cid 09821f54b84620afc728d67ab9683f0f261197843e25e2062b8f469ef9f61897 · original · via openalex
Title: Uncovering the Perfect Scalable Database for Hyperledger Fabric's Evolution Authors: Ivan Laishevskiy; Artem Barger; Vladimir Gorgadze Venue: IEEE International Conference on Omni-layer Intelligent Systems (COINS) Year: 2023 Canonical: doi:10.1109/coins57856.2023.10189331 URL: https://doi.org/10.1109/coins57856.2023.10189331 Collector: openalex OA: false ABSTRACT Hyperledger Fabric, a renowned permissioned blockchain platform, offers flexibility and customization through its simulate-ordervalidate paradigm and pluggable architecture. The world state database, a critical component in Hyperledger Fabric, captures the current state of blockchain applications. This study explores alternative state database implementations, including RocksDB, Boltdb, and BadgerDB, and evaluates their performance using diverse workloads, with a focus on memory and CPU utilization. Experimental results indicate that the evaluated alternatives surpass the default LevelDB and CouchDB provided by Hyperledger Fabric. Specifically, BadgerDB exhibits the highest throughput and lowest latency and bbolt demonstrates superior memory utilization efficiency. Moreover, we provide guidance on integrating new state databases into Hyperledger Fabric and present valuable insights for blockchain researchers and practitioners. By showcasing significant performance improvements, our research contributes to the advancement of enterprise blockchain technology and fosters scalability in permissioned blockchain systems.