Your primary problem will be NAT. Some routers can read the SIP or H323 used to setup the calls and dynamically do the corresponding NAT mappings.
Assuming your router cannot do this you are pretty much stuck placing the device in the DMZ on both ends. Video uses a random UDP port to stream the video each direction.
If you have any firewalls in the path you also must either put in very generic permission rules or have a firewall that understands the sip/h323.
Without reading the manuals for this device I cannot be sure they will even tolerate the NAT. Most newer stuff does. You need to form a connection using the outside internet addresses.
Once you get the call up how well it will work is gamble. If you try to run at too high a resolution you are likely to get random stalls but it all depends on how good the internet connections are.