Resilient aggregation scheme for confidential sensor reports

Aggregating sensor reports can significantly reduce sensor network traffics, but to provide confidentiality to these reports might disable meaningful aggregation. This paper tackles the problem by proposing a resilient aggregation scheme RAS that achieves 2-level aggregation at cluster and network respectively. Cluster aggregation collects local sensing reports, then authenticates and encrypts the aggregated report. Encrypted reports, while traversing the network to base station, are further processed by network aggregators. Upon receiving an encrypted report, a network aggregator matches it with the stored keywords and then processes the report based on certain policies such as forwarding to the next hop, updating it or simply dropping it on detecting duplicates. The reports are only decrypted at their destinations. We ensure the confidentiality of the reports, as the intermediate aggregators learn nothing about them except several encrypted keywords. We analyze the resilience and overhead of the RAS scheme with the simulation results showing that the scheme is resilient against node compromise attack with reasonable performance overhead. Our analysis show that the scheme is secure, resilient and efficient.