background image
340 Chapter 11: Scaling IP Addresses with NAT
Overlapping networks can occur for a number of reasons, such as a merger, the consolidation
of company resources that are tied with newly installed WAN components, and so on. Many
companies have chosen to use the private address space defined by RFC 1918, which reserves
the address ranges for the private network space shown in Table 11-2.
The overlapping of network numbering will probably continue to be a problem due to the
extensive use of private address space in the private sector and the current trend toward inter-
and intra-company connectivity. You can merge two companies using the same private address
space by using the NAT overlapping network feature; essentially, each network is translated to
the other. This double translation can take place on a single router.
The use of the limited number of addresses in the private space increases the odds dramatically
that an overlap will occur if two private networks are merged. It is with that in mind that most
design guidelines dictate that if using the private space, do not start with the 10.1.0.0 network
because others are likely to do just that as well. The recommended practice is to start in the
middle, such as 10.128.0.0, and work from there.
The drawback to this restriction is that most technical people read the same literature and go to
the same classes and talk to the same pundits. Therefore, the next time a merger occurs, they
will not have to worry about the overlap of the 10.1.0.0 network; they will have to worry about
the overlap of the 10.128.0.0 network.
Another area in which overlapping can occur is when a company elects to use a nonprivate
address for their own purposes with the idea that they will never connect to the Internet. This is
a very bad assumption in today's e-commerce driven world. Common sense would dictate that
an Internet connection would eventually be required in this e-age and consequently,
renumbering would be needed. With NAT, you have an interim fix for overlap.
TCP Load Distribution
NAT can be used for TCP load distribution. This works in a form that is somewhat reversed
from other translations. In the other three uses of NAT, the sender uses a nonlegitimate source
address in a packet destined for the outside world. In contrast, load distribution takes advantage
of the NAT function by allowing a site to advertise an address but when you send a packet to
the advertised address, it is rerouted to another set of addresses.
Table 11-2
Private Address Ranges
Class
Range
Number of Networks
A
10.0.0.0 1
B
172.16.0.0­172.31.0.0
16
C
192.168.0.0­192.168.255.0
255