In a simple example, I have a virtual server that directs inbound traffic to a load-balanced internal pool, say a couple of web servers.
When I add a 2nd ISP, I would create a 2nd self-ip, add it to the "external" vlan, create a 2nd virtual server with a new IP from the new ISP subnet, direct it to the same internal load-balanced pool.
What I'm a little confused by is: how is the response traffic going back to the client? In the manual above it mentions creating a SNAT list as automap. Or is this for something different? Does the default gateway of the Big IP get utilized here?
The manual also specifies creating an outbound VS directing traffic to the 2 ISP routers, but I'm guessing this is for use cases where the originating traffic is outbound from the internal network. Correct?