Infra in us-central1; on-prem in NYC (closest GCP region us-east4). Interconnect POP in Ashburn, VA. Need Cloud Interconnect to on-prem for backend traffic, expected bandwidth ≤500 Mbps; minimize latency and cost. What should you do?

This option is correct because it places the Cloud Interconnect components (Cloud Router and VLAN attachment) in us-east4, which is the closest Google Cloud region to the Ashburn, VA Interconnect POP and the on-premises network in NYC. This minimizes latency for the on-premises connection. Enabling global routing in the VPC allows the us-central1 infrastructure to communicate with the on-premises network via the us-east4 interconnect. The bandwidth for VLAN attachments is automatically provisioned up to 10 Gbps and is not explicitly set to 500 Mbps, making the other option with explicit bandwidth setting less accurate. The other options are incorrect because: Creating a Cloud Router in us-central1 while the VLAN attachment is in us-east4 is an invalid configuration; the Cloud Router and VLAN attachment must be in the same region. Creating all components in us-central1 would route traffic through us-central1 first, increasing latency and potentially cost, as the Interconnect POP is in Ashburn, VA, closer to us-east4.