WO2019179457A1 - 一种确定网络设备的状态的方法及装置 - Google Patents

一种确定网络设备的状态的方法及装置 Download PDF

Info

Publication number
WO2019179457A1
WO2019179457A1 PCT/CN2019/078832 CN2019078832W WO2019179457A1 WO 2019179457 A1 WO2019179457 A1 WO 2019179457A1 CN 2019078832 W CN2019078832 W CN 2019078832W WO 2019179457 A1 WO2019179457 A1 WO 2019179457A1
Authority
WO
WIPO (PCT)
Prior art keywords
early warning
network device
warning analysis
state
feature vector
Prior art date
Application number
PCT/CN2019/078832
Other languages
English (en)
French (fr)
Inventor
高云鹏
谢于明
肖欣
张亮
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to RU2020134150A priority Critical patent/RU2781813C2/ru
Priority to KR1020207029573A priority patent/KR102455332B1/ko
Priority to CA3094557A priority patent/CA3094557C/en
Priority to EP19770460.4A priority patent/EP3761566B1/en
Priority to JP2020550850A priority patent/JP7081741B2/ja
Publication of WO2019179457A1 publication Critical patent/WO2019179457A1/zh
Priority to US17/027,319 priority patent/US11405294B2/en

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14Network analysis or design
    • H04L41/142Network analysis or design using statistical or mathematical methods
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06Management of faults, events, alarms or notifications
    • H04L41/0631Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06Management of faults, events, alarms or notifications
    • H04L41/0631Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
    • H04L41/064Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis involving time analysis
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06Management of faults, events, alarms or notifications
    • H04L41/0681Configuration of triggering conditions
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14Network analysis or design
    • H04L41/145Network analysis or design involving simulating, designing, planning or modelling of a network
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50Network service management, e.g. ensuring proper service fulfilment according to agreements
    • H04L41/5003Managing SLA; Interaction between SLA and QoS
    • H04L41/5009Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0805Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability
    • H04L43/0817Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability by checking functioning
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/16Threshold monitoring

