Adaptive traitor tracing for large anonymous attack

Abstract

In this paper we focus on traitor tracing technologies for the anonymous re-broadcasting attack where the attackers redistribute the per-content encrypting key or the decrypted plain content. To defend against an anonymous attack, content is usually built with different variations. For example, content is divided into multiple segments, each segment comes with multiple variations, and each variation is differently encrypted. Each user/player can only play back one variation per segment through the content. A typical traitor tracing scheme for re-broadcasting attack involves two basic steps, assigning the key/variation to devices (the assignment step) and detecting at least one traitor in the coalition when a series of pirated key/content are recovered (the coalition detection step). The traceability of a traitor tracing scheme is defined to be the number of recovered pirate copies of the content/keys needed in order to detect traitors. In [1] we presented a traitor detection scheme that tries to detect the entire coalition all together. This significantly improved the traditional one-by-one detection approaches in the literature. However, the traceability of the traitor detection scheme has a up limit that is constrained by the number of variations q one can build into the content. We are motivated to improve the traceability on a larger collusion attack and lift the up-limit on traceability with a given q. In this paper we will show a new traitor tracing approach that will assign the variations with skewed probabilities. Our approach not only lifts the tracing up-limit but also enables the tracing agency to assign the variations so as to maximize the traceability for a given coalition size. Our traceability results show that it is possible to achieve good traceability when traitor size exceeds q, and continue doing well even after the coalition size reaches q log q. Copyright 2008 ACM.