Block-HRG: Block-based differentially private IoT networks release

With the rapid development of the Internet of Things (IoT), privacy-preserving IoT network (graph) analytic has attracted significant attention, where Local differential privacy (LDP) has been widely used. Current LDP-based network publishing approaches commonly require each device to remain online throughout the data collection. Nevertheless, many IoT devices have limited network connectivity, which restricts the applicability of these solutions. Besides, there are extensive collaborations among mutually trusted IoT devices that form natural communities, which is largely ignored by existing approaches, thereby compromising the maintenance of topological utility. To tackle these issues, we introduce block differential privacy, which extends differential privacy to IoT environment. On this basis, we propose a two-phase privacy-preserving IoT network release scheme. By dividing the graph into blocks and applying a hierarchical random graph (HRG)-based structure extraction approach on the inner-block edges, our scheme abandons the attempt to collect data from every single device and therefore can effectively capture the community structure. Furthermore, we propose an HRG aggregation approach that combines the block HRGs and the across-block edges to construct a global graph in a bottom-up manner, which significantly enhances the topological utility. Theoretical analysis and experimental results show that our proposed scheme achieves high-quality IoT graph release while ensuring individual privacy.