CN104901823A - Method and device for generating alarm threshold value, and method and device for monitoring service performance index - Google Patents
Method and device for generating alarm threshold value, and method and device for monitoring service performance index Download PDFInfo
- Publication number
- CN104901823A CN104901823A CN201410076801.6A CN201410076801A CN104901823A CN 104901823 A CN104901823 A CN 104901823A CN 201410076801 A CN201410076801 A CN 201410076801A CN 104901823 A CN104901823 A CN 104901823A
- Authority
- CN
- China
- Prior art keywords
- threshold
- value
- threshold value
- alarm threshold
- lower limit
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Monitoring And Testing Of Exchanges (AREA)
Abstract
The invention discloses a method and device for generating an alarm threshold value and a method and device for monitoring a service performance index. The method for generating the alarm threshold value comprises: forming a upper threshold value sample space with sample values greater than a predicted basic value and forming a lower threshold value sample space with sample values less than the predicted basic value on the basis of symmetric threshold values by using an asymmetric threshold value concept and using the predicted basic value as a reference; and according to the distribution of a sample with maximum degrees of deviation, performing secondary threshold value computation on the sample space without a maximum deviation sample so as to break a symmetric distribution characteristic of original dynamic threshold values and obtain a more reasonable alarm threshold value range. As a result, when being used for monitoring service performance, the method improves alarm accuracy and monitoring effectiveness and timely and effectively monitors a performance index abrupt change.
Description
Technical field
The present invention relates to technical field of network management, particularly relate to a kind of alarm threshold generation method, service feature index method for supervising and device.
Background technology
Current application of net is more and more extensive, and the routine work that people are a lot of and recreation all be unable to do without network.For improving the service quality of network, meet the demand of people to network, the operation maintenance of Virtual network operator to network is had higher requirement.
In traditional service mode, attendant mainly pays close attention to equipment state.Judge whether an equipment works well, and depends on equipment alarm.But the perception of client to business is highstrung, traditional centered by equipment state, there are problems in monitoring service mode that is Network Based, equipment alarm.Cannot quality of service have been caused to decline by Timeliness coverage, but the fault of equipment no alarm, cannot ensure to find fault prior to client.Therefore, in daily monitoring maintenance work, the monitoring of service feature index is become more and more important, become and helped attendant to judge the whether normal important guarantee of system cloud gray model.
In the monitoring to service feature index, the setting of alarm threshold is the key of performance index monitoring, how to arrange rational alarm threshold, directly determines the monitoring effect of performance index.Adopt static alarm threshold or static alarm threshold at times in prior art, namely for the static unique alarm threshold of monitored setup measures, now, concrete service feature index method for supervising as shown in Figure 1.
Step 101: extract monitoring initial data;
Step 102: the data of extraction and static alarm threshold are contrasted;
Step 103: judge whether the data extracted are greater than static alarm threshold; If so, then perform step 104, and return step 101; If not, then step 101 is returned;
Step 104: trigger alerts.
But, in practice, a lot of service feature indexs present the feature of the dynamic change with time cycle feature, when adopting static threshold method to set up or the performance index of static threshold method to set up to the feature presenting the dynamic change with time cycle feature at times to monitor, dynamic change cannot be embodied, if alarm threshold arranges lower, then cannot monitoring business mass change, alarm susceptibility is low; If alarm threshold arranges higher, though relatively can promote the alarm susceptibility of quality of service monitoring, in business idle, particularly night, too high alarm threshold can trigger a large amount of alarm by mistake.Static alarm threshold method to set up cannot monitor the sudden change of performance index timely and effectively, cannot trigger alerts in time.
Such as: access controller (Access Controller, AC) the online user number index had in the wireless lan (wlan) of obvious time cycle feature chosen as shown in Figure 2 is analyzed:
From Fig. 2, (its transverse axis is the time, the longitudinal axis is online user's number, obtain by carrying out sampling every 5 minutes to online user's number) can find out that WLAN AC online user number index has obvious time cycle feature, busy period, index value differed greatly, and had obvious index and rise and the decline stage.Often in order to avoid by mistake alarm occurs in monitoring, static threshold or at times threshold value are arranged all can be comparatively wide in range, cannot effective real-time tracking index variable condition.Such as when occurring in figure that state index occurs abnormal, because setup measures is too wide in range, cannot trigger alerts in time.
In order to the shortcoming that the alarm threshold method to set up overcoming above-mentioned static state exists, dynamic threshold is adopted in practical application, the basic skills of current calculating dynamic threshold technology is by historical data prediction current base threshold value, and then join probability algorithm obtains one and thinks rational threshold range.The selection of the upper lower threshold value of this threshold range is symmetrical relative to the baseline threshold of prediction.
Above-mentioned employing dynamic threshold is monitored service feature index, the dynamic tracking to index change can be realized, but the selection of the upper lower threshold value of above-mentioned threshold range is symmetrical relative to prediction fiducial value, there is the not high shortcoming of alarm accuracy in the monitoring of this threshold range to service feature, makes a concrete analysis of as follows:
Dynamic threshold algorithm mainly adopts symmetrical concept, and upper threshold value and lower threshold value are identical relative to the irrelevance of prediction base value.But in fact, from the distribution of sample space, with the baseline threshold predicted for baseline, the irrelevance of historical sample point to baseline of its upper and lower is different, and the sample point quantity of the baseline threshold upper and lower of prediction is also different.Also, namely for the baseline threshold of prediction, the possible space scope of sample should be asymmetrical relative to the baseline threshold of prediction.And the calculating of above-mentioned dynamic threshold drift gage count in be unified calculation.In all sample points, depart from maximum sample point comparatively large on the impact of threshold space scope, and do not consider its top being positioned at prediction base value or bottom.From prediction index value possible range angle, have definite meaning.But from formative dynamics threshold value, say for actual monitored work angle, the possible range of desired value will be expanded like this, cause because alarm threshold arranges excessive, cannot the situation of accurate alarm.
