Question

A company is designing a new AWS environment expected to grow to hundreds of VPCs across multiple accounts and Regions. The network team wants to avoid IP exhaustion and future re-addressing work. They plan to use dual-stack VPCs but want their internal east-west traffic and future workloads to primarily rely on IPv6 to eliminate address overlap concerns. When designing the IPv6 addressing scheme for the VPCs, which approach best meets these goals while allowing centralized, auditable allocation of address space?

Accepted Answer

Bring your own IPv6 (BYOIP) address ranges or use an IPAM pool with a provisioned IPv6 CIDR, then allocate contiguous, non-overlapping IPv6 blocks to each VPC from that centrally managed pool.

Answer

Let each VPC receive an Amazon-provided /56 IPv6 CIDR automatically at creation time, and manage overlap through Transit Gateway route tables.

Answer

Assign each VPC a unique IPv4 /16 and enable NAT64 so that IPv6 addressing is derived automatically and requires no central planning.

Answer

Use the same Amazon-provided /56 for all VPCs and rely on VPC peering to isolate the duplicate ranges.