CN116132330A

CN116132330A - Network detection method, network detection model training method and device

Info

Publication number: CN116132330A
Application number: CN202210817220.8A
Authority: CN
Inventors: 骆林; 唐蠡; 郭剑霓; 吴海英; 曾琳铖曦; 蒋宁
Original assignee: Mashang Xiaofei Finance Co Ltd
Current assignee: Mashang Xiaofei Finance Co Ltd
Priority date: 2022-07-12
Filing date: 2022-07-12
Publication date: 2023-05-16

Abstract

The embodiment of the disclosure discloses a network detection method, a network detection model training method and a device. The network detection method comprises the following steps: acquiring link information of a network link included in a network topology structure, wherein the link information comprises index information of equipment on the network link; inputting the link information into a pre-trained network detection model to detect the network condition, and obtaining a network condition detection result; the network condition detection result comprises at least one of the following: and detecting whether the network link is abnormal or not and detecting whether the equipment on the network link is abnormal or not. The technical scheme can improve the efficiency and accuracy of network detection.

Description

Network detection method, network detection model training method and device

Technical Field

The disclosure relates to the technical field of artificial intelligence, and in particular relates to a network detection method, a network detection model training method and a device.

Background

In the related art, network quality is usually detected by monitoring network indexes of an application or a host, such as APM (Application Performance Management, application performance monitoring tool) monitoring service, and transaction paths and processing times of each transaction of a service system are analyzed by aggregating real-time data of each processing link of the service system, so that full-link performance monitoring of the application is realized. However, such monitoring methods often only involve network metrics of the host, for example, monitoring traffic, bandwidth, and the number of data packets of the host, that is, only collecting, sorting, and visually displaying the network metrics. When a network fails, network anomalies between application level and service can be located only, and the failure cause and the failed network link are difficult to accurately locate. If the fault cause and the network link with the fault are to be located, other monitoring devices or manpower are combined, for example, a monitoring device is configured for each network link to monitor the network index of each network link; or, when network transmission and reception are problematic, the failed network link is manually checked. Obviously, the two fault locating modes are not only low in efficiency, but also low in accuracy.

Disclosure of Invention

The embodiment of the disclosure aims to provide a network detection method, a network detection model training method and a device, which improve the efficiency and accuracy of network detection.

In order to solve the above technical problems, embodiments of the present disclosure are implemented as follows:

in one aspect, an embodiment of the present disclosure provides a network detection method, including:

acquiring link information of a network link included in a network topology structure, wherein the link information comprises index information of equipment on the network link;

inputting the link information into a pre-trained network detection model to detect network conditions, and obtaining a network condition detection result; wherein the network condition detection result includes at least one of: and detecting whether the network link is abnormal or not and detecting whether the equipment on the network link is abnormal or not.

By adopting the technical scheme of the embodiment of the disclosure, the network condition detection is carried out on the network link by inputting the link information of the network link included in the network topology structure into a pre-trained network detection model and utilizing the network detection model to obtain the network condition detection result, wherein the network condition detection result comprises the detection result of whether the network link is abnormal or not and/or the detection result of whether equipment on the network link is abnormal or not. Therefore, the technical scheme provides a link monitoring and equipment monitoring mechanism below an application layer, and can accurately position an abnormal network link and/or equipment with an abnormal network link when the network is abnormal, so that the possible potential abnormal situation on the whole network link is accurately detected, and powerful data support is provided for network abnormality investigation and network maintenance. Further, compared with the scheme that the abnormal network needs to be detected by relying on manpower in the prior art, the technical scheme can automatically locate abnormal network links and/or abnormal equipment on the network links, does not need user participation, so that inaccurate locating caused by manpower factors is avoided, and the accuracy of network abnormal locating is improved. And moreover, the network anomaly positioning is realized by training the network detection model in advance and the network anomaly positioning is realized by the network detection model, so that the efficiency of network anomaly positioning is greatly improved, and the network anomaly positioning time is saved.

In another aspect, an embodiment of the present disclosure provides a method for training a network detection model, including:

acquiring sample link information of sample network links in a plurality of network topologies and label information of the sample network links; wherein the sample network link comprises: sample network links with normal network conditions and sample network links with abnormal network conditions; the sample link information includes index information of devices on the sample network link; the tag information is used to characterize at least one of: whether the sample network link is abnormal, whether a device on the sample network link is abnormal;

inputting the sample network link information into a network detection model to be trained to obtain a classification result corresponding to the sample network link, wherein the classification result comprises at least one of the following: whether the sample network link is abnormal, whether a device on the sample network link is abnormal;

and according to the classification result of the sample network link and the label information, adjusting the model parameters of the network detection model to be trained.

By adopting the technical scheme of the embodiment of the disclosure, sample link information (including sample network links with normal network conditions and sample network links with abnormal network conditions) of sample network links included in a plurality of network topological structures is input into a network detection model to be trained, classification results (including whether the sample network links are abnormal and/or whether equipment on the sample network links is abnormal) corresponding to the sample network links are obtained, and then model parameters of the network detection model to be trained are adjusted according to the classification results and label information of the sample network links. Therefore, the network detection model is trained in advance, so that the network detection model can be used for positioning network anomalies. Therefore, when the network anomaly location is realized by using the network detection model, the network link with the anomaly and/or equipment with the anomaly on the network link can be accurately located, so that the possible potential anomaly condition on the whole network link can be accurately detected, and powerful data support is provided for network anomaly investigation and network maintenance. Further, compared with the scheme that the abnormal network needs to be detected by relying on manpower in the prior art, the technical scheme can automatically locate abnormal network links and/or abnormal equipment on the network links, does not need user participation, so that inaccurate locating caused by manpower factors is avoided, and the accuracy of network abnormal locating is improved. In addition, based on the intellectualization and automation of the network detection model, the efficiency of network anomaly positioning is greatly improved, and the network anomaly positioning time is saved.

In still another aspect, an embodiment of the present disclosure provides a network detection apparatus, including:

the first acquisition module is used for acquiring link information corresponding to a network link included in a network topology structure, wherein the link information comprises index information of equipment on the network link;

the detection module is used for inputting the link information into a pre-trained network detection model to detect the network condition, so as to obtain a network condition detection result; wherein the network condition detection result includes at least one of: and detecting whether the network link is abnormal or not and detecting whether the equipment on the network link is abnormal or not.

In still another aspect, an embodiment of the present disclosure provides a network detection model training apparatus, including:

the second acquisition module is used for acquiring sample link information of sample network links included in the network topologies and label information of the sample network links; wherein the sample network link comprises: sample network links with normal network conditions and sample network links with abnormal network conditions; the sample link information includes index information of devices on the sample network link; the tag information is used to characterize at least one of: whether the sample network link is abnormal, whether a device on the sample network link is abnormal;

The classification module is used for inputting the sample link information into a network detection model to be trained to obtain a classification result corresponding to the sample network link, and the classification result comprises at least one of the following: whether the sample network link is abnormal, whether a device on the sample network link is abnormal;

and the model training module is used for adjusting the model parameters of the network detection model according to the classification result of the sample network link and the label information.

In yet another aspect, an embodiment of the present disclosure provides an electronic device, including a processor and a memory electrically connected to the processor, where the memory stores a computer program, and the processor is configured to call and execute the computer program from the memory to implement the network detection method described above, or the processor is configured to call and execute the computer program from the memory to implement the network detection model training method described above.

In yet another aspect, embodiments of the present disclosure provide a storage medium storing a computer program executable by a processor to implement the network detection method described above, or executable by a processor to implement the network detection model training method described above.

Drawings

In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present disclosure, and other drawings may be obtained according to these drawings without inventive effort to a person of ordinary skill in the art.

FIG. 1 is a schematic flow chart diagram of a network detection method according to an embodiment of the disclosure;

FIG. 2 is a schematic diagram of a network detection method according to an embodiment of the present disclosure;

FIG. 3 is a schematic block diagram of a network topology according to an embodiment of the present disclosure;

FIG. 4 is a schematic diagram of a network detection model according to an embodiment of the present disclosure;

FIG. 5 is a schematic output interface diagram of a network condition detection result according to an embodiment of the present disclosure;

FIG. 6 is a schematic flow chart diagram of a network detection model training method according to an embodiment of the present disclosure;

FIG. 7 is a schematic diagram of a network detection model training process according to an embodiment of the present disclosure;

FIG. 8 is a schematic block diagram of a network detection device according to an embodiment of the present disclosure;

FIG. 9 is a schematic block diagram of a network detection model training apparatus according to an embodiment of the present disclosure;

fig. 10 is a schematic block diagram of an electronic device in accordance with an embodiment of the present disclosure.

Detailed Description

The embodiment of the disclosure provides a network detection method, a network detection model training method and a device, which improve the efficiency and accuracy of network detection.

In order that those skilled in the art will better understand the technical solutions in the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present disclosure, and it is apparent that the described embodiments are only some embodiments of the present disclosure, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments in this disclosure without inventive faculty, shall fall within the scope of the present disclosure.

In the network condition detection aspect, in the prior art, network quality is usually detected by monitoring network indexes of an application or a host, and the monitoring mode often only relates to the network indexes of the host, for example, monitoring traffic, bandwidth and the number of data packets of the host, that is, only the network indexes can be collected, sorted and visually displayed. When a network fails, network anomalies between application level and service can be located only, and the failure cause and the failed network link are difficult to accurately locate. According to the network detection method, link information of the network links included in the network topology structure, including index information of equipment on the network links, is obtained, the network link information is input into the pre-trained network detection model, and the network detection model is utilized to detect network conditions of the network links, so that network condition detection results are obtained, including whether the network links are abnormal and/or whether equipment on the network links is abnormal, therefore, link monitoring and equipment monitoring mechanisms below an application layer are perfected, when the network is abnormal, the abnormal network links and/or equipment on the network links is accurately located, potential abnormal conditions on the whole network links are accurately detected, and powerful data support is provided for network abnormality investigation and network maintenance. In addition, if the network link is abnormal, even if the delay is slight, a certain time is required for spreading to the actual application service, the delay of the application service is still caused, and the client is affected to different degrees. Therefore, according to the embodiment of the disclosure, the network anomaly positioning is realized by pre-training the network detection model and by the network detection model, so that not only the network link with the anomaly but also the equipment with the anomaly can be positioned, and therefore, the problems and reasons can be timely checked under the condition that the network is abnormal, the network anomaly positioning efficiency is greatly improved, and the network anomaly positioning time is saved.