Definitions

  • the present application relates to the field of communications technologies, and in particular, to a method and apparatus for determining a state of a network device.
  • a threshold alarm system is generally established for the service performance data of the network device, and the state of the device on the network is monitored by setting different threshold ranges. Specifically, the current state of the network device is determined by determining a state corresponding to a set threshold range to which the service performance data indicator value of the network device belongs.
  • a network device in a faulty state may be in a normal working state at a certain time and a faulty state at another time. Therefore, it is obvious that in the above method, for a network device that has a fault of a fault, if the network device is only normal when the network device is in a normal working state, the network device may not be identified as having a fault, thereby causing omission. Early warning. Therefore, the above method determines the accuracy of the state of the network device is low.
  • the present application provides a method and apparatus for determining the state of a network device to solve the problem of low accuracy in determining the state of the network device in the prior art.
  • the present application provides a method for determining a state of a network device, the method comprising:
  • the early warning analysis device acquires multiple target key performance indicators (KPI) data of the network device in the preset duration, and acquires multiple feature information, and processes the multiple target KPI data according to each feature information to generate An element corresponding to each feature information, and the generated elements corresponding to the plurality of feature information are formed into the feature vector, and the feature vector is analyzed according to a preset early warning analysis model to determine the network device. a state; wherein any one of the feature information is used to represent a calculation manner of an element corresponding to the feature information in the feature vector.
  • KPI target key performance indicators
  • the above method determines the state of the network device by analyzing a plurality of target KPI data over a period of time, and not only determining the state of the network device by using data at one time, so that the accuracy of the determined network device is high. This can reduce the omission of warnings.
  • the early warning analysis device needs to generate the early warning analysis model before analyzing the feature vector according to the preset early warning analysis model
  • the specific method may be: acquiring the early warning analysis device
  • the feature vector samples corresponding to different states of the network device are subjected to logistic regression processing for each state and the feature vector samples corresponding to the network device state, to obtain the early warning analysis model.
  • the early warning analysis device may generate the early warning analysis model, so that the associated early warning analysis device subsequently determines the state of the network device according to the early warning analysis model.
  • the early warning analysis device analyzes the feature vector according to the preset early warning analysis model, and determines the state of the network device.
  • the specific method may be: the early warning analysis device according to the The early warning analysis model analyzes the feature vector, determines a probability value of the network device in each state, and multiplies each probability value by a reference value of a state corresponding to the preset probability value to obtain a plurality of products. And the warning analysis device adds the plurality of product values to obtain a state indication value, and determines a range of the setting indication value to which the state indication value belongs, and the state corresponding to the setting indication value range is The status of the network device.
  • the early warning analysis device can accurately determine the state of the network device, so that the subsequent maintenance is performed according to the state of the network device.
  • the early warning analysis device acquires a plurality of target KPI data of the network device in the predetermined duration
  • the specific method may be: the early warning analysis device receives the network device continuously sent by the network management device. KPI data, and from the received KPI data, obtain a plurality of target KPI data within the preset duration.
  • the early warning analysis may acquire a plurality of target KPI data within a preset duration according to actual requirements, so that the early warning analysis device subsequently determines the feature vector according to the plurality of KPI data.
  • the network device can be, but is not limited to, a wavelength division device, a router, a packet transport network device, and the like.
  • the early warning analysis device can determine the status of a plurality of network devices to enable corresponding network devices to be maintained accordingly.
  • the early warning analysis device may display the determined state of the network device to the user through the visual display device, so that the user can accurately identify the Determining the current state of the network device, thereby performing corresponding maintenance according to the state of the network device.
  • the present application further provides an early warning analysis device having the function of implementing the early warning analysis device in the above method example.
  • the functions may be implemented by hardware or by corresponding software implemented by hardware.
  • the hardware or software includes one or more modules corresponding to the functions described above.
  • the structure of the early warning analysis device includes an obtaining unit and a processing unit, and the units can perform the corresponding functions in the foregoing method examples.
  • the units can perform the corresponding functions in the foregoing method examples.
  • the structure of the early warning analysis device includes an obtaining unit and a processing unit, and the units can perform the corresponding functions in the foregoing method examples.
  • the structure of the early warning analysis device includes a memory and a processor, and optionally a communication interface for communicating with other devices in the network system, the processor It is configured to support the early warning analysis device to perform the corresponding function in the above method.
  • the memory is coupled to the processor, which stores program instructions and data necessary for the early warning analysis device.
  • the application further provides a network system, where the network system includes a network device layer, a network management layer, an early warning analysis layer, and a visual display layer, specifically including the early warning analysis device, the network device, and the network equipment mentioned in the foregoing design.
  • Network management equipment and visual display equipment are included in the network system.
  • the present application also provides a computer storage medium having stored therein computer executable instructions for causing the computer to perform the above-described tasks when invoked by the computer a way.
  • the present application also provides a computer program product comprising instructions which, when run on a computer, cause the computer to perform any of the methods described above.
  • the present application further provides a chip connected to a memory for reading and executing program instructions stored in the memory to implement any of the above methods.
  • FIG. 1 is a schematic structural diagram of a network system provided by the present application.
  • FIG. 2 is a flowchart of a method for determining a state of a network device according to the present application
  • FIG. 3 is a schematic diagram of removing an abnormal point of error correction error rate according to the present application.
  • FIG. 4 is a schematic diagram of a process for generating an early warning analysis model provided by the present application.
  • FIG. 5 is a schematic structural diagram of an early warning analysis device provided by the present application.
  • FIG. 6 is a structural diagram of an early warning analysis device provided by the present application.
  • the embodiment of the present invention provides a method and an apparatus for determining a state of a network device, which are used to solve the problem that the accuracy of determining the state of the network device in the prior art is low.
  • the method and the device of the present application are based on the same inventive concept. Since the principles of the method and the device for solving the problem are similar, the implementation of the device and the method can be referred to each other, and the repeated description is not repeated.
  • a network device is a device that provides a service for a user, and the network device can have multiple states, such as a normal working state, a low-risk state, a high-risk state, a fault state, and the like, wherein the low-risk state Both the high-risk state and the high-risk state can be considered as fault hazard states.
  • the network device can be, but is not limited to, a wavelength division device, a router, a packet transport network device, and the like.
  • the early warning analysis device is a device that determines the state of the network device by analyzing KPI data of the network device.
  • the early warning analysis device may be a server or a cluster composed of multiple servers.
  • the network management device is a device for collecting KPI data of the network device, and transmitting the collected KPI data to the early warning analysis device.
  • KPI data is data used to characterize the service performance of network devices.
  • the KPI data of different network devices is different.
  • a plurality of the embodiments of the present application refer to two or more.
  • FIG. 1 is a schematic diagram of a possible network system architecture for determining a state of a network device according to an embodiment of the present application, where the network system includes a network device layer, a network management layer, an early warning analysis layer, and a visual display. Layer, where:
  • the network device layer includes a plurality of network devices, respectively providing different service services for users.
  • the network device layer may include network devices such as a wavelength division device, a router, and a packet transport network device.
  • the network management layer includes a plurality of network management devices, configured to collect KPI data of any one of the network device layers, and transmit the collected KPI data to the early warning analysis layer, so that the early warning analysis layer
  • the early warning analysis device analyzes the KPI data to determine the status of the network device.
  • the network management layer may collect KPI data of the network device from the network device layer by using a standard common object request broker architecture (CORBA) northbound interface; then, the network management layer may pass the file.
  • CORBA common object request broker architecture
  • FTP file transfer protocol
  • the early warning analysis layer includes an early warning analysis device, and the early warning analysis device may be a server or a cluster composed of multiple servers.
  • the early warning analysis device in the early warning analysis layer determines the state of the corresponding network device by analyzing the KPI data transmitted by the network management layer, and displays the determined state of the network device to the user through the visual display layer.
  • the visual display layer is used to implement human-computer interaction.
  • the visual display layer includes at least one visual display device, and the user can identify the service reliability of the network device by using the status of any one of the network devices displayed by the at least one visual display device in the visual display layer. And performing corresponding maintenance on different states of the network device. For example, when it is determined that the network device is in a fault risk state, the network device may be correspondingly maintained in advance to prevent the network device from being faulty and causing service interruption, which may improve the user service experience.
  • a method for determining a state of a network device provided by an embodiment of the present application is applicable to a network system as shown in FIG. 1.
  • the specific process of the method includes:
  • Step 201 The early warning analysis device acquires multiple target KPI data of the network device in the preset duration.
  • the specific method may be: the early warning analysis device receives the KPI data of the network device continuously sent by the network management device, as shown in FIG. 2 200: The early warning analysis device acquires the plurality of KPI data within the preset duration from the received KPI data.
  • the preset duration may be a preset duration before the time when the early warning analysis device receives the KPI data.
  • the network devices may be any one of a wavelength division device, a router, a packet transport network device, and the like.
  • the KPI data of the wavelength division device may be a pre-correction error rate and a corrected error rate.
  • the bit error rate refers to the ratio of the number of bits in which the error occurs to the total number of bits transmitted.
  • the error rate may be a pre-correction error rate or a corrected error rate.
  • the forward error correction (FEC) algorithm can detect the number of erroneous bits and correct some of the errors. Therefore, the error rate before using the FEC algorithm is the error correction rate, and the error correction using the FEC algorithm. The error rate obtained after that is the error correction rate.
  • the bit error rate can be represented by an error parameter that is an integer. For example, if the error parameter is 6, it represents that the bit error rate is 10 to the power of -6; and when the bit error rate is 0, it represents There is no error. For the subsequent analysis, when the bit error rate is 0, it can be represented by the error parameter 13, that is, the bit error rate is 10 to the power of -10, indicating that it is close to zero.
  • the KPI data of other network devices are not listed here.
  • Step 202 The early warning analysis device acquires multiple feature information, and any one of the feature information is used to represent a calculation manner of an element corresponding to the feature information in the feature vector.
  • the early warning analysis device presets, for the network device, feature information corresponding to each element in the feature vector required to analyze the state of the network device, so the early warning analysis device obtains the multiple After the target KPI data, the plurality of feature information may be acquired, so that the early warning analysis device can accurately perform the subsequent step 203.
  • the description is made by taking the network device as a wavelength division device as an example.
  • the preset feature vector for the WDM device includes the number of service interruptions, the KPI trend degradation worst value, and the fluctuation value worst value. , mean value of fluctuation value, mean value of threshold distance and average value of threshold distance. Among them, the number of service interruptions, the worst value of KPI trend degradation, the worst value of fluctuation value, the average value of fluctuation value, the mean value of threshold distance and the average value of threshold distance can be regarded as multiple feature information, so that big data can be used.
  • the technology analyzes the KPI data to obtain corresponding elements according to each feature information, and finally constitutes a feature vector containing 6 elements.
  • the calculation manner of the corresponding element that can be characterized by each feature information may be as follows: the early warning analysis device may determine, according to multiple error correction error rates in the plurality of target KPI data of the wavelength division device. Describe whether each of the collection points in the preset time length is a fault, and then calculate the number of service interruptions within a preset duration, the number of service interruptions may be a positive integer; may be based on multiple error correction rates Calculating the fluctuation value, KPI trend value and threshold distance degree of each collection point of the wavelength division device within a preset duration, and calculating the fluctuation value mean value and the fluctuation value average value within the preset time length according to the fluctuation value According to the threshold distance, the threshold distance degree difference value and the threshold distance degree average value in the preset time length are calculated, and the four characteristic value values may be positive positive numbers in [0, 100], and are calculated according to the KPI trend value.
  • the worst value of the KPI trend value in the duration is obtained, that is, the KPI trend deterioration worst value
  • the preset duration of the feature information representation may be the same, that is, the preset duration; or may be a different duration, and may be a part of the preset duration, for example,
  • the preset duration of obtaining the plurality of target KPIs may be 30 days, and the preset duration of the number of service interruptions may be 30 days, the fluctuation value mean value, the fluctuation value average value, and the threshold distance degree difference value
  • the preset duration of the threshold and the threshold distance may be 1 day, and the preset duration of the KPI trend degradation may be 7 days, wherein 1 day and 7 days are part of the preset duration of 30 days. duration.
  • Step 203 The early warning analysis device processes the plurality of target KPI data according to each feature information, and generates an element corresponding to each feature information.
  • the early warning analysis device processes the multiple target KPI data according to each feature information, and generates an element corresponding to each feature information, which may be specifically characterized according to each feature information.
  • the element is calculated by the way the element corresponding to the feature information is calculated.
  • the network device is still taken as an example of the wavelength division device.
  • the calculation process of each element is specifically described as follows:
  • the early warning analysis device determines whether the wavelength division device is a fault by correcting the error rate (denoted as aft), and the fault is recorded as f, and the calculation method of f can be as shown in the following formula 1:
  • aft is equal to 13 (that is, the error correction rate is 0)
  • f is 1 to indicate that the wavelength division device is faulty, that is, the service is interrupted;
  • aft is not equal to 13
  • f is 0, indicating that the wavelength division device is not faulty, that is, the service is not interrupted.
  • the 13 described in Equation 2 is 13 for the error code when the error correction rate is 0;
  • the early warning analysis device calculates the number of times the service interruption occurs in the wavelength division device within a preset duration (which may be 30 days), and the number of service interruptions is fault, and the calculation method of the fault may be as shown in the following formula 2:
  • the early warning analysis device may first process the abnormal value and the noise portion of the multiple error correction rate of the wavelength division device, and calculate a plurality of error correction error rates within the first preset time duration ( It can be the steady state value of 30 days), and calculate the distance between the error correction rate and the steady state value at the current time point in real time, determine the fluctuation value according to the distance of the distance, and finally calculate the second preset time length (may be 1) Day) The difference and average of the fluctuation values of each point.
  • the specific calculation process can be as follows:
  • the early warning analysis device may specifically remove the abnormal points of the plurality of pre-correction error rate data by using a sigma ( ⁇ ) principle, as shown in FIG. 3, before the correction of [u-3 ⁇ , u+3 ⁇ ]
  • the error rate data is an abnormal point, where u is an expected value of the error correction rate; and the early warning analysis device can use the performance evaluation process algebra (FEPA) algorithm to remove the plurality of corrections The error rate noise, and then the early warning analysis device calculates an average value of the plurality of correction error rates after removing the abnormal point and the noise within the preset duration to obtain a steady state value; then the early warning analysis device is based on each sample Calculate the fluctuation value dev of each sampling point by the difference between the error correction rate and the steady state value of the point.
  • the following formula 3 can be used:
  • x is the pre-correction error rate of each sampling point, It is a steady-state value; finally, according to the obtained fluctuation value of each point, the fluctuation value maximum value dev_min within the preset duration and the fluctuation value average value dev_avg are calculated, and the following formula 4 and formula 5 can be respectively used:
  • Dev_min min(dev 1 , dev 2 , dev 3 ... dev n ) Equation 4;
  • n in the above formula 4 and formula 5 is the number of samples of the pre-correction error rate within the preset duration.
  • the early warning analysis device configures a pre-correction error rate threshold that the wavelength division device itself can support.
  • the error rate before correction is higher than a hardware characteristic threshold of the wavelength division device, the error rate before correction The closer the hardware characteristic threshold is, the worse the reliability of the wavelength division device is.
  • the threshold distance degree S of the wavelength division device is as shown in the following formula 6; when the error rate before correction is lower than the hardware
  • the threshold distance of the device is 0, as shown in Equation 6:
  • x is the error correction rate
  • x max is the maximum value of the error rate before the preset time length
  • v is the hardware characteristic threshold
  • the early warning analysis device calculates the threshold distance degree difference S_min and the threshold distance degree average value S_avg within the preset duration, respectively adopting the following formula 7 and formula 8:
  • Equation 8 and Equation 9 is the number of samples of the error rate before correction in the preset duration.
  • the early warning analysis device performs an exponentially weighted moving average (EWMA) process on the KPI data within a preset duration (which may be 7 days), and then performs a linear fitting process to obtain the KPI trend.
  • EWMA exponentially weighted moving average
  • Step 204 The early warning analysis device combines the generated elements corresponding to the plurality of feature information into the feature vector, and analyzes the feature vector according to a preset early warning analysis model to determine a state of the network device. .
  • a plurality of elements constituting the feature vector may be obtained by the foregoing step 203, and the early warning analysis may directly form the plurality of elements into the feature vector.
  • the network device is a wavelength division device
  • the elements corresponding to the plurality of feature information related to the wavelength division device obtained in the above step 203 are faults, slope, dev_avg, dev_min, S_avg, and S_min.
  • the early warning analysis device may analyze the feature vector according to a preset early warning analysis model to determine the state of the wavelength division device.
  • the early warning analysis device generates the early warning analysis model before analyzing the feature vector according to a preset early warning analysis model, and the specific early warning analysis device generation device
  • the process of the early warning analysis model may be: the early warning analysis device acquires feature vector samples corresponding to different states of the network device; the early warning analysis device performs logistic regression processing on each state and the feature vector samples corresponding to the state, The early warning analysis model is obtained.
  • the status of the network device may include a normal working state (also referred to as a healthy state), a low risk state, a high risk state, and a fault state.
  • a normal working state also referred to as a healthy state
  • a low risk state also referred to as a low risk state
  • a high risk state a high risk state
  • a fault state a normal working state
  • a fault occurs within a preset time period, a wave value maximum value and a wave value average value, a threshold distance degree difference value, and a threshold distance degree average
  • the values are all above 90, and the KPI trend degradation worst value is greater than 0, so that the feature vector corresponding to the normal working state may be ⁇ 0, 0.1, 100, 100, 100, 100 ⁇ ;
  • the value of the feature vector is different from the value of the feature vector in the corresponding feature vector, for example, there is no fault in the preset duration or only occurs once.
  • the fault, the mean value of the fluctuation value and the mean value of the fluctuation value, the maximum value of the threshold distance and the average value of the threshold distance are all below 90, but the average value of the fluctuation value and the average value of the threshold distance are both above 70, and the KPI trend is the most deteriorating.
  • the difference is a non-negative number, such that the eigenvector corresponding to the low-risk state may be ⁇ 0, 0.02, 81.52, 71.89, 83.46, 71 ⁇ ;
  • the value of the feature vector is lower than the value of the feature vector in the corresponding feature vector, for example, the number of failures in the preset duration is greater than 2, and the fluctuation value is The average value of the average value and the threshold distance are below 70, and the KPI trend deterioration worst value is less than 0, so that the feature vector corresponding to the high risk state may be ⁇ 5, -4.91, 66.1, 0, 24.43, 0 ⁇ ;
  • the threshold distance is the worst value and The average threshold distance is 0, and the KPI trend degradation worst value is less than 0, so the feature vector corresponding to the fault state may be ⁇ 8, -2.64, 28.01, 27.06, 0, 0 ⁇ .
  • the feature vector samples corresponding to each state can be known by the feature vectors corresponding to the different states described above, and the early warning analysis device performs model training on each state and the feature vector samples corresponding to the state based on a logistic regression algorithm.
  • the early warning analysis model can be generated.
  • the input of the early warning analysis model is a feature vector of the network device
  • the output result is a probability value that the network device is determined to be in each state according to the input feature vector, that is, multiple probability values may be obtained, thereby Determining a state of the network device based on the plurality of probability values.
  • the generating process of the early warning analysis model corresponding to the wavelength division device may be as shown in FIG. 4 .
  • the early warning analysis device analyzes the feature vector according to a preset early warning analysis model to determine a state of the network device
  • the specific method may be: the early warning analysis device is configured according to the The early warning analysis model analyzes the feature vector, determines a probability value of the network device in each state, and multiplies each probability value by a reference value of a state corresponding to the preset probability value to obtain a plurality of a product value; and adding the plurality of product values to obtain a state indication value; the early warning analysis device determines a range of the setting indication value to which the state indication value belongs, and the state corresponding to the setting indication value range As the state of the network device.
  • the early warning analysis device analyzes the determined feature vector by using the preset early warning analysis model, and the obtained multiple probability values are ⁇ g1, g2, g3, g4 ⁇ , where g1 is that the network device is faulty.
  • the probability of the state g2 is the probability that the network device is in a high-risk state
  • g3 is the probability that the network device is in a low-risk state
  • g4 is the probability that the network device is in a normal working state.
  • each of the states corresponds to a reference value
  • the reference values corresponding to the four states may be respectively recorded as h1, h2, h3, and h4, wherein each of the reference values corresponds to a range of values, for example, h1 corresponds to [9, 10], h2 corresponds to [6.5, 7.5], h3 corresponds to [2.5, 3.5], and h4 corresponds to [0, 0.5]. It is assumed that, in the process of determining the state of the network device, each of the preset reference values corresponding to each state is 10, 7, 3, 0, and then the device on the network obtained by combining the analysis feature vector is in each state.
  • Corresponding probability values g1, g2, g3, g4, the state indication value Z can be obtained by the following formula 9:
  • the range in which the status indication value belongs is different, and the corresponding status is different, wherein the value of the status indication value may be in [0, 10], wherein configuring the first intermediate value and the second intermediate value will be
  • the value of the status indication value is divided into three ranges, namely [0, the first intermediate value), [the first intermediate value, the second intermediate value], (the second intermediate value, 10], wherein the three set indication values
  • the ranges respectively correspond to different states.
  • the first intermediate value may be set in [6.8, 7.2]
  • the second intermediate value may be set in [8, 9], which may make the determination
  • the status of network devices is more accurate.
  • the foregoing warning device can determine the state corresponding to the range of the setting indication value by determining the range of the setting indication value to which the state indication value belongs, so that the state of the network device can be determined.
  • the early warning analysis device may pass the determined state of the network device through a visual display device, as shown in step 205 in FIG. 2 . Displayed to the user so that the user can accurately identify the current state of the network device, thereby performing corresponding maintenance according to the state of the network device.
  • the early warning analysis device may determine, according to actual requirements, whether the state of the network device that is finally determined is a state that the user needs to know, and only displays the state of interest to the user. For example, the user is concerned whether the network device is in a fault hazard state, and the alert analysis device determines whether the network device is in a fault hazard state after determining the state of the network device by using the method provided in this embodiment of the present application.
  • the network device is in a faulty state
  • the state of the network device is displayed to the user, so that the user can perform maintenance on the network device before the network device fails to prevent the network device from being faulty.
  • Business disruption which can improve the user's business experience.
  • the early warning analysis device acquires multiple target key performance indicator KPI data of the network device in the preset duration, and acquires multiple feature information, and according to each feature information,
  • the plurality of target KPI data are processed to generate an element corresponding to each feature information, and the generated elements corresponding to the plurality of feature information are formed into the feature vector, and the feature vector is compared according to a preset early warning analysis model An analysis is performed to determine the status of the network device.
  • the accuracy of the determined network device can be made high, thereby Can reduce the omission of warnings.
  • the embodiment of the present application further provides an early warning analysis device, which is applied to the early warning analysis device in the network system shown in FIG. 1 , and is used to implement the determined network device as shown in FIG. 2 .
  • the state of the method Referring to FIG. 5, the early warning analysis device 500 includes an acquisition unit 501 and a processing unit 502, wherein:
  • the acquiring unit 501 is configured to acquire a plurality of target KPI data of the network device in the preset duration, and acquire a plurality of feature information, where any one of the feature information is used to represent the calculation of the element corresponding to the feature information in the feature vector. the way;
  • the processing unit 502 is configured to process the multiple target KPI data according to each feature information, generate an element corresponding to each feature information, and form the generated elements corresponding to the multiple feature information into the feature vector. And analyzing the feature vector according to a preset early warning analysis model to determine a state of the network device.
  • the processing unit 502 analyzes the feature vector according to a preset early warning analysis model, and determines the state of the network device, specifically, according to the early warning analysis.
  • a model analyzing the feature vector, determining a probability value of the network device in each state; multiplying each probability value by a reference value of a state corresponding to the preset probability value to obtain a plurality of product values, And adding the plurality of product values to obtain a state indication value; determining a range of the setting indication value to which the state indication value belongs, and using the state corresponding to the setting indication value range as the state of the network device.
  • the acquiring unit 501 is further configured to acquire feature vector samples corresponding to different network device states of the network device; the processor 502 analyzes a model according to a preset early warning analysis model. Before the feature vector is analyzed, the method further generates the early warning analysis model: after the acquiring unit 501 acquires feature vector samples corresponding to different network device states of the network device, each network device state and the network device The feature vector samples corresponding to the state are subjected to logistic regression processing to obtain the early warning analysis model.
  • the early warning analysis device 500 further includes: a receiving unit, configured to receive KPI data of the network device that is continuously sent by the network management device; and the acquiring unit 501 is configured to acquire the preset duration When the plurality of target KPI data of the internal network device is used, the plurality of target KPI data within the preset duration is obtained from the KPI data received by the receiving unit.
  • the network device may be a wavelength division device, a router, a packet transport network device, or the like.
  • the early warning analysis device acquiring multiple target key performance indicator KPI data of the network device within the preset duration, and acquiring multiple feature information, and processing the multiple target KPI data according to each feature information. Generating an element corresponding to each feature information, and forming the generated element corresponding to the plurality of feature information into the feature vector, and analyzing the feature vector according to a preset early warning analysis model to determine the network The status of the device.
  • the accuracy of the determined network device can be made high, thereby Can reduce the omission of warnings.
  • the division of the unit in the embodiment of the present application is schematic, and is only a logical function division. In actual implementation, there may be another division manner.
  • the functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.
  • the integrated unit if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium.
  • a computer readable storage medium A number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) or a processor to perform all or part of the steps of the methods described in various embodiments of the present application.
  • the foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like, which can store program code. .
  • the embodiment of the present application further provides an early warning analysis device, where the early warning analysis device is applied to the early warning analysis device in the network system shown in FIG. 1 , and is used to implement the determined network as shown in FIG. 2 .
  • the method of the state of the device Referring to FIG. 6, the early warning analysis device 600 includes a processor 602 and a memory 603, wherein:
  • the processor 602 can be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP.
  • the processor 602 may further include a hardware chip.
  • the hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof.
  • the PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a general array logic (GAL), or any combination thereof.
  • the processor 602 and the memory 603 are connected to each other.
  • the processor 602 and the memory 603 are connected to each other by a bus 604;
  • the bus 604 may be a Peripheral Component Interconnect (PCI) bus or an extended industry standard architecture (Extended Industry Standard Architecture) , EISA) bus, etc.
  • PCI Peripheral Component Interconnect
  • EISA Extended Industry Standard Architecture
  • the bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in Figure 6, but it does not mean that there is only one bus or one type of bus.
  • the early warning analysis device 600 implements the method for determining the state of the network device as shown in FIG. 2:
  • the processor 602 is configured to obtain multiple target KPI data of the network device in the preset duration;
  • the early warning analysis device combines the generated elements corresponding to the plurality of feature information into the feature vector, and analyzes the feature vector according to a preset early warning analysis model to determine a state of the network device.
  • the memory 603 is configured to store a program or the like.
  • the program can include program code, the program code including computer operating instructions.
  • the memory 603 may include a RAM, and may also include a non-volatile memory such as at least one disk storage.
  • the processor 602 executes an application stored in the memory 603 to implement the above functions, thereby implementing a method for determining a state of the network device as shown in FIG. 2.
  • the processor 602 is configured to analyze the feature vector according to a preset early warning analysis model, and determine the state of the network device, specifically: according to the early warning An analysis model, analyzing the feature vector, determining a probability value of the network device in each state; multiplying each probability value by a reference value of a state corresponding to the preset probability value to obtain a plurality of product values And adding the plurality of product values to obtain a state indication value; determining a range of the setting indication value to which the state indication value belongs, and using the state corresponding to the setting indication value range as the state of the network device.
  • the processor 602 is further configured to generate the early warning analysis model before acquiring the feature vector according to a preset early warning analysis model: acquiring the network device Feature vector samples corresponding to different network device states; performing logical regression processing on each network device state and feature vector samples corresponding to the network device state to obtain the early warning analysis model.
  • the early warning analysis device 600 further includes: a communication interface 601, configured to receive data; and the processor 602, in acquiring multiple target KPI data of the network device within the preset duration Specifically, the method is: controlling the communication interface 601 to receive KPI data of the network device that is continuously sent by the network management device; and acquiring, from the received KPI data, a plurality of target KPI data within the preset duration.
  • the network device may be a wavelength division device, a router, a packet transport network device, or the like.
  • the early warning analysis device acquiring multiple target key performance indicator KPI data of the network device within the preset duration, and acquiring multiple feature information, and processing the multiple target KPI data according to each feature information. Generating an element corresponding to each feature information, and forming the generated element corresponding to the plurality of feature information into the feature vector, and analyzing the feature vector according to a preset early warning analysis model to determine the network The status of the device.
  • the accuracy of the determined network device can be made high, thereby Can reduce the omission of warnings.
  • the embodiment of the present application provides a method and apparatus for determining a state of a network device, and an early warning analysis device, which acquires multiple target KPI data of a network device within a preset duration, and acquires multiple feature information. And processing the multiple target KPI data according to each feature information, generating an element corresponding to each feature information, and forming the generated elements corresponding to the multiple feature information into the feature vector, and according to a preset
  • the early warning analysis model analyzes the feature vector to determine the state of the network device.
  • the accuracy of the determined network device can be made high, thereby Can reduce the omission of warnings.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment in combination of software and hardware.
  • the application can 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.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Environmental & Geological Engineering (AREA)
  • Physics & Mathematics (AREA)
  • Algebra (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Physics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Pure & Applied Mathematics (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Debugging And Monitoring (AREA)

Abstract

一种确定网络设备的状态的方法及装置,用以解决现有技术中确定网络设备的状态的准确性较低的问题。预警分析设备获取预设时长内网络设备的多个目标关键绩效指标KPI数据,以及获取多个特征信息,根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素,并将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。这样,通过分析一段时间内的多个目标KPI数据来确定网络设备的状态,而不仅仅通过一个时刻的数据来确定所述网络设备的状态,可以使得确定的网络设备的准确性较高,从而可以减少预警的遗漏。

Description

一种确定网络设备的状态的方法及装置
本申请要求于2018年03月22日提交中国国家知识产权局、申请号为201810241478.1、申请名称为“一种确定网络设备的状态的方法及装置”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信技术领域,尤其涉及一种确定网络设备的状态的方法及装置。
背景技术
随着通信技术的不断发展,网络系统的复杂性日益增加,网络设备的可靠性问题来带的维护成本变得越来越大,因此,为了减少网络设备出现故障之后再去维护而造成的成本较大的问题,可以在网络设备出现故障隐患即将出现故障时进行预警,因此监控网络设备的状态尤为重要。
目前,在现有的网络系统运维中,通常针对网络设备的业务性能数据建立阈值报警体系,通过设置不同的阈值范围监控网络上设备的状态。具体的,在某个时刻通过判断网络设备的业务性能数据指标值所属的设定阈值范围对应的状态,确定该网络设备的当前状态。
但是在实际中,例如,一个处于故障隐患状态的网络设备在工作时,可能在某一时刻处于正常工作状态,在另一时刻处于故障隐患状态。所以,显然,在上述方法中,对于一个存在故障隐患的网络设备,如果在该网络设备处于正常工作状态的时刻只会判断该网络设备正常,不能识别到该网络设备有故障隐患,而导致遗漏预警。因此,上述方法确定网络设备的状态的准确性较低。
发明内容
本申请提供一种确定网络设备的状态的方法及装置,用以解决现有技术中确定网络设备的状态的准确性较低的问题。
第一方面,本申请提供了一种确定网络设备的状态的方法,该方法包括:
预警分析设备获取预设时长内网络设备的多个目标关键绩效指标(key performance indicators,KPI)数据,以及获取多个特征信息,根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素,并将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态;其中,任一个特征信息用于表征特征向量中所述特征信息对应的元素的计算方式。
上述方法,通过分析一段时间内的多个目标KPI数据来确定网络设备的状态,而不仅仅通过一个时刻的数据来确定所述网络设备的状态,可以使得确定的网络设备的准确性较高,从而可以减少预警的遗漏。
在一种可能的设计中,所述预警分析设备在根据预设的预警分析模型,对所述特 征向量进行分析之前,需要生成所述预警分析模型,具体方法可以为:所述预警分析设备获取所述网络设备的不同状态对应的特征向量样本,并对每种状态以及该网络设备状态对应的特征向量样本进行逻辑回归处理,得到所述预警分析模型。
通过上述方法,所述预警分析设备可以生成所述预警分析模型,以使所属预警分析设备后续根据所述预警分析模型确定所述网络设备的状态。
在一种可能的设计中,所述预警分析设备根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态,具体方法可以为:所述预警分析设备根据所述预警分析模型,对所述特征向量进行分析,确定所述网络设备处于每种状态的概率值,将每个概率值与预设的该概率值对应的状态的基准值相乘,得到多个乘积值;然后所述预警分析设备将所述多个乘积值相加,得到状态指示值,并确定所述状态指示值所属的设定指示值范围,将所述设定指示值范围对应的状态作为所述网络设备的状态。
通过上述方法,所述预警分析设备可以准确地确定所述网络设备的状态,以使后续根据所述网络设备的状态进行相应地维护。
在一种可能的设计中,所述预警分析设备获取所述预设时长内网络设备的多个目标KPI数据,具体方法可以为:所述预警分析设备接收网管设备持续发送的所述网络设备的KPI数据,并从接收的所述KPI数据中,获取所述预设时长内的多个目标KPI数据。
通过上述方法,所述预警分析可以根据实际需求获取预设时长内多个目标KPI数据,以使所述预警分析设备后续根据所述多个KPI数据确定特征向量。
在一种可能的设计中,所述网络设备可以但不限于为波分设备、路由器、分组传送网设备等。这样,所述预警分析设备可以确定多种网络设备的状态,以使对不同的网络设备进行相应地维护。
在一种可能的设计中,所述预警分析设备在确定了所述网络设备的状态后,可以将确定的所述网络设备的状态通过可视化展示设备展示给用户,以使用户能准确地识别所述网络设备的当前状态,从而根据所述网络设备的状态进行相应维护。
第二方面,本申请还提供了一种预警分析设备,该预警分析设备具有实现上述方法实例中预警分析设备的功能。所述功能可以通过硬件实现,也可以通过硬件执行相应的软件实现。所述硬件或软件包括一个或多个与上述功能相对应的模块。
在一个可能的设计中,所述预警分析设备的结构中包括获取单元和处理单元,这些单元可以执行上述方法示例中的相应功能,具体参见方法示例中的详细描述,此处不做赘述。
在一个可能的设计中,所述预警分析设备的结构中包括存储器和处理器,可选的还包括通信接口,所述通信接口用于与网络系统中的其他设备进行通信交互,所述处理器被配置为支持所述预警分析设备执行上述方法中相应的功能。所述存储器与所述处理器耦合,其保存所述预警分析设备必要的程序指令和数据。
第三方面,本申请还提供了一种网络系统,所述网络系统中包括网络设备层、网管层、预警分析层和可视化展示层,具体包括上述设计中提及的预警分析设备、网络设备和网管设备和可视化展示设备。
第四方面,本申请还提供了一种计算机存储介质,所述计算机存储介质中存储有计算机可执行指令,所述计算机可执行指令在被所述计算机调用时用于使所述计算机执行上述任一种方法。
第五方面,本申请还提供了一种包含指令的计算机程序产品,当其在计算机上运行时,使得计算机执行上述任一种方法。
第六方面,本申请还提供了一种芯片,所述芯片与存储器相连,用于读取并执行所述存储器中存储的程序指令,以实现上述任一种方法。
附图说明
图1为本申请提供的一种网络系统的架构示意图;
图2为本申请提供的一种确定网络设备的状态的方法的流程图;
图3为本申请提供的一种去除纠前误码率异常点的示意图;
图4为本申请提供的一种预警分析模型的生成过程的示意图;
图5为本申请提供的一种预警分析设备的结构示意图;
图6为本申请提供的一种预警分析设备的结构图。
具体实施方式
下面将结合附图对本申请作进一步地详细描述。
本申请实施例提供一种确定网络设备的状态的方法及装置,用以解决现有技术中确定网络设备的状态的准确性较低的问题。其中,本申请所述方法和装置基于同一发明构思,由于方法及装置解决问题的原理相似,因此装置与方法的实施可以相互参见,重复之处不再赘述。
以下,对本申请中的部分用语进行解释说明,以便于本领域技术人员理解。
1)、网络设备,是为用户提供业务服务的设备,所述网络设备可以有多种状态,例如正常工作状态、低风险状态、高风险状态、故障状态等等,其中,所述低风险状态和所述高风险状态均可以认为是故障隐患状态。所述网络设备可以但不限于为波分设备、路由器、分组传送网设备等。
2)、预警分析设备,是通过分析网络设备的KPI数据来确定所述网络设备的状态的设备。所述预警分析设备可以为一个服务器,或者为多个服务器组成的集群。
3)、网管设备,是用于采集网络设备的KPI数据,并将采集的所述KPI数据传送给预警分析设备的设备。
4)、KPI数据,是用于表征网络设备的业务性能的数据,不同的网络设备的KPI数据不同。
5)、本申请实施例涉及的多个,是指两个或两个以上。
为了更加清晰地描述本申请实施例的技术方案,下面结合附图,对本申请实施例提供的确定网络设备的状态的方法及装置进行详细说明。
图1示出了本申请实施例提供的确定网络设备的状态的方法适用的一种可能的网络系统的架构,所述网络系统的架构中包括网络设备层、网管层、预警分析层和可视化展示层,其中:
所述网络设备层包括多个网络设备,分别为用户提供不同的业务服务。例如,所 述网络设备层可以包括波分设备、路由器、分组传送网设备等网络设备。
所述网管层包括多个网管设备,用于采集网络设备层中任一个网络设备的KPI数据,并将采集的所述KPI数据传输至所述预警分析层,以使所述预警分析层中的预警分析设备分析对所述KPI数据进行分析以确定所述网络设备的状态。其中,所述网管层可以通过标准的公共对象请求代理体系结构(common object request broker architecture,CORBA)北向接口从所述网络设备层中采集网络设备的KPI数据;然后,所述网管层可以通过文件传输协议(file transfer protocol,FTP)将采集的所述网络设备的所述KPI数据发送给所述预警分析层。
所述预警分析层包括预警分析设备,所述预警分析设备可以为一个服务器,或者为多个服务器组成的集群。所述预警分析层中的预警分析设备通过分析所述网管层传输的KPI数据来确定对应的网络设备的状态,并将确定的所述网络设备的状态通过所述可视化展示层向用户展示。
所述可视化展示层用于实现人机交互。所述可视化展示层中包括至少一个可视化展示设备,用户可以通过所述可视化展示层中的所述至少一个可视化展示设备展示的任一个网络设备的状态,来识别所述网络设备的业务可靠性,并针对所述网络设备的不同状态进行相应的维护。例如,当确定所述网络设备处于故障隐患状态时,可以提前对所述网络设备进行相应维护,以避免所述网络设备出现故障而导致业务中断,这样可以提高用户业务体验。
本申请实施例提供的一种确定网络设备的状态的方法,适用于如图1所示的网络系统。参阅图2所示,该方法的具体流程包括:
步骤201、预警分析设备获取预设时长内网络设备的多个目标KPI数据。
在一种可选的实施方式中,所述预警分析设备执行步骤201时,具体方法可以为:所述预警分析设备接收网管设备持续发送的所述网络设备的KPI数据,如图2中的步骤200所示;然后所述预警分析设备从接收的所述KPI数据中,获取所述预设时长内的所述多个KPI数据。具体的,所述预设时长可以是所述预警分析设备每次接收到KPI数据的时刻起之前的预设时长。
在具体实现时,由于网络设备有多种,因此不同的网络设备有不同的KPI数据,而所述网络设备可以是波分设备、路由器、分组传送网设备等等中的任一种。
例如,所述网络设备是波分设备时,所述波分设备的KPI数据可以是纠前误码率和纠后误码率。其中,误码率指的是发生误码的位数与传输的总位数之比。可选的,误码率可以是纠前误码率或者纠后误码率。
前向纠错(forward error correction,FEC)算法可以检测出错误比特的数量,并且纠正其中一部分的错误,因此,使用FEC算法之前的误码率就是纠前误码率,而使用FEC算法纠错之后得到的误码率就是纠后误码率。
具体的,误码率可以通过一个是整数的误码参数来体现,比如说误码参数是6,那么代表的是误码率是10的-6次方;而误码率是0的时候代表没有误码,为了后续的分析,误码率为0时,可以通过误码参数13表示,即表示误码率是10的-13次方,表示接近于0。在这里,其它网络设备的KPI数据这里不再一一列举。
步骤202、所述预警分析设备获取多个特征信息,任一个特征信息用于表征特征 向量中所述特征信息对应的元素的计算方式。
具体的,所述预警分析设备针对所述网络设备预设了分析所述网络设备的状态时所需的特征向量中每个元素对应的特征信息,因此所述预警分析设备在获取到所述多个目标KPI数据后,可以获取所述多个特征信息,以使所述预警分析设备可以准确执行后续步骤203。
仍以所述网络设备为波分设备为例进行说明,所述预警分析设备针对所述波分设备预设的特征向量中依次包括业务中断次数、KPI趋势劣化最差值、波动值最差值、波动值平均值、门限距离度最差值和门限距离度平均值。其中,业务中断次数、KPI趋势劣化最差值、波动值最差值、波动值平均值、门限距离度最差值和门限距离度平均值即可以认为是多个特征信息,这样可以采用大数据技术对KPI数据进行分析,来根据每个特征信息得到对应的元素,最终组成包含6个元素的特征向量。
例如,具体的,每个特征信息可以表征的对应元素的计算方式可以如下:所述预警分析设备可以根据所述波分设备的多个目标KPI数据中的多个纠后误码率判断出所述波分设备在预设时长内的每个采集点是否为故障,然后计算出在预设时长内的业务中断次数,所述业务中断次数可以为正整数;可以根据多个纠前误码率计算出所述波分设备在预设时长内的每个采集点的波动值、KPI趋势值和门限距离度,再根据波动值计算在预设时长内的波动值最差值和波动值平均值,根据门限距离度计算在预设时长内的门限距离度最差值和门限距离度平均值,这四个特征值取值可以是在[0,100]的正实数,根据KPI趋势值计算出在预设时长内的KPI趋势值最差值,即得到KPI趋势劣化最差值,所述KPI趋势劣化最差值可以取值为任意实数。
其中,每个特征信息表征的计算方式中所述的预设时长可以相同,即为所述预设时长;也可以是不相同的时长,具体可以是所述预设时长中的一部分时长,例如,获取所述多个目标KPI的所述预设时长可以是30天,而业务中断次数涉及的预设时长可以是30天,波动值最差值、波动值平均值、门限距离度最差值和门限距离度平均值涉及的预设时长可以是1天,KPI趋势劣化最差值涉及的预设时长可以是7天,其中1天和7天均是所述预设时长30天中的一部分时长。
步骤203、所述预警分析设备根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素。
在一种可选的实施方式中,所述预警分析设备根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素,具体可以根据每个特征信息表征的该特征信息对应的元素的计算方式,得到该元素。
例如,仍以所述网络设备为波分设备为例进行说明,结合步骤202中的描述,具体对每个元素的计算得到过程进行如下说明:
A、业务中断次数:
首先,所述预警分析设备通过纠后误码率(记为aft)来判断波分设备是否为故障,故障记为f,f的计算方法可以如以下公式一所示:
Figure PCTCN2019078832-appb-000001
其中,aft等于13(即纠后误码率为0)时,f为1表示波分设备故障,即业务中 断;aft不等于13时,f为0表示波分设备未故障,即业务未中断;其中,具体的,公式2中描述的13,为体现纠后误码率为0时的误码参数为13;
然后,所述预警分析设备计算出所述波分设备在预设时长内(可以为30天)发生业务中断的次数,记业务中断次数为fault,fault的计算方法可以如以下公式二所示:
Figure PCTCN2019078832-appb-000002
B、波动值最差值和波动值平均值:
具体的,所述预警分析设备可以先对所述波分设备的多个纠前误码率的异常值和噪声部分进行处理,计算出多个纠前误码率在第一预设时长内(可以为30天)的稳态值,并实时计算当前时刻点的纠前误码率与稳态值的距离,根据距离的远近判断波动值,最后计算出第二预设时长内(可以为1天)每个点的波动值的最差值和平均值。具体的计算过程可以如下:
所述预警分析设备具体可以采用3西格玛(sigma,σ)原则去除多个纠前误码率数据的异常点,如图3所示,在[u-3σ,u+3σ]之外的纠前误码率数据为异常点,其中u为纠前误码率的期望值;以及,所述预警分析设备可以采用性能评价进程代数(ferformance evaluation process algebra,FEPA)算法分析方法去除所述多个纠前误码率的噪声,然后所述预警分析设备对预设时长内去除异常点和噪声后的多个纠前误码率计算平均值,得到稳态值;之后所述预警分析设备根据每个采样点的纠前误码率与稳态值的差值计算每个采样点的波动值dev,具体可以采用以下公式三:
Figure PCTCN2019078832-appb-000003
其中,公式三中x为每个采样点的纠前误码率,
Figure PCTCN2019078832-appb-000004
为稳态值;最后根据得到的每个点的波动值,计算预设时长内的波动值最差值dev_min,和波动值平均值dev_avg,具体可以分别采用以下公式四和公式五:
dev_min=min(dev 1,dev 2,dev 3...dev n)                   公式四;
Figure PCTCN2019078832-appb-000005
其中上述公式四和公式五中的n为预设时长内纠前误码率的采样个数。
C、门限距离度最差值和门限距离度平均值:
具体的,所述预警分析设备配置所述波分设备本身所能支持的纠前误码率门限,当纠前误码率高于所述波分设备的硬件特性门限时,纠前误码率与所述硬件特性门限越近,所述波分设备的可靠性越差,此时所述波分设备的门限距离度S如以下公式六所示;当纠前误码率低于所述硬件特性门限时,该设备的门限距离度为0,如公式六所示:
Figure PCTCN2019078832-appb-000006
其中,公式六中x为纠前误码率,x max为预设时长内纠前误码率的最大值,v为所述硬件特性门限;
然后,所述预警分析设备计算预设时长内的门限距离度最差值S_min和门限距离度平均值S_avg,分别采用如下公式七和公式八:
S_min=min(S 1,S 2,S 3...S m)                  公式七;
Figure PCTCN2019078832-appb-000007
其中,公式八和公式九中的m为预设时长内纠前误码率的采样个数。
D、KPI趋势劣化最差值:
具体的,所述预警分析设备对预设时长内(可以为7天)的KPI数据进行指数加权移动平均值(exponentially weighted moving average,EWMA)处理,然后做线性拟合处理,得到所述KPI趋势劣化最差值,记为slope。
通过上述方法,可以得到每个特征信息对应的元素,从而可以得到所述特征向量。
步骤204、所述预警分析设备将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。
具体的,通过上述步骤203可以得到组成所述特征向量的多个元素,所述预警分析可直接将所述多个元素组成所述特征向量。例如,以所述网络设备为波分设备为例,通过上述步骤203中举例中得到的所述波分设备涉及的多个特征信息对应的元素fault、slope、dev_avg、dev_min、S_avg、S_min,所述预警分析设备可以组成特征向量T={fault,slope,dev_avg,dev_min,S_avg,S_min}。进而,所述预警分析设备可以根据预设的预警分析模型,对所述特征向量进行分析,确定所述波分设备的状态。
在一种可选的实施方式中,所述预警分析设备在根据预设的预警分析模型,对所述特征向量进行分析之前,还要生成所述预警分析模型,具体的所属预警分析设备生成所述预警分析模型的过程可以为:所述预警分析设备获取所述网络设备的不同状态对应的特征向量样本;所述预警分析设备对每种状态以及该状态对应的特征向量样本进行逻辑回归处理,得到所述预警分析模型。
具体的,所述网络设备的状态可以包括正常工作状态(也可以称之为健康状态)、低风险状态、高风险状态、故障状态。例如,以波分设备为例,通过经验值知道:
所述波分设备处于正常工作状态时,对应的特征向量中,在预设时长内出现过故障,波分值最差值和波分值平均值、门限距离度最差值和门限距离度平均值都是90以上,KPI趋势劣化最差值大于0,这样对应所述正常工作状态的特征向量可以是{0,0.1,100,100,100,100};
所述波分设备处于低风险状态时,对应的特征向量中,比所述波分设备处于正常 工作状态时特征向量的值差一些,比如在预设时长内没有出现过故障或者只发生过一次故障,波动值最差值和波动值平均值、门限距离度最差值和门限距离度平均值都低于90,但是波动值平均值和门限距离度平均值都在70以上,KPI趋势劣化最差值是非负数,这样对应所述低风险状态的特征向量可以是{0,0.02,81.52,71.89,83.46,71};
所述波分设备处于高风险状态时,对应的特征向量中,比所述波分设备处于低风险状态时特征向量的值差一些,比如在预设时长内发生故障的次数大于2,波动值平均值和门限距离度平均值都在70以下,KPI趋势劣化最差值小于0,这样对应所述高风险状态的特征向量可以是{5,-4.91,66.1,0,24.43,0};
所述波分设备处于故障状态时,对应的特征向量中,在预设时长内出现故障的次数大于5,波动值最差值和波动值平均值都在40以下,门限距离度最差值和门限距离度平均值都是0,KPI趋势劣化最差值小于0,这样对应所述故障状态的特征向量可以是{8,-2.64,28.01,27.06,0,0}。
通过上述描述的已知的不同状态对应的特征向量,可以得知每种状态对应的特征向量样本,所述预警分析设备对每种状态以及该状态对应的特征向量样本基于逻辑回归算法进行模型训练,可以生成所述预警分析模型。具体的,所述预警分析模型的输入为网络设备的特征向量,输出结果为根据输入的特征向量所述网络设备被判定为处于每种状态的概率值,即可以得到多个概率值,从而可以基于所述多个概率值判断所述网络设备的状态。例如,所述波分设备对应的所述预警分析模型的生成过程可以如图4所示。
在一种可选的实施方式中,所述预警分析设备根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态,具体方法可以为:所述预警分析设备根据所述预警分析模型,对所述特征向量进行分析,确定所述网络设备处于每种状态的概率值,将每个概率值与预设的该概率值对应的状态的基准值相乘,得到多个乘积值;并将所述多个乘积值相加,得到状态指示值;所述预警分析设备确定所述状态指示值所属的设定指示值范围,将所述设定指示值范围对应的状态作为所述网络设备的状态。
例如,所述预警分析设备通过所述预设的预警分析模型对确定的特征向量进行分析,得到的多个概率值为{g1,g2,g3,g4},其中g1为所述网络设备处于故障状态的概率,g2为所述网络设备处于高风险状态的概率,g3为所述网络设备处于低风险状态的概率,g4为所述网络设备处于正常工作状态的设备的概率。其中,每种状态对应一个基准值,上述四种状态对应的基准值可以分别记为h1,h2,h3,h4,其中每个基准值的均对应一个取值范围,例如,h1对应[9,10],h2对应[6.5,7.5],h3对应[2.5,3.5],h4对应[0,0.5]。假设,在确定所述网络设备的状态的过程中,预设的每种状态对应的基准值分别为10,7,3,0,然后结合分析特征向量得到的所述网络上设备处于每种状态对应的概率值g1,g2,g3,g4,可以通过以下公式九得到所述状态指示值Z:
Z=g1*10+g2*7+g3*3+g4*0                    公式九;
进一步的,状态指示值所属的范围不同,对应的状态也不同,其中,状态指示值的取值在可以在[0,10]中,其中,配置第一中间值和第二中间值将所述状态指示值的取值分成三个范围,即[0,第一中间值)、[第一中间值,第二中间值]、(第二中间值, 10],其中三个设定的指示值范围分别对应不同的状态。可选的,所述第一中间值可以在[6.8,7.2]中设定,所述第二中间值可以在[8,9]中设定,这样可以使得确定的网络设备的状态更加准确。
例如,如公式十所示:当得到Z后,当Z属于[0,第一中间值)时,确定所述网络设备处于正常工作状态;当Z属于[第一中间值,第二中间值]时,所述网络设备处于故障隐患状态(包括高风险状态和低风险状态);当Z属于(第二中间值,10]时,所述网络设备处于故障状态。
Figure PCTCN2019078832-appb-000008
这样通过上述方法,所述预警设备就可以通过确定所述状态指示值所属的设定指示值范围,进而确定所述设定指示值范围对应的状态,从而可以确定所述网络设备的状态。
在一种可选的实施方式中,所述预警分析设备在确定了所述网络设备的状态后,可以如图2中的步骤205所示,将确定的所述网络设备的状态通过可视化展示设备展示给用户,以使用户能准确地识别所述网络设备的当前状态,从而根据所述网络设备的状态进行相应维护。
在一种可选的实施方式中,所述预警分析设备可以根据实际需求判断最终确定的所述网络设备的状态是否是用户所急需了解的状态,而只向用户展示用户所关注的状态。例如,用户关注网络设备是否处于故障隐患状态,所述预警分析设备通过本申请实施例提供的方法确定了所述网络设备的状态之后,进一步确定所述网络设备是否处于故障隐患状态,当确定所述网络设备处于故障隐患状态时,将所述网络设备的状态展示给用户,这样可以使得用户提前在所述网络设备故障之前,对所述网络设备进行维护,避免所述网络设备出现故障而导致业务中断,这样可以提高用户业务体验。
采用本申请实施例提供的确定网络设备的状态的方法,预警分析设备获取预设时长内网络设备的多个目标关键绩效指标KPI数据,以及获取多个特征信息,根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素,并将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。这样,通过分析一段时间内的多个目标KPI数据来确定网络设备的状态,而不仅仅通过一个时刻的数据来确定所述网络设备的状态,可以使得确定的网络设备的准确性较高,从而可以减少预警的遗漏。
基于以上实施例,本申请实施例还提供了一种预警分析设备,该预警分析设备应用于如图1所示的网络系统中的预警分析设备,用于实现如图2所示的确定网络设备的状态的方法。参阅图5所示,该预警分析设备500包括:获取单元501和处理单元502,其中:
所述获取单元501,用于获取预设时长内网络设备的多个目标KPI数据,以及获取多个特征信息,其中,任一个特征信息用于表征特征向量中所述特征信息对应的元素的计算方式;
所述处理单元502,用于根据每个特征信息对所述多个目标KPI数据进行处理, 生成每个特征信息对应的元素,将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。
在一种可选的实施方式中,所述处理单元502在根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态时,具体用于:根据所述预警分析模型,对所述特征向量进行分析,确定所述网络设备处于每种状态的概率值;将每个概率值与预设的该概率值对应的状态的基准值相乘,得到多个乘积值,并将所述多个乘积值相加,得到状态指示值;确定所述状态指示值所属的设定指示值范围,将所述设定指示值范围对应的状态作为所述网络设备的状态。
在一种可选的实时方式中,所述获取单元501还用于获取所述网络设备的不同网络设备状态对应的特征向量样本;所述处理器502在根据预设的预警分析模型,对所述特征向量进行分析之前,还用于生成所述预警分析模型:在所述获取单元501获取所述网络设备的不同网络设备状态对应的特征向量样本后,对每种网络设备状态以及该网络设备状态对应的特征向量样本进行逻辑回归处理,得到所述预警分析模型。
在一种可选的实施方式中,所述预警分析设备500还包括:接收单元,用于接收网管设备持续发送的所述网络设备的KPI数据;所述获取单元501在获取所述预设时长内网络设备的多个目标KPI数据时,具体用于:从所述接收单元接收的所述KPI数据中,获取所述预设时长内的所述多个目标KPI数据。
在一种可选的实施方式中,所述网络设备可以为波分设备、路由器、分组传送网设备等。
采用本申请实施例提供的预警分析设备,获取预设时长内网络设备的多个目标关键绩效指标KPI数据,以及获取多个特征信息,根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素,并将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。这样,通过分析一段时间内的多个目标KPI数据来确定网络设备的状态,而不仅仅通过一个时刻的数据来确定所述网络设备的状态,可以使得确定的网络设备的准确性较高,从而可以减少预警的遗漏。
需要说明的是,本申请实施例中对单元的划分是示意性的,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式。在本申请的实施例中的各功能单元可以集成在一个处理单元中,也可以是各个单元单独物理存在,也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现,也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读取存储介质中。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)或处理器(processor)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括:U盘、移动硬盘、只读存储器(read-only memory,ROM)、随机存取存储器(random access  memory,RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
基于以上实施例,本申请实施例还提供了一种预警分析设备,所述预警分析设备应用于如图1所示的网络系统中的预警分析设备,用于实现如图2所示的确定网络设备的状态的方法。参阅图6所示,所述预警分析设备600包括:处理器602和存储器603,其中:
所述处理器602可以是中央处理器(central processing unit,CPU),网络处理器(network processor,NP)或者CPU和NP的组合。所述处理器602还可以进一步包括硬件芯片。上述硬件芯片可以是专用集成电路(application-specific integrated circuit,ASIC),可编程逻辑器件(programmable logic device,PLD)或其组合。上述PLD可以是复杂可编程逻辑器件(complex programmable logic device,CPLD),现场可编程逻辑门阵列(field-programmable gate array,FPGA),通用阵列逻辑(generic array logic,GAL)或其任意组合。
所述处理器602和所述存储器603之间相互连接。可选的,所述处理器602和所述存储器603通过总线604相互连接;所述总线604可以是外设部件互连标准(Peripheral Component Interconnect,PCI)总线或扩展工业标准结构(Extended Industry Standard Architecture,EISA)总线等。所述总线可以分为地址总线、数据总线、控制总线等。为便于表示,图6中仅用一条粗线表示,但并不表示仅有一根总线或一种类型的总线。
所述预警分析设备600在实现如图2所示的确定网络设备的状态的方法时:
所述处理器602,用于获取预设时长内网络设备的多个目标关键绩效指标KPI数据;
获取多个特征信息,任一个特征信息用于表征特征向量中所述特征信息对应的元素的计算方式;
根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素;
所述预警分析设备将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。
所述存储器603,用于存放程序等。具体地,程序可以包括程序代码,该程序代码包括计算机操作指令。所述存储器603可能包括RAM,也可能还包括非易失性存储器(non-volatile memory),例如至少一个磁盘存储器。所述处理器602执行所述存储器603所存放的应用程序,实现上述功能,从而实现如图2所示的确定网络设备的状态的方法。
在一种可选的实施方式中,所述处理器602,在根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态时,具体用于:根据所述预警分析模型,对所述特征向量进行分析,确定所述网络设备处于每种状态的概率值;将每个概率值与预设的该概率值对应的状态的基准值相乘,得到多个乘积值,并将所述多个乘积值相加,得到状态指示值;确定所述状态指示值所属的设定指示值范围,将所述设定指示值范围对应的状态作为所述网络设备的状态。
在一种可选的实施方式中,所述处理器602,在根据预设的预警分析模型,对所 述特征向量进行分析之前,还用于生成所述预警分析模型:获取所述网络设备的不同网络设备状态对应的特征向量样本;对每种网络设备状态以及该网络设备状态对应的特征向量样本进行逻辑回归处理,得到所述预警分析模型。
在一种可选的实施方式中,所述预警分析设备600还包括:通信接口601,用于接收数据;所述处理器602,在获取所述预设时长内网络设备的多个目标KPI数据时,具体用于:控制所述通信接口601接收网管设备持续发送的所述网络设备的KPI数据;从接收的所述KPI数据中,获取所述预设时长内的多个目标KPI数据。
在一种可选的实施方式中,所述网络设备可以为波分设备、路由器、分组传送网设备等。
采用本申请实施例提供的预警分析设备,获取预设时长内网络设备的多个目标关键绩效指标KPI数据,以及获取多个特征信息,根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素,并将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。这样,通过分析一段时间内的多个目标KPI数据来确定网络设备的状态,而不仅仅通过一个时刻的数据来确定所述网络设备的状态,可以使得确定的网络设备的准确性较高,从而可以减少预警的遗漏。
综上所述,通过本申请实施例提供一种确定网络设备的状态的方法及装置,预警分析设备,获取预设时长内网络设备的多个目标关键绩效指标KPI数据,以及获取多个特征信息,根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素,并将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。这样,通过分析一段时间内的多个目标KPI数据来确定网络设备的状态,而不仅仅通过一个时刻的数据来确定所述网络设备的状态,可以使得确定的网络设备的准确性较高,从而可以减少预警的遗漏。
本领域内的技术人员应明白,本申请的实施例可提供为方法、系统、或计算机程序产品。因此,本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个 方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
尽管已描述了本申请的优选实施例,但本领域内的技术人员一旦得知了基本创造性概念,则可对这些实施例做出另外的变更和修改。所以,所附权利要求意欲解释为包括优选实施例以及落入本申请范围的所有变更和修改。
显然,本领域的技术人员可以对本申请实施例进行各种改动和变型而不脱离本申请实施例的范围。这样,倘若本申请实施例的这些修改和变型属于本申请权利要求及其等同技术的范围之内,则本申请也意图包含这些改动和变型在内。

Claims (13)

  1. 一种确定网络设备的状态的方法,其特征在于,包括:
    预警分析设备获取预设时长内网络设备的多个目标关键绩效指标KPI数据;
    所述预警分析设备获取多个特征信息,任一个特征信息用于表征特征向量中所述特征信息对应的元素的计算方式;
    所述预警分析设备根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的元素;
    所述预警分析设备将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。
  2. 如权利要求1所述的方法,其特征在于,所述预警分析设备根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态,包括:
    所述预警分析设备根据所述预警分析模型,对所述特征向量进行分析,确定所述网络设备处于每种状态的概率值;
    所述预警分析设备将每个概率值与预设的该概率值对应的状态的基准值相乘,得到多个乘积值;
    所述预警分析设备将所述多个乘积值相加,得到状态指示值;
    所述预警分析设备确定所述状态指示值所属的设定指示值范围,将所述设定指示值范围对应的状态作为所述网络设备的状态。
  3. 如权利要求1或2所述的方法,其特征在于,所述预警分析设备在根据预设的预警分析模型,对所述特征向量进行分析之前,所述方法还包括:
    所述预警分析设备生成所述预警分析模型:
    所述预警分析设备获取所述网络设备的不同状态对应的特征向量样本;
    所述预警分析设备对每种状态以及该网络设备状态对应的特征向量样本进行逻辑回归处理,得到所述预警分析模型。
  4. 如权利要求1-3任一项所述的方法,其特征在于,所述预警分析设备获取所述预设时长内网络设备的多个目标KPI数据,包括:
    所述预警分析设备接收网管设备持续发送的所述网络设备的KPI数据;
    所述预警分析设备从接收的所述KPI数据中,获取所述预设时长内的多个目标KPI数据。
  5. 如权利要求1-4任一项所述的方法,其特征在于,所述网络设备为以下任一种设备:波分设备、路由器、分组传送网设备。
  6. 一种预警分析设备,其特征在于,包括:
    存储器,用于存储程序指令;
    处理器,用于调用所述存储器中的程序指令以执行下述方法:
    获取预设时长内网络设备的多个目标关键绩效指标KPI数据;
    获取多个特征信息,任一个特征信息用于表征特征向量中所述特征信息对应的元素的计算方式;
    根据每个特征信息对所述多个目标KPI数据进行处理,生成每个特征信息对应的 元素;
    将生成的所述多个特征信息对应的元素组成所述特征向量,并根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态。
  7. 如权利要求6所述的预警分析设备,其特征在于,所述处理器,在根据预设的预警分析模型,对所述特征向量进行分析,确定所述网络设备的状态时,具体用于:
    根据所述预警分析模型,对所述特征向量进行分析,确定所述网络设备处于每种状态的概率值;
    将每个概率值与预设的该概率值对应的状态的基准值相乘,得到多个乘积值;
    将所述多个乘积值相加,得到状态指示值;
    确定所述状态指示值所属的设定指示值范围,将所述设定指示值范围对应的状态作为所述网络设备的状态。
  8. 如权利要求6或7所述的预警分析设备,其特征在于,所述处理器,在根据预设的预警分析模型,对所述特征向量进行分析之前,还用于:
    生成所述预警分析模型:
    获取所述网络设备的不同状态对应的特征向量样本;
    对每种网络设备状态以及该状态对应的特征向量样本进行逻辑回归处理,得到所述预警分析模型。
  9. 如权利要求6-8任一项所述的预警分析设备,其特征在于,所述预警分析设备还包括:
    通信接口,用于接收数据;
    所述处理器,在获取所述预设时长内网络设备的多个目标KPI数据时,具体用于:
    控制所述通信接口接收网管设备持续发送的所述网络设备的KPI数据;
    从接收的所述KPI数据中,获取所述预设时长内的多个目标KPI数据。
  10. 如权利要求6-9任一项所述的预警分析设备,其特征在于,所述网络设备为以下任一种设备:波分设备、路由器、分组传送网设备。
  11. 一种计算机存储介质,其特征在于,所述计算机存储介质中存储有计算机可执行指令,所述计算机可执行指令在被所述计算机调用时用于使所述计算机执行权利要求1-5任一项所述的方法。
  12. 一种包含指令的计算机程序产品,其特征在于,当所述计算机程序产品在计算机上运行时,使得计算机执行权利要求1-5任一项所述的方法。
  13. 一种芯片,其特征在于,所述芯片与存储器相连,用于读取并执行所述存储器中存储的程序指令,以实现权利要求1-5任一项所述的方法。
PCT/CN2019/078832 2018-03-22 2019-03-20 一种确定网络设备的状态的方法及装置 WO2019179457A1 (zh)

Priority Applications (6)

Application Number Priority Date Filing Date Title
RU2020134150A RU2781813C2 (ru) 2018-03-22 2019-03-20 Способ и устройство для определения состояния сетевого устройства
KR1020207029573A KR102455332B1 (ko) 2018-03-22 2019-03-20 네트워크 장치의 상태를 결정하는 방법 및 장치
CA3094557A CA3094557C (en) 2018-03-22 2019-03-20 Method and apparatus for determining status of network device
EP19770460.4A EP3761566B1 (en) 2018-03-22 2019-03-20 Method and apparatus for determining state of network device
JP2020550850A JP7081741B2 (ja) 2018-03-22 2019-03-20 ネットワークデバイスの状態を判定するための方法及び装置
US17/027,319 US11405294B2 (en) 2018-03-22 2020-09-21 Method and apparatus for determining status of network device

