Two on-prem data centers connect to Google Cloud via Dedicated Interconnect in us-east1 and us-west1, each with a Cloud Router in its region. To maximize attachment availability diversity given no port SLA, how should you configure VLAN attachments and VPC routing?

To maximize attachment availability diversity without a port SLA, you need to connect each on-prem data center to both Google Cloud regions. This means each Dedicated Interconnect should have two VLAN attachments: one to the Cloud Router in its local region (e.g., us-east1 DI to us-east1 Cloud Router) and another to the Cloud Router in the remote region (e.g., us-east1 DI to us-west1 Cloud Router). This creates redundant paths across regions. Enabling global routing on the VPC ensures that routes learned by any Cloud Router are advertised to all regions in the VPC, allowing traffic to traverse these diverse paths. Regional routing would restrict route advertisement to the local region, limiting cross-region failover capabilities. Creating only one VLAN attachment per Dedicated Interconnect limits redundancy to within a single region.