Privacy preserving computation of trust-value in wireless sensor networks

Abstract

The traditional cryptographic mechanisms are not enough for wireless sensor networks (WSNs) when the control of sensor nodes are taken by an interior adversary. Trust manage ment system can solve this problem efficiently, and enhance the security and reliability of WSNs. One of the key problems in trust management system is computing trust values of nodes. However, previous works pay little attention to preserving data privacy during computing trust values. This paper focus on the problem that how to protect the data privacy while computing trust values, and proposes a privacy preserving trust value computation (PPTC) protocol to solve this problem. We develop a new private distributed scalar product protocol based on semi-honest third party to achieve the PPTC protocol. A great advantage of PPTC protocol is that it does not need any encryption or decryption operations, and the computation and communication overheads is low to O(n). What's more, we analyze the correctness and security of the PPTC protocol and prove that the protocol can resist to collusion of up to n -1 users.

Privacy‐friendly weighted‐reputation aggregation protocols against malicious adversaries in cloud services

“…The performance comparison is provided in Table I. In [15], the protocol is constructed by the arithmetics operations, and then, the computational cost is lowest; however, this protocol has to employ a trusted agent to distribute the credentials, while it is generally preferred to avoid trusted third party in secure multiparty computation protocols. Compared with the protocols in the semi-honest model, the extra computation/communication costs are added by the proofs of knowledge.…”

“…Asharov et al [14] presented a fair and efficient secure multiparty computation with reputation systems. Yuan et al [15] proposed a privacy-preserving trust value computation protocol to protect the data privacy while computing trust values, but their protocol is secure in the semi-honest model. The advantage of this protocol is that it does not need any encryption or decryption operations, and the computation and communication overhead is reduced to O.n/, while resisting a collusion of up to n 1 malicious users.…”

“…Compared with the protocols in the semi-honest model, the extra computation/communication costs are added by the proofs of knowledge. The performance comparison is provided in Table I. In [15], the protocol is constructed by the arithmetics operations, and then, the computational cost is lowest; however, this protocol has to employ a trusted agent to distribute the credentials, while it is generally preferred to avoid trusted third party in secure multiparty computation protocols. In [7], the protocol is executed in a decentralized network, and it is secure against malicious adversaries.…”