Summary of the invention
In view of this, embodiments provide a kind of alarm threshold generation method, service feature index method for supervising and device, the inaccurate problem of alarm that the alarm threshold obtained in order to solve employing dynamic threshold defining method of the prior art is brought.
A kind of alarm threshold generation method, described method comprises:
According to each sample point in the historical sample space of performance index and prediction algorithm, current service feature desired value is predicted, obtain predicting base value;
According to historical sample space, determine the distribution of current service feature desired value, and will the higher limit of this distribution and the higher limit of lower limit as threshold reference and the lower limit of threshold reference be represented; And
Be upper threshold value sample space and lower threshold value sample space by historical sample spatial division, wherein, the value of each sample point in upper threshold value sample space is more than or equal to described prediction base value, and the value of each sample point in lower threshold value sample space is less than described prediction base value;
Prediction base value maximum sample point is departed from when being present in described upper threshold value sample space in historical sample space, according to lower threshold value sample space, determine the distribution in lower threshold value space, will the higher limit of distribution in lower threshold value space and the higher limit of lower limit as lower threshold value and the lower limit of lower threshold value be represented;
Using the lower limit of the maximum in the lower limit of threshold reference and the lower limit of lower threshold value as alarm threshold, using the higher limit of the higher limit of threshold reference as alarm threshold.
Preferably, the maximum sample point of prediction base value is departed from when being present in described lower threshold value sample space in historical sample space, according to upper threshold value sample space, determine the distribution in upper threshold value space, using the higher limit of distribution that represents in upper threshold value space and the higher limit of lower limit as upper threshold value and the lower limit of upper threshold value;
Using the higher limit of the minimum value in the higher limit of threshold reference and the higher limit of upper threshold value as alarm threshold, using the lower limit of the lower limit of threshold reference as alarm threshold.
Preferably, described method also comprises:
The higher limit of following formula to alarm threshold is used to be optimized, the higher limit of the alarm threshold after being optimized;
The higher limit * (1+x) of the higher limit=alarm threshold of the alarm threshold after optimization;
Wherein, x represents the tolerance coefficient of the higher limit of the alarm threshold of setting.
Preferably, described method also comprises:
The lower limit of following formula to alarm threshold is used to be optimized, the lower limit of the alarm threshold after being optimized;
The lower limit * (1-y) of the lower limit=alarm threshold of the alarm threshold after optimization;
Wherein, y represents the tolerance coefficient of the lower limit of the alarm threshold of setting.
A kind of service feature index method for supervising, described method comprises:
Above-mentioned arbitrary described alarm threshold generation method is utilized to generate alarm threshold;
Current service feature desired value and described alarm threshold are compared;
When current service feature desired value is not between the higher limit and lower limit of alarm threshold, trigger alerts.
A kind of alarm threshold generating apparatus, described device comprises:
Prediction base value generation unit, for according to each sample point in the historical sample space of performance index and prediction algorithm, predicts current service feature desired value, obtains predicting base value;
Threshold reference determining unit, for according to historical sample space, determines the distribution of current service feature desired value, and will represent the higher limit of this distribution and the higher limit of lower limit as threshold reference and the lower limit of threshold reference;
Sample space division unit, for being upper threshold value sample space and lower threshold value sample space by historical sample spatial division, wherein, the value of each sample point in upper threshold value sample space is more than or equal to described prediction base value, and the value of each sample point in lower threshold value sample space is less than described prediction base value;
Distribution determining unit, prediction base value maximum sample point is departed from when being present in described upper threshold value sample space in historical sample space, according to lower threshold value sample space, determine the distribution in lower threshold value space, will the higher limit of distribution in lower threshold value space and the higher limit of lower limit as lower threshold value and the lower limit of lower threshold value be represented;
Alarm threshold determining unit, for using the lower limit of the maximum in the lower limit of threshold reference and the lower limit of lower threshold value as alarm threshold, using the higher limit of the higher limit of threshold reference as alarm threshold.
Preferably, distribution determining unit, when maximum sample point also for departing from prediction base value in historical sample space is present in described lower threshold value sample space, according to upper threshold value sample space, determine the distribution in upper threshold value space, using the higher limit of distribution that represents in upper threshold value space and the higher limit of lower limit as upper threshold value and the lower limit of upper threshold value;
Described alarm threshold determining unit, also for using the higher limit of the minimum value in the higher limit of threshold reference and the higher limit of upper threshold value as alarm threshold, using the lower limit of the lower limit of threshold reference as alarm threshold.
Preferably, described device also comprises:
Optimize unit, for using the higher limit of following formula to alarm threshold to be optimized, the higher limit of the alarm threshold after being optimized; The higher limit * (1+x) of the higher limit=alarm threshold of the alarm threshold after optimization; Wherein, x represents the tolerance coefficient of the higher limit of the alarm threshold of setting.
Preferably, described device also comprises:
Optimize unit, for using the lower limit of following formula to alarm threshold to be optimized, the lower limit of the alarm threshold after being optimized; The lower limit * (1-y) of the lower limit=alarm threshold of the alarm threshold after optimization; Wherein, y represents the tolerance coefficient of the lower limit of the alarm threshold of setting.
A kind of service feature index supervising device, described device comprises:
Above-mentioned arbitrary described alarm threshold generating apparatus;
Comparison module, compares for the alarm threshold current service feature desired value and described alarm threshold generating apparatus generated;
Trigger module, for when current service feature desired value is not between the higher limit and lower limit of alarm threshold, trigger alerts.
In the technical scheme of the embodiment of the present invention, by adopting asymmetric threshold concept, on the basis of symmetric thresholds, to predict that base value is for benchmark, the sample being greater than prediction base value is formed upper threshold value sample space, the sample point being less than prediction base value is formed lower threshold value sample space; Then according to the distribution of irrelevance maximum sample point, secondary threshold calculations is carried out to the sample space that there is not maximum deviation sample point, break the symmetrical characteristic of original dynamic threshold, obtain more reasonably alarm threshold span, therefore, when the monitoring for service feature, the accuracy of meeting alarm, strengthen the validity of monitoring, monitor performance index sudden change timely and effectively.
