SNAT descision based on client ip and link availability

Along these same lines... Don't mean to hijack your thread, but it seems related...

I am working on some link controllers handling traffic for a datacenter and office with 50 or so users. They have a main connection (about 30Mb/s) and a backup T1. I have this configured using priority group activation so the T1 is only used when the main connection is down, and everything seems to be working as expected.

Now, if/when the main connection goes down, they do NOT want normal user traffic to pass the outbound network VIP. They want to make sure all of the bandwidth on the backup T1 will only be used for "real service traffic", and not just local users surfing Yahoo!

My idea is to create an iRule on the outbound network VIP that looks at the IP of the node in the default_gateway_pool that was chosen, and if it's the backup T1's IP, then just drop the traffic. Does that sound like a reasonable way to handle this situation? Anyone have any better ideas? Here is what I have for the current iRule...

when LB_SELECTED {  
        if {[IP::addr [LB::server addr] equals            drop  
        }  
  }

Thanks in advance...

--brice