Failure Prediction Based On Failure Mode And Effective Early Warning Mechanism

With the increasing development of technology and the increasing complexity of the software,the frequency of software failure is gradually increasing.Although the software testing technology is relatively mature in the field of software reliability,some of the system which is already put into use still will be a lot of problem.Therefore,it is of great practical significance and research value to find a strategy which can effectively predict the failure behavior of online failure system.Based on the analysis of the current research situation on software failure prediction,this paper proposes a software failure prediction framework based on failure mode.First,the error log records and the corresponding failure events are collected,perform preprocessing of data.And then use the error log data set to construct the error propagation signature,which can represent the system real-time error status,based on these signatures can interpret the trend of failure.The regression analysis is used to construct the predictors for every failure mode.The predictor can monitor the system error log in real time to predict future failure behavior and make effective early warning at an appropriate time.The main work of this paper are as follows:1.At the beginning of this paper,the status quo of failure research at home and abroad is introduced,and the mechanism of software failure,forecasting method and evaluation measure are described.Then,the regression analysis method,construction principle and the prediction mechanism of the failure mode is recommended systematically.2.In view of the shortcomings of the existing scheme,our paper studies and improves the existing failure prediction mechanism,The dynamic failure pattern matching algorithm is used to match the failure mode of the current system and can still perform efficient failure prediction without being able to match the appropriate unknown failure mode.If the failure mode is matched to the single failure mode or the multiple failure modes,With the higher degree of completeness of the model library,the prediction accuracy will be significantly increased.3.Aiming at the shortcomings of the traditional failure mode prediction mechanism in the early warning work,we add the new maintenance time cost measure in the failure mode,and provide the strategy basis for the timely handling of the event when the high-risk probability event occurs.When the system is about to fail,making an effective early warning a reasonable time ahead.in the resource consumption of the minimum and sufficient time to prerequisites,to remind the relevant personnel to take emergency action,early prevention and avoid the occurrence of failure,effectively improve the failure to avoid the rate.4.In the complex software system with many components,thinking of the internal complex dependencies and the characteristics of error propagation,we can build a prediction model based on the failure mode,real-time monitoring and predicting the failure of internal components and the entire software system,which can improve the failure prediction rate and reduce the failure rate of the large complex structured component system.