Accompanying drawing explanation
Fig. 1 is the service feature index method for supervising flow chart in background technology of the present invention;
Fig. 2 is the AC online user number indicatrix in the WLAN in background technology of the present invention;
Fig. 3 is the flow chart of a kind of alarm threshold generation method in the embodiment of the present invention one;
Fig. 4 is the flow chart of a kind of service feature index method for supervising in the embodiment of the present invention two;
Fig. 5 is the structural representation of a kind of alarm threshold generating apparatus in the embodiment of the present invention four;
Fig. 6 is the structural representation of a kind of service feature index supervising device in the embodiment of the present invention five.
Embodiment
The inaccurate problem of alarm that the alarm threshold obtained to solve employing dynamic threshold defining method of the prior art is brought, embodiments provides a kind of alarm threshold generation method, service feature index method for supervising and device.
In order to the scheme of the embodiment of the present invention is clearly described, first analytic explanation is carried out to the general principle of the embodiment of the present invention:
The sample point that in sample space, irrelevance is maximum has the greatest impact to alarm threshold scope, but this sample point will only be positioned at the side of prediction base value, traditional dynamic threshold adopts symmetric thresholds to calculate in calculating, now this sample point will affect the threshold calculations of opposite side, and then can threshold range be expanded, reduce monitoring susceptibility.For overcoming this shortcoming, in the scheme of the embodiment of the present invention, adopt asymmetric threshold concept, on the basis of symmetric thresholds, to predict that base value is for benchmark, the sample being greater than prediction base value is formed upper threshold value sample space, the sample point being less than prediction base value is formed lower threshold value sample space.Then according to the distribution of irrelevance maximum sample point, secondary threshold calculations is carried out to the sample space that there is not maximum deviation sample point, break the symmetrical characteristic of original dynamic threshold, obtain more reasonably alarm threshold span.
Below in conjunction with Figure of description, the embodiment of the present invention is described in detail.
Embodiment one
The embodiment of the present invention one provides a kind of alarm threshold generation method, and its flow chart as shown in Figure 3, specifically comprises the following steps:
Step 300: according to each sample point in the historical sample space of performance index and prediction algorithm, predicts current service feature desired value, obtains predicting base value; And perform step 301.
Described prediction algorithm has multiple, such as the method for moving average, weighted mean method, correlation regression algorithm, mathematical modelings etc., in the embodiment of the present invention one, in conjunction with the feature of data and the feature of each prediction algorithm self needing the performance index monitored, a kind of suitable prediction algorithm can be selected.And then by the value of each sample point in the historical sample space of performance index and the prediction algorithm selected, just current service feature desired value can be predicted, obtain prediction base value.
Step 301: according to historical sample space, determines the distribution of current service feature desired value, and will represent the higher limit of this distribution and the higher limit of lower limit as threshold reference and the lower limit of threshold reference; And perform step 302.
The distribution of described current service feature desired value, can about prediction base value symmetry, also can be asymmetric.
In this step 301, according to historical sample space, determine the distribution of current service feature desired value, also namely according to historical sample space, determine the possible range of the service feature desired value in actual service feature index space.
In art of mathematics, sample value in sample space is the sample value part in real space, can, according to the sample value distribution of sample space (locally), utilize Probability Theory and Math Statistics knowledge to estimate the Distribution value scope in real space (overall situation).And current service feature desired value is a value in practical business performance index space, therefore, the possible distribution of each service feature desired value in actual service feature index space is the possible distribution of current service feature desired value.
Step 302: be upper threshold value sample space and lower threshold value sample space by historical sample spatial division; And perform step 303.
Wherein, the value of each sample point in upper threshold value sample space is more than or equal to described prediction base value, and the value of each sample point in lower threshold value sample space is less than described prediction base value;
It should be noted that, the position of step 301 and step 302 can exchange, and also can perform step 301 and step 302 simultaneously.
Step 303: judge whether the sample point departing from prediction base value maximum in historical sample space is present in described upper threshold value sample space, if so, then performs step 304; If not, then step 306 is performed.
Step 304: according to lower threshold value sample space, determines the distribution in lower threshold value space, obtains higher limit and the lower limit of lower threshold value spatial distribution scope, and performs step 305.
Step 305: using the lower limit of the maximum in the lower limit of threshold reference and the lower limit of lower threshold value as alarm threshold, using the higher limit of the higher limit of threshold reference as alarm threshold.
Step 304 and step 305 be consider based on a kind of like this: if the sample point departing from prediction base value maximum exists at upper threshold value sample space, when the distribution of current service feature desired value is symmetrical about prediction base value, the impact of its reasonability on the higher limit of the threshold reference determined in step 301 is little, and it is larger on the reasonability impact of the lower limit of the threshold reference determined in step 301, even if the lower limit of the threshold reference also determined in step 301 is less, expand the scope of alarm threshold, therefore, suitable increase (adopting step 304 and step 305 to come reasonably to increase in the present invention) is carried out to the lower limit of threshold reference, and then reach the object of the scope of tightening up alarm threshold.
Step 306: according to upper threshold value sample space, determines the distribution in upper threshold value space, obtains higher limit and the lower limit of the distribution in upper threshold value space, and performs step 307.
Because the sample point in upper threshold value sample space and lower threshold value sample space is combined the whole sample points formed in historical sample space, therefore, if the maximum sample point departing from prediction base value in historical sample space is not present in described upper threshold value sample space, then must be present in described lower threshold value sample space.
Step 307: using the higher limit of the minimum value in the higher limit of threshold reference and the higher limit of upper threshold value as alarm threshold, using the lower limit of the lower limit of threshold reference as alarm threshold.
Step 306 and step 307 be consider based on a kind of like this: if the sample point departing from prediction base value maximum exists in lower threshold value sample space, when the distribution of current service feature desired value is symmetrical about prediction base value, the impact of its reasonability on the lower limit of the threshold reference determined in step 301 is little, and it is larger on the reasonability impact of the higher limit of the threshold reference determined in step 301, even if the higher limit of the threshold reference also determined in step 301 is larger, expand the scope of alarm threshold, therefore, suitable reduction (adopting step 306 and step 307 to come reasonably to reduce in the present invention) is carried out to the higher limit of threshold reference, and then reach the object of the scope of tightening up alarm threshold.
