You're mixing apples and oranges with PPPoE and DHCP. Let me try to
simplify (well, at least clarify) a bit:
There are three separate LAN segments (networks) involved in what
you've described so far:
1) The network between your DSL modem and your internet service
provider. This is the place where PPPoE is involved. This segment
should also be set to automatically get an IP address from the ISP; I
believe but I'm not sure, that this is automatically implied in PPPoE;
it's similar to using DHCP. This will determine the IP address seen by
the outside world for your entire location.
2) The network segment between the DSL modem and the D-Link router.
The DSL modem acts as a DHCP server for this segment (even though it's
a DHCP client on the Internet side), and probably assigns itself an
address like 192.168.1.1. If everything is at factory defaults, it
will probably assign something like 192.168.1.50 as the WAN-side
address of the D-Link. This cable should be plugged to the WAN port of
the D-Link.
3) The network on the LAN side of the D-Link (the four similar plugs
on the back of it, plus the wireless network). Here's where you
probably need to change the factory settings. On the D-Link, try
setting the LAN side address to something like 172.16.0.1, subnet to
255.255.255.0. Now your PCs will be assigned starting at 172.16.0.50
or something like that and incrementing upward in the last number.
Both the DSL modem and the D-Link router act as NAT routers. Suppose I
use my PC at 172.16.0.50 to request a web page from
www.google.com.
DNS is used in a chain to find that the IP address is 64.233.169.104;
the chain thru the segments isn't of particular interest here. Now the
PC requests a web page from 64.233.169.104. The request goes to the
"gateway" configured into the PC via DHCP, 172.16.0.1 - thus the
request goes to the D-Link router. The router sees that we're looking
for a publicly routable address that isn't on our network segment, so
"translates" the address to the next gateway up the line, and sends it
to 192.168.1.1 - the DSL modem, while noting that the request came
from 172.16.0.50. The same NAT process takes place in the DSL modem,
which also notes that the request came from 192.168.1.50 (the WAN-side
address of the router). The request now goes out onto the public
internet. When the reply comes back from Google, the DSL modem looks
to see who requested it, and sends the reply to 192.168.1.50 (the
router). Similarly, the router then translates the address back to who
requested it, 172.16.0.50, and sends it to your PC.
The segment between the DSL modem and the router needs to be on a
different subnet than the network between your PCs and the router, and
the modem needs to be plugged to the WAN port of the router. This
whole process is why a NAT router inherently makes a pretty decent
firewall. If an unsolicited packet comes in from the internet, the
router has no place to send it, so the packet gets discarded. Port
forwarding is another whole topic, and is how you'd configure things
to allow access to a server you might wish to put up on your network.
Steve Hendrix