Supporting data privacy through fine-grain access control

许多数据密集型应用程序拥有大量具有不同访问权限的用户。也就是说，一些用户可能被授权查看所有数据，而其他用户应该只能查看特定的、可能重叠的子集。细粒度访问控制通常用于支持此类应用程序。使用这种机制，不同的用户(或用户组)可以被授予单个元组级别的访问权限，而不仅仅是整个文档或整个关系。实际上，细粒度访问控制通常在应用程序代码中实现。这种方法有很多局限性。首先，很难确保某些用户无法从他们有权访问的数据中推断或推断出私人数据。也就是说，我们一般不能提供隐私保障。其次，应用程序代码通常独立于数据库设计而发展，因此可能很难确保访问控制策略保持一致和正确。

Many data intensive applications have large numbers of users who have varying access privileges. That is, some users may be authorized to view all the data, while others should only be able to view specific, possibly overlapping, subsets. Fine-grained access control is typically used to support such applications. Using such a mechanism, different users (or user groups) can be granted access at the level of individual tuples, not just entire documents or entire relations. In practice, fine-grained access control is often implemented within application code. Such an approach has manylimitations. First, it is hard to ensure that certain users cannot infer or deduce private data from the data to which they have access. That is, in general we cannot provide an privacy guarantees. Second, application code often evolves independently of the database design so it may be hard to ensure that access control policies stayconsistent and correct.