In this solution, the primary and secondary sites both use the HCI architecture. Business systems are deployed on VMs on the HCI platform. Global load balancing servers are used to monitor the health status of VMs on the primary and secondary sites, and are used to balance the loads of links. Global load balancers are deployed at the egresses of the primary and secondary data centers, and sync policies are configured. ADC is deployed in bypass mode in the primary and secondary data centers. DR policies are configured on SCP to sync VM data from the HCI cluster on the primary site to the HCI cluster on the secondary site.

1. Configure MPLS VPN over the ISP network to connect the primary and secondary sites at Layer 3 (Layer 2 and Layer 3 are supported). This way, DR data can be transmitted between the sites, and the primary and secondary sites can be centrally managed. If the DR traffic is heavy, configure dedicated DR transmission interfaces.

2. Backup repositories use EDS. The external network of EDS and the internal network of HCI are connected to the same network plane through 10GE interfaces. This way, the storage network can be used for VM backup.
3. Configure management networks and VM network subnets for both the primary and secondary fault domains. Connect the VM networks of the primary and secondary fault domains and also their management networks. To avoid gateway conflicts, the network segments of the management networks of the primary and secondary fault domains must not overlap. If the VM networks of the primary and secondary fault domains are connected, make sure that the network segments of the VM networks do not overlap.
4. It is recommended to deploy SCP in different clusters to ensure its high availability. In this solution, SCP node 1 is deployed on the primary site, SCP node 2 is deployed on the secondary site, and SCP node 3 is deployed on a third-party site.

The figure below displays the network topology of the primary and secondary sites.