WO2013010404A1 - 设备性能预测处理方法及装置 - Google Patents

设备性能预测处理方法及装置 Download PDF

Info

Publication number
WO2013010404A1
WO2013010404A1 PCT/CN2012/076112 CN2012076112W WO2013010404A1 WO 2013010404 A1 WO2013010404 A1 WO 2013010404A1 CN 2012076112 W CN2012076112 W CN 2012076112W WO 2013010404 A1 WO2013010404 A1 WO 2013010404A1
Authority
WO
WIPO (PCT)
Prior art keywords
performance
data
module
performance data
time point
Prior art date
Application number
PCT/CN2012/076112
Other languages
English (en)
French (fr)
Inventor
燕红锁
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2013010404A1 publication Critical patent/WO2013010404A1/zh

Links

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
    • H04L41/147Network analysis or design for predicting network behaviour
    • 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 invention relates to the field of telecommunications network management technologies, and in particular, to a method and apparatus for predictive analysis processing of telecommunication equipment performance indicators. Background technique
  • the performance of telecommunication equipment is a management indicator that operators pay attention to.
  • the analysis and prediction of telecommunication network management on the future trend of telecommunication equipment performance is an important task of network management. It is especially important to use the existing performance data to predict the future performance trends of telecommunication equipment clearly and accurately, and provide practical guidance information for the subsequent operation of telecommunication equipment.
  • the existing telecommunication network management system predicts the performance of specific performance indicators, lacks practical and efficient means, and cannot provide important guidance and reference for the subsequent maintenance and processing of telecommunication equipment, thereby reducing network management efficiency.
  • the main object of the present invention is to provide a method and device for predicting device performance, which aims to improve the practicality and efficiency of telecommunication equipment performance prediction and network management efficiency.
  • the present invention provides a device performance prediction processing method, the method comprising:
  • the historical performance data after the screening is analyzed and calculated, and the time point of the performance warning value and the performance limit value of the device after the current time point is obtained, and the current time point is obtained. Performance data for the device.
  • the analyzing and calculating the historical performance data after the screening according to the preset performance early warning threshold and the performance exceeding threshold, according to the preset performance early warning threshold and the performance exceeding threshold, The historical performance data after the screening is analyzed and calculated, and the time point of the performance warning value and the performance limit value of the device after the current time point is obtained, and the performance data of the device after the current time point is obtained, including:
  • the historical performance data is analyzed by using a multiple curve fitting interpolation algorithm to calculate curve parameters
  • the time point of the performance warning value and the performance limit value of the device after the current time point is calculated, and the performance data of the device after the current time point is obtained.
  • the collecting historical performance data comprises: 10 to 100 sampling points.
  • the screening of historical performance data collected includes:
  • the performance warning value and the performance violation alarm value in the collected historical data are eliminated.
  • the screening of historical performance data collected includes:
  • the count is detected from the current time until the performance data of the detection point is the performance warning value or the performance limit alarm value.
  • the curve fitting interpolation algorithm is: a least squares method or a Lagrangian curve fitting algorithm.
  • the method further includes:
  • the performance data of the device after the current time point is displayed through the list and graph.
  • the method before the collecting the historical performance data from the device, the method further includes: setting the performance early warning threshold and the performance exceeding threshold.
  • the present invention also provides a device performance prediction processing device, including: a data acquisition module, a data filtering module, and a calculation module;
  • Data acquisition module used to periodically collect historical performance data from the device, for the data screen
  • the selection module provides the collected time performance data
  • a data filtering module configured to filter historical performance data collected by the data collection module, and send the filtered time performance data to the computing module
  • the calculation module is configured to analyze and calculate the filtered historical performance data sent by the data screening module according to the preset performance warning threshold and the performance threshold, and obtain the performance warning value and performance limit of the device after the current time point. The point in time of the value, and get the performance data of the device after the current point in time.
  • the calculation module includes: a curve parameter calculation unit and an alarm time calculation unit;
  • a curve parameter calculation unit configured to analyze the historical performance data by using a multiple curve fitting interpolation algorithm, calculate a curve parameter, and send the calculated curve parameter to the alarm time calculation unit;
  • the alarm time calculation unit is configured to calculate the performance warning value and the performance limit value of the device after the current time point according to the curve parameter sent by the curve parameter calculation unit and the preset performance warning threshold and the performance limit threshold. Time point, and get the performance data of the device after the current time point.
  • the device further includes: a prediction result display module, configured to display, by using a list and a graphic, performance data of the device after the current time point sent by the calculation module;
  • a prediction result display module configured to display, by using a list and a graphic, performance data of the device after the current time point sent by the calculation module;
  • the calculating module is further configured to send performance data of the device after the current time point to the prediction result display module.
  • the device further includes: a setting module, configured to provide a preset performance threshold and a performance threshold for the data screening module and the computing module;
  • the calculating module is further configured to receive and save the preset performance warning threshold and the performance limit threshold sent by the setting module;
  • the data screening module is further configured to receive and save the presetness sent by the setting module It can alert the threshold and the performance limit threshold.
  • the device performance prediction processing method and device provided by the invention analyzes and calculates the future performance of the telecommunication device by filtering the historical performance data and using a typical data prediction analysis algorithm based on the historical performance data of the telecommunication device.
  • the trend and the specific performance warning and performance time limit which provide important guidance and reference for the subsequent maintenance and processing of telecommunication equipment, improve the real-time and high efficiency of telecommunication equipment performance prediction, and further improve the network management efficiency.
  • FIG. 1 is a schematic flow chart of an embodiment of a device performance prediction processing method according to the present invention.
  • FIG. 2 is a graph showing a trend of device performance data in an embodiment of a device performance prediction processing method according to the present invention
  • FIG. 3 is a schematic diagram of a time point of obtaining a performance warning value and a performance limit value of a device after the current time point in an embodiment of the device performance prediction processing method of the present invention, and obtaining a performance data flow chart of the device after the current time point;
  • FIG. 4 is a schematic flow chart of another embodiment of a device performance prediction processing method according to the present invention.
  • FIG. 5 is a schematic structural diagram of an apparatus performance prediction processing device according to an embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a calculation module in an embodiment of a device performance prediction processing apparatus according to the present invention.
