Enforcing Policies in Content Manipulation Signature Schemes

Recently proposed content manipulation signatures schemes have a broad range of applications, such as authenticated content extraction, adaptive multimedia content delivery, and XML authentication. Most of the constructions are based on Merkle trees. We show that these constructions are unable to enforce manipulation (extraction) policies, i.e. rules defining what manipulations are permitted/prohibited. We propose a generalization of Merkle trees, so called Relaxed Merkle trees. We present their ability to enforce some useful classes of policies, analyze their expressive power, and discuss few enhancements.