You plan to support more tenant workloads by expanding capacity in another region or to protect tenant workloads across regions.

Figure 1. Adding a Region


Considerations

  • As you create each region, you must deploy at least two clusters.

    • The first cluster is for the management components such as Platform Services Controller, vCenter Server and NSX Manager instances. You can also configure this cluster as a target for disaster recovery of the Cloud Management Platform.

    • The second cluster is for shared edge and compute workloads.

    • If you are using a dedicated edge cluster model for your software-defined network, then a third cluster is required in each region.

  • Because adding vCenters Server instances in new regions requires additional local Platform Services Controller instances, consider the limit of 10 Platform Services Controller instances per vCenter Single Sign-On domain. If single pane of glass management is an important design goal for your environment, consider whether adding a new region requires creating a vCenter Single Sign-On domain.

  • When planning to protect some or all of your workloads between regions, allocate capacity for the failover of workloads without compromising the performance of the workloads already running in the recovery region.

  • As you deploy additional workload domains in each region, consider the need to deploy comparable workload domains in other regions if disaster recovery is a requirement.

  • Consider the maximum supported latency of 100 ms between linked vCenter Server instances.

  • Consider the maximum number of virtual machines that can be protected by using Site Recovery Manager.

    • The maximum number of virtual machines configured for protection by using array-based replication per Site Recovery Manager instance is 5,000.

    • You can extend this limit by deploying additional Site Recovery Manager appliances and associating them with the same source vCenter Server.

  • Consider the maximum number of virtual machines that can be replicated using vSphere Replication

    • The maximum number of virtual machines per vCenter Server configured for protection by using vSphere Replication is 2,000.

    • You cannot extend this limit by deploying additional vSphere Replication appliances and associating them with the same source vCenter Server.

    • Each vSphere Replication appliance must be associated with a different source vCenter Server.