  • FIG. 7 is a schematic structural diagram of another embodiment of a device performance prediction processing apparatus according to the present invention.
  • the solution of the embodiment of the present invention mainly includes: collecting historical performance data of the telecommunication device, filtering the historical performance data, and using a typical data prediction analysis algorithm to analyze and calculate the future performance trend of the telecommunication device and the specific performance warning and performance limit.
  • Time point, for telecommunication equipment Subsequent maintenance processing provides important guidance and reference to improve the real-time and high efficiency of telecommunication equipment performance prediction.
  • an embodiment of the present invention provides a device performance prediction processing method, including: Step S101: Timing collects historical performance data from a telecommunication device (hereinafter referred to as a device). In order to perform real-time analysis and prediction on the performance of the device, this embodiment first needs to periodically collect historical performance data of specific performance indicators from the device.
  • the performance granularity can be collected from the device by using the performance granularity of 15 minutes or 24 hours, and the data is collected as the original sampling point data, that is, the data is collected every 15 minutes or 24 hours. Therefore, the sample point data is performance data corresponding to a certain point in time.
  • the number of sampling points of historical performance data should be strictly limited. It is necessary to ensure that the number of sampling points is between 10 and 100 sampling points.
  • Step S102 Filter the collected historical performance data.
  • the collected historical performance data needs to be selected, the sampling point data that does not meet the requirements are filtered, and the filtered historical performance data is selected as a subsequent analysis and calculation for performance prediction.
  • Basic data After the historical performance data of the device is collected, the collected historical performance data needs to be selected, the sampling point data that does not meet the requirements are filtered, and the filtered historical performance data is selected as a subsequent analysis and calculation for performance prediction.
  • Basic data After the historical performance data of the device is collected, the collected historical performance data needs to be selected, the sampling point data that does not meet the requirements are filtered, and the filtered historical performance data is selected as a subsequent analysis and calculation for performance prediction.
  • the principle of the screening is: if the performance warning value and the performance exceeding the alarm value in the historical performance data, the culling is required; the performance warning value and the performance violation alarm value exceed a preset performance warning. Performance data for threshold and performance thresholds.
  • the principle of the screening may also be determined according to the number of historical performance data collected, specifically: for a series of historical performance data collected, counting from the current time until the count reaches the required number; or from the current The time is forwarded to detect the count until the performance data of the detection point is the performance warning value or the performance violation alarm value.
  • Step S103 Perform analysis and calculation on the filtered historical performance data according to the preset performance warning threshold and the performance limit threshold, and obtain the performance warning value of the device after the current time point and The point in time when the performance exceeds the limit, and the performance data of the device after the current time point is obtained.
  • the historical performance data after the screening is analyzed and calculated as: using a plurality of curve fitting interpolation algorithms for analysis, and using the time points and performance data of the collected historical performance data, the fitting curve parameters are calculated.
  • the performance warning threshold and the performance limit threshold are inversely solved, and the performance warning value of the device after the current time point is calculated. And the time point when the performance exceeds the limit, and the performance data of the device after the current time point is obtained.
  • the curve fitting interpolation algorithm can adopt existing relatively advanced algorithms, such as: least squares method and Lagrangian curve fitting algorithm.
  • the calculation result is indefinite, indicating that the performance index of the device is not degraded.
  • FIG. 2 is a trend graph of device performance data in an embodiment of the device performance prediction processing method of the present invention.
  • the abscissa is the time point
  • the ordinate is the performance data
  • the 0 point is the origin, that is, the current time.
  • the data between point A and point 0 is historical performance data.
  • Point B is the time point at which the performance warning threshold occurs.
  • the corresponding performance data is the performance warning threshold.
  • the curve between BC indicates the threshold warning threshold, which can be displayed in yellow.
  • C point is the time point when the performance exceeds the limit.
  • the corresponding performance data is the performance limit threshold.
  • the curve after C point indicates the performance limit threshold segment, which can be displayed in red.
  • step 103 includes:
  • Step 1031 The historical performance data is analyzed by using a multiple curve fitting interpolation algorithm, and the curve parameters are calculated.
  • the curve fitting interpolation algorithm may adopt a least square method or a Lagrangian curve fitting algorithm or the like.
  • Step 1032 Calculate the performance warning value and the performance limit value of the device after the current time point according to the curve parameter and the preset performance warning threshold and the performance limit threshold, and obtain Performance data for devices after the current point in time.
  • the present embodiment Based on the historical performance data of the collection device, the present embodiment analyzes and calculates the future performance trend of the telecommunication equipment and the specific performance warning and performance time limit by filtering the historical performance data and using a typical data prediction analysis algorithm. Therefore, it provides important guidance and reference for the subsequent maintenance and processing of telecommunication equipment, improves the real-time and high-efficiency of telecommunication equipment performance prediction, and further improves the efficiency of network management.
  • another embodiment of the present invention provides a device performance prediction processing method.
  • the method further includes:
  • Step S100 Set a performance early warning threshold and a performance limit threshold.
  • step 103 the method further includes:
  • Step S104 Display performance data of the device after the current time point by using a list and a graph.
  • the difference between this embodiment and the foregoing embodiment is that in this embodiment, the performance warning threshold and the performance limit threshold need to be set and saved locally.
  • the prediction result of the performance prediction trend of the device is displayed to the user through the list and the graphic.
  • the prediction results are displayed in a list. If you select a row of data, you can use the "Performance Trend" to display the graph window to further view the performance trend of specific performance indicators.
  • the abscissa is the time point information
  • the ordinate is the performance data
  • the origin is the performance data corresponding to the current time and the current time.
  • the data before the origin is the performance data of the historical performance data sampling point
  • the data after the origin is The time performance data in the predicted performance period
  • the performance data exceeding the performance warning threshold is shown in yellow
  • the curve of the performance data exceeding the performance limit threshold is indicated in red, so that the overall effect of the display is very obvious.
  • the user can predict the performance of the device, performance warning, performance violation, and other different sections through the performance prediction trend of the device, and know the current performance status of the device in real time, so that it can be better
