CN109672548B

CN109672548B - Long term evolution voice VoLTE network fault detection method, device and server

Info

Publication number: CN109672548B
Application number: CN201710966390.1A
Authority: CN
Inventors: 左一平; 梁燕萍
Original assignee: Research Institute of China Mobile Communication Co Ltd; China Mobile Communications Corp
Current assignee: China Mobile Communications Group Co Ltd; Research Institute of China Mobile Communication Co Ltd
Priority date: 2017-10-17
Filing date: 2017-10-17
Publication date: 2022-04-01
Anticipated expiration: 2037-10-17
Also published as: CN109672548A

Abstract

The invention provides a VoLTE network fault detection method, a VoLTE network fault detection device and a VoLTE network fault detection server. The method comprises the following steps: acquiring packet loss statistical data of each monitoring network point; according to the packet loss statistical data, the corresponding relation between the user label corresponding to the user with the packet loss and the packet loss information is counted; and determining the user label of the user with the probable packet loss according to the corresponding relation. By adopting the method, the user label of the user with the packet loss can be determined at a high probability by associating the user label with the packet loss information, so that an accurate basis is provided for defining the fault reason of the link with the VoLTE network fault.

Description

Long term evolution voice VoLTE network fault detection method, device and server

Technical Field

The invention relates to the technical field of wireless, in particular to a VoLTE network fault detection method, a VoLTE network fault detection device and a VoLTE network fault detection server.

Background

With the increasing application of Long Term Evolution (LTE for short) Voice over LTE (Voice over LTE for short), it is important to improve network performance, and an important task is network problem location. However, the voice problem involves numerous network elements and complex scenes, and even because cross-department and cross-operator can also generate the phenomenon of liability deniability, how to accurately delimit the problem becomes the key point and difficulty in improving the voice quality.

Currently, there is a method for delimiting problems such as Voice over Internet Protocol (VoIP) or the like by counting the number of lost packets of Real-time Transport Protocol (RTP) and the number of lost packets of RTP carried in RTP Control Protocol (RTCP) with a monitoring point. For example, a monitoring point may be set in each network segment from a source end to an opposite end, and whether packet loss exists in each network segment is monitored to find out which segment of the whole link the failure occurs in, so as to perform VoLTE network failure detection.

However, in the above manner, the number of RTP packet drops can only roughly define which segment the network problem occurs in, but cannot define the specific reason of the Volte network failure, so it is necessary to provide a method capable of accurately detecting the Volte network failure.

Disclosure of Invention

The technical scheme of the invention aims to provide a VoLTE network fault detection method, a VoLTE network fault detection device and a VoLTE network fault detection server, which are used for solving the problems that only a link with a VoLTE network fault can be determined and the specific fault reason cannot be defined by adopting a VoLTE network fault detection mode in the prior art.

The specific embodiment of the invention provides a method for detecting a long-term evolution voice Voice (Volte) network fault, which comprises the following steps:

acquiring packet loss statistical data of each monitoring network point;

according to the packet loss statistical data, the corresponding relation between the user label corresponding to the user with the packet loss and the packet loss information is counted;

and determining the user label of the user with the probable packet loss according to the corresponding relation.

Preferably, the Volte network fault detection method further includes:

and analyzing the network fault according to the user label of the user with the determined approximate probability packet loss.

Preferably, after the step of obtaining packet loss statistical data of each monitoring node, the method further includes:

according to the packet loss statistical data, extracting specific users of which the total packet number of data transmitted in each monitoring network point is greater than or equal to a first numerical value and the packet loss rate is greater than or equal to a second numerical value;

wherein, in the step of counting the corresponding relationship between the user tag corresponding to the user with the packet loss and the packet loss information according to the packet loss statistical data:

and according to the packet loss statistical data, counting the corresponding relation between the user label corresponding to the specific user and the packet loss information.

Preferably, the method for detecting a fault in a Volte network, wherein the step of counting a correspondence between a user tag corresponding to a user who has lost a packet and packet loss information according to the packet loss statistical data includes:

counting the corresponding relation between the user labels with different dimensionalities of the users with packet loss and the packet loss information;

and establishing a Hash tree according to the corresponding relation.

Preferably, the method for detecting a fault in a Volte network, wherein the step of determining the user tag of the user with the probable packet loss according to the correspondence includes:

deleting the user tags of which the corresponding packet loss information does not meet a first preset condition in the Hash tree to obtain the pruned Hash tree;

and determining the user label of which the corresponding packet loss information meets a second preset condition in the pruned Hash tree, wherein the user label is the user label of the user with the packet loss with the high probability.

