Scaling distributed ledgers: Achieving arbitrary throughput and sub-second latency

This talk describes two techniques to improve the throughput and latency of blockchain technologies.

First, this talk presents the design, implementation and evaluation of some key sharded blockchains. Sharding is one of the main approaches to address blockchain scalability, the key idea is to create groups (called shards) of nodes that handle only a subset of all transactions and system state. Sharding allows blockchains to achieve linear scalability; that is, the throughput of the system is theoretically unbounded as it can be arbitrarily increased by adding new shards. Secondly, this talk presents a novel distributed settlement system for pre-funded payments that can be used as a financial side-infrastructure to support retail payments of a primary system. This side-infrastructure complements scalable (high-throughput) blockchains by achieving extremely low latency by foregoing the expenses of consensus, making the system applicable to point of sale payments.