Upgrading Security Properties for Updatable Public-Key Encryption through Modular Transformations

Abstract

Updatable public-key encryption (UPKE) serves as a key tool for ensuring compromise resilience in dynamic group messaging systems such as MLS and TreeKEM. Prior work has introduced a variety of security notions for UPKE, including chosen-randomness and chosen-update security, chosen ciphertext attack (CCA) security, adaptive corruption resilience, public verifiability, and adversarially generated forking update structures. However, these notions make the security definition complex, and there is very little work on instantiating schemes that satisfy it.

In this work, we develop generic transformations that enable a sequence of upgrades from weaker to stronger UPKE notions, each introducing additional security properties. These security properties include straightline and forking update structures, selective and adaptive corruptions, public verifiability, chosen randomness and chosen-update security, and CPA and CCA notions. We define intermediate security notions that serve as useful stepping stones while adding the aforementioned properties. Some transformations rely only on standard tools such as random oracles, while others additionally exploit natural structural properties of schemes such as update uniformity and ciphertext updatability, which we identify and formalize in this work. This modular structure allows one to isolate and focus on specific combinations of properties of interest, facilitating easier security analysis.