Design And Implementation Of Database Encryption Algorithm Recognition Tool Based On Machine Learning

With the advent of the big data era,the use of databases to manage massive amounts of data has become the first choice in most application scenarios,and the security of data storage is also more important.The research and application of database encryption storage technology have solved the problem of data security to a great extent,and detecting the encryption function of the database has also become one of the standard requirements for evaluating the security level of the database.In order to detect the security level of the database,a database encryption algorithm identification tool needs to be developed to identify the encryption algorithm provided by the encryption interface of the tested database,and to detect whether the encryption function provided by the database product meets the requirements of the security standard.Combining the different key lengths and working modes of encryption algorithms,the 4 common encryption algorithms provided by the database product encryption interface,such as DES,AES,and RC4,are expanded to 9 types,and the extraction method of general database ciphertext columns is studied.In the existing research on encryption algorithm recognition,the text analysis method based on machine learning has rich theoretical basis,therefore,try multiple types of feature extraction methods of ciphertext,such as randomness test feature,entropy feature,etc.,and combine with decision trees,random forests and other machine learning algorithms to design experiments and train recognition models.Designed and implemented a database encryption algorithm identification tool based on machine learning.The system is divided into three modules: pre-processing module,encryption algorithm identification module and extension module.The pre-processing module configures the test information and uses the optional plaintext conditions during database encryption to realize algorithm pre-classification and ciphertext extraction.The encryption algorithm recognition module generates a feature file,calls the machine learning recognition model,and obtains the recognition result.The expansion module updates and expands the algorithm library and model library of the tool to expand the recognition range and improve the recognition accuracy.Through the actual test of Dameng database and My SQL database,it is shown that the recognition tool can complete the recognition of common database encryption algorithms.Compared with existing methods,this recognition tool can support the expansion of encryption algorithms and recognition models,and has higher applicability and recognition accuracy.