The Study And Examination Of Relational Database Ciphertext Query In DBAAS

In the DBAAS (Database as a Service) model, the data owner outsource give his data to a DBSP (Database Service Provider), and the database plays the role of service.DBSP is responsible for the user to access, management and query data without being granted. That means, the database security threats are not only from the network assaultant, but from the delegated authority to the security implications. To do with this problem, it's usually using the database encryption to protect data security. But in this method, it decrypts the ciphertext one by one and then retrieve plaintext, which would lead to a sharply decline in the performance, especially for structured relational data in the data encryption, it will destroy some of the inherent characteristics of the original data, such as the location of the data sequences, the numerical size of the transform and so on. It has more difficult to follow the explicit retrieval method.As a response, in this paper, it analyzes the structure of the database from the structural features and encryption methods with the DBAAS model to study the relationship between common database numeric and CHAR/VARCHAR or CLOB characters. While retaining the feature extraction, it extracts the character data's location in the sequences and proposes an improved relationship algorithm based on a dictionary and the characters adjacent. For CHAR/VARCHAR, it proposes two filtering algorithms, and for CLOB, the two-step filtering one is taken. From the strategy, it can narrow the scope of the ciphertext quickly to retrieve the data, improve the matching accuracy of the character data.For the security, as the ciphertext characters data set is extracted in the algorithm, the client can not be decrypted without the key operations, so data encryption algorithm there is in the level of security. Finally, from the ciphertext experiments in character data CHAR/VARCHAR and CLOB types, it verifies that the proposed dictionary-based segmentation and grouping characters algorithm is effective in the retrieval time and the accuracy also has a good performance.