中文摘要：

针对云中数据隐私保护问题,采用加密存储是一可行的选择。而在未解密的情况下,如何对密文进行计算成了近年来研究的热点。为提高密文计算效率,节约密文数据的存储空间,将选择加密策略与CryptDB密文数据库系统相结合,设计并实现了支持选择加密策略的CryptDB密文数据库系统。由于CryptDB使用多个洋葱加密模型,需将数据加密成多份以适用不同场景,对不敏感字段加密增加了计算时间及存储空间。针对这些问题,提出一种用户自定义的敏感字段检测算法,并且在原系统中创建和注册明文洋葱,修改元数据表,以及对自定义SQL语句的拦截和改写,实现选择加密策略。实验结果表明,在满足用户对数据表安全性要求的前提下,随着加密字段减少,密文计算时间和存储空间显著减少,提高了CryptDB系统的效率和实用性。

英文摘要：

In order to protect privacy of sensitive data stored on the cloud,encrypting data is a feasible way. Computing on the encrypted data without decrypting is becoming a research hotspot. Combining selective encryption strategy with CryptDB, an encrypted database system which supports selective encryption strategy is proposed to improve the performance of enerypfing and decrease the storage space. Since the use of multiple onion encryption models, data should be encrypted into several copies by CryptDB to adapt different situations. Encrypting non-sensitive fields will experience high computation time and increase storage space. A user-defined sensitive field detection algorithm is proposed and implemented through creating and registering the plain onion within original system, modifying the metadata ta- bles, and intercepting and revising the SQL queries. With the context meeting the user＇ s requirements of data security, the experimental results show that the scheme decreases the time of computation and storage space with the number of encrypted fields decreasing and makes the system more efficient and practical.