A Data Center Fault Prediction Method Based on Data Augmentation

Abstract

The invention discloses a data center fault prediction method based on data enhancement, which normalizes the data points to obtain the input and output data pairs of the fault prediction model, and determines the input vector x related to the fault to form a real fault data set D fault_real ; Then build a data augmentation model, for the real fault data set D fault_real For data enhancement, the generator generates samples, and the discriminator is updated with generated samples and real samples. If the discriminator can distinguish between generated samples and real samples, the discriminator is fixed, and the generator is updated to regenerate samples. If the discriminator cannot distinguish between generated samples and real samples samples, the data is merged to generate a data set D after data enhancement full ;Finally use data set D full Train the fault prediction model until the model loss cannot be reduced, according to the data set D full The data format of the model requires the data of the current time point to be input into the model, and the output is to predict the probability of failure at the time point behind the horizon, so as to realize the failure prediction. This method effectively improves the accuracy of fault prediction.

Description

technical field [0001] The invention belongs to the technical field of data center fault prediction, and in particular relates to a data enhancement-based data center fault prediction method. Background technique [0002] In recent years, with the development of the mobile Internet, the amount of Internet data has shown explosive growth, and more and more Internet services are also based on the analysis of big data. These have led to a rapid increase in the demand for computing resources. The computing power of a single computer can no longer meet the demand. So cloud computing came into being. Cloud computing is the product of the integration of traditional computer and network technologies such as distributed computing, parallel computing, virtualization, and load balancing. Cloud computing virtualizes a large number of servers into computing resource nodes through virtual machine technology. Users do not need to care about hardware implementation and maintenance. They ...

AI Technical Summary

Problems solved by technology

There are few methods for time series data, and there are methods such as adding Gaussian noise, which are not effective

[0008] To sum up, although the data-driven node fault prediction technology is suitable for high-complexity data centers, it is often affected by the lack of fault-related data and the unbalanced training set, resulting in low accuracy.

Images