Preferably, in the method for detecting a fault in a Volte network, in the step of deleting a user tag in the Hash tree whose corresponding packet loss information does not meet a first preset condition, and obtaining the pruned Hash tree, the packet loss information includes a total packet number and a packet loss rate of the counted transmission data, where the first preset condition is that the total packet number is smaller than a third numerical value, and the packet loss rate is smaller than a fourth numerical value.

Preferably, the method for detecting a fault in a Volte network, wherein the step of determining that the corresponding packet loss information in the pruned Hash tree conforms to a user tag of a second preset value includes:

sequencing the corresponding user tags according to the sequence of the packet loss numbers recorded in the packet loss information from high to low in the pruned Hash tree;

extracting the first N user tags in the sorted user tags;

analyzing whether at least two user tag groups which have inclusion relation and the difference value between the corresponding packet loss numbers is smaller than a preset numerical value exist in the previous N user tags or not;

if the user tag group exists, retaining one user tag in the user tag group, and deleting other tags in the user tag group;

and supplementing the deleted user tags in the first N user tags according to the sorted user tags, and determining that the N user tags after the user tags are supplemented are the user tags of the users with the probability of packet loss.

The specific embodiment of the present invention further provides a device for detecting a long term evolution voice network fault, where the device includes:

the acquisition module is used for acquiring the packet loss statistical data of each monitoring network point;

the first analysis module is used for counting the corresponding relation between the user label corresponding to the user with the packet loss and the packet loss information according to the packet loss statistical data;

and the second analysis module is used for determining the user label of the user with the probable packet loss according to the corresponding relation.

In another aspect, an embodiment of the present invention further provides a server, including a memory, a processor, and a computer program stored in the memory and executable on the processor; the processor, when executing the program, implements the Volte network failure detection method as described in any of the above.

Yet another aspect of the embodiments of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the program, when executed by a processor, implements the Volte network fault detection method as described in any one of the above.

One or more embodiments of the invention have at least the following beneficial effects:

according to the Volte network fault detection method, the user tags of the users with the packet loss at a high probability can be determined by associating the user tags with the packet loss information, so that an accurate basis is provided for defining the fault reasons of the links with the VoLTE network fault.

Drawings

Fig. 1 is a schematic flowchart of a Volte network fault detection method according to an embodiment of the present invention;

fig. 2 is a schematic flow chart of a Volte network fault detection method according to another embodiment of the present invention;

FIG. 3 is a flowchart illustrating the execution of step S120;

FIG. 4 is a schematic structural diagram of a Hash tree established by the Volte network fault detection method according to the embodiment of the present invention;

FIG. 5 is a flowchart illustrating the execution of step S130;

FIG. 6 is a flowchart illustrating the execution of step S132;

fig. 7 is a schematic structural diagram of a voltage network fault detection apparatus according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

As shown in fig. 1, the method for detecting a long term evolution voice network fault according to the specific embodiment of the present invention includes:

s110, obtaining packet loss statistical data of each monitoring network point;

s120, according to the packet loss statistical data, counting the corresponding relation between the user label corresponding to the user with packet loss and the packet loss information;

and S130, determining the user label of the user with the probable packet loss according to the corresponding relation.

Referring to fig. 2, preferably, after step S130, the method for detecting a fault in a Volte network according to the embodiment of the present invention further includes:

and S140, analyzing the network fault according to the user label of the user with the determined approximate probability packet loss.

In addition, in another aspect of the embodiment of the present invention, after the step S110, the method further includes:

in step S120, according to the packet loss statistical data, the step of counting a correspondence between a user tag corresponding to a user with a packet loss and packet loss information is:

By adopting the steps, in order to avoid randomness, a total packet number threshold (a first numerical value) and a packet loss rate threshold (a second numerical value) of data transmitted by a user of each monitoring node are preset, and only statistical data, which is obtained from each monitoring node and has the total packet number of the transmitted data being greater than or equal to the first numerical value (for example, m) and the packet loss rate being greater than or equal to the second numerical value (for example, n%), is used as a data basis for calculating a user tag of a user with the approximate packet loss rate.

In addition, specifically, in step S110, by setting a monitoring point between network elements in a certain area, for example, setting a monitoring point between two Session Border Controllers (SBC), packet loss statistical data is obtained, and packet loss statistics can be performed. The data transmitted between the network elements can be judged whether the data is the packet loss statistical data according to the continuity of the Sequence number carried in the RTP.

