When to use and not use divert/natd ...

[John Nielsen]

On Sep 5, 2014, at 9:15 PM, John Case <case at SDF.ORG> wrote:

> For many years I would build FreeBSD firewalls and they would be very, very simple - I just set gateway_enable="yes" in rc.conf and everything just worked.
> 
> However, these firewalls *always* had real, routable IPs no both sides. Both interfaces had real, routable IPs.
> 
> Now I have a firewall that has two non-routable IPs for its interfaces, and is connected to a internet router with the real IP.  When I try to builda  very simple firewall  it does not work, and I am forced to use ipdivert and natd.
> 
> If I use ipdivert and natd, it works just fine.
> 
> So, am I correct that I can create a simple gateway without natd/divert as long as both interfaces are real IPs, but if both interfaces are non-routable IPs, I am forced to use divert/natd ?

Just think about the 'routing' aspect. In your current scenario it sounds like the Internet-connected device is doing NAT. It knows about its public IP and its private subnet. It sounds like you have a second private subnet behind your FreeBSD machine about which the Internet-connected device knows nothing. For packets to get from the Internet-connected device to your second subnet one of two things needs to happen:
 1) The Internet-connected device has a static route to the second subnet (so it knows to use your FreeBSD machine as the gateway), or
 2) The FreeBSD machine performs NAT (a second time), so the Internet-connected device send traffic to it even though it knows nothing about the subnet behind it.

I would prefer 1) as it's simpler and double-NAT isn't generally a good thing. However, if you don't have a way to add a route to the Internet-connected device then 2) isn't necessarily bad.

In your previous all-routable-IPs setups something was presumably advertising the route for you. The new setup isn't much different in principle.

JN

PS: Using the in-kernel NAT with IPFW is simpler and more efficient than using natd...