Rootconf 2025 Annual Conference - 16th and 17th May

Valkey, the community-driven fork of Redis, an in-memory data store which has introduced several innovations while maintaining compatibility with Redis APIs. In this session, we deep-dive into the internals of Valkey Cluster, focusing on its distributed architecture, failure detection, partition handling, and data sharding mechanisms.

We will dissect key components like the cluster bus protocol, slot-based sharding, gossip-based health monitoring, and failover orchestration. The talk will also explore recent improvements in efficient message transfer, failover and extensions propagation.

Attendees will gain a practical understanding of:

• How Valkey achieves horizontal scalability beyond standalone setup
• Internals of failure detection and quorum-based failover

I’m Harkrishn Patro, Software Engineer @ AWS ElastiCache / Valkey Maintainer. I’ve overall 10 years of experience mostly spent around data systems. In the recent years, I spend most of my time on improving Valkey for the community. I’ve worked on improving memory efficiency, scaling clustering and pub/sub system of Valkey.

Helpful links:

• https://valkey.io
• https://github.com/valkey-io/valkey/
• https://www.linkedin.com/in/harkrishn-patro/