At the Hydra Conference 2021, Alex and Noah discussed with Alexey Fyodorov the process and insights gained from building a new storage engine specifically designed for the Kafka API, incorporating a thread-per-core execution model alongside the Raft data replication model. This innovative storage solution eliminates the use of virtual memory and page cache, opting instead for custom read-ahead and write-behind strategies to achieve predictable latencies. The discussion emphasized the co-design approach that integrates these strategies with Raft, aiming to optimize data handling and replication efficiency.