Detection method and device

Abstract

The invention discloses a detection method and device, which are used for enhancing the detection flexibility. The method comprises the following steps: acquiring traffic statistics data; constructing a Gaussian mixture model according to the traffic statistics data; calculating the threshold value of the traffic statistics data through the Gaussian mixture model; and when test data exceeds the threshold value, determining that the test data is abnormal data.

Description

technical field [0001] The invention relates to the field of communication technology, in particular to a detection method and device. Background technique [0002] With the complexity of networking, more and more attention has been paid to fault location in the transmission domain. The current fault location methods are all fault location based on existing expert experience. impact, it is necessary to discover potential network failures as early as possible, and find out the correlation between potential failures and performance indicators through massive historical data mining. [0003] Currently, FMA (FaultManagerAnalyzer, Fault Management Analysis) is used to perform fault diagnosis on the performance of the transmission domain, which specifically includes: judging whether the index is abnormal according to the set threshold, and generating a fault alarm if the index exceeds the threshold; according to the generated fault alarm, Statistics of fault ratios, including ala...

AI Technical Summary

Problems solved by technology

[0005] 1. Fault detection is based on fixed thresholds. Once the deployment is completed, the default thresholds will be used to detect indicators. It does not support flexible and automatic setting of thresholds based on scenarios, traffic, busy or idle, etc., which cannot meet the fine-grained requirements. maintenance needs

[0006] 2. Most of the network faults are performance faults, which need to go through a gradual process from normal to abnormal indicators. However, because these faults cannot be detected in advance, these faults are "post-processing", and emergency troubleshooting is passively carried out, which increases troubleshooting. failure cost

[0007] 3. Fault diagnosis is based on the experience of experts to judge which indicators are abnormally related to the fault. It cannot quickly and accurately locate the fault and does not meet the purpose of efficient operation and maintenance.

Images