Hello Alex H
I see you're looking to clarify how routing works between VNets connected to the same ExpressRoute circuit, particularly in relation to VNet peering and remote gateway configurations.Yes, your understanding is absolutely correct.
- Without VNet peering, traffic between VNET1 and VNET2 would go through the ExpressRoute circuit to on-premises, which isn't ideal because of increased latency and costs. Without VNET peering VNET1 and VNET2 wouldn't communicate.
- And your understanding of peering is also correct. With peering enabled and the remote gateway disabled, traffic between the VNets travels directly instead of being routed back through the ExpressRoute circuit, resulting in more efficient routing. But you can't communicate with VNET2 from your on-premises.
- Each VNet uses its own gateway for on-premises connectivity, which helps prevent asymmetric routing and keeps traffic paths straightforward.
- Regarding address space advertisement, you're spot-on. When each VNet advertises its address space to the ER circuit, it helps avoid routing issues like asymmetric routing. If both VNets advertise default routes to the ER circuit, express rote will receive two 0.0.0.0/0 routes. ECMP will be used unless weights are set.
- Setting a higher BGP weight for VNET1 will direct on-premises traffic to use VNET1 for Internet bound routes. This approach is a standard and effective design for managing outbound path selection.
Hope the above answer helps! Please let us know do you have any further queries.
Please do consider to “up-vote” wherever the information provided helps you, this can be beneficial to other community members