  • the telecommunication equipment performs subsequent maintenance and processing, provides practical guidance and suggestions for network operations, and improves the efficiency of network operations.
  • an embodiment of the present invention provides a device performance prediction processing device, including: a data collection module 501, a data filtering module 502, and a calculation module 503, where:
  • the data collection module 501 is configured to periodically collect historical performance data from the device, and provide the collected data performance function for the data filtering module 502.
  • the data filtering module 502 is configured to filter the collected historical performance data sent by the data collecting module 501, and send the filtered time performance data to the computing module 503;
  • the calculating module 503 is configured to analyze and calculate the filtered historical performance data sent by the data screening module 502 according to the preset performance early warning threshold and the performance exceeding threshold, and obtain the performance warning value and performance of the device after the current time point. The time point of the limit value is obtained, and the performance data of the device after the current time point is obtained.
  • the embodiment first collects historical performance data of specific performance indicators from the device through the data collection module 501.
  • the performance granularity can be collected from the device by using the 15-minute or 24-hour performance granularity, and the data is collected as the original sampling point data, that is, the data is collected every 15 minutes or 24 hours. Therefore, the sample point data is performance data corresponding to a certain point in time.
  • the number of sampling points of historical performance data should be strictly limited. It is necessary to ensure that the number of sampling points is between 10 and 100 sampling points.
  • the collected historical performance data is selected by the data screening module 502, and the sampling point data that does not meet the requirements is filtered, and the filtered historical performance data is selected as a subsequent analysis and calculation performance.
  • the underlying data of the forecast is selected by the data screening module 502, and the sampling point data that does not meet the requirements.
  • the principle of the screening is: if the performance warning value and the performance exceeding the alarm value in the historical performance data, the culling is needed; the performance warning value and the performance exceeding the alarm value are super Performance data for pre-set performance warning thresholds and performance limit thresholds.
  • the principle of the screening may also be determined according to the number of historical performance data collected, specifically: for a series of historical performance data collected, counting from the current time until the count reaches the required number; or from the current The time is forwarded to detect the count until the performance data of the detection point is the performance warning value or the performance violation alarm value.
  • the calculated historical performance data is analyzed and calculated by the calculation module 503, and the analysis and calculation process is specifically as follows: using a multi-curve fitting interpolation algorithm for analysis, using the time point and performance data of the collected historical performance data, Fit the specific parameters of the curve. After that, based on the calculated curve parameters, combined with the pre-set thresholds and the thresholds of the preset performance indicators, the performance warning threshold and the performance limit threshold are inversely solved, and the performance warning value of the device after the current time point is calculated. And the time point when the performance exceeds the limit, and the performance data of the device after the current time point is obtained.
  • the curve fitting interpolation algorithm can adopt existing relatively advanced algorithms, such as: least square method and Lagrangian curve fitting algorithm.
  • the calculation result is indefinite, indicating that the performance index of the device is not degraded.
  • the trend graph of the device performance data is shown in Figure 2.
  • the abscissa is the time point, the ordinate is the performance data, and the 0 point is the origin, that is, the current time.
  • the data between point A and point 0 is historical performance data.
  • Point B is the time point at which the warning threshold occurs.
  • the corresponding performance data is the performance warning threshold.
  • the curve between BC indicates the threshold warning threshold, which can be displayed in yellow.
  • C point is the time point when the performance exceeds the limit.
  • the performance data is the performance limit threshold, and the curve after the C point indicates the performance limit threshold segment, which can be displayed in red.
  • the calculation module 503 includes: a curve parameter calculation unit 5031 and an alarm time calculation unit 5032;
  • Curve parameter calculation unit 5031 which is used for multi-curve fitting interpolation algorithm for historical performance The data is analyzed, the curve parameters are calculated, and the calculated curve parameters are sent to the alarm time calculation unit 5032;
  • the alarm time calculation unit 5032 is configured to calculate, according to the curve parameter sent by the curve parameter calculation unit 5031, the preset performance warning threshold and the performance limit threshold, and calculate the performance warning value and performance limit of the device after the current time point. The point in time of the value, and get the performance data of the device after the current point in time.
  • another embodiment of the present invention provides a device performance prediction processing device.
  • the device further includes: a setting module 500, configured to provide a preset for the data filtering module 502 and the computing module 503. The performance warning threshold and the performance threshold are exceeded; correspondingly, the calculating module 503 is further configured to receive and save the preset performance warning threshold and the performance threshold threshold sent by the setting module 500; The module 502 is further configured to receive and save the preset performance warning threshold and the performance limit threshold sent by the setting module 500.
  • the method further includes: a prediction result display module 504, configured to display, by using a list and a graphic, a device after the current time point sent by the calculation module 503.
  • the calculation module 503 is further configured to send the performance data of the device after the current time point to the prediction result display module 504.
  • the difference between this embodiment and the foregoing embodiment is that the setting module 500 sets the performance warning threshold and the performance limit threshold, and saves it locally.
  • the prediction result display module 504 passes the prediction result of the performance prediction trend of the device through the list and the graphic two. Ways to show the user.
  • the prediction results are displayed in a list mode. If the user selects a certain row of data, the performance window can be displayed through the "performance trend" to further view the performance trend of the specific performance indicators.
  • the abscissa is time point information
  • the ordinate is Performance data
  • the origin is the performance data corresponding to the current time and the current time.
  • the data before the origin is the performance data of the historical performance data sampling point
  • the data after the origin is the performance data of the time in the predicted performance period, exceeding the performance of the performance warning threshold.
  • the curve of the data is shown in yellow, and the curve of performance data that exceeds the performance limit threshold is shown in red, making the overall effect of the display very obvious.
  • the device performance prediction processing device provided by the present invention can be installed in a network management device as a logic module or software.
  • the device performance prediction processing method and device analyzes and calculates the future performance trend of the telecommunication device by filtering historical performance data and using a typical data prediction analysis algorithm based on the historical performance data of the telecommunication device.
  • the specific performance warning and performance time limit which provides important guidance and reference for the subsequent maintenance and processing of telecommunication equipment, improves the real-time and high efficiency of telecommunication equipment performance prediction, and further improves the network management efficiency.

