Two network virtual appliances in us-central1 provide connectivity to on‑prem 10.0.0.0/8. Requirements: all on‑prem traffic must traverse the appliances, tolerate single appliance failure, and use both appliances simultaneously. Which routing method meets these requirements?

The correct option is to use an internal TCP/UDP load balancer. This method allows for active-active operation, distributing traffic across both network virtual appliances simultaneously, which satisfies the requirement to use both appliances. If one appliance fails, the load balancer automatically directs traffic to the healthy appliance, ensuring fault tolerance. A custom static route can then point the on-premise destination (10.0.0.0/8) to the internal TCP/UDP load balancer's IP address as the next hop, ensuring all traffic traverses the appliances. Creating two routes with different priorities would result in an active-passive setup, where only one appliance is used at a time, failing the simultaneous usage requirement. HTTP(S) load balancers are for HTTP/S traffic and are not suitable for general network appliance routing. Network load balancers (external) are for external traffic, not internal routing within a VPC.