Design and Performance Evaluation of a Hybrid Blockchain Credential Verification Architecture for Universities

Abstract

Learning institutions are increasingly vulnerable to the theft of sensitive information facilitated by the sophisticated capabilities of Large Language Models (LLM) in social engineering attacks. The LLMs are usually aided by the use of social engineering through pretexting, impersonation, and even document-forgery activities. This research focuses on developing a hybrid Blockchain login information verification system that combines permissioned ledger governance with public rooting to ensure tamper evidence and fast revocation propagation, while keeping credentials off-chain. The system is designed for operational verification in learning institutions. It focuses on low-latency API verification, auditable issuance, and secure private information disclosure. Performance is measured by using queuing-based workload modeling and benchmarking parameters to compare the efficiency and security of centralized verification systems against permissioned-only ledgers. The system uses a hybrid model under baseline and adversarial traffic, modelled on LLM-driven phishing attacks. Results show reduced peak verification latency relative to permissioned-only designs, while preserving stronger integrity than centralised systems. The integration of identity management, effective dispute handling, and robust incident response mechanisms is a proof of concept.