Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn about Chitu, a novel Byzantine-Fault Tolerant consensus protocol that minimizes unnecessary fallback mechanisms in distributed systems through this 18-minute conference presentation from USENIX ATC '25. Explore how researchers from Shanghai Jiao Tong University reconsidered consensus fundamentals to develop a framework that avoids relying on timing assumptions, designated leaders, or randomization as intrinsic requirements. Discover how Chitu's asynchronous DAG-based approach enables nodes to first attempt consensus through pure message exchange, only falling back to mechanisms like random coins or leader election when correct nodes have divergent opinions on proposals. Examine the protocol's performance characteristics, including its ability to commit proposals in just four message delays in optimal conditions while maintaining O(1) time complexity in expectation during worst-case scenarios. Review experimental results from Amazon EC2 deployments demonstrating Chitu's significant latency improvements compared to existing DAG-based protocols that always incorporate leaders or randomization in their execution paths, making this essential viewing for distributed systems researchers and practitioners working on Byzantine fault tolerance and consensus mechanisms.