The network detection method provided by the embodiment of the present disclosure may be performed by a network detection device or performed by software installed in the network detection device, and in particular, the network detection device may be a terminal device or a server device.

Fig. 1 is a schematic flow chart of a network detection method according to an embodiment of the disclosure, as shown in fig. 1, the method includes:

s102, acquiring link information of a network link included in a network topology structure, wherein the link information comprises index information of equipment on the network link.

Alternatively, the link information may include link related information of the network link in addition to index information of the device on the network link. The link-related information may include the number of devices included on the network link, device information, device location information, link identification information (e.g., a link name), and a link flow direction, wherein the device information may include at least one of device identification information (e.g., a device name), a device MAC address (hardware address, physical address, or link address), a device type, and the like.

When acquiring the link related information, firstly acquiring a network topology structure, wherein the network topology structure comprises one or more network links. And secondly, analyzing the network topology structure to obtain the link related information corresponding to each network link.

When the index information of the device is acquired, the device itself optionally has the capability of acquiring the index information of the device itself, so that the device can report the respective acquired index information to the network detection device. Alternatively, an index collection probe may be installed on the device, where the index collection probe is connected to the network detection device, and the index collection probe has the capability of collecting index information of the device, so as to transmit the collected index information to the network detection device.

S104, inputting the network link information into a pre-trained network detection model to detect the network condition, and obtaining a network condition detection result; wherein the network condition detection result includes at least one of: the method comprises the following steps of detecting whether a network link is abnormal or not and detecting whether equipment on the network link is abnormal or not.

In this embodiment, the output mode of the network condition detection result by the network detection model is not limited. Optionally, the network detection model may output at least one of the following information: link identification information of an abnormal network link, identification information of an abnormal device, a link condition score of the abnormal network link, and a performance score of the abnormal device. Wherein the link condition score is used for representing the link condition of the network link, and the higher the score is, the better the link condition of the network link is. The performance score of the device is used to characterize the performance of the device, the higher the score, the better the performance of the device. The calculation of the link status score and the device performance score will be described in detail in the following embodiments, which will not be described herein.

Fig. 2 is a schematic diagram of a network detection method according to an embodiment of the present disclosure. As can be seen from fig. 2, after obtaining the link information of the network link, the network condition detection result can be obtained only by inputting the link information into a pre-trained network detection model, so that the possible potential abnormal situation on the whole network link can be detected in time. The following describes in detail a network detection method provided by the embodiments of the present disclosure, first introducing a network topology structure and corresponding network link information thereof.

Fig. 3 is a schematic block diagram of a network topology according to an embodiment of the present disclosure. As shown in fig. 3, the network topology includes a machine room a and a machine room B, where the machine room a includes devices router1, router2, router3, switch1, switch2 and gateway1, and the machine room B includes devices router4, router5, router6, switch3, switch4 and gateway2, where router1, router2, router3, router4, router5 and router6 are routers, and switches 1, switch2, switch3 and switch4 are gateways. By parsing the network topology, link-related information including device names, device location information, device MAC addresses, and device types of the respective devices can be obtained as shown in table 1 below.

TABLE 1

Device name	Device location information	Device MAC address	Device type
				router1	A machine room-3 cabinet 1 row	aa-bb-xx-xx-23-12	Router
router2	A machine room-3 cabinet 2 rows	aa-bb-xx-xx-23-11	Router
				router3	A machine room-3 cabinet 3 rows	aa-bb-xx-xx-23-13	Router
switch1	A machine room-2 cabinet 1 row	aa-bb-xx-xx-24-12	Switch board
				switch2	A machine room-1 cabinet 1 row	aa-bb-xx-xx-25-11	Switch board
gateway1	A machine room-1 cabinet 2 rows	aa-bb-xx-xx-26-11	Gateway (GW)
				router4	B machine room-2 cabinet 1 row	cc-bb-xx-xx-26-13	Router
router5	B machine room-2 cabinet 2 rows	cc-bb-xx-xx-26-12	Router
				router6	B machine room-2 cabinet 4 rows	cc-bb-xx-xx-26-15	Router
switch3	B machine room-1 cabinet 2 rows	cc-bb-xx-xx-27-13	Switch board
				switch4	B machine room-1 cabinet 1 row	cc-bb-xx-xx-27-12	Switch board
gateway2	B machine room-3 cabinet 1 row	cc-bb-xx-xx-36-12	Gateway (GW)

As can be seen from table 1, by analyzing the network topology result, the link related information from the gateway of the root node to the router of the leaf node can be known, so as to analyze a plurality of unidirectional network links, and table 2 below schematically lists the link related information of a part of the network links, including the link names, the number of devices in the network links, and the link flow direction.