Preferably, in order to strengthen the reasonability of alarm threshold further, described method also comprises:
The higher limit of following formula to alarm threshold is used to be optimized, the higher limit of the alarm threshold after being optimized;
The higher limit * (1+x) of the higher limit=alarm threshold of the alarm threshold after optimization;
Wherein, x represents the tolerance coefficient of the higher limit of the alarm threshold of setting.
Preferably, in order to strengthen the reasonability of alarm threshold further, described method also comprises:
The lower limit of following formula to alarm threshold is used to be optimized, the lower limit of the alarm threshold after being optimized;
The lower limit * (1-y) of the lower limit=alarm threshold of the alarm threshold after optimization;
Wherein, y represents the tolerance coefficient of the lower limit of the alarm threshold of setting.
In the scheme of the embodiment of the present invention one, due to the distribution according to irrelevance maximum sample point, secondary threshold calculations is carried out to the sample space that there is not maximum deviation sample point, break the symmetrical characteristic of original dynamic threshold, obtain more reasonably alarm threshold span, therefore, when the monitoring for service feature, the accuracy of meeting alarm, strengthens the validity of monitoring, monitors performance index sudden change timely and effectively.
The situation that scheme below by embodiment two pairs of embodiment of the present invention one is applied to when service feature index is monitored is described.
Embodiment two
The embodiment of the present invention two provides a kind of service feature index method for supervising, and its flow chart as shown in Figure 4, specifically comprises the following steps:
Step 401: utilize arbitrary described alarm threshold generation method in embodiment one to generate alarm threshold.
Step 402: current service feature desired value and described alarm threshold are compared.
Step 403: judge current service feature desired value whether between the higher limit and lower limit of alarm threshold, if so, then performs step 404; If not, then step 405 is performed.
Step 404: trigger alerts, and jump to step 401.
Step 405: described current service feature desired value is added historical sample data, and jumps to step 401.
Perform this step 405 when above-mentioned judged result is no, can effectively monitor dynamic performance index, real-time tracking performance index variation tendency, and then the threshold value of dynamic conditioning monitoring in real time.
By the scheme of the embodiment of the present invention two, because the alarm threshold used is relatively more rational, therefore, the accuracy of alarm can be improved preferably, strengthen the validity of business monitoring.
Below by a concrete service feature index historical sample data, the embodiment of the present invention one and embodiment two are described in more detail.
Embodiment three
The embodiment of the present invention three is tested based on choosing WLAN performance index AC online user number, for example is described the alarm threshold generation method of the embodiment of the present invention and service feature index method for supervising, specifically comprises the following steps:
The first step: predict current performance desired value in conjunction with Forecasting Methodology according to historical sample, obtains predicting base value.
In order to improve the accuracy of computational prediction benchmark, when calculating, preliminary treatment can be carried out to data, such as, rejecting the maximum in historical sample data and minimum value.
Obtain predicting that the Forecasting Methodology of base value has multiple, the such as method of moving average, weighted mean method, correlation regression algorithm, mathematical modeling etc., illustrate with the method for moving average below:
The method of moving average is a kind of arithmetic mean method of improvement, is adaptive prediction model.It is comparatively large on predicted value impact according to Recent data, and the fact that data at a specified future date are less on predicted value impact, average was moved by the phase.
The weight of each element of rolling average is all equal.
The computing formula of rolling average is as follows:
Ft={(At-1)+（At-2）+（At-3）+…+（At-n)}/n
Ft represents the predicted value to lower first phase;
N represents the number in period of rolling average;
At-1 represents actual value in early stage;
At-2, At-3 and At-n represent that front two phases, first three phase are until the actual value of front n phase respectively.
Choose WLAN performance index AC online user number to test, get the history sample points same period according to 40, concrete distribution is as shown in table (1):
| 464 | 446 | 478 | 491 | 491 | 541 | 531 | 532 | 488 | 489 |
| 476 | 489 | 521 | 511 | 530 | 453 | 465 | 547 | 548 | 468 |
| 475 | 467 | 458 | 578 | 489 | 491 | 550 | 538 | 441 | 467 |
| 524 | 478 | 458 | 468 | 531 | 568 | 557 | 441 | 467 | 467 |
Table (1)
To the prediction base value that above-mentioned 40 historical sample data utilize the above-mentioned method of moving average to obtain performance index be: 496.8.
Second step: according to historical sample space, determines the distribution of current service feature desired value.
Because the service feature desired value in the embodiment of the present invention is static, therefore, also can be described as dynamic threshold.The various ways such as dynamic threshold benchmark can adopt data modeling, probability theory obtain, and adopt the confidential interval algorithm in probability theory here, draw current point in time, the possible distribution of service feature desired value, formative dynamics threshold reference.
Described confidential interval refers to the estimation interval of the population parameter constructed by sample statistic.In statistics, the confidential interval (Confidence interval) of a probability sample is the interval estimation of certain population parameter to this sample.The actual value of this parameter that what confidential interval represented is has certain probability to drop on the degree of the surrounding of measurement result.The credibility of what confidential interval provided the is measured value of measured parameter, namely required above " certain probability ".This probability is called as confidence level.For example, if someone supporting rate is 55% in once general election, and the confidential interval in confidence level 0.95 is (50%, 60%), so his true supporting rate has the probability of 95 percent to drop between 60 50 percent and percent, and therefore the possibility of the not enough half of his true supporting rate is less than percent 2.5.As in example, confidence level generally represents with percentage, and the confidence space therefore in confidence level 0.95 also can be expressed as: 95% confidential interval.The two ends of confidential interval are called as fiducial limit.Concerning the estimation of a given situation, confidence level is higher, and corresponding confidential interval will be larger.