In step S120, the packet loss information in the packet loss statistical data is associated with the user tag, so as to calculate which users, cells and/or network elements have packet losses, and establish a corresponding relationship between the user tag corresponding to the user having packet losses and the packet loss information.

Preferably, in an embodiment of the present invention, the packet loss information may include a total packet number, a packet loss number, and/or a packet loss rate of data transmitted within a preset time.

In addition, in the specific embodiment of the present invention, the user tag may include multiple dimensions, for example, as follows, the multiple dimensions included in the user tag may be:

the transmission direction is as follows: ascending or descending;

the session type is as follows: voice or video

Network element: central Cell, base station, public data network gateway PGW, SBC

Time: 0 to 24 points

The place attribute is as follows: grid (C)

Equipment manufacturers: zhongxing, Huashi, or Nokia, etc. (version number may be included if necessary)

The user: IMSI, mobile phone number, etc.;

the terminal model: apple, millet, three stars (if necessary, also can contain OS version number)

The network type: 4G-4G, 4G-2G, 2G-4G

According to the multi-dimensional user tag, the association shown in table 1 below can be established between the user tag corresponding to the user who has lost the packet and the packet loss information:

TABLE 1

User' s

Terminal model

Cell

……

Total number of packets

Number of lost packets

Packet loss rate

A

Apple (Malus pumila)

Cell1

……

1000

10

1％

B

Millet

Cell2

……

2000

1

<1％

C

Three stars

Cell1

……

3000

15

0.5％

In the method for detecting a fault in a Volte network according to the embodiment of the present invention, according to the association between the user tag corresponding to the user with the packet loss and the packet loss information, statistics of the corresponding relationship between the user tag and the packet loss information is performed, and preferably, step S120 is shown in fig. 3, where the step of performing statistics of the corresponding relationship between the user tag corresponding to the user with the packet loss and the packet loss information according to the packet loss statistical data includes:

s121, counting the corresponding relations between the user tags with different dimensionalities of the users with packet loss and the packet loss information;

and S122, establishing a Hash tree according to the corresponding relation.

For example, the corresponding relationship between the user tag of a single dimension such as a user, a network element, and a session type and the packet loss information is respectively counted, the corresponding relationship between the user tag of a user and a cell combination dimension and the packet loss information is further counted, the corresponding relationship between the user tag of a user, a cell, and a session type that is a voice combination dimension and the packet loss information is further counted, by using this way, the corresponding relationship between the user tag of different dimensions and the packet loss information for defining a user with a packet loss is established, and a Hash tree as shown in fig. 4 is formed, so as to be used for determining the user tag of a subsequent user with a packet loss at a high probability.

Preferably, in the method for detecting a fault in a Volte network according to the specific embodiment of the present invention, in step S130, the step of determining the user tag of the user with the probable packet loss according to the correspondence relationship includes, as shown in fig. 5:

s131, deleting the user tags, corresponding to the packet loss information, of which the packet loss information does not meet the first preset condition, in the Hash tree to obtain the pruned Hash tree;

and S132, determining that the corresponding packet loss information in the pruned Hash tree meets a user tag of a second preset condition, and the user tag is a user tag of a user with a high probability of packet loss.

Specifically, in step S131, deleting the user tag in the Hash tree whose corresponding packet loss information does not meet a first preset condition, and obtaining the pruned Hash tree, where the packet loss information includes a total packet number and a packet loss rate of the counted transmission data, and the first preset condition is that the total packet number is smaller than a third numerical value, and the packet loss rate is smaller than a fourth numerical value.

Through the step S131, thresholds are respectively set for the total packet number and the packet loss rate of the transmission data counted in the preset time, and based on the set thresholds, the Hash tree is pruned, and the non-critical user tags are removed, so that the calculation amount is reduced, and the result of the method for detecting the fault in the Volte network according to the embodiment of the present invention is obtained more quickly.

In the embodiment of the present invention, preferably, for a user tag with a single dimension, the threshold retained in the Hash tree is set according to the user tag with each dimension, and preferably, the retained user tag with a single dimension includes: user, network element, location attribute, set manufacturer and terminal model. The decision rule for determining whether a certain user tag appears in a user tag of a single dimension is to decide whether it is likely to be "pruned". Generally, for a relatively large number of application labels, for example, thousands of application labels may exist in a device, so that some devices are inevitably subjected to "pruning" due to insufficient traffic; on the contrary, for the user tags of the "uplink/downlink" type, since the uplink and downlink traffic volume is generally equivalent in a certain time period, the situation that the uplink or the downlink is "pruned" generally does not occur, and therefore, the user tags in a single dimension are not recommended to be present.

