When redistribution is configured between OMP and BGP at two Data Center sites that have Direct Connection Interlink, which step avoids learning the same routes on WAN Edge routers of the DCs from…

The 'down-bit' prevents redistribution of underlay routes back into the same VPN via OMP, which is different from preventing loops when redistributing between OMP and BGP across two DC sites.

Defining different VRFs would segment traffic but would not inherently prevent routes from being redistributed and learned redundantly if the redistribution policy allows it.

Setting OMP admin distance lower than BGP influences route preference on the local router but does not prevent the advertisement of routes that could lead to a loop across the interlink.

Configuring the same overlay Autonomous System (AS) number on both Data Center WAN Edge routers leverages OMP's built-in loop prevention mechanism, which prevents a vEdge from advertising a route back into the same AS from which it was learned, thereby avoiding routing loops and redundant route learning.