Abstract: This paper presents an effective scheme for encrypting and querying private character sequences in CloudDB. To ensure the security of character sequences in the untrusted CloudDB, the scheme would encrypt the private data in a client-side, and then construct feature index for them. The feature index is constructed based on such steps as subfield partition, subfield domain partition and partition identified assignment, which can capture the key feature information of character sequences and discard non-key feature information, thus has good security and effectiveness. When querying data, the query operations over the private fields would be translated into the new query operations over the corresponding index fields, making them able to be executed correctly over the CloudDB, without the need to decrypt the data, consequently improving the query efficiency. Finally, both theoretical analysis and experimental evaluation validate the security, effectiveness, efficiency and flexibility of the proposed scheme.

Key words: CloudDB, enterprise information management, private character sequence

摘要： 面向云数据库的字符串型隐私信息的加密查询问题，提出了一个有效解决方案。为了确保字符串数据在不可信云数据库中安全性，在客户端，首先对其进行加密并为其构建特征索引。特征索引通过子字段划分、子字段值域分区、分区标识符分配等步骤，能有效地获取字符串的关键特征，而丢弃非关键特征，具有很好的安全性和有效性。然后，数据查询时，方案将定义在隐私字段上的用户查询转换为定义在索引字段上的新查询，使其可以在云数据库上正确执行，而无需解密数据，有效地提高了查询效率。最后，理论分析和实验评估验证了方案的安全性、有效性、高效性和灵活性。

关键词: 云数据库, 企业信息管理, 隐私字符串