Chaotic test optimization method and system based on k-means algorithm

Abstract

The invention provides a chaotic test optimization method based on a k-means algorithm. The optimization method comprises the following steps: S1, storing a large amount of experimental data through a local SQLite database; S2, after the experiment data are stored, creating a chaos experiment through a create command; S3, clustering and classifying the experimental data through a K-Means algorithm before the experimental data are transmitted into the experimental component of the chaos experiment, and storing the classified data in a data set mode; S4, transmitting the data set into a to-be-experimented component of a chaos experiment, and starting the chaos experiment; S5, after the experiment is finished, recovering all data in the data set, including variation data and newly added data; and S6, recycling the experimental data, perfecting the content of the experimental data, and executing the step S3 again. According to the invention, experimental data can be classified, and variation data and dirty data can be recovered.

Description

technical field [0001] The invention relates to the technical field of chaos testing, in particular to a chaos testing optimization method and system based on a k-means algorithm. Background technique [0002] The current chaos test solution is to store a large amount of experimental data in the local database, call the relevant chaos experiment executor to create or destroy the experiment by calling the create and destroy commands, and query the preparation records of the chaos experiment and the chaos experiment environment through the status command . Among them, a large amount of experimental data in the database will flood into the system at the beginning of the chaos test to attack the system and find the weak parts of the system, thereby exposing the problems of the system. [0003] Chaos testing actually uses a large amount of data to impact the system and find the weak link of the system, although the influx of a large amount of data will generate a large load on a...

AI Technical Summary

Problems solved by technology

However, a large amount of data has not been classified, and it is difficult to direct the influx of data to a single node, resulting in most data not being able to influx and resulting in data waste; and in the destruction experiment, the destroy command issues a destruction command, which needs to be influx The input data is clear one by one. If the data is mutated under the action of the code, it will cause the data to be left behind and generate dirty data, which will have a potential impact on the system.

Images