Choose carrying out sample to historical data, and after adopting prediction algorithm to obtain corresponding index value mathematic expectaion, utilize probability confidential interval algorithm to carry out dynamic threshold calculating.Such as, obtain predicting base value according to rolling average algorithm, also namely obtain " sample average " (prediction base value is sample average) of desired value, then calculate sampling error.Performance data sample space is a kind of data random sequence, meets normal distribution characteristic, adds, subtracts " sampling error " of calculating by " sample average " obtained, and draws two end points of confidential interval.That is:
for prediction base value, σ is the variance calculated according to sample space and prediction base value, and n is sample space number, z
α/2for confidence level coefficient (table look-up and obtain).
Probability calculation is carried out to the sample space in the first step, adopt confidential interval algorithm, getting confidence level is 95%, the confidential interval obtained is: (423.3,570.3), the distribution of current service feature desired value is also namely determined, and using when representing the higher limit 570.3 of this distribution and lower limit 423.3 as the lower limit of the higher limit of threshold reference and threshold reference, this threshold reference is (423.3,570.3).
3rd step: by second step prediction base value based on, sample space is divided into two parts, is respectively: as table (2) shown in upper threshold value sample space (comprising 15 sample points altogether) and as table (3) lower threshold value sample space (comprising 25 sample points altogether):
| 541 | 531 | 532 | 521 | 511 |
| 530 | 547 | 548 | 538 | 524 |
| 531 | 568 | 557 | 578 | 550 |
Table (2)
| 464 | 446 | 478 | 491 | 491 |
| 488 | 489 | 476 | 489 | 465 |
| 468 | 475 | 467 | 458 | 489 |
| 491 | 441 | 467 | 458 | 468 |
| 441 | 467 | 467 | 453 | 478 |
Table (3)
4th step: by all samples and prediction base value being contrasted, obtaining irrelevance maximum sample point is 578, is positioned at upper threshold value sample space.Therefore carry out secondary probability calculation to lower threshold value space, adopt confidential interval mode, getting confidence level is 95%, obtains lower threshold value space possible range to be: (500.6,440.6).
5th step: the lower threshold value lower limit 440.6 that the 4th step is tried to achieve by secondary probability calculation, threshold reference lower limit 423.3 comparison of trying to achieve with second step, get the lower limit of relative larger value 440.6 as alarm threshold, using the higher limit of the higher limit 570.3 of threshold reference as alarm threshold, thus obtain alarm threshold (570.3,440.6), can see not about 496.8 symmetries.
6th step: according to the alarm threshold obtained (also can be described as asymmetric dynamic threshold value), be aided with tolerance degree coefficient, thus obtain final dynamic threshold.If upper threshold value tolerance coefficient is x, lower threshold value tolerance coefficient is y, and the asymmetric dynamic threshold value finally obtained is [570.3 × (1+x), 440.6 × (1-y)].
7th step: the WLAN online user number desired value supposing up-to-date collection is Z.Through comparing, if Z value size is outside [570.3 × (1+x), 440.6 × (1-y)] interval in asymmetric dynamic threshold value, then trigger alerts, otherwise, be judged as normal.
Embodiment four
The embodiment of the present invention four provides a kind of alarm threshold generating apparatus, its structural representation as shown in Figure 5, described device comprises: prediction base value generation unit 501, threshold reference determining unit 502, sample space division unit 503, distribution determining unit 504 and alarm threshold determining unit 505, wherein:
Prediction base value generation unit 501, for according to each sample point in the historical sample space of performance index and prediction algorithm, predicts current service feature desired value, obtains predicting base value;
Threshold reference determining unit 502, for according to historical sample space, determines the distribution of current service feature desired value, and will represent the higher limit of this distribution and the higher limit of lower limit as threshold reference and the lower limit of threshold reference;
Sample space division unit 503, for being upper threshold value sample space and lower threshold value sample space by historical sample spatial division, wherein, the value of each sample point in upper threshold value sample space is more than or equal to described prediction base value, and the value of each sample point in lower threshold value sample space is less than described prediction base value;
Distribution determining unit 504, prediction base value maximum sample point is departed from when being present in described upper threshold value sample space in historical sample space, according to lower threshold value sample space, determine the distribution in lower threshold value space, will the higher limit of distribution in lower threshold value space and the higher limit of lower limit as lower threshold value and the lower limit of lower threshold value be represented;
Alarm threshold determining unit 505, for using the lower limit of the maximum in the lower limit of threshold reference and the lower limit of lower threshold value as alarm threshold, using the higher limit of the higher limit of threshold reference as alarm threshold.
Preferably, described distribution determining unit 504, when maximum sample point also for departing from prediction base value in historical sample space is present in described lower threshold value sample space, according to upper threshold value sample space, determine the distribution in upper threshold value space, using the higher limit of distribution that represents in upper threshold value space and the higher limit of lower limit as upper threshold value and the lower limit of upper threshold value;
Preferably, described alarm threshold determining unit 505, also for using the higher limit of the minimum value in the higher limit of threshold reference and the higher limit of upper threshold value as alarm threshold, using the lower limit of the lower limit of threshold reference as alarm threshold.
Preferably, described device also comprises:
Optimize unit 506, for using the higher limit of following formula to alarm threshold to be optimized, the higher limit of the alarm threshold after being optimized; The higher limit * (1+x) of the higher limit=alarm threshold of the alarm threshold after optimization; Wherein, x represents the tolerance coefficient of the higher limit of the alarm threshold of setting, and * represents multiplication symbol.
Preferably, described device also comprises:
Optimize unit 506, for using the lower limit of following formula to alarm threshold to be optimized, the lower limit of the alarm threshold after being optimized; The lower limit * (1-y) of the lower limit=alarm threshold of the alarm threshold after optimization; Wherein, y represents the tolerance coefficient of the lower limit of the alarm threshold of setting, and * represents multiplication symbol.
Embodiment five
The embodiment of the present invention five provides a kind of service feature index supervising device, and as shown in Figure 6, described device comprises its structural representation: alarm threshold generating apparatus 601, comparison module 602 and trigger module 603, wherein:
Alarm threshold generating apparatus 601, it is arbitrary described alarm threshold generating apparatus in the embodiment of the present invention four;
Comparison module 602, compares for the alarm threshold current service feature desired value and described alarm threshold generating apparatus generated;
Trigger module 603, for when current service feature desired value is not between the higher limit and lower limit of alarm threshold, trigger alerts.
