A Key Encapsulation Mechanism from Low Density Lattice Codes

TL;DR

This paper introduces a new key encapsulation mechanism based on low density lattice codes that reduces key size while maintaining security, through innovative matrix representations and security analyses.

Contribution

It presents a novel KEM-LDLC design that decreases key size and improves efficiency by using Latin square LDLC matrices and Hermite normal form, with comprehensive security and performance evaluations.

Findings

Reduced key size via Latin square LDLC parity-check matrix

Enhanced security and efficiency over previous code-based KEMs

Successful security analysis against known cryptographic attacks

Abstract

Key Encapsulation Mechanisms (KEMs) are a set of cryptographic techniques that are designed to provide symmetric encryption key using asymmetric mechanism (public key). In the current study, we concentrate on design and analysis of key encapsulation mechanism from low density lattice codes (KEM-LDLC) to go down the key size by keeping an acceptable level of security. The security of the proposed KEM-LDLC relies on the difficulty of solving the closest vector problem (CVP) and the shortest basis problem (SBP) of the lattices. Furthermore, this paper discusses other performance analyses results such as key size, error performance, and computational complexity, as well as conventional security analysis against applied attacks. Reducing the key size is performed by two approaches: (i) saving the generation sequence of the latin square LDLCs parity-check matrix of as a part of the secret key set; (ii) using the hermite normal form (HNF) of the latin square LDLCs generator matrix as part of the public key set. These enhancements enable us to attain greater efficiency and security compared to earlier code-based KEMs.