Taking the data statistics of a day in a certain area as an example, the threshold of the user tag of each single dimension can be shown in the following table 2:

TABLE 2

In addition, the determination method for the multi-dimensional user tag threshold can adopt the following steps: and dividing the threshold of the single-dimension user label by the corresponding statistical number of other dimension user labels.

By adopting the mode, only the total packet number of the counted transmission data is kept in the Hash tree and is larger than or equal to the third data, the user tags with the packet loss rate larger than or equal to the fourth numerical value are kept, and other user tags are deleted.

Preferably, in step S132, the step of determining that the corresponding packet loss information in the pruned Hash tree corresponds to the user tag of the second preset value, and is the user tag of the user with the probability of packet loss, includes, as shown in fig. 6:

s1321, sorting the corresponding user tags according to the sequence of the packet loss numbers recorded in the packet loss information from high to low in the pruned Hash tree;

s1322, extracting the first N user tags in the sorted user tags;

s1323, whether at least two user label groups which have an inclusion relationship and the difference value between the corresponding packet loss numbers is smaller than a preset numerical value exist in the previous N user labels is analyzed;

s1324, if the user tag group exists, reserving one user tag in the user tag group, and deleting other tags in the user tag group;

s1325, according to the sorted user tags, the deleted user tags in the first N user tags are supplemented, and the N user tags after the user tags are supplemented are determined to be the user tags of the users with the probability of packet loss.

In the above manner, the first N user tags with the higher packet loss number recorded in the packet loss information are obtained through step S1321 and step S1322; analyzing the extracted first N user tags through steps S1323 and S1324, screening out the user tags with the inclusion relationship, reserving the user tags with the inclusion relationship and keeping one of the user tags in the user tag group with the difference value between the corresponding packet loss numbers smaller than a preset value, deleting other user tags, and supplementing the other user tags into the extracted first N user tags through step S1325, thereby determining that the finally obtained N user tags after the user tags are supplemented are the user tags of the users with the packet loss at the approximate rate.

For example, when the multidimensional user tags a2 and B1 are completely contained in the one-dimensional user tag a2 and the difference between the corresponding packet loss numbers is smaller than the preset value, it can be considered that the user tags at the lower level are completely contained in the user tags at the higher level, and one of the user tags can be deleted.

The N user tags of the users with the packet loss of the large probability determined by the method are more accurate, the repetition is avoided, and accurate data are provided for the follow-up network fault analysis.

In the method for detecting a Volte network fault according to the specific embodiment of the present invention, in step S140, when performing network fault analysis according to the user tag of the user with the determined approximate probability packet loss, preferably, the specific way of performing network fault analysis may be:

for the determined N user tags, each user tag is prioritized, and the user tags can be roughly classified into a tag type (such as network element) of a network problem needing to be located and a tag type (such as user and terminal type) of a network problem needing not to be located;

for each user label, if the corresponding user label contains the label type without positioning the network problem, the investigation focus can be concentrated on factors such as users, time periods, even version numbers and the like;

for each user tag, if the corresponding user tag only contains the tag type of the network problem to be located, processing is carried out according to the property of the tag type.

It can be understood that, by using the method for detecting a Volte network fault according to the embodiment of the present invention, after obtaining the user tags of N users with a large probability of packet loss, the user tags can be used as data bases for analyzing various network faults, and for different network faults, the network fault analysis may be performed in different manners, and may be specifically determined according to analysis requirements.

Specifically, after the process of performing network fault analysis on the user tags of N users with approximate packet loss starts, the process may respectively perform the determination on whether the time tags are included, whether the time tags belong to busy hours or special periods, whether the user/terminal types are included, whether the cell/grid is included, whether the uplink or downlink fields are included, whether the device is included, whether the device manufacturer is included, and the like, and respectively perform different processing according to different determination results. For example, when the cell/grid is judged to be included, the radio problem corresponding to the cell/grid is proposed to be checked; when judging that the unidirectional link contains the uplink or downlink field, suggesting to check whether the unidirectional link has a fault; when the device manufacturer is judged to be included, the software version of the device of the manufacturer is suggested to be checked, whether bug occurs is determined, and upgrading is carried out according to the situation; when judging that the equipment comprises a plurality of pieces of equipment, suggesting to check whether a link between the equipment has a fault; when it is judged that only one device is included, it is recommended to investigate whether or not there is a failure or the like in the device itself.

According to the method for detecting the Volte network fault, the user tag of the user with the packet loss at a high probability can be determined by associating the user tag with the packet loss information, so that the fault reason of the link with the Volte network fault can be defined.

Another aspect of the specific embodiment of the present invention further provides a device for detecting a long term evolution voice over lte network failure, as shown in fig. 7, where the device includes:

According to the Volte network fault detection device, the user tags of the users with the packet loss at a high probability can be determined by associating the user tags with the packet loss information, so that an accurate basis is provided for defining the fault reasons of the links with the VoLTE network fault.

Preferably, as shown in fig. 7, the voltage network fault detection apparatus further includes:

and the third analysis module is used for carrying out network fault analysis according to the user label of the user with the determined approximate probability packet loss.

In addition, the apparatus further comprises:

the data extraction module is used for extracting specific users of which the total packet number of the data transmitted in each monitoring network point is greater than or equal to a first numerical value and the packet loss rate is greater than or equal to a second numerical value according to the packet loss statistical data;

and the first analysis module counts the corresponding relation between the user label corresponding to the specific user and the packet loss information according to the packet loss statistical data.

By adopting the above module, in order to avoid randomness, a total packet number threshold (a first numerical value) and a packet loss rate threshold (a second numerical value) of data transmitted by a user of each monitoring node are preset, and only statistical data, which is obtained from each monitoring node and has the total packet number of the transmitted data being greater than or equal to the first numerical value (for example, m) and the packet loss rate being greater than or equal to the second numerical value (for example, n%), is used as a data basis for calculating a user tag of a user with a large probability of packet loss.

Specifically, the first analysis module includes:

the statistical unit is used for counting the corresponding relations between the user labels with different dimensionalities of the users with the packet loss and the packet loss information;

and the Hash tree establishing unit is used for establishing the Hash tree according to the corresponding relation.

Additionally, the second analysis module includes:

a deleting unit, configured to delete a user tag in the Hash tree, where corresponding packet loss information does not meet a first preset condition, to obtain a pruned Hash tree;

and the analysis unit is used for determining the user label of the user with the probable packet loss, of which the corresponding packet loss information meets the second preset condition, in the pruned Hash tree.

Specifically, the packet loss information includes a total packet number of the counted transmission data and a packet loss rate, where the first preset condition is that the total packet number is smaller than a third numerical value, and the packet loss rate is smaller than a fourth numerical value.

Preferably, the analysis unit includes:

a sorting subunit, configured to sort, according to a sequence that packet loss numbers recorded in the packet loss information in the pruned Hash tree are from high to low, corresponding user tags;

the extracting subunit is used for extracting the first N user tags in the sorted user tags;

the analysis subunit is used for analyzing whether at least two user tag groups which have an inclusion relationship and the difference between the corresponding packet loss numbers is smaller than a preset value exist in the previous N user tags;

a deleting subunit, configured to, if the user tag group exists, reserve one of the user tags in the user tag group, and delete another tag in the user tag group;

and the supplementing subunit is used for supplementing the deleted user tags in the first N user tags according to the sorted user tags, and determining that the N user tags after the user tags are supplemented are the user tags of the users with the probability of packet loss.

In another aspect, an embodiment of the present invention further provides a server, including a memory, a processor, and a computer program stored in the memory and executable on the processor; wherein the processor implements the method for detecting a fault in a Volte network as described in any of the above when executing the program.

The invention also provides a computer-readable storage medium, on which a computer program is stored, wherein the program, when executed by a processor, implements a Volte network failure detection method as described in any one of the above.

Based on the above detailed description, those skilled in the art should understand specific embodiments of the server and the computer-readable storage medium using the voltage network failure detection method according to the embodiments of the present invention, and therefore, detailed descriptions thereof are omitted. The method and the device for detecting the Volte network fault, which are provided by the embodiment of the invention, can more accurately position the reason of the Volte network fault by using the algorithm of big data correlation analysis, can greatly reduce the problem of misjudgment of the network fault caused by random factors such as environment, user behaviors and the like, and can provide correct troubleshooting suggestions according to the positioning result of the network fault.

While the preferred embodiments of the present invention have been described, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined in the following claims.

