Xml-Based Privacy Model in Pervasive Computing
暂无分享,去创建一个
The years coming promise to bring new area of information technology, transferring it
from scientists minds into reality, on one hand a new paradigm known as pervasive calm,
ubiquitous computing, or pervasive computing has the ability to overcome a lot of
insufficiencies of the current information systems while on the other hand central blocks
of pervasive computing are in direct conflicts with privacy protection fundamentals.
Considerable efforts have been taken to cope with this problem but each one had its own
shortage. Some just provide one privacy type like location privacy or just identity privacy,
some of them were not platform independence, and some resulted to a lot of privacy
alarms.
In this thesis we proposed a new privacy model in pervasive computing that provides all
four privacy types (ID, Location, Time, and content) for the user with high control over private information (User Control over Private Information) and as less privacy warnings
as possible (Unobtrusiveness of Privacy Mechanism). To complete the proposed model
we showed model privacy policies with XML tags and distributed decision making
processes in different layers to provide high scalability.
To validate the model, through implementation we showed that model provides “Privacy
Policy Expressiveness” with supporting mandatory and discretionary rules, uncertainty
handling and conflict resolution. We showed model unobtrusiveness with experimenting
and measuring the time user wastes on dealing with privacy sub-system. We showed that
our model provides content, identity, location and time privacy that leads to a high level
of user control over private information. The model scalability would be granted by using
XML as a platform independent format to describe privacy policies with addition of
distributed decision making processes.
The validation results confirmed that the model supports all four metrics of
“expressiveness of privacy policies”, all four metrics of “user control over private
information”, and both factors of “scalability”, with less than 10% “unobtrusiveness”.