| Suitable workload | Analysis, reporting, big data | Transaction processing |
| Raw data | Normalized data from various sources | Data captured as is in a single source, such as a transaction system |
| Data capture | Bulk writes operations, usually based on a pre-determined bulk batch schedule | Optimized for continuous write operations with new data available to maximize transaction throughput |
| Data normalization | Denormalized schemas, such as star schemas or snowflake schemas. | Highly normalized static schema |
| Data storage | Optimized for simple access and fast query performance using columnar storage | Optimized for high throughput write operations to single-row oriented physical blocks. |
| Data access | Optimized to minimize I/O and maximize data throughput | Large number of small read operations |