Revisiting unpredictability-based RFID privacy models

Abstract

Recently, there have been several attempts in establishing formal RFID privacy models in the literature. These models mainly fall into two categories: one based on the notion of indistinguishability of two RFID tags, denoted as ind-privacy, and the other based on the unpredictability of the output of an RFID protocol, denoted as unp-privacy. Very recently, at CCS'09, Ma et al. proposed a modified unp-privacy model, referred to as unp′ -privacy. In this paper, we first revisit the existing RFID privacy models and point out their limitations. We then propose a new RFID privacy model, denoted as unp*-privacy, based on the indistinguishability of a real tag and a virtual tag. We provide justification for the new model and formally clarify its relationship with ind-privacy model. Finally, we modify Ma et al.'s 2-round RFID protocol to a 3-round mutual authentication RFID protocol and prove that it is of unp * -privacy. © 2010 Springer-Verlag Berlin Heidelberg.