TABLE 2

It should be noted that, fig. 3 only schematically illustrates a network topology, in practical application, the number of network links that can be included in the network topology may be one or more, and the number of devices included on each network link may be one or more.

In one embodiment, as shown in fig. 4, the network detection model includes a device score layer, a network link score layer, and a classification layer, based on which network link information is input into a pre-trained network detection model, and when network condition detection is performed on a network link, the following actions can be specifically performed through the respective layers of the network detection model.

And the device grading layer of the network detection model is used for calculating the performance scores of the devices on the network link according to the index information of the devices on the network link.

Wherein, when calculating the performance score of the device, a first weight corresponding to each piece of index information of the device may be determined first, and the index information of the device may include at least one piece of information of network jitter, network packet loss, network delay and network bandwidth. And secondly, determining the performance score of the equipment according to the index information of the equipment and the first weight corresponding to each index information.

Alternatively, the manner in which the performance score of the device is calculated may be expressed as the following equation (1):

wherein score (device) represents the performance score of the device, weight _k Representing a first weight, metric, corresponding to the kth index information _k Represents the kth index information (or device index value), and n represents the number of index information of the device.

Since different devices take different roles in the network link, their core metrics are different for different types of devices. For example, the core indexes corresponding to the router are network jitter and network packet loss, the core indexes corresponding to the switch are network bandwidth, network delay and network packet loss, and the core indexes corresponding to the gateway are network delay and network jitter. The first weights corresponding to the same index information may be the same or different for different types of devices, as shown in table 3 below.

TABLE 3 Table 3

It should be noted that the weight values listed in table 3 are merely exemplary, and in practical application, each first weight value may be defined and set in a targeted and flexible manner according to different device types.

And the network link grading layer of the network detection model is used for calculating the link condition score of the network link according to the performance score of the equipment on the network link.

When calculating the link condition score, the second weight corresponding to each device on the network link may be determined first, and then the link condition score of the network link may be determined according to the second weight corresponding to each device and the performance score of each device. The second weight corresponding to the device is used to represent the importance of the corresponding device on the network link.

Alternatively, the way the link condition score is calculated may be expressed as the following equation (2):

wherein Score (link) represents the link condition Score, weight (device) _i ) Representing a second weight corresponding to an ith device on the network link, score (device _i ) Representing the performance score of the ith device, m is the number of devices included on the network link. Alternatively, if the calculated link condition score according to formula (2) is not an integer, the calculated score may be rounded, such as rounded up or rounded down.

Optionally, the second weight corresponding to the device on the network link may be determined according to the device type corresponding to the device. If the device types are different, the second weights corresponding to the devices are also different, as shown in table 4.

TABLE 4 Table 4

Device type	Second weight
		Router	0.6
Switch board	0.7
		Gateway (GW)	0.9

The classification layer of the network detection model is used for determining whether the network link is abnormal according to the corresponding link condition score of the network link; and/or determining whether the device on the network link is abnormal based on the performance score of the device on the network link.

And if the link condition score of the network link is smaller than a first preset threshold value, determining that the network link is abnormal. If the link condition score of the network link is greater than or equal to a first preset threshold value, determining that the network link is normal.

Along the network links shown in table 2 above, the link condition scores of the respective network links are shown in table 5 below. Assuming that the first preset threshold is 8, the network Link8 shown in table 5 is a network Link with abnormal network conditions.

TABLE 5

/>

In one embodiment, after determining that an abnormal network link (hereinafter referred to simply as an abnormal network link) occurs, the classification layer of the network detection model may further determine whether the device on the abnormal network link is abnormal according to a performance score of each device on the abnormal network link. Specifically, if the performance score of the device on the abnormal network link is less than a second preset threshold, determining that the device is abnormal. And if the performance score of the equipment on the abnormal network link is greater than or equal to a second preset threshold value, determining that the equipment is normal.

Therefore, when the network is abnormal, the technical scheme provided by the embodiment not only can accurately locate the abnormal network link, but also can locate the abnormal equipment on the abnormal network link, so that the problem and the reason can be timely checked under the condition that the network is abnormal, the network abnormality locating efficiency is greatly improved, and the network abnormality locating time is saved.

In one embodiment, the link information further includes location information of devices on the network link. After the equipment is determined to be abnormal, the position information of the abnormal equipment can be determined, so that the position information of the abnormal equipment is provided to the front end, and a front end maintainer can accurately position the abnormal equipment according to the position information of the abnormal equipment, so that the abnormal equipment is maintained, and the problem of network abnormality is rapidly solved.

In one embodiment, if the abnormal network link includes a plurality of abnormal network links, the classification layer of the network detection model may further sort the link information of the plurality of abnormal network links according to the link status score of each abnormal network link, and output the link information of the plurality of abnormal network links according to the sorting result.

In one embodiment, when an abnormal network link is detected, early warning information can be sent out, and the early warning information is used for identifying that the abnormal network link exists in the network topology structure. The early warning information may include at least one of link identification information (e.g., a link name) of the abnormal network link, identification information (e.g., a device name) of the abnormal device on the abnormal network link, location information of the abnormal device, and the like.

