| 要約: | Blockchain provides a secure, decentralized, and distributed ledger system. However, scalability is a key limitation of blockchain, particularly for high-throughput applications. Direct Acyclic Graphs (DAGs) offer a promising solution by enabling the simultaneous processing of multiple transactions. However, achieving efficient and stable consensus in DAG-based systems is a challenge, as conventional linear ordering mechanisms do not fully address the complexities of DAG structures. In this article, we present a novel consensus protocol, specifically designed for DAG-based distributed ledgers, known as the Score-Based Periodic Ordering and Consensus Protocol. Our approach builds on the protocols of Phantom and GHOSTForge, focusing on enhanced order stability and scalability. It introduces score-based ordering at periodic checkpoints for individual minors. This selective checkpoint mechanism reduces computational overhead by limiting block reevaluation. The protocol also uses a score-based consistency check to ensure finality and resilience against double-spending. Our experiments show significant efficiency gains with stable order and reduced reconsideration across various checkpoint intervals. Furthermore, the experimental results demonstrate that global consensus and order convergence are achieved among miners. Stable order converges without extra communication, even under network heterogeneity. In addition, the evaluation further highlights resilience, with minimal uncommitted blocks even under varying network conditions.
|
|---|
