KR20230128953A - Predictive maintenance device and method for network monitoring - Google Patents

Abstract

Disclosed is a predictive maintenance apparatus and method for preventing failure of network equipment in advance by collecting alarms generated from network equipment and predicting the life of the network equipment. According to one aspect, a predictive maintenance device for network control includes a collection unit for collecting alarms generated from network equipment; and a life predicting unit that calculates the survival probability of the network equipment by inputting the free A/S period and the number of alarms of the network equipment to a survival probability model having a free A/S period and the number of alarms as parameters.

Description

Predictive maintenance device and method for network monitoring

The present invention relates to an apparatus and method for predictive maintenance for network control, and more particularly, to a predictive maintenance apparatus and method for network control capable of preventing failure by predicting the lifespan of network equipment.

In October 2021, due to a work error within the domestic telecommunications company, there was a major failure that prevented wired and wireless Internet access for about an hour during lunchtime. Such major failures are caused by unannounced failures of some network equipment. In order to prevent large-scale failure, the state of network equipment is checked through EMS (Element Management System) or by directly accessing network equipment, but it is not easy to predict failure, especially when it is not directly related to network communication. It is more difficult to predict the failure of unattended components. Therefore, there is a need for a method to predict a network failure in advance and prevent a large-scale failure.

The present invention has been proposed to solve the above problems, to provide a predictive maintenance apparatus and method for preventing failure of network equipment in advance by predicting the life of network equipment by collecting alarms generated from network equipment. It has a purpose.

According to one aspect, a predictive maintenance device for network control includes a collection unit for collecting alarms generated from network equipment; and a life predicting unit that calculates the survival probability of the network equipment by inputting the free A/S period and the number of alarms of the network equipment to a survival probability model having a free A/S period and the number of alarms as parameters.

The survival probability model may apply a worn out weight to the free A/S period.

The consumption weight may be determined according to the characteristics of the free A/S period and the network equipment.

The consumption weight may be 1 or less.

The life predicting unit may count the number of occurrences of the alarm by applying a weight for each type of alarm.

The weight of some alerts may be greater than 1.

The sum of the weights of all types of alerts may be 1.

The life expectancy unit may input a ratio of the number of occurrences of the alarm to a predetermined limited number of alarms to the survival probability model.

The predictive maintenance device may further include an alarm unit outputting an alarm when the calculated survival probability is less than or equal to a preset threshold.

According to another aspect, a predictive maintenance method performed by a predictive maintenance device includes collecting an alert generated from a network device; and calculating the survival probability of the network equipment by inputting the free A/S period and the number of alarms of the network equipment into a survival probability model having a free A/S period and the number of alarms as parameters.

The survival probability model may apply a worn out weight to the free A/S period.

The consumption weight may be determined according to the characteristics of the free A/S period and the network equipment.

The consumption weight may be 1 or less.

In the calculating, the number of occurrences of the alarm may be counted by applying a weight for each type of alarm.

The weight of some alerts may be greater than 1.

The sum of the weights of all types of alerts may be 1.

In the calculating, a ratio of the number of occurrences of the alarm to a predetermined limited number of alarms may be input to the survival probability model.

The predictive maintenance method may further include outputting an alarm when the calculated survival probability is equal to or less than a preset threshold.

According to the present invention, by calculating the life expectancy of network equipment, that is, the probability of survival, based on the free A/S period and the number of alarms, the operator can recognize the need for repair or replacement even before the free A/S period arrives. Therefore, large-scale failures of wired/wireless networks can be prevented in advance.

1 is a diagram showing the configuration of a predictive maintenance system for network control according to an embodiment of the present invention.
2 is a graph showing the survival probability according to the number of months according to an embodiment of the present invention.
3 is a flowchart illustrating a predictive maintenance method according to an embodiment of the present invention.

Hereinafter, the embodiments disclosed in this specification will be described in detail with reference to the accompanying drawings, but the same or similar components are given the same reference numerals regardless of reference numerals, and redundant description thereof will be omitted. The suffix "part" for components used in the following description is given or used interchangeably in consideration of ease of writing the specification, and does not itself have a meaning or role distinct from each other. In addition, in describing the embodiments disclosed in this specification, if it is determined that a detailed description of a related known technology may obscure the gist of the embodiment disclosed in this specification, the detailed description thereof will be omitted. In addition, the accompanying drawings are only for easy understanding of the embodiments disclosed in this specification, the technical idea disclosed in this specification is not limited by the accompanying drawings, and all changes included in the spirit and technical scope of the present invention , it should be understood to include equivalents or substitutes.