Abstract

本发明公开了一种设备性能预测处理方法,包括:定时从设备上采集历史性能数据;对采集的历史性能数据进行筛选;根据预设的性能预警门限及性能越限门限,对筛选后的历史性能数据进行分析计算,得到当前时间点之后的设备出现性能预警值及性能越限值的时间点,并得到当前时间点之后的设备的性能数据。本发明还公开了一种设备性能预测处理装置,采用本发明能提高电信设备性能预测的实时与高效性,进一步提高网管效率。

Description

设备性能预测处理方法及装置
技术领域
本发明涉及电信网管技术领域, 尤其涉及一种用于电信设备性能指标 的预测分析处理的方法及装置。 背景技术
目前, 电信设备的性能是运营商重点关注的管理指标, 电信网管对电 信设备性能未来趋势的分析与预测是网络管理的重要工作。 如何清晰、 准 确的利用现有性能数据对电信设备的未来性能趋势进行预测 , 为电信设备 的后续运营提供实用的指导信息, 显得尤为重要。
现有的电信网管对于具体性能指标的性能预测, 缺乏实用与高效的手 段, 无法为电信设备的后续维护处理提供重要指导和参考, 从而降低网管 效率。 发明内容
本发明的主要目的在于提供一种设备性能预测处理方法及装置, 旨在 提高电信设备性能预测的实用与高效性以及网管效率。
为了达到上述目的, 本发明提出一种设备性能预测处理方法, 该方法 包括:
定时从设备上采集历史性能数据;
对采集的历史性能数据进行筛选;
根据预设的性能预警门限及性能越限门限, 对筛选后的历史性能数据 进行分析计算, 得到当前时间点之后的设备出现性能预警值及性能越限值 的时间点, 并得到当前时间点之后的设备的性能数据。
优选地, 所述根据预设的性能预警门限及性能越限门限, 对筛选后的 历史性能数据进行分析计算, 根据预设的性能预警门限及性能越限门限, 对筛选后的历史性能数据进行分析计算, 得到当前时间点之后的设备出现 性能预警值及性能越限值的时间点, 并得到当前时间点之后的设备的性能 数据, 包括:
采用多次曲线拟合插值算法对所述历史性能数据进行分析, 计算曲线 参数;
根据所述曲线参数及预设的性能预警门限及性能越限门限, 计算得到 当前时间点之后的设备出现性能预警值及性能越限值的时间点, 并得到当 前时间点之后的设备的性能数据。
优选地, 所述采集历史性能数据包括: 10到 100个采样点。
优选地, 所述对采集的历史性能数据进行筛选, 包括:
剔除采集的历史性数据中的性能预警值及性能越限告警值。
优选地, 所述对采集的历史性能数据进行筛选, 包括:
从当前时间往前计数, 直到计数值达到要求的个数;
或者, 从当前时间往前检测计数, 直到检测点的性能数据为性能预警 值或性能越限告警值。
优选地, 所述曲线拟合插值算法为: 最小二乘法或拉格朗日曲线拟合 算法。
优选地, 所述得到当前时间点之后的设备的性能数据之后, 该方法还 包括:
通过列表及图形展示当前时间点之后的设备的性能数据。
优选地, 所述定时从设备上采集历史性能数据之前, 该方法还包括: 设置所述性能预警门限及性能越限门限。
本发明还提出一种设备性能预测处理装置, 包括: 数据采集模块、 数 据筛选模块和计算模块; 其中,
数据采集模块, 用于定时从所在设备上采集历史性能数据, 为数据筛 选模块提供采集的历时性能数据;
数据筛选模块, 用于对数据采集模块发来的采集的历史性能数据进行 筛选, 将筛选后的历时性能数据发送给计算模块;
计算模块, 用于根据预设的性能预警门限及性能越限门限, 对数据筛 选模块发来的筛选后的历史性能数据进行分析计算, 得到当前时间点之后 的设备出现性能预警值及性能越限值的时间点, 并得到当前时间点之后的 设备的性能数据。
优选地, 所述计算模块包括: 曲线参数计算单元和告警时间计算单元; 其中,
曲线参数计算单元, 用于采用多次曲线拟合插值算法对所述历史性能 数据进行分析, 计算曲线参数, 将计算得出的曲线参数发送给告警时间计 算单元;
告警时间计算单元, 用于根据曲线参数计算单元发来的所述曲线参数 及预设的性能预警门限及性能越限门限, 计算得到当前时间点之后的设备 出现性能预警值及性能越限值的时间点, 并得到当前时间点之后的设备的 性能数据。
优选地, 该装置还包括: 预测结果展示模块, 用于通过列表及图形展 示计算模块发来的当前时间点之后的设备的性能数据;
相应的, 所述计算模块, 还用于将当前时间点之后的设备的性能数据 发给预测结果展示模块。
优选地, 该装置还包括: 设置模块, 用于为数据筛选模块和计算模块 提供预设的所述性能预警门限及性能越限门限;
相应的, 所述计算模块, 还用于接收并保存设置模块发来的预设的所 述性能预警门限及性能越限门限;
所述数据筛选模块, 还用于接收并保存设置模块发来的预设的所述性 能预警门限及性能越限门限。
本发明提出的一种设备性能预测处理方法及装置, 在采集电信设备历 史性能数据的基础上, 通过对历史性能数据进行过滤, 并采用典型的数据 预测分析算法, 分析计算出电信设备未来的性能趋势以及具体的性能预警 和性能越限时间点, 从而为电信设备的后续维护处理提供重要指导和参考, 提高电信设备性能预测的实时与高效性, 进一步提高网管效率。 附图说明
图 1是本发明设备性能预测处理方法一实施例流程示意图;
图 2是本发明设备性能预测处理方法一实施例中设备性能数据的趋势 曲线图;
图 3是本发明设备性能预测处理方法一实施例中得到当前时间点之后 的设备出现性能预警值及性能越限值的时间点, 并得到当前时间点之后的 设备的性能数据流程示意图;
图 4是本发明设备性能预测处理方法另一实施例流程示意图; 图 5是本发明设备性能预测处理装置一实施例结构示意图;
图 6是本发明设备性能预测处理装置一实施例中计算模块的结构示意 图;
图 7是本发明设备性能预测处理装置另一实施例结构示意图。
为了使本发明的技术方案更加清楚、 明了, 下面将结合附图作进一步 评述。 具体实施方式
