Uncoordinated Pseudonym Changes for Privacy Preserving in Distributed Networks

Pseudonyms have been adopted to preserve identity privacy of nodes in distributed networks. Frequent and unlinkable changes of pseudonyms need to be enabled by having at least k nodes change together to confuse potential eavesdroppers. Existing approaches either depend on the coordination from central controllers, or involve interactive signaling between the nodes. This can potentially compromise privacy. This paper proposes a fully uncoordinated approach to change pseudonyms in distributed networks, where each node uses a pseudonym until its expiration and then changes after a random delay. We develop a new model to analyse the time-varying population of changing pseudonyms. Critical conditions are analytically established, under which individual nodes can independently change their pseudonyms while their identity privacy is preserved. The conditions are validated by illustrative examples. Corroborated by simulations, the accuracy of the analytical model improves, as the number of nodes increases. The analysis confirms that, the k-anonymity can be achieved at a negligible throughput loss in the case of large networks.