage of information. Information may be computer readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Flash Memory or other memory technology, Compact Disc Read Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, Magnetic tape cartridges, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, excludes transitory computer-readable media, such as modulated data signals and carrier waves.

It should be noted that, herein, the terms "comprising", "comprising" or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements, It also includes other elements not expressly listed or inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in a process, method, article or apparatus that includes the element.

The above-mentioned serial numbers of the embodiments of the present invention are only for description, and do not represent the advantages or disadvantages of the embodiments.

From the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general hardware platform, and of course hardware can also be used, but in many cases the former is better implementation. Based on this understanding, the technical solutions of the present invention can be embodied in the form of software products in essence or the parts that make contributions to the prior art, and the computer software products are stored in a storage medium (such as ROM/RAM, magnetic disk, CD-ROM), including several instructions to make a service classification device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in the various embodiments of the present invention.

The above are only the preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, without departing from the principles of the present invention, several improvements and modifications can be made. It should be regarded as the protection scope of the present invention.

Claims (14)

1. A method for locating a fault of a video service is characterized by comprising the following steps:

acquiring data of a first index parameter and data of a second index parameter of a video service in a multi-domain network;

analyzing to obtain user experience quality information of the video service based on a pre-trained service quality perception model and the data of the first index parameter; the service quality perception model is used for representing the incidence relation between the data of the first index parameter and the user experience quality information;

evaluating the service quality of the video service according to the user experience quality information of the video service;

under the condition that the service quality of the video service does not meet the preset requirement, fault location is carried out based on a pre-trained fault location model set and the data of the second index parameter;

wherein the set of fault location models includes a plurality of fault location models, each corresponding to a network domain.

2. The method of claim 1, wherein the fault locating based on the pre-trained fault location model set and the data of the second index parameter comprises:

and respectively inputting the index parameter data of each network domain in the data of the second index parameters into the fault positioning model corresponding to each network domain to carry out fault reasoning.

3. The method of claim 1, wherein the first metric parameter comprises: index parameters in a wireless network, index parameters in a core network and index parameters in a bearer network.

4. The method of claim 1, wherein the second indexing parameter comprises: index parameters in a wireless network, index parameters in a transmission network, index parameters in a core network, index parameters in a bearer network, and index parameters at a user data center.

5. The method according to claim 1, wherein the obtaining of the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network is preceded by:

for each network domain, respectively executing the following processes to obtain the fault location model set:

obtaining a first sample dataset of the network domain; wherein the first sample dataset comprises metric parameter data for the network domain;

and performing model training by using the first sample data set to obtain a fault positioning model corresponding to the network domain.

6. The method according to claim 1, wherein the obtaining of the data of the first index parameter and the data of the second index parameter of the video service in the multi-domain network is preceded by:

acquiring a second sample data set; the second sample data set comprises index parameter data of the video service in the multi-domain network and user experience quality information;

preprocessing the second sample data set to obtain a target data set;

and performing model training by using the target data set to obtain the service quality perception model.

7. The method of claim 6, wherein said pre-processing said second set of sample data to obtain a target data set comprises:

calculating a maximum information coefficient between the data of each type of index parameter in the index parameter data and the associated user experience quality information by using a maximum information coefficient method;

selecting the target data set from the second sample data set according to the maximum information coefficient obtained by calculation; wherein the target data set includes data of a first index parameter in the index parameter data and user experience quality information associated with the data of the first index parameter, and a maximum information coefficient between the data of the first index parameter and the user experience quality information is greater than a preset threshold.

8. The method of claim 1, wherein after fault locating, the method further comprises:

and sending the obtained fault information to the network domain with the fault.

9. A video service fault location apparatus, comprising:

the first acquisition module is used for acquiring data of a first index parameter and data of a second index parameter of the video service in the multi-domain network;

the analysis module is used for analyzing and obtaining user experience quality information of the video service based on a pre-trained service quality perception model and the data of the first index parameter; the service quality perception model is used for representing the incidence relation between the data of the first index parameter and the user experience quality information;

an evaluation module for evaluating the quality of the video service according to the user experience quality information of the video service

The fault positioning module is used for positioning faults based on a pre-trained fault positioning model set and the data of the second index parameter under the condition that the service quality of the video service does not meet the preset requirement;

wherein the set of fault location models includes a plurality of fault location models, each corresponding to a network domain.

10. The apparatus of claim 9,

the fault location module is specifically configured to: and under the condition that the service quality of the video service does not meet the preset requirement, respectively inputting the index parameter data of each network domain in the data of the second index parameter into a fault positioning model corresponding to each network domain for fault reasoning.

11. The apparatus of claim 9, wherein the first metric parameter comprises: index parameters in a wireless network, index parameters in a core network and index parameters in a bearer network.

12. The apparatus of claim 9, wherein the second indexing parameter comprises: index parameters in a wireless network, index parameters in a transmission network, index parameters in a core network, index parameters in a bearer network, and index parameters at a user data center.

13. An electronic device comprising a processor, a memory and a program or instructions stored on the memory and executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method of fault location of a video service according to any one of claims 1 to 8.

14. A computer-readable storage medium, characterized in that a program or instructions are stored on the computer-readable storage medium, which program or instructions, when executed by a processor, implement the steps of the method for fault localization of video services according to any of claims 1 to 8.

Images