Skip to content

Cross Region Replication🔗

🔑 1. Built-in Redundancy Options (Storage Account Level)🔗

When you create an ADLS Gen2 account, you choose a redundancy type. Azure handles replication under the hood.

✅ Options:🔗

  1. LRS (Locally Redundant Storage)

  2. 3 copies within a single datacenter in one region.

  3. Cheapest but no cross-region replication.

  4. ZRS (Zone-Redundant Storage)

  5. 3 copies across 3 availability zones in one region.

  6. Provides resilience against zone failures, but not region-wide outages.

  7. GRS (Geo-Redundant Storage)

  8. 6 copies: 3 in primary region + 3 in a paired secondary region.

  9. Replication is asynchronous.

  10. RA-GRS (Read-Access Geo-Redundant Storage)

  11. Same as GRS, but read access to the secondary region is available.

  12. Useful for disaster recovery and read-heavy workloads.

  13. GZRS (Geo-Zone Redundant Storage)

  14. Combines ZRS (in primary region) + asynchronous replication to a paired secondary region.

  15. High durability + regional disaster recovery.

  16. RA-GZRS (Read-Access GZRS)

  17. Same as GZRS but with read access to secondary region.

  18. Best option for mission-critical cross-region replication.

🔑 2. Asynchronous Data Movement (Custom Replication)🔗

Sometimes built-in GRS/RA-GRS may not meet your latency, cost, or compliance requirements. In that case, you build custom cross-region replication:

Approaches:🔗

  • Azure Data Factory (ADF) Copy Activity

  • Schedule pipelines to copy data from ADLS in Region A → ADLS in Region B.

  • Flexible (supports filtering, transformations, scheduling).

  • Good for batch replication.

  • Azure Data Share

  • For sharing snapshots of datasets across regions.

  • Suited for read-only scenarios.

  • AzCopy / Azure Storage Sync Service

  • AzCopy CLI can sync containers across regions.

  • Best for ad-hoc or bulk replication.

  • Event-Driven Replication with Event Grid + Functions

  • Trigger on BlobCreated/Updated events → replicate object to another region.

  • Provides near real-time cross-region replication.

  • Third-party tools (Databricks, Apache NiFi, Informatica, etc.)

  • For more complex pipelines involving CDC (Change Data Capture) or multi-cloud replication.

🔑 3. Strategy Selection (When to Use What)🔗

  • Mission-critical, zero-downtime needs → Use RA-GZRS for built-in cross-region replication with read access.
  • Compliance requirements (regulatory control over region) → Use custom ADF/Event Grid pipelines to replicate only selected data.
  • Disaster recovery only (cold standby) → Use GRS/RA-GRS with failover enabled.
  • Performance optimization (local reads in multiple geographies) → Use RA-GRS or RA-GZRS so consumers in another geography can read from the secondary.

🔑 4. Failover Considerations🔗

  • For GRS/RA-GRS/GZRS/RA-GZRS, failover to secondary is manual.
  • When you trigger failover, the secondary becomes primary.
  • Important: failover breaks replication (you must reconfigure after).

✅ Summary:

  • Use RA-GZRS if you want the best balance of availability + performance with cross-region reads.
  • Use ADF/Event Grid/AzCopy if you need fine-grained control over what, when, and how data is replicated across regions.
  • Always align your choice with RPO (Recovery Point Objective), RTO (Recovery Time Objective), and compliance needs.