ScyllaDB Cluster
ScyllaDB is a wide-column NoSQL database engineered as a close drop-in replacement for Apache Cassandra and DynamoDB workloads. It uses a shard-per-core C++ architecture to extract higher throughput and more predictable latency from the same hardware than JVM-based alternatives.
Baseline topologies
| Topology | Minimum nodes | Roles |
|---|---|---|
| ring-3-node | 3 | scylladb-node |
| ring-scaled | 6 | scylladb-node |
What drives the cost
- Dataset size (GB)
- Total dataset size before replication. Combined with replication_factor to determine raw storage requirements.
- Replication factor
- Number of data replicas across the cluster ring. Multiplies storage and write amplification.
- Peak write operations per second
- Peak sustained write operations per second. Primary driver for CPU core count due to shard-per-core architecture.
Where it runs
bare_metal, dedicated, cloud_vm
Size ScyllaDB Cluster and get a cost estimate