Edit

Share via


Network topology and connectivity for Oracle Database@Azure - BCDR connectivity design

This article focuses on high availability (HA) and disaster recovery (DR) connectivity considerations for Oracle Exadata Database@Azure deployments. It begins by addressing single-region redundancy, then expands to cross-region architectures to help ensure that your environments remain resilient and meet low recovery point objective (RPO) and recovery time objective (RTO) requirements.

You must address the following key requirements:

  • Low latency: Minimize latency for database traffic for optimal performance.

  • Business continuity and disaster recovery (BCDR): HA of your Oracle Exadata Database@Azure and related applications is crucial to maintain business operations that have low RTO or RPO.

Network connectivity between primary and standby Oracle Exadata Database@Azure instances can be established via Azure networking and Oracle Cloud Infrastructure (OCI) networking. The default route for this connectivity is through Azure. This article focuses on architecting for networking within Azure to meet BCDR requirements. For more information, see BCDR for Oracle Database@Azure.

Redundancy within the same region

For single-region redundancy, you need two separate Oracle Exadata Database@Azure deployments. Each deployment should be in a different virtual network and aligned with a different availability zone.

The following diagram illustrates a single-region redundant setup that uses two Oracle Exadata Database@Azure deployments in separate virtual networks to span multiple availability zones.

Diagram that shows in-region redundancy for Oracle Exadata Database@Azure.

Note

Oracle Data Guard uses the underlying Azure network. Data replication occurs through a peered Azure virtual network within the Azure region and doesn't use the OCI network. Virtual network peering costs apply.

Redundancy across regions

To achieve cross-region redundancy with Oracle Database@Azure, deploy your primary and standby Exadata infrastructures in separate virtual networks in different Azure regions. You can establish direct connectivity by using global virtual network peering, which enables low-latency, high-throughput replication between regions.

The following diagram shows the recommended cross-region configuration that uses global virtual network peering.

Diagram that shows Oracle Database@Azure deployed in two regions connected via global virtual network peering.

Design considerations

  • Networking options for BCDR: You can establish network connectivity between the primary and standby Oracle Database@Azure instances via Azure networking and OCI networking. The default route for this connectivity is through Azure.

  • Availability zones: Ensure that the two Oracle Exadata Database@Azure deployments span different availability zones to isolate failures at the availability zone level.

Design recommendations

  • Network virtual appliances (NVAs): Minimize the use of NVAs and other appliances in the network path because each extra hop adds latency that can degrade Data Guard replication performance.

  • Asynchronous replication for cross-region DR: Because of higher latency between geographically distant regions, consider using asynchronous Data Guard replication. This mode can handle increased latency without affecting production performance.

Next step

To prepare for your migration, consider the connection from Oracle Database on-premises to Oracle Exadata Database@Azure. For more information, see Migration connectivity design.