Modeling of Security and Privacy Architecture for Protecting Databases in Cloud Computing Infrastructure

In order to prevent data leakage and ensure the security of tenant’s private data, and to enable tenants to have a precise understanding of the security level of their private data, the author proposes a modeling of the security and privacy architecture for protecting databases in cloud computing infrastructure. The author proposes a document database privacy protection architecture, which builds upon the existing architecture by adding a privacy protection layer between the application layer and the storage layer, forming a new service deployment architecture. Then, the author introduced the basic methods of privacy protection based on facial document databases. In order to adapt to the data structure system based on document storage for document databases, the author designed a basic method of privacy protection for document databases based on segmentation and obfuscation. By utilizing the free nature of document oriented database patterns, privacy protection data can be achieved through appropriate segmentation. For nested document structures, the author designed a document structure tree to retain document structure information. The results show that by comparing the experimental data of the 50w and 100w groups horizontally, it can be found that under the same cutoff score, as the number of database data increases, the time for data queries also increases accordingly, the query time of database A has increased by nearly 300ms compared to database B, and the additional time of the other groups is also roughly the same. By vertically comparing the experimental data of the 50w and 100w groups, it can be found that the query time of the C database is nearly 200ms longer than that of the A database, and as the sharding factor increases, the query time also increases accordingly, but the proportion of increase begins to slow down. By comparing data with the same segmentation factor but different data volumes, it can be seen that the impact of data volume on query time is positively correlated. This model can ensure that the database system is transparent to users at the view layer after privacy protection, and ensure the correctness and integrity of privacy data.