In this embodiment, the network condition detection result may be output on the display screen, and early warning information may be sent out, where the early warning information may be output in text and/or voice form, for example, output in the display window "abnormality occurs in network Link 8-! ", while ringing.

Fig. 5 is a schematic output interface diagram of a network condition detection result according to an embodiment of the present disclosure, and as shown in fig. 5, in a detection result display window, the display content may include: abnormal network links, link flows, abnormal devices, and abnormal device locations (i.e., abnormal device location information). The early warning window is used for displaying early warning information. The detection result display window and the early warning window can be displayed on the same interface, or the early warning window is displayed above the detection result in a popup window mode.

As can be seen from fig. 5, with the network detection method provided by the embodiment, front-end maintenance personnel can intuitively check link related information of an abnormal network link, so that abnormal equipment can be accurately positioned according to the position of the abnormal equipment, and further, the abnormal equipment can be maintained, and the problem of network abnormality can be rapidly solved.

FIG. 6 is a schematic flow chart diagram of a network detection model training method, as shown in FIG. 6, according to one embodiment of the disclosure, the method comprising:

s602, acquiring sample link information of sample network links included in a plurality of network topologies and label information of the sample network links; wherein the sample network link comprises: sample network links with normal network conditions and sample network links with abnormal network conditions; the sample link information includes index information of devices on the sample network link; the tag information is used to characterize at least one of: whether the sample network link is abnormal or not, and whether the device on the sample network link is abnormal or not.

Alternatively, the sample link information may include link related information of the sample network link in addition to the index information of the device. The link-related information may include the number of devices included in the sample network link, device information, device location information, link identification information (e.g., a link name), and a link flow direction, wherein the device information may include at least one of device identification information (e.g., a device name), a device MAC address, a device type, and the like.

When acquiring the link related information, firstly acquiring a network topology structure, wherein the network topology structure comprises one or more sample network links. And secondly, analyzing the network topology structure to obtain the link related information corresponding to the sample network link.

When the index information of the device is acquired, the device itself optionally has the capability of acquiring the index information of the device itself, so that each device can report the index information acquired by each device. Alternatively, an index collection probe may be installed on the device, where the index collection probe has the ability to collect index information of the device, so as to report the collected index information.

S604, inputting the sample link information into a network detection model to be trained to obtain a classification result corresponding to the sample network link, wherein the classification result comprises at least one of the following: whether the sample network link is abnormal or not, and whether the device on the sample network link is abnormal or not.

How to classify the sample network links by using the network detection model to be trained will be described in detail in the following embodiments, which are not described herein.

S606, according to the classification result and the label information of the sample network link, the model parameters of the network detection model to be trained are adjusted.

In this embodiment, the iterative training of the network detection model to be trained is achieved by adjusting the model parameters of the network detection model to be trained for multiple times, so as to obtain the trained network detection model.

By adopting the technical scheme of the embodiment of the disclosure, sample link information (including sample network links with normal network conditions and sample network links with abnormal network conditions) of sample network links included in a plurality of network topological structures is input into a network detection model to be trained, classification results (including whether the sample network links are abnormal and/or whether equipment on the sample network links is abnormal) corresponding to the sample network links are obtained, and then model parameters of the network detection model are adjusted according to the classification results and label information of the sample network links. Therefore, the network detection model is trained in advance, so that the network detection model can be used for positioning network anomalies. Therefore, when the network anomaly location is realized by using the network detection model, the network link with the anomaly and/or equipment with the anomaly on the network link can be accurately located, so that the possible potential anomaly condition on the whole network link can be accurately detected, and powerful data support is provided for network anomaly investigation and network maintenance. Further, compared with the scheme that the abnormal network needs to be detected by relying on manpower in the prior art, the technical scheme can automatically locate abnormal network links and/or abnormal equipment on the network links, does not need user participation, so that inaccurate locating caused by manpower factors is avoided, and the accuracy of network abnormal locating is improved. In addition, based on the intellectualization and automation of the network detection model, the efficiency of network anomaly positioning is greatly improved, and the network anomaly positioning time is saved.

In one embodiment, the model parameters of the network detection model include at least one of:

(1) The first weight corresponding to each index information of the device may include at least one of information of network jitter, network packet loss, network delay and network bandwidth.

(2) And the second weight corresponding to each device on the network link is used for indicating the importance degree of the corresponding device on the network link.

(3) A first preset threshold associated with the link condition score corresponding to the abnormal network link.

(4) And a second preset threshold value corresponding to the abnormal equipment and related to the performance score of the equipment.

In one embodiment, the network detection model to be trained comprises: a device scoring layer, a network link scoring layer, and a classification layer. Based on the above, when the sample link information is input into the network detection model to be trained to obtain the classification result corresponding to the sample network link, the following actions can be specifically executed through each layer of the network detection model to be trained.

The device scoring layer of the network detection model to be trained is used for calculating the performance score of the device on the sample network link according to the index information of the device on the sample network link.

When calculating the performance score of the device, a first weight corresponding to the index information of the device may be first determined, where the index information of the device may include at least one of network jitter, network packet loss, network delay, and network bandwidth. And secondly, determining the performance score of the equipment according to the index information of the equipment and the first weight corresponding to each index information. The calculation manner of the performance score of the device may be expressed as the above formula (1), and since the formula (1) is described in detail in the above embodiment, a detailed description thereof is omitted here.

The network link scoring layer of the network detection model to be trained is used for calculating the link condition score of the sample network link according to the performance score of the equipment on the sample network link.

When calculating the link condition score, the second weight corresponding to each device on the sample network link may be determined first, and then the link condition score of the sample network link may be determined according to the second weight corresponding to each device and the performance score of each device. The calculation manner of the link condition score may be expressed as the above formula (2), and since the formula (2) is described in detail in the above embodiment, a detailed description thereof is omitted here.

The classification layer of the network detection model to be trained is used for determining whether the sample network link is abnormal or not according to the link condition score of the sample network link; and/or determining whether the device on the sample network link is abnormal based on the performance score of the device on the sample network link.

The specific process of determining whether the sample network link is abnormal and/or whether the device is abnormal by using the classification layer of the network detection model to be trained is similar to the process of determining whether the network link to be detected is abnormal and/or whether the device is abnormal by using the network detection model in the above embodiment, and is not repeated here.

In one embodiment, if the classification result of the sample network link does not meet the preset classification condition, adjusting the model parameters of the network detection model to be trained according to the classification result, and inputting the sample link information into the network detection model after the adjustment parameters again to classify the sample network link; and if the classification result meets the preset classification condition, stopping iteration to obtain the trained network detection model.

Wherein, the preset classification condition can comprise at least one of the following: the accuracy of the classification result is larger than or equal to a preset accuracy threshold, and the iteration times reach a preset times threshold. The preset accuracy threshold may be set according to accuracy requirements for the network detection model. The higher the preset accuracy threshold, the higher the accuracy of the network detection model.

Fig. 7 is a schematic diagram of a network detection model training method according to an embodiment of the present disclosure. In fig. 7, sample link information is input into a network detection model to be trained, and each layer of the network detection model to be trained is utilized to process the sample link information, so as to obtain a classification result of the sample network link. If the classification result meets the preset classification condition, stopping iteration to obtain a trained network detection model; and if the classification result does not meet the preset classification condition, adjusting the model parameters of the network detection model to be trained, classifying the sample network link again based on the network detection model after the model parameters are adjusted, and finally obtaining the trained network detection model.

In summary, particular embodiments of the present subject matter have been described. Other embodiments are within the scope of the following claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may be advantageous.

The above is a network detection method and a network detection model training method provided by the embodiments of the present disclosure. Based on the same thought, the embodiment of the disclosure also provides a network detection device and a network detection model training device.

Fig. 8 is a schematic block diagram of a network detection device according to an embodiment of the present disclosure, as shown in fig. 8, the device includes:

a first obtaining module 81, configured to obtain link information of a network link included in a network topology, where the link information includes index information of a device on the network link;

the detection module 82 is configured to input the link information into a pre-trained network detection model to perform network condition detection, so as to obtain a network condition detection result; wherein the network condition detection result includes at least one of: and detecting whether the network link is abnormal or not and detecting whether the equipment on the network link is abnormal or not.

In one embodiment, the network detection model includes: a device scoring layer, a network link scoring layer and a classification layer;

the device grading layer is used for calculating the performance score of the device on the network link according to the index information of the device on the network link;

The network link scoring layer is used for calculating the link condition score of the network link according to the performance score of the equipment on the network link;

the classification layer is used for determining whether the network link is abnormal according to the link condition score of the network link; and/or determining whether the device on the network link is abnormal according to the performance score of the device on the network link.

In one embodiment, the detection module 82 includes:

a first determining unit, configured to determine a first weight corresponding to each piece of index information of the device, where the index information includes at least one piece of information of network jitter, network packet loss, network delay, and network bandwidth;

and the second determining unit is used for determining the performance score of the equipment according to the index information of the equipment and the first weight corresponding to each index information.

In one embodiment, the detection module 82 includes:

a third determining unit, configured to determine a second weight corresponding to each device on the network link, where the second weight is used to represent an importance level of the corresponding device on the network link;

and a fourth determining unit, configured to determine a link status score of the network link according to the second weight corresponding to each device and the performance score of each device.

In an embodiment, the third determining unit is further configured to:

and determining a second weight corresponding to the equipment according to the equipment type corresponding to the equipment.

In one embodiment, if the abnormal network link includes a plurality of abnormal network links, the classification layer is further configured to sort the link information of the plurality of abnormal network links according to the link status score of each abnormal network link, and output the link information of the plurality of abnormal network links according to the sorting result.

It should be understood by those skilled in the art that the above network detection device can be used to implement the network detection method described above, and the detailed description thereof should be similar to that of the method described above, so as to avoid complexity, and is not repeated herein.

By adopting the device of the embodiment of the disclosure, the link information of the network link included in the network topology structure is acquired and is input into the pre-trained network detection model, and the network detection model is utilized to detect the network condition of the network link, so as to obtain the network condition detection result, including the detection result of whether the network link is abnormal or not and/or the detection result of whether the equipment on the network link is abnormal or not. Therefore, the device provides a link monitoring and equipment monitoring mechanism below an application layer, and can accurately position an abnormal network link and/or equipment with an abnormal network link when the network is abnormal, so that the potential abnormal situation on the whole network link can be accurately detected, and powerful data support is provided for network abnormality investigation and network maintenance. Further, compared with the scheme that the abnormal network needs to be detected by relying on manpower in the prior art, the technical scheme can automatically locate abnormal network links and/or abnormal equipment on the network links, does not need user participation, so that inaccurate locating caused by manpower factors is avoided, and the accuracy of network abnormal locating is improved. And moreover, the network anomaly positioning is realized by training the network detection model in advance and the network anomaly positioning is realized by the network detection model, so that the efficiency of network anomaly positioning is greatly improved, and the network anomaly positioning time is saved.

Fig. 9 is a schematic block diagram of a network detection device according to another embodiment of the present disclosure, as shown in fig. 9, the device includes:

a second obtaining module 91, configured to obtain sample link information of sample network links in a plurality of network topologies and tag information of the sample network links; wherein the sample network link comprises: sample network links with normal network conditions and sample network links with abnormal network conditions; the sample link information includes index information of devices on the sample network link; the tag information is used to characterize at least one of: whether the sample network link is abnormal, whether a device on the sample network link is abnormal;

the classification module 92 is configured to input the sample link information into a network detection model to be trained, and obtain a classification result corresponding to the sample network link, where the classification result includes at least one of the following: whether the sample network link is abnormal, whether a device on the sample network link is abnormal;

and the model training module 93 is configured to adjust model parameters of the network detection model according to the classification result of the sample network link and the tag information.

In one embodiment, the network detection model to be trained comprises: a device scoring layer, a network link scoring layer and a classification layer;

the device scoring layer is used for calculating the performance score of the device on the sample network link according to the index information of the device on the sample network link;

the network link scoring layer is used for calculating the link condition score of the sample network link according to the performance score of the equipment on the sample network link;

the classification layer is used for determining whether the sample network link is abnormal or not according to the link condition score of the sample network link; and/or determining whether the device on the sample network link is abnormal based on the performance score of the device on the sample network link.

By adopting the device of the embodiment of the disclosure, the sample link information (including the sample network links with normal network conditions and the sample network links with abnormal network conditions) of the sample network links included in the multiple network topological structures is acquired and input into the network detection model to be trained, so as to obtain the classification result (including whether the sample network links are abnormal and/or whether the equipment on the sample network links is abnormal) corresponding to the sample network links, and further, the model parameters of the network detection model are adjusted according to the classification result and the label information of the sample network links. It can be seen that the apparatus enables the network detection model to be used for network anomaly localization by pre-training the network detection model. Therefore, when the network anomaly location is realized by using the network detection model, the network link with the anomaly and/or equipment with the anomaly on the network link can be accurately located, so that the possible potential anomaly condition on the whole network link can be accurately detected, and powerful data support is provided for network anomaly investigation and network maintenance. Further, compared with the scheme that the abnormal network needs to be detected by relying on manpower in the prior art, the technical scheme can automatically locate abnormal network links and/or abnormal equipment on the network links, does not need user participation, so that inaccurate locating caused by manpower factors is avoided, and the accuracy of network abnormal locating is improved. In addition, based on the intellectualization and automation of the network detection model, the efficiency of network anomaly positioning is greatly improved, and the network anomaly positioning time is saved.

It should be understood by those skilled in the art that the above-mentioned network detection model training device can be used to implement the above-mentioned network detection model training method, and the detailed description thereof should be similar to that of the above-mentioned method section, so as to avoid complexity, and is not repeated here.

Based on the same thought, the embodiment of the disclosure further provides an electronic device, as shown in fig. 10. The electronic device may vary considerably in configuration or performance and may include one or more processors 1001 and memory 1002, where the memory 1002 may store one or more stored applications or data. Wherein the memory 1002 may be transient storage or persistent storage. The application programs stored in the memory 1002 may include one or more modules (not shown), each of which may include a series of computer-executable instructions for use in an electronic device. Still further, the processor 1001 may be configured to communicate with the memory 1002 and execute a series of computer executable instructions in the memory 1002 on an electronic device. The electronic device may also include one or more power supplies 1003, one or more wired or wireless network interfaces 1004, one or more input/output interfaces 1005, and one or more keyboards 1006.

