Detection Method on the Privacy Leakage for Composite Services

Abstract

The Internet of Things (IoT), as a emerging strategic industry, is widely used in digital homes, smart manufacturing, smart cities, and other fields. However since service capability of a single IoT device is limited, it is difficult to meet the complex service requirements of users. Multi-domain sharing, dynamic integration, and collaborative development have become new tendencies of IoT service development. This paper focuses on the privacy leakage detection of such composite services. Firstly, due to the heterogeneity of various privacy information, we study the dependencerelationamong sensitivity of privacy information. Secondly, dynamic changes of third-party service-calling rights can cause unauthorized access to privacy information during the sharing of shared services. To avoid illegal access, we propose a dynamic privacy information leakage detection method based on security label. Specifically, our research includes the dependencies between services, services and privacy information, constructing a knowledge and privacy information map, generating a private information security label for the combined service, and determining whether the call will cause privacy information disclosure by dynamically verifying the security label during the service invocation process. Finally, we use extensive experiments to validate our methods efficacy and performance.