Those skilled in the art should understand, embodiments of the invention can be provided as method, system or computer program.Therefore, the present invention can adopt the form of complete hardware embodiment, completely software implementation or the embodiment in conjunction with software and hardware aspect.And the present invention can adopt in one or more form wherein including the upper computer program implemented of computer-usable storage medium (including but not limited to magnetic disc store, CD-ROM, optical memory etc.) of computer usable program code.
The present invention describes with reference to according to the flow chart of the method for the embodiment of the present invention, equipment (system) and computer program and/or block diagram.Should understand can by the combination of the flow process in each flow process in computer program instructions realization flow figure and/or block diagram and/or square frame and flow chart and/or block diagram and/or square frame.These computer program instructions can being provided to the processor of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing device to produce a machine, making the instruction performed by the processor of computer or other programmable data processing device produce device for realizing the function of specifying in flow chart flow process or multiple flow process and/or block diagram square frame or multiple square frame.
These computer program instructions also can be stored in can in the computer-readable memory that works in a specific way of vectoring computer or other programmable data processing device, the instruction making to be stored in this computer-readable memory produces the manufacture comprising command device, and this command device realizes the function of specifying in flow chart flow process or multiple flow process and/or block diagram square frame or multiple square frame.
These computer program instructions also can be loaded in computer or other programmable data processing device, make on computer or other programmable devices, to perform sequence of operations step to produce computer implemented process, thus the instruction performed on computer or other programmable devices is provided for the step realizing the function of specifying in flow chart flow process or multiple flow process and/or block diagram square frame or multiple square frame.
Although describe the preferred embodiments of the present invention, those skilled in the art once obtain the basic creative concept of cicada, then can make other change and amendment to these embodiments.So claims are intended to be interpreted as comprising preferred embodiment and falling into all changes and the amendment of the scope of the invention.
Obviously, those skilled in the art can carry out various change and modification to the present invention and not depart from the spirit and scope of the present invention.Like this, if these amendments of the present invention and modification belong within the scope of the claims in the present invention and equivalent technologies thereof, then the present invention is also intended to comprise these change and modification.
Claims (10)
1. an alarm threshold generation method, is characterized in that, described method comprises:
According to each sample point in the historical sample space of performance index and prediction algorithm, current service feature desired value is predicted, obtain predicting base value;
According to historical sample space, determine the distribution of current service feature desired value, and be upper threshold value sample space and lower threshold value sample space by historical sample spatial division, wherein, the value of each sample point in upper threshold value sample space is more than or equal to described prediction base value, and the value of each sample point in lower threshold value sample space is less than described prediction base value;
Prediction base value maximum sample point is departed from when being present in described upper threshold value sample space in historical sample space, according to lower threshold value sample space, determine the distribution in lower threshold value space, will the higher limit of distribution in lower threshold value space and the higher limit of lower limit as lower threshold value and the lower limit of lower threshold value be represented;
Using the lower limit of the maximum in the lower limit of threshold reference and the lower limit of lower threshold value as alarm threshold, using the higher limit of the higher limit of threshold reference as alarm threshold.
2. the method for claim 1, it is characterized in that, the maximum sample point of prediction base value is departed from when being present in described lower threshold value sample space in historical sample space, according to upper threshold value sample space, determine the distribution in upper threshold value space, using the higher limit of distribution that represents in upper threshold value space and the higher limit of lower limit as upper threshold value and the lower limit of upper threshold value;
Using the higher limit of the minimum value in the higher limit of threshold reference and the higher limit of upper threshold value as alarm threshold, using the lower limit of the lower limit of threshold reference as alarm threshold.
3. method as claimed in claim 1 or 2, it is characterized in that, described method also comprises:
The higher limit of following formula to alarm threshold is used to be optimized, the higher limit of the alarm threshold after being optimized;
The higher limit * (1+x) of the higher limit=alarm threshold of the alarm threshold after optimization;
Wherein, x represents the tolerance coefficient of the higher limit of the alarm threshold of setting.
4. the method as described in as arbitrary in claim 1 or 2, it is characterized in that, described method also comprises:
The lower limit of following formula to alarm threshold is used to be optimized, the lower limit of the alarm threshold after being optimized;
The lower limit * (1-y) of the lower limit=alarm threshold of the alarm threshold after optimization;
Wherein, y represents the tolerance coefficient of the lower limit of the alarm threshold of setting.
5. a service feature index method for supervising, is characterized in that, described method comprises:
Arbitrary described alarm threshold generation method in claim 1-4 is utilized to generate alarm threshold;
Current service feature desired value and described alarm threshold are compared;
When current service feature desired value is not between the higher limit and lower limit of alarm threshold, trigger alerts.
6. an alarm threshold generating apparatus, is characterized in that, described device comprises:
Prediction base value generation unit, for according to each sample point in the historical sample space of performance index and prediction algorithm, predicts current service feature desired value, obtains predicting base value;
Threshold reference determining unit, for according to historical sample space, determines the distribution of current service feature desired value, and will represent the higher limit of this distribution and the higher limit of lower limit as threshold reference and the lower limit of threshold reference;
Sample space division unit, for being upper threshold value sample space and lower threshold value sample space by historical sample spatial division, wherein, the value of each sample point in upper threshold value sample space is more than or equal to described prediction base value, and the value of each sample point in lower threshold value sample space is less than described prediction base value;
Distribution determining unit, prediction base value maximum sample point is departed from when being present in described upper threshold value sample space in historical sample space, according to lower threshold value sample space, determine the distribution in lower threshold value space, will the higher limit of distribution in lower threshold value space and the higher limit of lower limit as lower threshold value and the lower limit of lower threshold value be represented;
Alarm threshold determining unit, for using the lower limit of the maximum in the lower limit of threshold reference and the lower limit of lower threshold value as alarm threshold, using the higher limit of the higher limit of threshold reference as alarm threshold.
