"Nick" <(E-Mail Removed)> wrote in message
news:1A20D5E0-1056-4D97-9687-(E-Mail Removed)...
> The problem is, we are unable to access them by their public IP address
from our intranet. From within our intranet we can access the machines by
their private addresses (10.x.x.x) just fine, as these packets are not
routed to our RRAS box.
Since you are accessing by IP# there isn't any DNS involved here (sorry,
guys). What you describe is exactly the way it is supposed to behave if you
are "reverse-NATting" (Static NATing) from publich IP#s bound to the
external Interface of the Router back to these machines on your internal
LAN.
Contrary to popular misconception, Ethernet runs on MAC addresses not on
IP#s. The role of the IP# in Ethernet is only to provide a Layer3 routing
mechanism and to provide a means to resolve the MAC address (ARP). The
reason intranet host must use the private addresses to access the servers is
because NAT can't make "u-turns". When you send a packet to the external
IP# the "NAT" process takes it and creates a situation where the source and
destination MAC addresses in the packet headers are the same address. It
can't go from itself to istself and shoots itself in the head.
These types of situations will work with other types of "processing" like
the "Web Publishing" or "Server Publishing" features of ISA & Proxy2 because
the internal methodology is different, but it will not work with a NAT
Device.
So when outside your system use the public IP# and when inside the system
use the private IP#. If you want to access by "FQDN" then make sure your
DNS functions properly to resolve to the proper IP# as the other guys are
describing.
--
Phillip Windell [MCP, MVP, CCNA]
www.wandtv.com