UC Davis

With the advent of Bitcoin, the first real-world and practical blockchain system, researchers and practitioners have turned their eyes and started developing new cryptocurrencies and blockchains. There are two types of blockchains: Permissioned and Permissionless. Public or permissionless blockchains are highly decentralized so that replicas in the system can join and leave anytime, and they are computation-intensive. Permissioned or private blockchains require a set of known replicas, and they are communication-intensive. Blockchain applications have use cases in various fields such as food supply, medical records, healthcare, supply chain, trade finance, manufacturing.Every blockchain system maintains an immutable ledger of transactions and works in a trustless environment to process and replicate transactions among a set of replicas. In permissioned blockchains, replicas and identities are known, and yet they could act Byzantine (behave maliciously). At the core of each blockchain system, there is a consensus protocol, which is the main focus of our work. Due to the nature of these consensus protocols (being communication-intensive), they are poorly scalable, especially when participants of this distributed system are far apart geographically. In the first part of our work, we introduce GEOBFT, a global-scale resilient permissioned blockchain system. GEOBFT scales Byzantine Fault-Tolerant (BFT) protocols in the WAN environment when the system’s nodes are in different continents in the world and far apart. It uses the notation of clustering by grouping replicas in different geographical locations and requires all nodes to maintain the full ledger. Limitations of a fully replicated ledger in a WAN environment led us to step into the Sharding world. So many researchers and works have tried to use sharding to scale BFT protocols. In the second part of this work, we present RINGBFT a high throughput resilient sharding protocol. RINGBFT introduces a topology of a Ring among clusters of replicas to reduce the communication and prevent deadlocks. While designing a performant BFT protocol is crucial, implementing and building a real system that uses the protocol is even more important. A poorly build fabric can mask all the advantages of a creatively designed BFT protocol. As the third part of our work, we present a permissioned blockchain fabric called ResilientDB. We have fully designed and implemented a well-crafted fabric to show the inherent ability and limitations of all BFT protocols. As we explored deeply in the permissioned environment, we project to step into the permissionless world as our next step, and we propose our idea for Hybrid-Blockchains and BFT with Trusted Components in the last chapter.