Customer Story
Metabit Trading manages over $1 billion in assets. By switching from a traditional OLAP database to RisingWave, they cut computing costs by 95% and improved data timeliness by 3x.
The Challenge
Metabit Trading's existing OLAP-based monitoring system was limited to 100 queries per second and could not handle concurrent queries from multiple users. The system lacked strong consistency guarantees, and scaling to meet growing demands only degraded overall performance further.
Horizontal scaling paradoxically worsened query performance, while hundreds of CPU cores were consumed for basic monitoring workloads. For a quantitative investment firm managing over $1 billion in assets, these limitations posed real operational risk.
The Solution
RisingWave replaced the OLAP query pattern with a streaming architecture: Kafka ingests raw data, materialized views perform continuous incremental computation, and threshold-based alerts fire in real time. Strong consistency and user-defined functions provide the reliability and flexibility Metabit requires.
Incremental computation replaces full table scans on every query. User-defined functions (UDFs) enable custom business logic in SQL, and threshold-based alerting triggers sub-second notifications for critical trading events.
Results
Metabit achieved a 95% cost reduction compared to their no-MV baseline and a 70% reduction versus their MV-enabled baseline. Data timeliness improved by 3x, compute nodes were halved, and the system now handles tens of thousands of queries per second with sub-second alert latency.
| Metric | Before | After (RisingWave) |
|---|---|---|
| Cost vs no-MV baseline | 100% | 95% reduction |
| Cost vs MV baseline | 100% | 70% reduction |
| Data timeliness | Baseline | 3x improvement |
| Compute nodes | Full cluster | 50% fewer |
| Alert latency | Seconds+ | Sub-second |
| Query throughput | 100 QPS | Tens of thousands QPS |
See how RisingWave can cut your infrastructure costs by up to 95%.
Reduce Your Streaming Costs →