An Intelligent Comparative Evaluation of the MDDM Framework for Secure Data Encryption Using Hill Cipher, Playfair and AES-128 -GCM

The paper discusses a systematic comparative evaluation of MDDM (Matrix based Dominating set Driven Message Cipher) with three reference algorithms: Playfair Cipher, Hill Cipher, and AES-128-GCM. The aim of this study is not to place MDDM as a substitute for verified standard encryption,
but to analyse how the developed graph-algebraic framework acts compared to classical matrix/substitution based cipher approach and an advanced engineering reference model. 57 plaintext cases, including both single-block and decomposed block inputs with each case repeats 10 times,
are evaluated for the comparison. MDDM is evaluated along with two shared
key modes: the unimodular/triangular key mode and the random invertible
key mode. In the comprehensive run MDDM attains 14.38% average bit
avalanche under triangular/unimodular key policy and 15.84% under
random-invertible key policy, which is above Hill Cipher ( 8.45%) and
Playfair (5.90%) but remains far below from AES-128-GCM (49.97%).
MDDM also provides two internal integrity checks IC1 & IC2 that detected
all 2,400 performed tampering tests, while Hill cipher and Playfair does not
have such an integrated tamper protection. These contributions come with
a clear observations that: In Python/SymPy Implementation MDDM is
slower and the cipher text expansion is highest, approximately 20-21x. And
the analysis support placing MDDM as a Cipher framework demonstrating
graph-algebraic pipeline along with the integrity verification rather than a
competing/reference method to AES-GCM.