This talk presents Conflux, a scalable and decentralized blockchain system that achieves near-optimal throughput and latency, with respect to underlying P2P gossip network bandwidth and propagation delay. Conflux operates with a novel consensus protocol which optimistically processes concurrent blocks without discarding any as forks and adaptively assigns weights to blocks based on their topologies in the Conflux ledger structure (called Tree-Graph). The adaptive weight mechanism enables Conflux to detect and thwart liveness attack by automatically switching between a fast mode in normal scenarios and a structured mode to ensure consensus progress during liveness attacks. We evaluated Conflux on Amazon EC2 clusters with up to 12k full nodes. The consensus protocol of Conflux achieves a block throughput of 9.38Mbps with 20Mbps network bandwidth limit per node while confirming transactions under half a minute.
Ming Wu is a co-founder and the CTO of Conflux which is a blockchain startup. Before joining Conflux, he was a senior researcher at Systems Research Group in Microsoft Research Asia. Ming earns his Ph.D. from the Institute of Computing Technology, Chinese Academy of Sciences in 2007, and his bachelor degree from University of Science and Technology of China. His main research interests include distributed transaction processing systems, graph computing engines, artificial intelligence platforms, and blockchain systems. He has many publications on top conferences in computer system area, such as SOSP, OSDI, NSDI, ATC, EuroSys, SoCC, VLDB, etc. He served as Program Committee Members of OSDI, ASPLOS, HotDep, and MiddleWare, and also the Publication Chair of SOSP'17.