本发明实施例解决方案主要是: 采集电信设备历史性能数据, 对历史 性能数据进行过滤, 并采用典型的数据预测分析算法, 分析计算出电信设 备未来的性能趋势以及具体的性能预警和性能越限时间点, 为电信设备的 后续维护处理提供重要指导和参考, 提高电信设备性能预测的实时与高效 性。
如图 1所示, 本发明一实施例提出一种设备性能预测处理方法, 包括: 步驟 S101 : 定时从电信设备(以下简称设备)上采集历史性能数据。 为了对设备的性能进行实时的分析与预测, 本实施例首先需要定时从 设备上采集具体性能指标的历史性能数据。
具体地, 根据当前设备的实际应用场景, 可以采用 15分钟或 24小时 性能粒度, 定时从设备上采集历史性能数据, 做为原始的采样点数据, 即 每隔 15分钟或 24小时采集一次数据, 因此, 该采样点数据为对应某一时 间点的性能数据。
为保证后续性能预测算法的准确性, 应严格限制历史性能数据的采样 点数量, 必须保证采样点数量在 10 ~ 100个采样点之间。
步驟 S102: 对采集的历史性能数据进行筛选。
具体地, 在采集到设备的历史性能数据后, 需要对采集的历史性能数 据进行歸选, 对不符合要求的采样点数据进行过滤, 歸选过滤后的历史性 能数据作为后续分析计算进行性能预测的基础数据。
这里, 所述筛选的原则是: 如果历史性能数据中性能预警值和性能越 限告警值, 则需要进行剔除; 所述性能预警值和所述性能越限告警值为超 过预先设定的性能预警门限及性能越限门限的性能数据。
另外, 所述筛选的原则还可以根据采集的历史性能数据个数确定, 具 体为: 对于采集的一系列历史性能数据, 从当前时间往前计数, 直到计数 值达到要求的个数; 或者从当前时间往前检测计数, 直到检测点的性能数 据为性能预警值或性能越限告警值。
步驟 S103: 根据预设的性能预警门限及性能越限门限, 对筛选后的历 史性能数据进行分析计算, 得到当前时间点之后的设备出现性能预警值及 性能越限值的时间点, 并得到当前时间点之后的设备的性能数据。
本步驟中, 所述对筛选后的历史性能数据进行分析计算为: 采用多次 曲线拟合插值算法进行分析, 利用采集的历史性能数据的时间点和性能数 据, 计算出拟合曲线参数。 之后, 根据计算出的曲线参数, 结合预先设定 的性能指标的预警门限和越限门限, 对性能预警门限、 性能越限门限进行 反向求解, 计算得到当前时间点之后的设备出现性能预警值及性能越限值 的时间点, 并得到当前时间点之后的设备的性能数据。
其中, 所述曲线拟合插值算法可以采用现有的较为成为的算法, 比如: 最小二乘法及拉格朗日曲线拟合算法等。
若采用上述算法一直无法计算出设备具体的性能预警和性能越限的时 间点, 则计算结果为无期限, 表明该设备的性能指标基本不会劣化。
如图 2所示, 图 2是本发明设备性能预测处理方法一实施例中设备性 能数据的趋势曲线图。 横坐标为时间点, 纵坐标为性能数据, 0点为原点, 即当前时间。 A点到 0点之间的数据为历史性能数据。 B点为性能预警门 限出现的时间点, 其对应的性能数据为性能预警门限, B-C之间的曲线表 示越限预警门限段, 可以采用黄色显示; C 点为性能越限出现的时间点, 其对应的性能数据为性能越限门限, C点之后的曲线表示性能越限门限段, 可以采用红色显示。
如图 3所示, 步驟 103包括:
步驟 1031: 采用多次曲线拟合插值算法对历史性能数据进行分析, 计 算曲线参数。
其中, 所述曲线拟合插值算法可以采用最小二乘法或拉格朗日曲线拟 合算法等。
步驟 1032: 根据曲线参数及预设的性能预警门限及性能越限门限, 计 算当前时间点之后的设备出现性能预警值及性能越限值的时间点, 并得到 当前时间点之后的设备的性能数据。
本实施例在采集设备历史性能数据的基础上, 通过对历史性能数据进 行过滤, 并采用典型的数据预测分析算法, 分析计算出电信设备未来的性 能趋势以及具体的性能预警和性能越限时间点, 从而为电信设备的后续维 护处理提供重要指导和参考, 提高电信设备性能预测的实时与高效性, 进 一步提高网管效率。
如图 4所示, 本发明另一实施例提出一种设备性能预测处理方法, 在 上述实施例的基础上, 在步驟 S101之前还包括:
步驟 S100: 设置性能预警门限及性能越限门限。
在步驟 103之后还包括:
步驟 S104: 通过列表及图形展示当前时间点之后的设备的性能数据。 本实施例与上述实施例的区别在于, 本实施例中需要设置性能预警门 限及性能越限门限, 并保存在本地。
同时, 在通过多次曲线拟合插值算法对历史性能数据进行分析计算, 得到设备的性能预测趋势的预测结果之后, 将设备的性能预测趋势的预测 结果通过列表及图形两种方式向用户进行展示。
首先将预测结果采用列表方式进行展示, 若选择某行数据, 可以通过 "性能趋势" 显示图形窗口, 进一步查看特定性能指标的性能趋势情况。 如图 2所示, 横坐标为时间点信息, 纵坐标为性能数据, 原点为当前时间 及当前时间对应的性能数据, 原点之前的数据为历史性能数据采样点的性 能数据, 原点之后的数据为预测的性能周期内的时刻性能数据, 超过性能 预警门限的性能数据的曲线以黄色表示, 超过性能越限门限的性能数据的 曲线以红色表示, 使得展示图的整体效果非常明显。
用户可以通过设备的性能预测趋势展示图中显示性能正常、 性能预警、 性能越限等不同区段, 实时了解设备当前的性能状态, 从而可以更好的对 电信设备进行后续的维护处理, 为网络运营提供实用的指导和建议, 提高 了网络运营的效率。
如图 5所示, 本发明一实施例提出一种设备性能预测处理装置, 包括: 数据采集模块 501、 数据筛选模块 502以及计算模块 503 , 其中:
数据采集模块 501 , 用于定时从所在设备上采集历史性能数据, 为数据 筛选模块 502提供采集的历时性能数据;
数据筛选模块 502,用于对数据采集模块 501发来的采集的历史性能数 据进行筛选, 将筛选后的历时性能数据发送给计算模块 503;
计算模块 503 , 用于根据预设的性能预警门限及性能越限门限,对数据 筛选模块 502发来的筛选后的历史性能数据进行分析计算, 得到当前时间 点之后的设备出现性能预警值及性能越限值的时间点, 并得到当前时间点 之后的设备的性能数据。
