BLOCKCHAIN-ORCHESTRATED CYBER-PHYSICAL SUPPLY CHAIN NETWORKS WITH BYZANTINE FAULT TOLERANCE FOR MANUFACTURING ROBUSTNESS

Abstract

This study quantitatively examined how blockchain orchestration and Byzantine fault tolerance (BFT) were associated with manufacturing robustness in cyber-physical supply chain networks under varying workload and fault conditions. A controlled, scenario-based experimental design was implemented using network-run–level simulation outputs and system log data, enabling systematic comparison across three coordination regimes: centralized coordination, non-BFT blockchain coordination, and BFT-enabled blockchain orchestration. The analytic sample comprised 360 validated network runs spanning low, medium, and high event loads and low, moderate, and elevated fault intensities. Manufacturing robustness was operationalized using downtime probability, throughput stability, schedule deviation, recovery-time behavior, service level variance, and inventory oscillation indicators, while consensus performance and data integrity were modeled as explanatory mechanisms. Descriptive findings showed that centralized coordination achieved the highest mean throughput (920.3 TPS) and lowest mean confirmation latency (1.12 seconds) but exhibited higher downtime probability (5.20%) and longer recovery time (49.0 minutes) under fault stress. BFT-enabled orchestration demonstrated lower throughput (547.0 TPS) and higher confirmation latency (4.71 seconds) but achieved superior robustness outcomes, including lower downtime probability (2.80%), reduced schedule deviation (14.3 minutes), and faster recovery (31.3 minutes). Correlation analysis indicated strong associations between consensus performance and robustness, with deadline adherence negatively correlated with downtime probability (r = −0.66) and recovery time (r = −0.65). Data integrity metrics were also strongly related to robustness, as data completeness showed a negative correlation with downtime probability (r = −0.61). Hierarchical regression results revealed that coordination regime and BFT configuration explained 31% of the variance in manufacturing robustness, which increased to 63% when consensus performance and data integrity were included. Mediation analysis showed that consensus performance and data integrity jointly accounted for a substantial portion of the architecture effect, with a total indirect effect of −0.21. Interaction models further indicated that the robustness advantages of BFT-enabled orchestration strengthened under higher workload and fault intensity. Overall, the findings demonstrated that manufacturing robustness in cyber-physical supply chains was shaped by coordination architecture through measurable performance and integrity mechanisms, providing quantitative evidence on how distributed trust and fault tolerance influenced operational stability under stress.