OptiSecure Framework: NIST based 6G-native A nonymity-Preserving I
dentity A ttestation Framework for AI-Enabled IIoTs in Smart
Manufacturing Industry 5.0
Abstract
In the transformative landscape of B5G/6G-based Artificial
Intelligence enabled-Industrial Internet of Things (AIeIIoTs), deployed
in an adverse and hostile environment of the Smart Manufacturing
Industry 5.0 environment(SMIE), the convergence of Cognitive
Intelligence has revolutionized the real-time sensing, critical
environment cognizance-cum-monitoring, imperative decision-making,
seamless remote cloud connectivity and efficient spectrum utilization
for dynamic adaptability in 6G network architecture of IIoTs. In this
inimical and agile digital realm, secure communication in 6G-AIeIIoTs
plays an inevitable role, and to ensure this, the proposed research
unveils a novel NIST-Zero Trust Architecture (NIST-ZTA)-based
anonymity-preserving identity attestation-cum-mutual authenticated key
agreement framework especially tailored for Split Learning in
6G-SMIE/AIeIIoTs encompassing NIST-POLP (Principles of Least Privilege).
By embracing NIST-ZTA, framework adeptly eliminates the inherent trust
assumption of a covert communication channel between the edge-enabled
Access Point (EAP) and cloud-enabled-remote Authentication Server (CAS)
by introducing cryptographically agile authentication between EAP and
CAS. The proposed framework contrives a resilient anonymity-preserving
cryptographic framework fortified by AES-128 in Galois-Counter Mode
(GCM) and ingeniously incorporates only Symmetric Key Cryptography for
mutual authentication of server-side and client-side sub-models
assimilating Session, User-Session and Network Identifiers. The proposed
work also proficiently counters a spectrum of system-critical threats
ranging from Ephemeral Secret Leakage to Rogue Access Point Attacks.
Furthermore, network simulation and real-time test-bed performance
analysis of the proposed framework demonstrates that it significantly
optimizes the computational cost by 12.5% and communicational cost by
14.29%, as compared to the state-of-the-art. The proposed
framework models the dynamic interaction between adversaries and the
authentication system as a strategic game. Using Nash equilibrium and
Steepest descent convex optimization, the system optimizes resource
allocation to defend against both known and emerging attack vectors.
This game-theoretic formulation not only enhances the system’s ability
to anticipate adversarial strategies but also ensures robust and
adaptive security for AI-enabled IIoT environments. Lastly, the security
posture of the proposed framework in 6G- AIeIIoTs landscape is
established by both formal viz. BAN Logic and Scyther-Formal Security
Analysis Tool and informal cryptographic-cum-network security analysis.
