Data Lifecycle, Architecture, and Storage
Lifecycle stages
Plan, capture, store, use/share, archive, and dispose. Each stage has governance obligations: retention schedules, legal hold, and defensible disposition. Know the difference between retention (how long to keep) and disposition (how to destroy), and that litigation holds suspend destruction.
Concepts to master
- Architecture layers: source systems, integration/ETL, data warehouse, data lake, and marts; operational vs. analytical stores.
- Schema-on-write vs. schema-on-read and the implications for quality and governance.
- Master data management and reference data.
- Storage choices: transactional databases, warehouses, object storage, and imaging archives; hot/warm/cold tiers.
- Backup, redundancy, disaster recovery, RPO and RTO.
- Data lineage and provenance across the pipeline.
- Archiving clinical data when systems are retired, and the need to preserve readability and legal record status.
Traps
Questions often confuse a data lake with a warehouse (raw, schema-on-read vs. curated, schema-on-write) or ask about RPO vs. RTO-RPO is acceptable data loss, RTO is acceptable downtime. Another trap: assuming deletion is always correct; legal hold and retention requirements override routine disposal.
Rule of thumb
For any data question, trace the lineage: where did it originate, how was it transformed, where is it stored, who can use it, and when will it be destroyed? If any link is undocumented, the architecture has a gap.
Worked example
A legacy EHR is retired. Before shutdown, data must be migrated or archived in a readable, retrievable form with an index, and retention schedules and legal holds must be honored-deleting the system without archiving can violate record-retention obligations.