Supporting a high-throughput distributed database like VoltDB in long-term, real-world deployments quickly demonstrates that many assumptions about traditional database systems or failure models are invalid. This talk will focus on determinism and the role it plays in VoltDB applications, and will answer the following questions: What is determinism in the context of distributed databases?
What are some basic sources of non-determinism and how can they be addressed? In what awesome ways can determinism be leveraged in distributed systems? How can non-determinism risks be mitigated and what are some interesting related engineering tradeoffs?