Picodata Sirin
A Cassandra-CQL-compatible plugin on the Picodata shard-per-core distributed engine. Drop-in for Apache Cassandra and Scylla — no application code changes. Strong consistency by default, faster LWT via Raft, optional in-memory storage.
Download Picodata Sirin documentation
Key capabilities
-
Compatibility with CQL and Apache Cassandra drivers -
Strong data consistency and high LWT performance -
Automatic data compression with two storage modes: in-memory or disk -
Add nodes online with background rebalancing -
Shard-per-core throughput on the same cluster as the SQL core -
Drop-in replacement for Cassandra and Scylla without application code changes
Why teams choose Sirin
-
Smart data management
- Automatic Zstd compression at the page level.
- True in-memory tables — all data lives in memory, not just the write-ahead buffer.
- Picodata's LSM-tree compacts metadata and cleans up automatically. No more REPAIR.
-
Distributed query execution
- A unified data space — every node sees the full schema and data.
- Tier architecture removes per-node token assignment.
- Consistent secondary indexes — efficient under writes.
-
Production-grade operations
- Tested on industrial MES/IIoT workloads.
- Horizontal scaling with built-in fault tolerance.
- Single operational footprint shared with the rest of the Picodata cluster.
-
Extreme performance
- Strong consistency by default — no trade-off between speed and reliability.
- Lightweight transactions run on Raft, not Paxos — and that makes them faster than Cassandra.
- In-memory tables — tens of times faster than Cassandra on point reads.
Sirin vs Apache Cassandra
| Capability | Apache Cassandra | Picodata Sirin |
|---|---|---|
| Consistency model | Eventual Consistency | Strong Consistency |
| Storage architecture | LSM-tree only | Hybrid (in-memory + LSM-tree) |
| In-memory tables | No | Yes |
| Data compression | Manual configuration | Automatic (Zstd) |
| Concurrency model | JVM threads, GC pauses | Shard-per-core, no GC |
| Lightweight transactions | Paxos | Raft |
| Open source | Apache 2.0 | BSD (core); commercial plugin |
Sirin on Picodata
Picodata is an open-source distributed SQL database with PostgreSQL wire compatibility, shard-per-core architecture, and a Rust plugin SDK. Sirin is a commercial plugin from the Picodata team. It proves the model — a Cassandra-CQL-compatible layer that runs against the same shared cluster as the SQL core.
Public benchmark of Sirin vs Apache Cassandra 5.0.6 on a comparable CPU budget: 8.9× more point reads and 89× lower read amplification. Migration methodology is on the migration page.