Abstract
The main purpose of this work is to raise awareness about a primitive that can provide large efficiency gains in post-quantum cryptography: multi-recipient KEMs, ormKEMs. In a nutshell, when encapsulating a key to N parties, an mKEM generates a single ciphertext that can be decapsulated by all parties. The size of an mKEM ciphertext can be significantly smaller than the sum of the sizes of N KEM ciphertexts. Moreover, individual receivers only need a small part of the mKEM ciphertext to run decapsulation. We then propose mKyber, a very compact mKEM based on Kyber. Asymptotically, the size of an mKyber multi-recipient ciphertext is 16 times smaller than the sum of the sizes of N Kyber ciphertexts. The algorithmic description and parameters of mKyber and Kyber are very similar, which facilitates the re-use of existing security analyses, implementations, and formal verification tools that have been developed for Kyber. Finally, we showcase some selected applications. mKEMs can be used to greatly reduce the bandwidth cost of the group key agreement protocol underlying the Message Layer Security (MLS) secure group messaging standard. Reducing bandwidth is one of the primary design considerations for MLS. More fundamentally, mKEMs reduce the cost of broadcasting private information to groups of recipients (e.g. a fleet of Cloud Hardware Security Modules).