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810241478.1A CN110300008B (zh) 2018-03-22 2018-03-22 一种确定网络设备的状态的方法及装置
CN201810241478.1 2018-03-22

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/027,319 Continuation US11405294B2 (en) 2018-03-22 2020-09-21 Method and apparatus for determining status of network device

Publications (1)

Publication Number Publication Date
WO2019179457A1 true WO2019179457A1 (zh) 2019-09-26

Family

ID=67986753

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/078832 WO2019179457A1 (zh) 2018-03-22 2019-03-20 一种确定网络设备的状态的方法及装置

Country Status (7)

Country Link
US (1) US11405294B2 (zh)
EP (1) EP3761566B1 (zh)
JP (1) JP7081741B2 (zh)
KR (1) KR102455332B1 (zh)
CN (1) CN110300008B (zh)
CA (1) CA3094557C (zh)
WO (1) WO2019179457A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2022028633A (ja) * 2020-07-30 2022-02-16 ジオ プラットフォームズ リミティド 重要業績評価指標の階層的計算のためのシステム及び方法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113535444B (zh) * 2020-04-14 2023-11-03 中国移动通信集团浙江有限公司 异动检测方法、装置、计算设备及计算机存储介质
US11799568B2 (en) * 2020-12-10 2023-10-24 Verizon Patent And Licensing Inc. Systems and methods for optimizing a network based on weather events
CN115242669B (zh) * 2022-06-30 2023-10-03 北京华顺信安科技有限公司 一种网络质量监测方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015057119A1 (en) * 2013-10-18 2015-04-23 Telefonaktiebolaget L M Ericsson (Publ) Alarm prediction in a telecommunication network
CN104881436A (zh) * 2015-05-04 2015-09-02 中国南方电网有限责任公司 一种基于大数据的电力通信设备性能分析方法及装置
CN106952029A (zh) * 2017-03-15 2017-07-14 中国电力科学研究院 一种用于对变电设备状态监测装置进行评价的方法及系统

Family Cites Families (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100504835B1 (ko) 2003-05-15 2005-07-29 엘지전자 주식회사 이동 통신 단말기의 지역시간 보정 방법
EP1869937B1 (en) 2005-04-14 2018-03-28 LG Electronics Inc. -1- A method of reconfiguring an internet protocol address in handover between heterogeneous networks
US7590733B2 (en) 2005-09-14 2009-09-15 Infoexpress, Inc. Dynamic address assignment for access control on DHCP networks
KR100656358B1 (ko) 2005-10-25 2006-12-11 한국전자통신연구원 Mobile IP 환경에서의 핸드오버 수행 방법
WO2007078663A2 (en) 2005-12-16 2007-07-12 Interdigital Technology Corporation Mobility middleware architecture for multiple radio access technology apparatus
KR100739807B1 (ko) 2006-02-06 2007-07-13 삼성전자주식회사 Dhcp를 이용한 핸드오버 정보 검색 및 획득 방법 및장치
US7647283B2 (en) * 2006-12-31 2010-01-12 Ektimisi Semiotics Holdings, Llc Method, system, and computer program product for adaptively learning user preferences for smart services
CN101729127A (zh) 2008-10-17 2010-06-09 中兴通讯股份有限公司 时间同步节点和方法
CN101442799B (zh) 2008-12-17 2010-12-01 广州市新邮通信设备有限公司 一种实现终端同步网络侧系统时间的方法、系统及装置
CN102056283B (zh) 2009-10-27 2013-10-16 电信科学技术研究院 移动通信系统中时间同步的方法、系统及装置
CN102056284A (zh) 2009-10-27 2011-05-11 大唐移动通信设备有限公司 时间同步方法、系统和设备
CN102083194B (zh) 2009-11-30 2014-12-03 电信科学技术研究院 时间信息发送与时间同步方法、系统和设备
JP5051252B2 (ja) 2010-02-18 2012-10-17 沖電気工業株式会社 ネットワーク障害検出システム
CN102026230A (zh) * 2010-12-20 2011-04-20 中兴通讯股份有限公司 Cdma网络数据业务质量监控的方法及装置
US20120179283A1 (en) * 2011-01-10 2012-07-12 International Business Machines Corporation Managing a performance of solar devices throughout an end-to-end manufacturing process
US9167463B2 (en) 2011-09-02 2015-10-20 Telcordia Technologies, Inc. Communication node operable to estimate faults in an ad hoc network and method of performing the same
CN103178990A (zh) * 2011-12-20 2013-06-26 中国移动通信集团青海有限公司 一种网络设备性能监控方法及网络管理系统
CN103796296B (zh) 2012-11-02 2018-05-08 中国移动通信集团公司 长期演进系统中的时间同步方法、用户设备及基站
CN104010382B (zh) 2013-02-25 2019-02-01 中兴通讯股份有限公司 数据传输方法、装置及系统
EP3108685A4 (en) 2014-02-17 2017-02-15 Telefonaktiebolaget LM Ericsson (publ) Service failure in communications networks
CN104469833B (zh) 2015-01-08 2018-08-21 西安电子科技大学 一种基于用户感知的异构网络运维管理方法
US10261851B2 (en) * 2015-01-23 2019-04-16 Lightbend, Inc. Anomaly detection using circumstance-specific detectors
US10402511B2 (en) * 2015-12-15 2019-09-03 Hitachi, Ltd. System for maintenance recommendation based on performance degradation modeling and monitoring
RU2630415C2 (ru) 2016-02-20 2017-09-07 Открытое Акционерное Общество "Информационные Технологии И Коммуникационные Системы" Способ обнаружения аномальной работы сетевого сервера (варианты)
CN107360048A (zh) 2016-05-09 2017-11-17 富士通株式会社 节点性能评估方法、装置和系统
WO2018011742A1 (en) * 2016-07-13 2018-01-18 Incelligent P.C. Early warning and recommendation system for the proactive management of wireless broadband networks
US11223668B2 (en) * 2017-01-12 2022-01-11 Telefonaktiebolaget Lm Ericsson (Publ) Anomaly detection of media event sequences

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015057119A1 (en) * 2013-10-18 2015-04-23 Telefonaktiebolaget L M Ericsson (Publ) Alarm prediction in a telecommunication network
CN104881436A (zh) * 2015-05-04 2015-09-02 中国南方电网有限责任公司 一种基于大数据的电力通信设备性能分析方法及装置
CN106952029A (zh) * 2017-03-15 2017-07-14 中国电力科学研究院 一种用于对变电设备状态监测装置进行评价的方法及系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2022028633A (ja) * 2020-07-30 2022-02-16 ジオ プラットフォームズ リミティド 重要業績評価指標の階層的計算のためのシステム及び方法
JP7279119B2 (ja) 2020-07-30 2023-05-22 ジオ プラットフォームズ リミティド 重要業績評価指標の階層的計算のためのシステム及び方法

Also Published As

Publication number Publication date
US20210006481A1 (en) 2021-01-07
EP3761566A4 (en) 2021-04-07
JP2021516511A (ja) 2021-07-01
RU2020134150A3 (zh) 2022-04-22
CA3094557A1 (en) 2019-09-26
US11405294B2 (en) 2022-08-02
CN110300008B (zh) 2021-03-23
EP3761566A1 (en) 2021-01-06
KR102455332B1 (ko) 2022-10-14
KR20200128144A (ko) 2020-11-11
RU2020134150A (ru) 2022-04-22
CA3094557C (en) 2024-03-26
CN110300008A (zh) 2019-10-01
EP3761566B1 (en) 2023-10-25
JP7081741B2 (ja) 2022-06-07

Similar Documents

Publication Publication Date Title
WO2019179457A1 (zh) 一种确定网络设备的状态的方法及装置
US11792217B2 (en) Systems and methods to detect abnormal behavior in networks
WO2018103453A1 (zh) 检测网络的方法和装置
EP3211831A1 (en) N-tiered eurt breakdown graph for problem domain isolation
US9692671B2 (en) Method and apparatus for automatically determining causes of service quality degradation
EP3863223A1 (en) Method and device for training service quality evaluation model
US9524223B2 (en) Performance metrics of a computer system
US10467087B2 (en) Plato anomaly detection
CN109088775B (zh) 异常监控方法、装置以及服务器
EP3595347B1 (en) Method and device for detecting health state of network element
JP2019507454A (ja) アプリケーションの実行中に観察される問題の根本原因を特定する方法
US10616040B2 (en) Managing network alarms
US11501106B2 (en) Anomaly factor estimation device, anomaly factor estimation method, and storage medium
JP2018148350A (ja) 閾値決定装置、閾値決定方法及びプログラム
WO2019041870A1 (zh) 实现故障原因定位的方法、装置及存储介质
CN115529595A (zh) 一种日志数据的异常检测方法、装置、设备及介质
CN111654405B (zh) 通信链路的故障节点方法、装置、设备及存储介质
US20170302506A1 (en) Methods and apparatus for fault detection
CN112380073A (zh) 一种故障位置的检测方法、装置及可读存储介质
US20200213203A1 (en) Dynamic network health monitoring using predictive functions
RU2781813C2 (ru) Способ и устройство для определения состояния сетевого устройства
Priovolos et al. Using anomaly detection techniques for securing 5G infrastructure and applications
US9311210B1 (en) Methods and apparatus for fault detection
US11093315B2 (en) Systems and methods for detecting a fault or a model mismatch
CN117692300A (zh) 一种故障根因定位方法、装置、电子设备及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19770460

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2020550850

Country of ref document: JP

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 3094557

Country of ref document: CA

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2019770460

Country of ref document: EP

Effective date: 20201001

ENP Entry into the national phase

Ref document number: 20207029573

Country of ref document: KR

Kind code of ref document: A