Research On Privacy-preserving Aggregation For MIN/MAX Computation In WSN

Data Aggregation mechanisms in Wireless sensor networks (WSNs) is one of the most efficientmethods of reducing data communication. And privacy-preserving is the fundamental securityrequirement of the users’ data. Efficient privacy-preserving aggregation algorithms play animportant role in applications of the WSNs. The computation of data aggregation functions likeMIN/MAX is widely used in production and living. So it is necessary to do some research on theprivacy-preserving aggregation for MIN/MAX computation.This thesis proposes two new privacy-preserving aggregation algorithms for MIN/MAXcomputation based on the analysis of the unique of the MIN/MAX computation. The first one isSMCPART: a privacy-preserving aggregation algorithm based on the secure multiparty computation.And the second one is DCPART: a privacy-preserving aggregation algorithm based on the disjointedcluster. Regarding SMCPART, the cluster leader gets the node number that its value is MIN/MAXthrough the secure multiparty computation within the cluster, and then the node sends its value tothe sink node. For DCPART, it slices the node value into two pieces and each piece is transmitted tothe sink node through the different paths with operations of aggreagation and computation duringtransmission.These two algorithms are simulated on TinyOS platform with the tool of TOSSIM. Based onthe simulation result and the theory, we make some comparison between SMCPART and DCPART.And the result shows that these two algorithms are available for privacy-preserving. But in detail,DCPART can protect the privacy that whether a node is a MIN/MAX node and SMCPART does not.Moreover, when the number of the node is determinated, the commucation of the DCPART is stablebut SMCPART fluctuates wildly. DCPART demonstrates better performance in privacy-preservingand communication effiency compred with SMCPART.