How data tiering works
- Tier categories:
- Hot: For active, high-velocity data requiring lowest latency (e.g., recent sensor readings, real-time events).
- Warm: For less frequently accessed data where cost-performance balance matters (e.g., recent logs, near-real-time archives).
- Cold: For archival or seldom-used data where retention is required but access is infrequent (e.g., historical records, compliance archives).
- Automatic or manual tiering: Administrators can define policies to move data automatically between tiers based on age, size, value or retention rules, or trigger movements manually when needed.
- Resource optimisation: By aligning storage class with workload priority, CrateDB allows you to allocate premium resources to critical data and more economical storage to older or less-active data.
- Unified access: Regardless of tier, you can query across Hot, Warm and Cold data using the same SQL interface, preserving analytics continuity without separate silos.
.png?width=800&height=331&name=Data%20Tiering%20(1).png)
Why it matters
- Cost savings: Storing older or less-used data on lower-cost tiers dramatically reduces infrastructure spending while preserving ability to query when needed.
- Performance where it matters: Your most active workloads (Hot tier) get the fastest storage and compute, ensuring high throughput and low latency.
- Storage flexibility: As data grows, you don’t need to treat all data equally — you can tier with intention.
- Data retention & compliance: Meets needs for long-term retention, audit trails, or regulatory archives without sacrificing budget or performance.
- Simplicity at scale: With CrateDB’s unified architecture, you avoid maintaining separate systems for hot vs cold data, queries span tiers transparently.
Typical use cases
- IoT & Sensor Analytics: Keep last 30 days of high-velocity telemetry in the Hot tier, move older data to Warm, and archive multi-year history in Cold.
- Log & event data platforms: Store recent logs for real-time monitoring in Hot, older logs for trend analysis in Warm, and compliance required data in Cold.
- Enterprise retention policies: For regulated industries, store recent operational data on-site (Hot/Warm) and archive legacy data to cost-efficient Cold tier while maintaining query capability.
- Hybrid Deployments / Edge Scenarios: On edge or constrained hardware, keep only Hot data locally while syncing Warm or Cold data to central storage or cloud.
Best practices
- Define clear policies (age, size, access frequency) for data movement between tiers.
- Use partitioning or timestamp-based sharding to make tiering decisions efficient and predictable.
- Monitor access patterns and adjust tier definitions according to changing workload behaviour.
- For Cold tier data, review availability vs cost trade-offs and ensure that query performance expectations are aligned.
- Document your data lifecycle and retention rules, especially for compliance and audit purposes.
