Engineering solutions for end-to-end quality traceability in supply chains based on distributed ledgers and intelligent sensors

Abstract

The globalization of supply chains and the increasing complexity of logistics routes raise requirements for transparency and data protection regarding product quality, rendering traditional centralized record-keeping systems insufficient and highlighting the urgency of end-to-end traceability. This study aims to develop and quantitatively assess the engineering architecture of a cyber-physical system for end-to-end quality monitoring in supply chains based on the synergy of distributed ledger technologies and smart sensor nodes. A multi-level architecture is proposed, including a physical layer with autonomous sensor nodes based on ARM Cortex-M4 microcontrollers equipped with thermohygrometers, accelerometers, and GLONASS/GPS modules; a communication layer using LoRaWAN and NB-IoT/MQTT; edge gateways for data preprocessing; and a private blockchain layer on the Hyperledger Fabric v2.2 platform with smart contracts for the automatic validation and recording of product lifecycle events. Validation was conducted via simulation modeling of a complex pharmaceutical supply chain in the AnyLogic environment, varying the number of validating nodes (4-16), block size (1-4 MB), and consensus algorithms (Raft, PBFT), as well as through bench testing with 20 sensor nodes and real LoRaWAN and NB-IoT networks. Statistical processing was performed in R-project using analysis of variance (ANOVA) and correlation-regression analysis. It was found that the private blockchain provides acceptable latency and throughput compared to a centralized DBMS while significantly increasing resilience to Data Tampering, Replay, and Man-in-the-Middle threats, demonstrating an attack detection rate close to 100%. In contrast, a public blockchain with PoW consensus proved unsuitable for operational tasks due to high latencies. The study demonstrates the non-linear influence of block size, node count, and consensus type on TPS and latency, as well as the inverse effect of network scaling on performance, establishing a trade-off between decentralization and transaction processing speed. Power consumption measurements show that combining an event-driven monitoring mode with LoRaWAN ensures maximum autonomous operation time for sensor nodes (up to one year on a standard battery), imposing additional constraints on permissible data flow intensity. Comprehensive analysis confirms the engineering feasibility of end-to-end traceability systems based on private blockchains and IIoT, and provides a quantitatively grounded basis for the multi-criteria optimization of the architecture to meet specific requirements and constraints of industrial supply chains.