Claims

1. a long-term evolution voice Volte network fault detection method, is characterized in that, comprises:

Obtain the packet loss statistics of each monitoring network;

According to the packet loss statistical data, collect statistics on the correspondence between user labels corresponding to users with packet loss and packet loss information;

According to the corresponding relationship, determine the user label of the user with a high probability of packet loss; the user label includes at least one of the following dimensions: transmission direction, session type, network element, time, location attribute, equipment manufacturer, user, terminal model, network type ;

The step of collecting statistics on the correspondence between user labels corresponding to users with packet loss and packet loss information according to the packet loss statistical data includes:

Count the correspondence between user labels of different dimensions and packet loss information of users who have lost packets;

According to the corresponding relationship, establish a hash Hash tree;

The step of determining the user label of the user with a high probability of packet loss according to the corresponding relationship includes:

Delete the user label whose corresponding packet loss information does not meet the first preset condition in the Hash tree, and obtain the pruned Hash tree;

Determine that in the pruned Hash tree, the corresponding user label whose packet loss information meets the second preset condition is the user label of the user with a high probability of packet loss;

The method also includes:

Perform network fault analysis according to the user labels of users with high probability of packet loss.

2. The Volte network fault detection method according to claim 1, wherein after the step of acquiring the packet loss statistics of each monitoring network point, the method further comprises:

According to the packet loss statistical data, extract specific users whose total number of packets of data transmitted in each of the monitoring network points is greater than or equal to a first value, and whose packet loss rate is greater than or equal to a second value;

Wherein, according to the packet loss statistical data, in the step of counting the correspondence between the user label corresponding to the packet loss user and the packet loss information:

According to the packet loss statistical data, the correspondence between the user label corresponding to the specific user and the packet loss information is counted.

3. Volte network fault detection method according to claim 1, is characterized in that, described deleting the user label that the corresponding packet loss information in described Hash tree does not meet the first preset condition, obtains Hash tree after pruning In the step of , the packet loss information includes the total number of packets and the packet loss rate of the statistics transmitted data, wherein the first preset condition is that the total number of packets is less than the third value, and the packet loss rate is less than the fourth value numerical value.

4. Volte network fault detection method according to claim 1, is characterized in that, in described Hash tree after described pruning, the corresponding packet loss information conforms to the user label of the second preset value, which is large. The steps of probabilistically dropping the user tag of the user include:

According to the order of the number of lost packets recorded in the packet loss information in the Hash tree after pruning, the corresponding user tags are sorted;

Extract the top N user tags in the sorted user tags;

Analyze whether there are at least two user tag groups with an inclusive relationship in the first N user tags, and the difference between the corresponding number of lost packets is less than the preset value;

If there is the user tag group, keep one of the user tags in the user tag group, and delete the other tags in the user tag group;

According to the sorted user tags, the deleted user tags in the first N user tags are supplemented, and it is determined that the N user tags after the supplemented user tags are user tags of users with a high probability of packet loss.

5. A long-term evolution voice Volte network fault detection device is characterized in that, comprising:

The acquisition module is used to acquire the packet loss statistics of each monitoring network;

A first analysis module, configured to count the correspondence between user labels corresponding to users with packet loss and packet loss information according to the packet loss statistics; the user labels include at least one of the following dimensions: transmission direction, session Type, network element, time, location attributes, equipment manufacturer, user, terminal model, network type;

a second analysis module, configured to determine the user label of the user with a high probability of packet loss according to the corresponding relationship;

The third analysis module is used to perform network fault analysis according to the determined user labels of users with high probability of packet loss;

The first analysis module includes:

A statistical unit, used to count the correspondence between the user labels of different dimensions of the users with packet loss and the packet loss information;

A Hash tree establishment unit is used to establish a hash Hash tree according to the corresponding relationship;

The second analysis module includes:

a deletion unit, configured to delete the user label whose corresponding packet loss information does not meet the first preset condition in the Hash tree, and obtain the pruned Hash tree;

The analyzing unit is configured to determine, in the pruned Hash tree, the corresponding user labels whose packet loss information meets the second preset condition are the user labels of users with a high probability of packet loss.

6. A server, characterized in that it comprises a memory, a processor and a computer program stored on the memory and running on the processor; when the processor executes the program, the program according to claim 1 to 4 Any one of the Volte network fault detection methods.

7. A computer-readable storage medium on which a computer program is stored, characterized in that, when the program is executed by a processor, the Volte network fault detection method according to any one of claims 1 to 4 is implemented.