| Big Data is a hot area of the research in internet industry. With the coming into the Cloud Times, Big Data has an explosive development and it attracts an increasing number of attentions. In fact, Big Data is entering every aspect of daily life, changing the way we work and influencing the way we think and it has created great value. Obviously, Big Data will lead a revolution both in material life and in spiritual life in our society. However, Big Data which contains massive amount as well as potential value also has potential risks, including privacy problem, fake data and broken down of data integrity. These problems limit the development of Big Data and make it a big challenge.At this stage, however, research of Big Data is mainly in convergence of the traditional security strategy such as user authentication and access control, and most of the security solutions come out from the traditional platforms. At the same time, the privacy policy has already become the concern of the industry. Therefore, this article will focus on two problems: how to avoid the illegal user to utilize the data of the platform for data security; how to avoid the legal user to abuse the data of the platform by the purpose of curiosity for privacy policy. This article concerns two aspects above to ensure that data platform can provide stable and reliable legal data service under the precondition of its privacy can’t be invaded. To solve these problems, this article provides two strategies of the Big Data system, and the main work is as follows:(1)The model of privacy policy presented in this article has two aspects: protecting the privacy of user query so it can refrain the disclosure of the searching content; protecting the privacy of the data center as it can resist the legal user by sending the query to the data center for stealing the extra information of the center, then the result is the disclosure of the privacy. In this article, the author advocates a model based on key encryption to protect the privacy in the situation of data outsourcing. By means of encryption with key to these data, no user is allowed to get these data, this model use only a few encryption schemes to ensure the privacy policy, while not influencing the operation of the platform. And for user query privacy policy, this article bases on anonymous solution and query resolution by trusted third party. This model lets the data platform have no way to relate directly with the specific user on account of the searching query.(2)The article presents a data-access model aimed at unstructured data platform that also contains two small models: the user management model of the Big Data platform and the access control model of that. This article realizes the user authentication in form of the token in the platform. In addition, the author takes account of the model of traditional access control to administrate authority and control access which hence forbids the user from accessing the data platform without permission. The result of this model is protecting the data of the platform from revealing and increasing security level of the data.(3)The article establishes a prototype system to achieve the two security enhancement model above. This article chooses Hadoop-Hive Big Data platform, and adds specific realization of privacy preserving model and data access model based on that platform. The whole design including architecture, models and logic flow is described in the article, and then this article builds up the prototype system, obtaining some data by testing the system. By analyzing the data, this article ensures the feasibility of the two security enhancement model above, and assures the security of Big Data. |
PDF Full Text Request