7. device as claimed in claim 6, it is characterized in that, distribution determining unit, when maximum sample point also for departing from prediction base value in historical sample space is present in described lower threshold value sample space, according to upper threshold value sample space, determine the distribution in upper threshold value space, using the higher limit of distribution that represents in upper threshold value space and the higher limit of lower limit as upper threshold value and the lower limit of upper threshold value;
Described alarm threshold determining unit, also for using the higher limit of the minimum value in the higher limit of threshold reference and the higher limit of upper threshold value as alarm threshold, using the lower limit of the lower limit of threshold reference as alarm threshold.
8. device as claimed in claims 6 or 7, it is characterized in that, described device also comprises:
Optimize unit, for using the higher limit of following formula to alarm threshold to be optimized, the higher limit of the alarm threshold after being optimized; The higher limit * (1+x) of the higher limit=alarm threshold of the alarm threshold after optimization; Wherein, x represents the tolerance coefficient of the higher limit of the alarm threshold of setting.
9. the device as described in as arbitrary in claim 6 or 7, it is characterized in that, described device also comprises:
Optimize unit, for using the lower limit of following formula to alarm threshold to be optimized, the lower limit of the alarm threshold after being optimized; The lower limit * (1-y) of the lower limit=alarm threshold of the alarm threshold after optimization; Wherein, y represents the tolerance coefficient of the lower limit of the alarm threshold of setting.
10. a service feature index supervising device, is characterized in that, described device comprises:
Arbitrary described alarm threshold generating apparatus in claim 6-9;
Comparison module, compares for the alarm threshold current service feature desired value and described alarm threshold generating apparatus generated;
Trigger module, for when current service feature desired value is not between the higher limit and lower limit of alarm threshold, trigger alerts.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410076801.6A CN104901823B (en) | 2014-03-04 | 2014-03-04 | A kind of alarm threshold generation method, service feature index monitoring method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410076801.6A CN104901823B (en) | 2014-03-04 | 2014-03-04 | A kind of alarm threshold generation method, service feature index monitoring method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104901823A true CN104901823A (en) | 2015-09-09 |
| CN104901823B CN104901823B (en) | 2018-09-25 |
Family
ID=54034227
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410076801.6A Active CN104901823B (en) | 2014-03-04 | 2014-03-04 | A kind of alarm threshold generation method, service feature index monitoring method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104901823B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106209432A (en) * | 2016-06-30 | 2016-12-07 | 中国人民解放军国防科学技术大学 | Network equipment subhealth state method for early warning based on dynamic threshold and device |
| CN106201829A (en) * | 2016-07-18 | 2016-12-07 | 中国银联股份有限公司 | Monitoring Threshold and device, monitoring alarm method, Apparatus and system |
| CN106557401A (en) * | 2016-10-13 | 2017-04-05 | 中国铁道科学研究院电子计算技术研究所 | A kind of dynamic threshold establishing method and system of information technoloy equipment monitor control index |
| CN106656583A (en) * | 2016-12-02 | 2017-05-10 | 郑州云海信息技术有限公司 | Dynamic threshold alarming method and device |
| CN106856442A (en) * | 2015-12-09 | 2017-06-16 | 北京神州泰岳软件股份有限公司 | A kind of performance indications monitoring method and device |
| CN107124290A (en) * | 2016-02-25 | 2017-09-01 | 中国移动通信集团上海有限公司 | A kind of alarm threshold generation method and device |
| CN107357714A (en) * | 2016-05-09 | 2017-11-17 | 腾讯科技（深圳）有限公司 | A kind of failure analysis methods and device based on monitor supervision platform |
| CN107888435A (en) * | 2016-09-29 | 2018-04-06 | 杭州华为数字技术有限公司 | The threshold setting method and device of a kind of alarm monitoring |
| CN108038040A (en) * | 2017-12-08 | 2018-05-15 | 上海市信息网络有限公司 | Computer cluster performance indicator detection method, electronic equipment and storage medium |
| CN109391486A (en) * | 2017-08-07 | 2019-02-26 | 中国移动通信集团浙江有限公司 | The method for early warning and server to interoperability Developing Tactics based on user experience |
| CN110262947A (en) * | 2018-03-12 | 2019-09-20 | 腾讯科技（深圳）有限公司 | Threshold alarm method, apparatus, computer equipment and storage medium |
| CN111239346A (en) * | 2020-03-26 | 2020-06-05 | 泛测（北京）环境科技有限公司 | Method and device for identifying pollution event in air quality monitoring |
| CN111327449A (en) * | 2018-12-17 | 2020-06-23 | 中国移动通信集团北京有限公司 | Method, device, equipment and medium for determining network abnormity |
| CN111427801A (en) * | 2020-04-24 | 2020-07-17 | 上海新炬网络技术有限公司 | Service performance evaluation method based on load test tool |
| CN111488562A (en) * | 2020-04-07 | 2020-08-04 | 中国移动通信集团江苏有限公司 | Threshold determination method, device, equipment and medium based on face comparison |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030079160A1 (en) * | 2001-07-20 | 2003-04-24 | Altaworks Corporation | System and methods for adaptive threshold determination for performance metrics |
| CN101964997A (en) * | 2009-07-21 | 2011-02-02 | 中国移动通信集团黑龙江有限公司 | Method and device for carrying out early warning on network performance |
| CN102111307A (en) * | 2009-12-29 | 2011-06-29 | 亿阳信通股份有限公司 | Method and device for monitoring and controlling network risks |
-
2014
- 2014-03-04 CN CN201410076801.6A patent/CN104901823B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030079160A1 (en) * | 2001-07-20 | 2003-04-24 | Altaworks Corporation | System and methods for adaptive threshold determination for performance metrics |
| CN101964997A (en) * | 2009-07-21 | 2011-02-02 | 中国移动通信集团黑龙江有限公司 | Method and device for carrying out early warning on network performance |