Terms including ordinal numbers, such as first and second, may be used to describe various components, but the components are not limited by the terms. These terms are only used for the purpose of distinguishing one component from another.

Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, terms such as "comprise" or "have" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

In implementing the present invention, components may be subdivided for convenience of description, but these components may be implemented in one device or module, or one component may be divided into multiple devices or modules may be implemented in

1 is a diagram showing the configuration of a predictive maintenance system for network control according to an embodiment of the present invention. Referring to FIG. 1 , the predictive maintenance system according to the present embodiment includes a network equipment 110 and a predictive maintenance device 140 .

The network equipment 110 is equipment constituting a wired/wireless network, and includes equipment directly involved in network communication and equipment not directly involved in network communication. For example, equipment directly involved in network communication is a physical network device, such as a centralized unit (CU), a distributed unit (DU), a radio unit (RU), a switch, a router, and a packet optical transport network (POTN) constituting a wireless network. ) equipment, PTN (Packet Transport Network) equipment, etc. Equipment not directly involved in network communication includes, for example, an air conditioner in a data center, various sensors, and a compressor. However, it is not limited thereto. When a failure occurs, the network equipment 110 generates various alarms according to the failure type.

The predictive maintenance device 140 predicts the life of the network equipment 110 by collecting alarms generated from the network equipment 110 . As shown in FIG. 1, the predictive maintenance device 140 may collect alarms based on various communication protocols for some network equipment 110, and IoT adapter 120 for some network equipment 110. And an alert may be collected through the IoT gateway (GW) 130 . The IoT adapter 120 has a built-in communication modem to transmit alarms generated from network equipment such as air conditioners or sensors to the IoT gateway 130, and the IoT gateway 130 transmits the received alarm to the predictive maintenance device 140. . The IoT adapter 120 may be a device supporting serial communication such as Modbus.

For example, the predictive maintenance device 140 may collect alarms through a telemetry collector, and a protocol capable of real-time/periodic push such as NETCONF/RESTCONF/gRPC and traditional SNMP, CLI, and Syslog. Alerts can be collected using In addition, the predictive maintenance device 140 may collect alerts through well-known open source data collectors such as Telegraf, Fluentd, and Logstash. The predictive maintenance device 140 may store the collected alerts in a time series database.

The predictive maintenance device 140 predicts the life of the network equipment 110 based on the alert collected from the network equipment 110 and the free A/S period of the network equipment 110 . Specifically, the predictive maintenance device 140 calculates the survival probability of the network equipment 110 based on the survival probability model. The predictive maintenance device 140 may transmit a related alarm to a network control system, for example, a network management system (NMS), or visualize and display a related alarm when the calculated survival probability is less than or equal to a preset threshold. The survival probability model and the like will be described in more detail below.

On the other hand, in this embodiment, the survival probability is calculated for the network equipment 110, where the network equipment 110 may mean one physical equipment itself, or a part constituting the physical equipment (eg, memory, CPU). That is, the predictive maintenance device 140 may calculate the survival probability for one physical equipment itself, or if an alarm is generated for each individual component constituting one physical equipment, the survival probability for each individual component. can be calculated. Therefore, for convenience of explanation, in the present specification and claims, individual components are comprehensively referred to as the network equipment 110 and described.

The predictive maintenance device 140 may include a memory, one or more processors (CPUs), a display device, an input device, and a communication circuit. The memory may include high-speed random access memory, and may also include one or more magnetic disk storage devices, non-volatile memory such as flash memory devices, or other non-volatile semiconductor memory devices. The memory may store various types of information and program instructions, and the program is executed by a processor. One or more processors execute various software programs and/or sets of instructions stored in memory to perform various functions and process data for the system. The communication circuit performs communication through an external port or communication by an RF signal. The communication circuitry converts electrical signals to RF signals and vice versa, and communicates with the communication network, other mobile gateway devices, and communication devices via the RF signals. As shown in FIG. 1, the predictive maintenance device 140 includes a collection unit 141, a life prediction unit 142, an alarm unit 143, and a visualization unit 144, which are implemented as programs and stored in memory. It can be stored in and executed by at least one processor, implemented and operated as a combination of hardware and software.

The collection unit 141 collects alerts generated by the network devices 110 . The collection unit 141 can collect alarms through a telemetry collector, and collects alarms using a protocol capable of real-time/periodic push such as NETCONF/RESTCONF/gRPC and traditional SNMP, CLI, and Syslog. can do. In addition, the collection unit 141 may collect alerts through well-known open source data collectors such as Telegraf, Fluentd, and Logstash.

The collection unit 141 may collect alarms for some network equipment 110 through the IoT adapter 120 and the IoT gateway (GW) 130 . The IoT adapter 120 has a built-in communication modem to transmit an alarm generated from network equipment such as an air conditioner or a sensor to the IoT gateway 130, and the collection unit 141 removes the IoT adapter 120 from the IoT gateway 130. You can collect alerts received through The IoT adapter 120 may be a device supporting serial communication such as Modbus.

The collecting unit 141 may store the collected alerts in a time series database. Here, the database may be a network access storage device, or may be a local storage device such as a hard disk. In addition, the collection unit 141 may perform preprocessing to remove outlier data from among collected alerts.

The life prediction unit 142 predicts the life of the network devices 110 based on the alerts collected from the network devices 110 and the free A/S period of the network devices 110 . Specifically, the life predicting unit 142 calculates the survival probability of the network equipment 110 based on a survival probability model having the number of alarms and the free A/S period as parameters. The survival probability model has a negative exponential distribution that gradually decreases, and the slope changes as the number of alarms increases and the free A/S period approaches. An example of a survival probability model is as follows (Equation 1).

(Equation 1)

Here, y _i is the survival probability of network equipment i, x _i is the free A / S period (number of months) of network equipment i, and α _i,WOP is the weight of the worn out according to the free A / S period , T _i,Allow is the number of alarm occurrences of network device i, and T _i,Limit is the limit number of alarms of network device i.

Among the parameters of (Equation 1) above, αi _,WOP is a consumption weight according to the free A/S period, and is determined according to the free A/S period and the characteristics of the corresponding network equipment i. For example, when there are two different network equipment, even if the free A/S period of the two network equipment is the same, in the case of frequently used network equipment, some core parts even before the free A/S period arrives. replacement may be required. α _i,WOP is used to reflect this characteristic, and its value can be set by an operator. For example, even if the free A / S period of both the first network equipment and the second network equipment is 5 years, if the first network equipment is used more often than the second network equipment or generally fails more frequently, α _i,WOP of the first network equipment may be set to 0.8, for example, and α _i,WOP of the first network equipment may be set to 1, for example. In this case, assuming that the number of occurrences of an alarm is the same, the expected survival probability of the first network device is reduced compared to that of the second network device.

Among the parameters of Equation 1 above, T _i,Allow means the number of alarm occurrences of network equipment i. For T _i,Allow , each occurrence of an alert can be counted as one time regardless of the type of alert, or the number of occurrences of an alert can be counted by applying a weight according to the type of alert. For example, if a weight of 1 is applied to the first type of alarm, a weight of 0.5 is applied to the second type of alarm, the first type of alarm occurs twice and the second type of alarm occurs twice, The total number of alarms may be counted as 1*2 + 0.5*2 = 3. For some important alarms, the weight is set much higher than 1, so that the survival probability can be lowered and replacement can be induced when a corresponding important alarm occurs. Counting the number of alarm occurrences by applying a weight according to the type of alarm is expressed as the following (Equation 2).

(Equation 2)

Here, β _i,j ^alarm is the weight for the j-th type alarm of network device i.

In one embodiment, the weight of each alert may be set such that the sum of the weights of all types of alerts of network equipment i is 1. The case where the sum of the weights of all types of alerts of network equipment i is 1 is expressed as Equation 3 as follows. In this case, the weight does not exceed 1 even for some important alarms.

(Equation 3)

Among the parameters of Equation 1 above, T _i,Limit is the limit number of alarms of network equipment i. This value can be set appropriately by the operator. The number of occurrences of the limit alarm is a fixed value, and as the number of occurrences of an alarm (T _i,Allow ) generated by network equipment i increases, the number of occurrences of the alarm (T i,Allow ) for the number of limit alarms (T _{i, Limit} ) The ratio increases, and the survival probability according to (Equation 1) above decreases. That is, the survival probability gradually decreases according to the number of alarms.

If the survival probability according to (Equation 1) is conceptually described as a graph, it is shown in FIG. 2. Figure 2 is a graph showing the survival probability according to the number of months according to an embodiment of the present invention, the x-axis is the number of months and the y-axis is the survival probability (%), showing an example when the free A / S period is 3 years. As shown in FIG. 2, according to the survival probability of the present invention, the maximum survival probability is 100%, and the survival probability decreases as the free A/S period approaches. At this time, as described above, the shape of the graph may vary according to values such as a worn out weight (α _{i, WOP} ) and a limited number of alerts (T _{i, Limit} ).

Referring back to FIG. 1 , the alarm unit 143 generates an alarm when the survival probability calculated by the life predicting unit 142 is equal to or less than a preset threshold value. The alarm unit 143 transmits the generated alarm to a network control system, for example, a Network Management System (NMS), or to the visualization unit 144 . Alternatively, the alarm unit 143 may output an alarm through a speaker. The alarm unit 143 may classify alarm levels according to survival probability values. For example, it may be divided into critical, major, minor, etc., but is not limited thereto.

The visualization unit 144 may visualize and display the alarm transmitted from the alarm unit 143 . The visualization unit 144 may visualize and display the alarm through a separate web page. In addition, the visualization unit 144 may visualize and display the survival probability calculated by the life expectancy unit 142 . At this time, the visualization unit 144 may visualize and display the survival probability for each network device together with the network topology. The visualization unit 144 may use an open source based visualization technology such as Grafana. The visualization unit 144 may superimpose and display the survival probability on a performance graph of major core parts (eg, CPU, memory) of network equipment.

3 is a flowchart illustrating a predictive maintenance method according to an embodiment of the present invention. The method described with reference to FIG. 3 may be performed by a processor of the predictive maintenance device 140 or a combination of hardware and software.

Referring to FIG. 3 , in step S301 , the predictive maintenance device 140 collects alarms generated from the network devices 110 . The predictive maintenance device 140 may collect alarms through a telemetry collector, and send alarms using protocols capable of real-time/periodic push such as NETCONF/RESTCONF/gRPC and traditional SNMP, CLI, and Syslog. can be collected In addition, the predictive maintenance device 140 may collect alerts through well-known open source data collectors such as Telegraf, Fluentd, and Logstash.

In one embodiment, the predictive maintenance device 140 may collect alarms for some network equipment 110 through the IoT adapter 120 and the IoT gateway (GW) 130 . The IoT adapter 120 has a built-in communication modem to transmit an alarm generated from network equipment such as an air conditioner or a sensor to the IoT gateway 130, and the predictive maintenance device 140 transmits an alarm from the IoT gateway 130 to the IoT adapter 120. You can collect alerts received through The IoT adapter 120 may be a device supporting serial communication such as Modbus.

The predictive maintenance device 140 may store the collected alerts in a time series database. Here, the database may be a network access storage device, or may be a local storage device such as a hard disk. In addition, the predictive maintenance device 140 may perform preprocessing to remove outlier data from among collected alerts.

In step S302, the predictive maintenance device 140 calculates the survival probability of each network equipment using a survival probability model for each network equipment having the free A/S period and the number of alarms as parameters. The survival probability model is the same as the example of (Equation 1) described above.

In step S303, the predictive maintenance device 140 compares the probability of survival of each network device calculated in step S302 with a threshold value, and determines whether there is a network device having a probability of survival less than or equal to the threshold value. In this case, the thresholds may be different for each network equipment or all may be the same.

When there is a network device having a survival probability below the threshold value, in step S304, the predictive maintenance device 140 outputs an alarm. The predictive maintenance device 140 may transmit an alarm to a network control system, for example, a Network Management System (NMS), visualize and display the alarm, or output an alarm through a speaker. At this time, the predictive maintenance device 140 may divide the alarm level according to the survival probability value. For example, it may be divided into critical, major, minor, etc., but is not limited thereto. The predictive maintenance device 140 can visualize and display alarms through a separate web page, visualize and display the survival probability of each network equipment together with the network topology, and provide key core parts of network equipment (e.g., The survival probability can be superimposed on the performance graph of CPU and memory) and displayed.

According to the above embodiment of the present invention, by calculating the expected life of the network equipment 110, that is, the probability of survival, based on the free A/S period and the number of alarms, the operator repairs the network equipment 110 even before the free A/S period arrives. Alternatively, the need for replacement can be recognized, and large-scale failures of the wired/wireless network can be prevented in advance.

While this specification contains many features, such features should not be construed as limiting the scope of the invention or the claims. Also, features described in separate embodiments in this specification may be implemented in combination in a single embodiment. Conversely, various features that are described in this specification in a single embodiment may be implemented in various embodiments individually or in combination as appropriate.

Although actions are described in a particular order in the drawings, it should not be understood that such actions are performed in the specific order as shown, or that the actions are performed in a series of sequential order, or that all described actions are performed to achieve a desired result. . Multitasking and parallel processing can be advantageous in certain circumstances. In addition, it should be understood that the division of various system components in the above-described embodiments does not require such division in all embodiments. The program components and systems described above may generally be implemented as a package in a single software product or multiple software products.

The method of the present invention as described above may be implemented as a program and stored in a recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.) in a computer-readable form. Since this process can be easily performed by a person skilled in the art to which the present invention belongs, it will not be described in detail.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention to those skilled in the art to which the present invention belongs, and thus the above-described embodiments and It is not limited by drawings.

110: network equipment
120: IoT adapter
130: IoT gateway
140: predictive maintenance device
141: collection unit
142: life prediction unit
143: alarm unit
144: visualization unit

Claims (19)

In the predictive maintenance device for network control,
a collection unit that collects alarms generated from network equipment; and
A predictive maintenance device including a lifespan prediction unit that calculates the survival probability of the network equipment by inputting the free A/S period and the number of alarms of the network equipment into a survival probability model having the free A/S period and the number of alarms as parameters . According to claim 1,
The survival probability model,
A predictive maintenance device, characterized in that for applying a worn out weight to the free A / S period. According to claim 2,
The consumption weight is,
The predictive maintenance device, characterized in that determined according to the free A / S period and the characteristics of the network equipment. According to claim 2,
The predictive maintenance device, characterized in that the consumption weight is 1 or less. According to claim 1,
The life prediction unit,
A predictive maintenance device characterized in that counting the number of occurrences of the alarm by applying a weight for each type of alarm. According to claim 5,
A predictive maintenance device, characterized in that the weight of some alarms is greater than 1. According to claim 5,
A predictive maintenance device, characterized in that the sum of the weights of all types of alarms is 1. According to claim 1,
The life prediction unit,
and inputting a ratio of the number of occurrences of the alarm to a predetermined limited number of alarms to the survival probability model. According to claim 1,
The predictive maintenance device further comprises an alarm unit outputting an alarm when the calculated survival probability is less than or equal to a preset threshold. As a predictive maintenance method performed by a predictive maintenance device,
Collecting alarms generated from network equipment; and
and calculating the survival probability of the network equipment by inputting the free A/S period and the number of alarms of the network equipment into a survival probability model having the free A/S period and the number of alarms as parameters. According to claim 10,
The survival probability model,
The predictive maintenance method characterized in that for the free A / S period to apply a worn out (worn out) weight. According to claim 11,
The consumption weight is,
The predictive maintenance method, characterized in that determined according to the free A / S period and the characteristics of the network equipment. According to claim 11,
The predictive maintenance method, characterized in that the consumption weight is 1 or less. According to claim 10,
The calculating step is
A predictive maintenance method comprising counting the number of occurrences of the alarm by applying a weight for each type of alarm. According to claim 14,
A predictive maintenance method, characterized in that the weight of some alerts is greater than 1. According to claim 14,
A predictive maintenance method, characterized in that the sum of the weights of all types of alerts is 1. According to claim 10,
The calculating step is
and inputting a ratio of the number of alarm occurrences to a predetermined limited number of alarms to the survival probability model. According to claim 10,
The predictive maintenance method further comprising outputting an alarm when the calculated survival probability is less than or equal to a preset threshold. A computer program stored in a computer-readable recording medium as a computer program for executing the method according to any one of claims 10 to 18 through a computer system.