A company has a 2 Gbps Direct Connect hosted connection to a VPC in ap-southeast-2 and has added a 5 Gbps hosted connection from a different Direct Connect location in the same Region. The hosted connections terminate on different routers in the office with an iBGP session between those routers. The engineer wants the VPC to prefer the 5 Gbps connection for sending traffic to the office and fail over to the 2 Gbps connection only if the 5 Gbps link fails. Which approach satisfies this?

The correct approach is to configure an outbound BGP policy on the router attached to the 2 Gbps connection to advertise routes with a longer ASPATH to AWS. BGP prefers routes with a shorter ASPATH. By artificially lengthening the ASPATH for routes advertised over the 2 Gbps connection, you make those routes less preferred by AWS, ensuring the 5 Gbps connection is used as the primary path. Advertising a longer prefix from the 2 Gbps connection would make those routes more specific, and BGP prefers more specific routes, which is the opposite of the desired outcome. Advertising a less-specific route from the 5 Gbps connection would make it less preferred. Configuring a longer ASPATH on the 5 Gbps connection would make it less preferred, again, the opposite of the goal.