Researcher: Peter Steenkiste
Access Control for Pervasive Computing Information
Pervasive computing environments automatically and proactively help users with a variety of tasks. These tasks can be fairly simple (e.g. “Find the nearest printer”) or quite sophisticated (e.g. a context-aware message delivery service that selects how to deliver services based on the user’s current context). In order to help users, the pervasive computing environment needs access to a lot of information, including information about the user’s physical context, computing context, and personal information. Since some of this information is sensitive, it is important that we develop appropriate access control mechanisms so users can control what users and services can obtain what information and under what circumstances. In fact, without proper access control, pervasive computing environments are unlikely to be deployed. We propose to develop a set of access control mechanism for pervasive computing information that will make it possible to deploy pervasive computing applications.
Providing access control for pervasive computing information differs in a number of fundamental ways from more traditional access control (e.g. for files and databases). First, the information comes in different forms (e.g. granularities) and can often be derived from diverse sources in the environment, so the access control mechanism must focus on the nature of the information and not its source. Second, policies can be context-dependent. For example, users may only give people access to information at certain times of the day.