QuestionQ18

VerifiedNetwork Design

A company runs its IT services in a multi-site hybrid infrastructure. The company deploys resources on AWS in the us-east-1 Region and the eu-west-2 Region. It also deploys resources in its own data centers in the United States (US) and the United Kingdom (UK). In both AWS Regions, the company uses a transit gateway to interconnect 15 VPCs. The company has established a transit gateway peering connection between the two transit gateways. The VPC CIDR blocks do not overlap with one another or with IP addresses used in the data centers. The VPC CIDR prefixes can be aggregated either at a Regional level or across the company's entire AWS environment.

The data centers connect to each other through a private WAN connection. IP routing information is dynamically exchanged through Interior BGP (iBGP) sessions. The data centers retain connectivity to AWS by using one AWS Direct Connect connection in the US and one Direct Connect connection in the UK. Each Direct Connect connection terminates on a Direct Connect gateway and is associated with its local transit gateway through a transit VIF.

Traffic takes the shortest geographic path from source to destination. For example, packets from the UK data center that target resources in eu-west-2 traverse the local Direct Connect connection. For cross-Region data transfers, such as from the UK data center to VPCs in us-east-1, the private WAN connection must be used to minimize AWS costs. A network engineer has configured each transit gateway association on the Direct Connect gateway to advertise VPC-specific CIDR IP prefixes only for the local Region. Routes to the other Region must be learned through BGP from routers in the other data center in their original, non-aggregated form.

The company recently had a cross-Region data-transfer issue because of problems with its private WAN connection. The network engineer must modify the routing configuration to avoid similar disruptions in the future. The solution must not change the original traffic-routing goal during normal operation.

Which modifications meet these requirements?

Choose two
Explanation

During normal operation, cross-Region traffic must keep using the private WAN (to avoid inter-Region charges over the transit gateway peering), so the Direct Connect + TGW-peering path may only be a failover. Both correct answers rely on Regional aggregation, not entire-environment aggregation. Option C uses longest-prefix match: advertise the other Region's aggregate over the local Direct Connect while the WAN still carries the original, more-specific (non-aggregated) prefixes, so the specific WAN routes always win and the aggregate is used only if the WAN fails. Option E uses BGP communities: advertise both Regional aggregates on both Direct Connect sides and have the data-center routers prefer WAN-learned routes via community-based policy, falling back to Direct Connect when the WAN drops. Options A and D are wrong because aggregating across the company's entire AWS environment also covers the local Region and can hairpin local traffic across Regions; B abandons aggregation and just advertises specifics.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!