I am a part of P2P file sharing system hosted by a web site.

My client has a "registered" IP address (my ISP1's IP address) and a port number. The IP and port number are used by the site to check if my client is authenticated and the site will put my client "online" for its file sharing.

The client is an app that runs on one of my host machines on the LAN.

I'm not sure exactly how they work it, but randomly the IP address of their "keep alive" site will change ( <- This is important).

It might be using a proxy like nginx for the front end and the replies come from actual server's IP address instead of the frontend IP. 

I don't think it's the replies that are hurting me.

I think it's the ER7206 have 2 ISP IP addresses (ISP1 is registered) for load balancing and redundancy (WAN and WAN/LAN1). The "outbound" traffic from the client will go out on either interface (WAN or WAN/LAN1) and the site will never authenticate the client and put it online

My work around has been to use Static Routes.

The site's keep alive server's IP address shows up in the app's log file, and I've been adding static routes to match them when I see that my client is offline. 

I'm up to 7 static IP addresses. And I don't know how many IP addresses they use for a keep alive server. And, I only know if the client is offline when I manually open the site's status page for my client (I've been checking it 2 or 3 times per day).

I've tried using a Virtual Server, and assigning "WAN" (ISP1) as the "Interface". But that didn't seem to work. 

I did disable all the Static Routes when the Virtual Server was online, but after about 15 min client was still "offline" (normally, it takes a minute or less for the client to come online using static routes). 

Suggestions?