| CN102111307A (en) * | 2009-12-29 | 2011-06-29 | 亿阳信通股份有限公司 | Method and device for monitoring and controlling network risks |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106856442A (en) * | 2015-12-09 | 2017-06-16 | 北京神州泰岳软件股份有限公司 | A kind of performance indications monitoring method and device |
| CN107124290A (en) * | 2016-02-25 | 2017-09-01 | 中国移动通信集团上海有限公司 | A kind of alarm threshold generation method and device |
| CN107357714A (en) * | 2016-05-09 | 2017-11-17 | 腾讯科技（深圳）有限公司 | A kind of failure analysis methods and device based on monitor supervision platform |
| CN107357714B (en) * | 2016-05-09 | 2020-09-08 | 腾讯科技（深圳）有限公司 | Fault analysis method and device based on monitoring platform |
| CN106209432B (en) * | 2016-06-30 | 2019-04-19 | 中国人民解放军国防科学技术大学 | Network equipment inferior health method for early warning and device based on dynamic threshold |
| CN106209432A (en) * | 2016-06-30 | 2016-12-07 | 中国人民解放军国防科学技术大学 | Network equipment subhealth state method for early warning based on dynamic threshold and device |
| CN106201829A (en) * | 2016-07-18 | 2016-12-07 | 中国银联股份有限公司 | Monitoring Threshold and device, monitoring alarm method, Apparatus and system |
| CN106201829B (en) * | 2016-07-18 | 2019-01-22 | 中国银联股份有限公司 | Monitor Threshold and device, monitoring alarm method, apparatus and system |
| CN107888435A (en) * | 2016-09-29 | 2018-04-06 | 杭州华为数字技术有限公司 | The threshold setting method and device of a kind of alarm monitoring |
| CN106557401A (en) * | 2016-10-13 | 2017-04-05 | 中国铁道科学研究院电子计算技术研究所 | A kind of dynamic threshold establishing method and system of information technoloy equipment monitor control index |
| CN106656583A (en) * | 2016-12-02 | 2017-05-10 | 郑州云海信息技术有限公司 | Dynamic threshold alarming method and device |
| CN109391486A (en) * | 2017-08-07 | 2019-02-26 | 中国移动通信集团浙江有限公司 | The method for early warning and server to interoperability Developing Tactics based on user experience |
| CN108038040A (en) * | 2017-12-08 | 2018-05-15 | 上海市信息网络有限公司 | Computer cluster performance indicator detection method, electronic equipment and storage medium |
| CN108038040B (en) * | 2017-12-08 | 2021-05-11 | 上海市信息网络有限公司 | Computer cluster performance index detection method, electronic equipment and storage medium |
| CN110262947A (en) * | 2018-03-12 | 2019-09-20 | 腾讯科技（深圳）有限公司 | Threshold alarm method, apparatus, computer equipment and storage medium |
| CN110262947B (en) * | 2018-03-12 | 2022-05-17 | 腾讯科技（深圳）有限公司 | Threshold warning method and device, computer equipment and storage medium |
| CN111327449A (en) * | 2018-12-17 | 2020-06-23 | 中国移动通信集团北京有限公司 | Method, device, equipment and medium for determining network abnormity |
| CN111239346A (en) * | 2020-03-26 | 2020-06-05 | 泛测（北京）环境科技有限公司 | Method and device for identifying pollution event in air quality monitoring |
| CN111488562A (en) * | 2020-04-07 | 2020-08-04 | 中国移动通信集团江苏有限公司 | Threshold determination method, device, equipment and medium based on face comparison |
| CN111427801A (en) * | 2020-04-24 | 2020-07-17 | 上海新炬网络技术有限公司 | Service performance evaluation method based on load test tool |
| CN111427801B (en) * | 2020-04-24 | 2024-03-19 | 上海新炬网络技术有限公司 | Service performance evaluation method based on load test tool |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104901823B (en) | 2018-09-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104901823A (en) | Method and device for generating alarm threshold value, and method and device for monitoring service performance index | |
| CN107086944B (en) | Anomaly detection method and device | |
| CN108038040B (en) | Computer cluster performance index detection method, electronic equipment and storage medium | |
| Loukatou et al. | Stochastic wind speed modelling for estimation of expected wind power output | |
| CN103577268B (en) | Adaptive resource Supply Method based on application load | |
| ElNozahy et al. | A probabilistic load modelling approach using clustering algorithms | |
| CN112131441B (en) | Method and system for rapidly identifying abnormal electricity consumption behavior | |
| CN110633194B (en) | Performance evaluation method of hardware resources in specific environment | |
| CN108228412A (en) | A kind of method and device based on system health degree faults of monitoring system and hidden danger | |
| CN107909508A (en) | A kind of distribution transformer load abnormality alarming method | |
| Hua et al. | Fast unscented transformation-based transient stability margin estimation incorporating uncertainty of wind generation | |
| WO2019071438A1 (en) | Method and apparatus for monitoring state of device in process industry and medium | |
| CN107623924A (en) | It is a kind of to verify the method and apparatus for influenceing the related Key Performance Indicator KPI of Key Quality Indicator KQI | |
| Pinceti et al. | Detection and localization of load redistribution attacks on large-scale systems | |
| Özaltın et al. | Predicting the solution time of branch-and-bound algorithms for mixed-integer programs | |
| Chen et al. | Real-time forecasting and visualization toolkit for multi-seasonal time series | |
| CN116663747B (en) | Intelligent early warning method and system based on data center infrastructure | |
| CN106779443B (en) | Operation risk determination method and device in power dispatching | |
| Straka et al. | Simulation as a means of activity streamlining of continuously and discrete production in specific enterprise | |
| Sun et al. | Performance comparison of distributed state estimation algorithms for power systems | |
| US20130268230A1 (en) | Processing distributions | |
| CN114117786A (en) | Stadium power supply system maintenance method, system, equipment and medium | |
| CN105760956A (en) | Method for predicting power grid reliability based on linear-regression analysis | |
| CN113285847A (en) | Communication network anomaly detection method and system of intelligent converter station monitoring system | |
| CN114239952A (en) | Method, device, equipment and medium for detecting abnormity of key performance index |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |