The lecture states at a few different points that VPC Peering is a hub-and-spoke model. I didn’t understand how this could be the case given that transitive peering isn’t possible.
Consider the following topology:
VPC_A --- VPC_HUB --- VPC_B
A hub-and-spoke model would imply that traffic between the spokes flows via the hub. So if VPC A has to talk to VPC B, traffic would flow via the Hub VPC. However transitive peering isn’t supported, which means VPC A will need to have its own independent peering relationship with VPC B.
That doesn’t sound like a hub-and-spoke, rather more like point-to-point connectivity.
1 Answers
For VPC peering the hub is from the perspective of the individual VPC itself and the spokes are its peered VPCs. This is slightly different from the Transit Gateway where the hub is the centralised routing table(s). The AWS documentation recommends either creating the independent peering connection (as you stated) or using Transit Gateway. https://docs.aws.amazon.com/vpc/latest/peering/invalid-peering-configurations.html
I completely agree with this point. It’s very confusing. The answer from "Learning AWS" of positioning the hub on each node does not really seem to me to be correct.