In this thesis, we look at definitions and black-box constructions with efficient instantiations for broadcast encryption and traitor tracing. We begin by looking at the security notions for broadcast encryption found in the literature. Since there is no easy way to compare these existing notions, we propose a framework of security notions for which we establish relationships. We then show where existing notions fit within this framework. Second, we present a black-box construction of a decentralized dynamic broadcast encryption scheme. This scheme does not rely on any trusted authorities, and new users can join at any time. It achieves the strongest security notion based on the security of its components and has an efficient instantiation that is fully secure under the DDH assumption in the standard model. Finally, we give a black-box construction of a message-based traitor tracing scheme, which allows tracing not only based on pirate decoders but also based on watermarks contained in a message. Our scheme is the first one to obtain the optimal ciphertext rate of 1 symptotically. We then show that at today's data rates, the scheme is already practical for standard choices of values.
