Data Security and Accountable Privacy in Cloud Computing Systems

云计算正在成为大规模、持久或广泛传播的数据的标准计算方式。因此，数据云已成为计算版图中的重要元素。有了数据云的重要性，我们开始见证将重要数据外包给第三方带来的数据安全问题。 我的研究将通过两种方式解决数据安全问题：(I)通过限制云服务器访问数据的机会和(Ii)设计审计设施，供用户检查其数据在云系统中的传播状况。我将研究将计算外包给云的新方法，同时限制与云共享的信息量。近年来，安全外包成为云计算背景下的热门话题。然而，加密数据后允许的计算范围是一个主要问题。我的研究将使用视觉加密技术来开发编码方案，允许在云中比较数据值。通过扩展可以在伪装数据的同时完成的计算类型，我们可以提供一个适应广泛的云应用程序的框架。此外，我还将调查数据云中的隐私审计。越来越多的用户对大型SaaS(软件即服务)云(如Facebook)持有的用户信息过多感到失望。此外，用户还担心数据的删除和放置没有考虑到用户的最佳利益(即隐私)。在隐私审计工作中，我想创建一个框架，它将生成一个独立创建的关于用户信息在云中的传播情况的报告。在实现上述云中数据安全的方法之前，有许多有趣的研究挑战需要解决。调查和解决这些挑战将是我未来几年的研究重点。

Cloud computing is becoming the standard way of computing with large-scale, long-lasting, or widely disseminated data. As a result, data clouds have become important elements of the computing landscape. With that importance of data clouds, we are beginning to witness the data safety problems that come with outsourcing important data to third parties. My research will address the data safety problems in two ways: (i) by limiting the opportunities for cloud servers to gain access to data and (ii) devising auditing facilities for users to check on the status of the spread of their data within the cloud system. I will investigate new ways of outsourcing computation to clouds while limiting the amount of information shared with them. Recently, secure outsourcing has become a hot topic in the context of cloud computing. However, the range of computations allowed after encrypting the data is a major issue. My research will use visual cryptographic techniques to develop encoding schemes that will allow comparisons of data values within clouds. By extending the types of computations that can be done while disguising the data, we can provide a framework that accommodates a wide range of cloud applications. In addition, I will investigate privacy auditing in data clouds. There is growing frustration among the users that large SaaS (software as a service) clouds (e.g., Facebook) are holding more information about their users than necessary. Also, users are concerned that data deletions and placement are not carried with the users' best interests (i.e., privacy) in mind. In the privacy auditing work I want to create a framework that will produce an independently created report on the spread of a user's information in the cloud. There are many interesting research challenges that should be addressed before realizing the above mentioned ways to data safety in clouds. Investigating and solving these challenges will be the focus of my research for next several years.