In the DRM setting, the attacker is a very powerful adversary, owning the software as well as the underlying hardware. This context is far different from the black-box attacker commonly considered in conventional cryptography. Therefore, cryptographers have tried to design new cryptographic tools fitting the DRM requirements. A related issue in cryptography is that of side-channel attacks, where the attacker is stronger than the black-box attacker, but usually weaker than a DRM attacker. In this paper, we aim to show that the study of side-channel attacks can benefit from DRM research, and in particular from the attacker models and solutions tailored to this specific setting. We focus on a specific issue, namely the cache attacks against the AES, and show how current counter-measures can be seen as restricted versions of a previous protection proposed in the DRM setting. We demonstrate that those kind of counter-measures are weak against cache-based side-channel attacks by reusing results from the DRM setting.