This white paper on HA Best Practices and DBaaS is pretty interesting, especially for its definitions of “HA Service Level Tiers”:
- Single-instance database.
- DR: Requires restore from the last backup, potential for data loss.
- RAC database.
- DR: Requires restore from the last backup (with potential for data loss).
- RAC database with Data Guard and GoldenGate.
- DR: Real-time failover to standby database (with zero or near-zero data loss).
- RAC database with Data Guard, GoldenGate and Application Continuity.
- DR: In-flight transactions are preserved (with zero data loss).
The “Platinum” service level, naturally, requires some seriously impressive kit. I wonder what “Titanium” or “Palladium” will eventually turn out to need 🙂