Database Access Control and Privacy

访问控制是说谁可以对什么数据执行什么操作的能力；例如，在线书店中的发货人可以读取客户地址，但不能更新它。访问控制在用户身份验证后接管，决定允许该用户访问、读取、写入和更新哪些内容(例如，发货人不应插入新的库存)。客户可能会被允许说，他们的地址可以用于发货，但不能用于广告；这就是通常所说的数据隐私。航运和广告是隐私目的的例子。有关公司库存的数据通过访问控制进行控制。另一方面，客户数据也受到隐私问题的影响。实施这些隐私要求，以及表达什么应该是隐私的工具，都是计算机科学家关心的方面。 我们的访问控制研究基于基于角色的访问控制(RBAC)模型，该模型将权限收集到角色中，并在一次操作中将权限分配给用户。RBAC简化了复杂环境中的访问控制管理，帮助安全系统设计人员更好地完成工作。存在一些度量标准来说明角色设计何时是“好的”(例如，最少可能的角色)；我们将搜索其他角色。我们将研究RBAC设计的更新，并测试所需的属性。我们最近研究了一些数据上带有隐私标签的RBAC。我们建议进一步进行这项研究，检查具有这些标签的数据的存储，以及在查询和更新隐私标签数据的同时，不同物理分区的运行时权衡。 科学的方法包括开发模型，证明模型的性质，展示其对真实世界情况的适用性，开发原型和工具，以及性能分析比较。这项研究的贡献将是更好地理解保护隐私，在存在访问控制机制的情况下保护隐私，在变化中维护这些系统，衡量我们何时有一个好的设计，以及更有效地实施访问控制和隐私保护系统。

Access control is the ability to say who can do what operations on what data; e.g., the shipper in an on-line book store can read a customer address but not update it. Access control takes over after user authentication, deciding what that user is allowed to access, read, write and update (e.g. the shipper should not be inserting new inventory). Customers may be allowed to say that their address can be used for shipping but not for advertising; this is what is commonly referred to as data privacy. Shipping and advertising are examples of privacy purposes. Data concerning the company's inventory is controlled by access control. Customer data, on the other hand, is also subject to privacy concerns. Implementing these privacy requirements, and the tools to express what should be private, are aspects which concern Computer Scientists. Our access control research is based on the role-based access control (RBAC) model where permissions are collected in roles and assigned to users in one operation. RBAC simplifies the management of access control in complex environments, helping the security system designer to do a better job. Some metrics exist to say when a role design is "good" (e.g. fewest possible roles); we will search for others. We will study updates to an RBAC design, and tests for desired properties. We have recently studied RBAC with privacy labels on some of the data. We propose to further this research, examining the storage of data with these labels, and the run-time tradeoffs of different physical partitions of the privacy-labeled data, while querying and updating it. The scientific approach involves developing a model, proving properties of the model, showing its applicability to real-world situations, developing prototypes and tools, and performance analysis comparisons. The contributions of the research will be a better understanding of preserving privacy, preserving privacy in the presence of an access control mechanism, maintaining these systems under change, measuring when we have a good design, and more efficiently implementing access control and privacy- preserving systems.