International Journal of Science, Engineering and TechnologyAuthors: Aadit Kapare, Rachit Singh, Tanish Venkatesh, Reshma Sonar
Abstract: MerkleChain is a recently proposed owner-oriented blockchain architecture in which data owners keep their own separate Merkle trees and participating miners store only the structures and cryptographic commitments relevant to their subscriptions. The design aims to support privacy, controlled redaction, and parallel processing. This study examines a narrower question: does dividing a fixed number of records into owner-specific Merkle trees reduce the cryptographic work needed for construction, proof verification, and updates, if the owner roots must still be fed into a common root? A reproducible Python microbenchmark compares one large balanced global tree with sixteen balanced owner trees and a top-level tree of owner roots, at four data sizes from 256 to 16,384 records. The implementation uses SHA-256 with domain-separated inputs for leaves and internal nodes. In all tested power-of-two workloads, both arrangements require N − 1 internal-node hashes to build the full commitment structure, and the total proof path remains log₂(N) levels. Measured verification and update times are generally higher for the partitioned model at these small scales, while construction times stay nearly the same. These results do not refute MerkleChain’s wider claims about privacy and parallelism, because the experiment includes no chameleon hashes, distributed consensus, or concurrent miners. They show instead that ownership partitioning is a systems-design decision and should not be assumed to reduce hash complexity
DOI:
