Hydra's Heads – Petr Kuznetsov
summary
Argues that global total-order consensus is overkill for many block‑style systems and shows a concrete alternative that adds partial trust to the model, uses reliable broadcast primitives, and drops full consensus to cut cost. Presents a prototype implementation, benchmarks against the 2019/2020 popular Byzantine consensus baseline, and reports up to two orders of magnitude improvement in latency/throughput while preserving necessary safety properties. Explains tradeoffs, when partial ordering suffices, and which applications benefit from the cheaper protocol.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 3:11Multithreading to distributed systems
- 8:12Blockchain myths
- 12:03Blockchain science state
- 18:42Trustless consensus alternatives
- 22:34Improving blockchain new approaches
- 26:51Byzantine protocols applications
- 33:33Other speaker research
- 34:14Quantum computers consensus
- 36:22Consensus in neural networks
- 40:16Optimizing sync in concurrency
- 44:13Call to young researchers
related talks
46mDistributed teams that actually work
31mDistributed teams that actually work!
47mDistributed Teams — Can They Work?
46mLeading Distributed Teams
08mHow to run distributed teams
08mWhat does the Future of Events Look Like?
32mDistributed DevOps
15mBeyond Code: Building trust in multinational and (not only) remote teams