Research On Privacy Prseserving Technology In The MIT Platform

The Mobile Internet era has changed the way people work, live, and it also brings new opportunities and challenges to the insurance industry. Using Mobile Integrated Terminal(MIT) to carry on insurance business has become a new insurance marketing model, which can help agents recommend insurance products, insure, survey, and fulfill other insurance processes quickly. However, mobile applications bring in convenience as well as many security risks, of which users’ privacy problem is particularly serious.Academic research on privacy issues exists for a long time, which generally focuses on the areas of database applications. It includes privacy preserving technologies based on data mining and data publishing. And relevant theoretical models such as k-anonymity, information loss has been practiced. However, most of the current privacy preserving technologies still in the theoretical study and has not been widely applied to practical applications. One the one hand the relevant legal requirements is still not perfect, one the other hand users and application developers don’t pay enough emphasis on privacy preserving. So how to improve and innovate the existing security and privacy preserving techniques so as to satisfy the actual needs of applications is significant.This thesis is based on the MIT project, and it describes the traditional data publishing privacy theory and technology and briefly describes the Android platform and its data privacy preserving technology. Besides, a fine-grained privacy preserving scheme is designed according to the specific needs of the MIT application, which includes the backup of privacy files, the password protection method based on the Hash salt,and a personalized anonymity model and algorithm for data publishing.For the backup need of customer data and other privacy data in the local MIT application, a backup method based on byte splitting and recombination is proposed. It splits original files into multiple encrypted files,and backup them locally at first and then backup them into cloud. The Byte transforming process is easier than traditional encryption algorithms and doesn’t need to manage secret keys, so it has better performance and applies for mobile applications.Considering insurance companies may publish or share users’ data with others, this paper proposed an improved personalized(α,L) anonymous model which based on the existing anonymous model and an anonymous algorithm based on clustering. Besides, experiments was carried out with the use of public data set, and results shows that the algorithm has better performance on both the time cost and data loss than the existing clustering anonymous algorithm.Finally, the work of this thesis is summarized, and point that the next work should focus on improving the security of file backup method and dynamic data sets’ privacy preserving etc, to improve the above scheme.