为了对设备的性能进行实时的分析与预测, 本实施例首先通过数据采 集模块 501定时从设备上采集具体性能指标的历史性能数据。
具体地, 根据当前电信设备的实际应用场景, 可以采用 15 分钟或 24 小时性能粒度, 定时从设备上采集历史性能数据, 做为原始的采样点数据, 即每隔 15分钟或 24小时采集一次数据, 因此, 该采样点数据为对应某一 时间点的性能数据。
为保证后续性能预测算法的准确性, 应严格限制历史性能数据的采样 点数量, 必须保证采样点数量在 10 ~ 100个采样点之间。
在采集到设备的历史性能数据后, 通过数据筛选模块 502对采集的历 史性能数据进行歸选, 对不符合要求的采样点数据进行过滤, 歸选过滤后 的历史性能数据作为后续分析计算进行性能预测的基础数据。
其中, 所述筛选的原则是: 如果历史性能数据中性能预警值和性能越 限告警值, 则需要进行剔除; 所述性能预警值和所述性能越限告警值为超 过预先设定的性能预警门限及性能越限门限的性能数据。
另外, 所述筛选的原则还可以根据采集的历史性能数据个数确定, 具 体为: 对于采集的一系列历史性能数据, 从当前时间往前计数, 直到计数 值达到要求的个数; 或者从当前时间往前检测计数, 直到检测点的性能数 据为性能预警值或性能越限告警值。
之后, 通过计算模块 503对筛选后的历史性能数据进行分析计算, 其 分析计算过程具体为: 采用多次曲线拟合插值算法进行分析, 利用采集的 历史性能数据的时间点和性能数据, 计算出拟合曲线的具体参数。 之后, 根据计算出的曲线参数, 结合预先设定的性能指标的预警门限和越限门限, 对性能预警门限、 性能越限门限进行反向求解, 计算得到当前时间点之后 的设备出现性能预警值及性能越限值的时间点, 并得到当前时间点之后的 设备的性能数据。
其中, 曲线拟合插值算法可以采用现有的较为成为的算法, 比如: 最 小二乘法及拉格朗日曲线拟合算法等。
若采用上述算法一直无法计算出设备具体的性能预警和性能越限的时 间点, 则计算结果为无期限, 表明该设备的性能指标基本不会劣化。
设备性能数据的趋势曲线图如图 2所示, 横坐标为时间点, 纵坐标为 性能数据, 0点为原点, 即当前时间。 A点到 0点之间的数据为历史性能 数据。 B点为预警门限出现的时间点, 其对应的性能数据为性能预警门限, B-C之间的曲线表示越限预警门限段,可以采用黄色显示; C点为性能越限 出现的时间点, 其对应的性能数据为性能越限门限, C 点之后的曲线表示 性能越限门限段, 可以采用红色显示。
如图 6所示, 计算模块 503包括: 曲线参数计算单元 5031及告警时间 计算单元 5032; 其中,
曲线参数计算单元 5031 , 用于采用多次曲线拟合插值算法对历史性能 数据进行分析, 计算曲线参数, 将计算得出的曲线参数发送给告警时间计 算单元 5032;
告警时间计算单元 5032,用于根据曲线参数计算单元 5031发来的所述 曲线参数及预设的性能预警门限及性能越限门限, 计算得到当前时间点之 后的设备出现性能预警值及性能越限值的时间点, 并得到当前时间点之后 的设备的性能数据。
如图 7 所示, 本发明另一实施例提出一种设备性能预测处理装置, 在 上述实施例的基础上还包括: 设置模块 500, 用于为数据筛选模块 502和计 算模块 503提供预设的所述性能预警门限及性能越限门限; 相应的, 所述 计算模块 503 ,还用于接收并保存设置模块 500发来的预设的所述性能预警 门限及性能越限门限; 所述数据筛选模块 502,还用于接收并保存设置模块 500发来的预设的所述性能预警门限及性能越限门限。
本发明另一实施例提出一种设备性能预测处理装置, 在上述实施例的 基础上还包括: 预测结果展示模块 504, 用于通过列表及图形展示计算模块 503发来的当前时间点之后的设备的性能数据;相应的,所述计算模块 503 , 还用于将当前时间点之后的设备的性能数据发给预测结果展示模块 504。
本实施例与上述实施例的区别在于, 本实施例通过设置模块 500设置 性能预警门限及性能越限门限, 并保存在本地。
同时, 在通过多次曲线拟合插值算法对历史性能数据进行分析计算, 得到设备的性能预测趋势的预测结果之后, 通过预测结果展示模块 504将 设备的性能预测趋势的预测结果通过列表及图形两种方式向用户进行展 示。
首先将预测结果采用列表方式进行展示, 若用户选择某行数据, 可以 通过 "性能趋势" 显示图形窗口, 进一步查看特定性能指标的性能趋势情 况。 如图 2所示, 在性能趋势展示图中, 横坐标为时间点信息, 纵坐标为 性能数据, 原点为当前时间及当前时间对应的性能数据, 原点之前的数据 为历史性能数据采样点的性能数据, 原点之后的数据为预测的性能周期内 的时刻性能数据, 超过性能预警门限的性能数据的曲线以黄色表示, 超过 性能越限门限的性能数据的曲线以红色表示 , 使得展示图的整体效果非常 明显。
用户可以通过设备的性能预测趋势展示图中显示性能正常、 性能预警、 性能越限等不同区段, 实时了解设备当前的性能状态, 从而可以更好的对 电信设备进行后续的维护处理, 为网络运营提供实用的指导和建议, 提高 了网络运营的效率。
本发明提供的设备性能预测处理装置可以作为逻辑模块或软件安装于 网络管理设备中。
本发明实施例设备性能预测处理方法及装置, 在采集电信设备历史性 能数据的基础上, 通过对历史性能数据进行过滤, 并采用典型的数据预测 分析算法, 分析计算出电信设备未来的性能趋势以及具体的性能预警和性 能越限时间点, 从而为电信设备的后续维护处理提供重要指导和参考, 提 高电信设备性能预测的实时与高效性, 进一步提高网管效率。
以上所述仅为本发明的优选实施例, 并非因此限制本发明的专利范围 , 凡是利用本发明说明书及附图内容所作的等效结构或流程变换, 或直接或 间接运用在其它相关的技术领域, 均同理包括在本发明的专利保护范围内。

