PDA: Privacy-preserving Data Aggregation in Wireless Sensor Networks

A wireless sensor network (WSN) is an ad-hoc network composed of small sensor nodes deployed in large numbers to sense the physical world. Wireless sensor networks have very broad application prospects including both military and civilian usage. Sensors are usually resource-limited and power-constrained. They suffer from restricted computation, communication, and power resources. Sensors can provide fine-grained raw data. Alternatively, they may need to collaborate on in-network processing to reduce the amount of raw data sent, thus conserving resources such as communication bandwidth and energy. We refer to such in-network processing generically as data aggregation. In many sensor network applications, the designer is usually concerned with aggregate statistics such as SUM, AVERAGE, or MAX/MIN of data readings over a certain region or period. As a result, data aggregation in WSNs has received substantial attention. As sensor network applications expand to include increasingly sensitive measurements of everyday life, preserving data privacy becomes an increasingly important concern. For example, a future application might measure household details such as power and water usage, computing average trends and making local recommendations. Without providing proper privacy protection, such applications of WSNs will not be practical, since participating parties may not allow tracking their private data. We present two privacy-preserving data aggregation schemes called Cluster-based Private Data Aggregation (CPDA) and Slice-Mix-AggRegaTe (SMART) respectively, for additive aggregation functions in WSNs. The goal of our work is to bridge the gap between collaborative data aggregation and data privacy in wireless sensor networks. When there is no packet loss, in both CPDA and SMART, the sensor network can obtain a precise aggregation result while guaranteeing that no private sensor reading is released to other sensors. Observe that this is a stronger result than previously proposed protocols that are able to compute approximate aggregates only (without violating privacy). Our presented schemes can be built on top of existing secure communication protocols. Therefore, both security and privacy are supported by the proposed data aggregation schemes. In the CPDA scheme, sensor nodes are formed randomly into clusters. Within each cluster, our design leverages algebraic properties of polynomials to calculate the desired aggregate value. At the same time, it guarantees that no individual node knows the data values of other nodes. The intermediate aggregate values in each cluster will be further aggregated (along an aggregation tree) on their way to the data sink. In the SMART scheme, each node hides its private data by slicing it into pieces. It sends encrypted data slices to different intermediate aggregation nodes. After the pieces are received, intermediate nodes calculate intermediate aggregate values and further aggregate them to the sink. In both schemes, data privacy is preserved while aggregation is carrying out. We evaluate the two schemes in terms of efficacy of privacy preservation, communication overhead, and data aggregation accuracy, comparing them with a commonly used data aggregation scheme TAG, where no data privacy is provided. Simulation results demonstrate the efficacy and efficiency of our schemes