With the development and progress of network technology, hidden dangers of network security have also become a threat to network storage systems. This research mainly discusses the secure storage of big data based on hybrid large-scale databases. The user sends an authentication request to the SSP, and the SSP verifies the user’s time stamp ts to ensure that it is not a replayed message, and checks whether the user ID is legal. The virtualization layer logically divides the hardware devices at the facility layer through virtualization technology, provides unified services to the outside world, and supports the operation of the platform layer; the platform layer builds a file system and database to provide an operating environment for the operation layer; the operation layer provides users with a big data environment resource access behavior and platform traffic are reasonably allocated and managed, platform business deployment, etc.; the application layer provides users with big data computing services. During the file download process, the big data storage module is relatively fast (300-600k/s), while the speed of Seccloud and invisible cloud is relatively slow (200-400k/s). This research helps to improve the data access efficiency of the big data secure storage module.
The defects of machine learning prediction technology can be more comprehensive and automatic learning model to find the defects in software has become the main method of defect prediction, selection and study of algorithm is the key to improve the accuracy and efficiency of machine learning. Comparing different machine learning defect prediction methods reveals that the algorithms have different advantages in different evaluation indicators, the use of these advantages and combining the stacking ensemble learning method in machine learning is put forward different prediction algorithm of prediction results. As software metrics and again the prediction model of software defect prediction combined machine learning algorithm is based on the experiment with the model of Eclipse, the data sets show the effectiveness of the model.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.