Claims

1、 一种设备性能预测处理方法, 其特征在于, 该方法包括: 定时从设备上采集历史性能数据;
对采集的历史性能数据进行筛选;
根据预设的性能预警门限及性能越限门限, 对筛选后的历史性能数据 进行分析计算, 得到当前时间点之后的设备出现性能预警值及性能越限值 的时间点, 并得到当前时间点之后的设备的性能数据。
2、 根据权利要求 1所述的方法, 其特征在于, 所述根据预设的性能预 警门限及性能越限门限, 对筛选后的历史性能数据进行分析计算, 得到当 前时间点之后的设备出现性能预警值及性能越限值的时间点, 并得到当前 时间点之后的设备的性能数据, 包括:
采用多次曲线拟合插值算法对所述历史性能数据进行分析, 计算曲线 参数;
根据所述曲线参数及预设的性能预警门限及性能越限门限, 计算得到 当前时间点之后的设备出现性能预警值及性能越限值的时间点, 并得到当 前时间点之后的设备的性能数据。
3、 根据权利要求 1所述的方法, 其特征在于, 所述采集历史性能数据 包括: 10到 100个采样点。
4、 根据权利要求 1所述的方法, 其特征在于, 所述对采集的历史性能 数据进行筛选, 包括:
剔除采集的历史性数据中的性能预警值及性能越限告警值。
5、 根据权利要求 1所述的方法, 其特征在于, 所述对采集的历史性能 数据进行筛选, 包括:
从当前时间往前计数, 直到计数值达到要求的个数;
或者, 从当前时间往前检测计数, 直到检测点的性能数据为性能预警 值或性能越限告警值。
6、 根据权利要求 2所述的方法, 其特征在于, 所述曲线拟合插值算法 为: 最小二乘法或拉格朗日曲线拟合算法。
7、 根据权利要求 1所述的方法, 其特征在于, 所述得到当前时间点之 后的设备的性能数据之后, 该方法还包括:
通过列表及图形展示当前时间点之后的设备的性能数据。
8、 根据权利要求 1所述的方法, 其特征在于, 所述定时从设备上采集 历史性能数据之前, 该方法还包括:
设置所述性能预警门限及性能越限门限。
9、 一种设备性能预测处理装置, 其特征在于, 包括: 数据采集模块、 数据筛选模块和计算模块; 其中,
数据采集模块, 用于定时从所在设备上采集历史性能数据, 为数据筛 选模块提供采集的历时性能数据;
数据筛选模块, 用于对数据采集模块发来的采集的历史性能数据进行 筛选, 将筛选后的历时性能数据发送给计算模块;
计算模块, 用于根据预设的性能预警门限及性能越限门限, 对数据筛 选模块发来的筛选后的历史性能数据进行分析计算, 得到当前时间点之后 的设备出现性能预警值及性能越限值的时间点, 并得到当前时间点之后的 设备的性能数据。
10、 根据权利要求 9所述的装置, 其特征在于, 所述计算模块包括: 曲线参数计算单元和告警时间计算单元; 其中,
曲线参数计算单元, 用于采用多次曲线拟合插值算法对所述历史性能 数据进行分析, 计算曲线参数, 将计算得出的曲线参数发送给告警时间计 算单元;
告警时间计算单元, 用于根据曲线参数计算单元发来的所述曲线参数 及预设的性能预警门限及性能越限门限, 计算得到当前时间点之后的设备 出现性能预警值及性能越限值的时间点, 并得到当前时间点之后的设备的 性能数据。
11、根据权利要求 9或 10所述的装置, 其特征在于, 所述装置还包括: 预测结果展示模块, 用于通过列表及图形展示计算模块发来的当前时间点 之后的设备的性能数据;
相应的, 所述计算模块, 还用于将当前时间点之后的设备的性能数据 发给预测结果展示模块。
12、 根据权利要求 11所述的装置, 其特征在于, 所述装置还包括: 设 置模块, 用于为数据筛选模块和计算模块提供预设的所述性能预警门限及 性能越限门限;
相应的, 所述计算模块, 还用于接收并保存设置模块发来的预设的所 述性能预警门限及性能越限门限;
所述数据筛选模块, 还用于接收并保存设置模块发来的预设的所述性 能预警门限及性能越限门限。
PCT/CN2012/076112 2011-07-18 2012-05-25 设备性能预测处理方法及装置 WO2013010404A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201110202055.7A CN102891761B (zh) 2011-07-18 2011-07-18 设备性能预测处理方法及装置
CN201110202055.7 2011-07-18