In particular, in this embodiment, an electronic device includes a memory, and one or more programs, where the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the electronic device, and the one or more programs configured to be executed by one or more processors include instructions for:

By adopting the technical scheme of the embodiment of the disclosure, the network condition of the network link is detected by utilizing the network detection model by inputting the link information of the network link included in the network topology structure into the pre-trained network detection model, so as to obtain the network condition detection result, including the detection result of whether the network link is abnormal or not and/or the detection result of whether equipment on the network link is abnormal or not. Therefore, the technical scheme provides a link monitoring and equipment monitoring mechanism below an application layer, and can accurately position an abnormal network link and/or equipment with an abnormal network link when the network is abnormal, so that the possible potential abnormal situation on the whole network link is accurately detected, and powerful data support is provided for network abnormality investigation and network maintenance. Further, compared with the scheme that the abnormal network needs to be detected by relying on manpower in the prior art, the technical scheme can automatically locate abnormal network links and/or abnormal equipment on the network links, does not need user participation, so that inaccurate locating caused by manpower factors is avoided, and the accuracy of network abnormal locating is improved. And moreover, the network anomaly positioning is realized by training the network detection model in advance and the network anomaly positioning is realized by the network detection model, so that the efficiency of network anomaly positioning is greatly improved, and the network anomaly positioning time is saved.

In particular, in another embodiment, an electronic device includes a memory, and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the electronic device, and the one or more programs configured to be executed by one or more processors comprise instructions for:

The presently disclosed embodiments also provide a storage medium storing one or more computer programs comprising instructions that, when executed by an electronic device comprising a plurality of application programs, enable the electronic device to perform the various processes of the network detection method embodiments described above, and are specifically configured to perform:

The disclosed embodiments also provide a storage medium storing one or more computer programs, the one or more computer programs comprising instructions, which when executed by an electronic device comprising a plurality of application programs, enable the electronic device to perform the various processes of the network detection model training method embodiments described above, and in particular to perform:

acquiring sample link information of sample network links in a plurality of network topologies and label information of the sample network links; wherein the sample network link comprises: sample network links with normal network conditions and sample network links with abnormal network conditions; the sample link information includes index information of devices on the sample network link; the tag information is used to characterize at least one of: whether the network condition of the sample network link is abnormal and whether the equipment on the sample network link is abnormal;

And adjusting model parameters of the network detection model according to the classification result of the sample network link and the label information.

The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For convenience of description, the above devices are described as being functionally divided into various units, respectively. Of course, the functions of the various elements may be implemented in the same one or more software and/or hardware when implementing the present disclosure.

It will be apparent to those skilled in the art that embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer 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 Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.

The disclosure may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The disclosure may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.

The foregoing is merely exemplary of the present disclosure and is not intended to limit the present disclosure. Various modifications and variations of this disclosure will be apparent to those skilled in the art. Any modifications, equivalent substitutions, improvements, or the like, which are within the spirit and principles of the present disclosure, are intended to be included within the scope of the claims of the present disclosure.

Claims

1. A network detection method, comprising:

2. The method of claim 1, wherein the network detection model comprises: a device scoring layer, a network link scoring layer and a classification layer;

3. The method of claim 2, wherein calculating the performance score of the device on the network link based on the metric information of the device on the network link comprises:

determining a first weight corresponding to each piece of index information of the equipment, wherein the index information comprises at least one piece of information of network jitter, network packet loss, network delay and network bandwidth;

and determining the performance score of the equipment according to the index information of the equipment and the first weight corresponding to each index information.

4. The method of claim 2, wherein calculating a link condition score for the network link based on the performance scores of the devices on the network link comprises:

determining a second weight corresponding to each device on the network link, wherein the second weight is used for representing the importance degree of the corresponding device on the network link;

and determining a link condition score of the network link according to the second weight corresponding to each device and the performance score of each device.

5. The method of claim 4, wherein for each device on the network link, the second weight corresponding to the device is determined by:

6. The method of claim 2, wherein if the abnormal network link includes a plurality of abnormal network links, the classification layer is further configured to sort the link information of the plurality of abnormal network links according to the link condition score of each abnormal network link, and output the link information of the plurality of abnormal network links according to the sorting result.

7. A network detection model training method, comprising:

Acquiring sample link information of sample network links included in a plurality of network topologies and label information of the sample network links; wherein the sample network link comprises: sample network links with normal network conditions and sample network links with abnormal network conditions; the sample link information includes index information of devices on the sample network link; the tag information is used to characterize at least one of: whether the sample network link is abnormal, whether a device on the sample network link is abnormal;

inputting the sample link information into a network detection model to be trained to obtain a classification result corresponding to the sample network link, wherein the classification result comprises at least one of the following: whether the sample network link is abnormal, whether a device on the sample network link is abnormal;

8. The method of claim 7, wherein the network detection model to be trained comprises: a device scoring layer, a network link scoring layer and a classification layer;

9. A network detection device, comprising:

the first acquisition module is used for acquiring link information of a network link included in a network topology structure, wherein the link information comprises index information of equipment on the network link;

10. A network detection model training device, comprising:

11. An electronic device comprising a processor and a memory electrically connected to the processor, the memory storing a computer program, the processor being configured to call and execute the computer program from the memory to implement the network detection method of any one of claims 1-6, or the processor being configured to call and execute the computer program from the memory to implement the network detection model training method of any one of claims 7-8.

12. A storage medium storing a computer program executable by a processor to implement the network detection method of any one of claims 1-6, or executable by a processor to implement the network detection model training method of any one of claims 7-8.