Performance Analysis of Dynamic Scaling in Hyperledger Fabric-Based Academic Credential Verification Systems

Academic credential fraud represents a growing global challenge, with diploma mills generating billions of US dollars annually and large-scale cases of counterfeit degrees being reported across multiple countries. Blockchain technology offers a promising technical foundation for addressing this problem due to its inherent properties of immutability, decentralization, and cryptographic verifiability. However, empirical evidence on how blockchain-based academic credential verification systems behave under extreme and highly variable workloads—such as those observed during graduation periods—remains limited.In this study, a Hyperledger Fabric network deployed on Kubernetes is designed and experimentally evaluated under four different scaling configurations, including one static baseline and three auto-scaling variants. The system is subjected to six workload scenarios ranging from 50 to 1000 concurrent users. Experimental results demonstrate that an aggressive auto-scaling configuration achieves approximately 115% higher throughput and reduces latency by nearly 67% compared to a fixed-replica deployment under peak load conditions. Even at maximum load, response times remain below 500 ms at the 99th percentile, while sudden traffic surges are accommodated through scaling actions completed in approximately 50 seconds. The findings highlight clear trade-offs between responsiveness and resource efficiency and provide practical deployment guidelines for institutions planning to adopt blockchain-based academic credential verification systems.