Publications (1)

Publication Number Publication Date
WO2013010404A1 true WO2013010404A1 (zh) 2013-01-24

Family

ID=47535137

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2012/076112 WO2013010404A1 (zh) 2011-07-18 2012-05-25 设备性能预测处理方法及装置

Country Status (2)

Country Link
CN (1) CN102891761B (zh)
WO (1) WO2013010404A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103701637A (zh) * 2013-12-16 2014-04-02 国家电网公司 一种电力通信传输网运行趋势分析方法
CN112203320A (zh) * 2019-07-08 2021-01-08 中国移动通信集团贵州有限公司 一种基于灰色模型预测目标网络参数的方法和装置

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105468885A (zh) * 2014-09-04 2016-04-06 秦皇岛中科百捷电子信息科技有限公司 一种能耗数据的可视化分析方法
CN104636874B (zh) * 2015-02-12 2019-04-16 北京嘀嘀无限科技发展有限公司 检测业务异常的方法及设备
CN105447758A (zh) * 2015-12-09 2016-03-30 上海银天下科技有限公司 预警值设置方法和装置
CN115103415B (zh) * 2022-07-14 2023-08-18 中国联合网络通信集团有限公司 基站算力调度方法、装置及存储介质
CN117035160A (zh) * 2023-07-07 2023-11-10 华能核能技术研究院有限公司 一种核电厂设备标定参数趋势分析预测方法及系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1617512A (zh) * 2004-11-25 2005-05-18 中国科学院计算技术研究所 一种自适应网络流量预测和异常报警方法
CN101170375A (zh) * 2007-11-30 2008-04-30 中兴通讯股份有限公司 用于sdh设备的性能管理方法及装置
CN101267362A (zh) * 2008-05-16 2008-09-17 亿阳信通股份有限公司 一种性能指标值正常波动范围的动态确定方法及其装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2834409B1 (fr) * 2002-01-03 2005-01-14 Cit Alcatel Systeme de gestion de reseaux de transport base sur l'analyse des tendances des donnees acquises sur le reseau
CN102006613B (zh) * 2010-11-19 2013-07-03 暨南大学 一种移动核心网故障数据二重联合线性判别方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1617512A (zh) * 2004-11-25 2005-05-18 中国科学院计算技术研究所 一种自适应网络流量预测和异常报警方法
CN101170375A (zh) * 2007-11-30 2008-04-30 中兴通讯股份有限公司 用于sdh设备的性能管理方法及装置
CN101267362A (zh) * 2008-05-16 2008-09-17 亿阳信通股份有限公司 一种性能指标值正常波动范围的动态确定方法及其装置

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103701637A (zh) * 2013-12-16 2014-04-02 国家电网公司 一种电力通信传输网运行趋势分析方法
CN112203320A (zh) * 2019-07-08 2021-01-08 中国移动通信集团贵州有限公司 一种基于灰色模型预测目标网络参数的方法和装置
CN112203320B (zh) * 2019-07-08 2023-04-28 中国移动通信集团贵州有限公司 一种基于灰色模型预测目标网络参数的方法和装置

Also Published As

Publication number Publication date
CN102891761B (zh) 2017-08-11
CN102891761A (zh) 2013-01-23

Similar Documents

Publication Publication Date Title
WO2013010404A1 (zh) 设备性能预测处理方法及装置
JP6184270B2 (ja) 将来のネットワーク攻撃を検知及び予測するために、様々な指標と過去の攻撃事例を相関させ、攻撃に関する指標のプロファイルを作成するシステム及び方法
KR101068862B1 (ko) 모니터링을 통한 실시간 에너지 관리 시스템
CN106227167A (zh) 一种油气生产设备运行状态的预警方法
CN110213125A (zh) 一种云环境下基于时序数据的异常检测系统
CN112418557B (zh) 基于云服务的数据分析和预测系统及方法
CN102148701A (zh) 用于监控系统中监控指标的趋势分析方法及装置
CN112903815A (zh) 桥梁伸缩缝状态的监测方法及监测系统
CN109408386B (zh) 一种软件老化流式监测系统及其监测方法
WO2015019299A1 (en) Method for measuring dynamics of a flow of energy and relevant device and system
CN114283590B (zh) 车流量高峰预测方法及装置、电子设备
CN115372816A (zh) 基于数据分析的配电开关设备运行故障预测系统及方法
CN110019175B (zh) 一种基于行为轨迹库的区域碰撞分析方法
JP6369895B2 (ja) モータ異常検知システム、モータ異常検知方法、及びモータ異常検知プログラム
CN116520236B (zh) 一种智能电表的异常检测方法和系统
CN114662058B (zh) 无线站点监测方法及装置
CN116611953A (zh) 基于物联网的电能表生产制造实时数据管理方法和系统
CN108920655B (zh) 一种道路天气信息系统时空覆盖范围量化方法及装置
CN116823220A (zh) 一种电缆运行状态监测平台及设备
CN202798762U (zh) 用于电力通信故障信息分析告警装置
CN116108376A (zh) 一种反窃电的监测系统、方法、电子设备及介质
CN109388512A (zh) 针对大规模计算机集群异常程度的评估与分析系统
CN110120893B (zh) 一种定位网络系统安全问题的方法及装置
CN113034732B (zh) 一种牵引变流器滤网堵塞的诊断方法和装置
JP5402071B2 (ja) 省エネルギー支援システム

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: 12814194

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 12814194

Country of ref document: EP

Kind code of ref document: A1