Scale to arbitrarily large database sizes and avoid rewrites or migrations. Say no to manually sharded databases that slow you down exactly when your business is growing.
Make data available whenever and wherever without any manual scripting or ad-hoc firefighting. Place primary replicas across fault-domains to ensure high write availability. Optionally extend the cluster with read replicas in remote regions for low-latency reads as well as for failover on active region failure.
Written ground-up in C++ to deliver avoid JVM garbage collection pauses and memory tuning issues. No quorum at read time and no anti-entropy/read repairs ensure high throughput and low p99 latency operations.
Use as the mission-critical, stateful complement to your stateless microservices. Kubernetes-like intent-based orchestration lets you port your database to new infrastructure reliably and effortlessly.
Run fewer nodes with larger disks to save on compute costs. Add tiered storage to efficiently store colder data on cheaper storage while remaining completely transparent to the application code.