A Data Integration Platform with Identifiable Falsification Detection and its Evaluation in Automotive Parts Manufacturing

To promote a decarbonized society, transparency in carbon footprint of products (CFP) data across supply chains has become critical. This paper proposes a blockchain-based platform for the automotive parts manufacturing industry that enables secure management and representation of each component's CFP while providing precise falsification localization capabilities. Unlike existing methods that handle CFP data in plaintext and cannot identify the exact location of falsification, our approach leverages Merkle tree structures and XOR operations to achieve both pinpoint identification of falsified components and secure CFP data management. In particular, by exploiting the commutative property of XOR operation, we eliminate the need for ordering information in hash aggregation, and by utilizing its self-inversion property, we can accurately distinguish between the components where falsification occurred and those merely affected by it. We implement and evaluate a prototype system using Hyperledger Iroha, blockchain-based open-source Distributed Ledger Technology framework. The execution time of falsification localization increased as the falsification rate grew. Specifically, when verifying a dataset corresponding to the components of an automobile, the process required up to 27.88 seconds, compared to 9.52 seconds in the previous work. The evaluation confirmed that execution time was sensitive to the number of parts and to duplication and falsification rates, while CPU and memory usage were largely unaffected by these two rates. Although the verification is slower than methods that cannot localize falsification, the proposed system provides the critical advantage of precise falsification detection while maintaining secure CFP data management.