Title: Rethinking Large-Scale Consensus
Abstract: Although distributed consensus has been studied by for three decades, they were not deployed at a large scale until decentralized cryptocurrencies like Bitcoin. In this talk, I will explain why the classical theoretical foundation for distributed systems is insufficient for capturing the robustness and game theoretic properties we care about for new decentralized environments. Specifically, I will demonstrate why almost all classical "synchronous" consensus protocols are underspecified and thus unimplementable in practice. I will then describe a new model called "best-possible partition tolerance" that allows us to achieve honest-majority consensus while providing resilience to network partitions (the combination of which was classically deemed impossible due to a well-known lower bound by Dwork, Lynch, and Stockmeyer).
Bio: Elaine Shi is an Associate Professor in Cornell University. Her research interests include cryptography, distributed systems, algorithms, foundations of blockchains, and language-based security. She is a recipient of the Packard Fellowship, the Sloan Fellowship, the ONR YIP award, the NSF CAREER award, the NSA Best Scientific Security paper, and various other best-paper and research awards. Elaine obtained her Ph.D. from Carnegie Mellon University. Before joining Cornell, she was an Assistant Professor in UMD.