Article from petri.co.il (click to read the whole article).
Αν δε λειτουργεί ο παραπάνω σύνδεσμος, ανοίξτε το Introduction to Network Address Translation
Almost anyone with a modern Internet connection has used Network Address Translation (NAT). NAT has been a big part of large scale IP network deployments since the IPv4 address space began to dwindle. At its most basic, NAT enables an internal IP address (or addresses) to be mapped to an external IP address (or addresses); this enables the use of private RFC 1918 IP addresses on internal networks while only using a single or very few public IP addresses.
The first thing to understand is the different types of NAT, which include: static, dynamic and overloaded.
With static NAT a single specific internal IP address is mapped to a single specific external IP address; an example of this is shown in Figure 1.
Figure 1 – Static NAT
In Figure 1, a PC on the internal network needs to communicate to a host on the external network, which in this case is the Internet. Private RFC 1918 are not routable on the public Internet and would not be allowed as a source or destination address. To get around this, NAT can be statically configured to map the internal IP address 192.168.1.10 (the host) to the external IP address 203.0.113.10; if this was with all communications on the external network it would look like traffic would be coming from 203.0.113.10 instead of 192.168.1.10. In this situation NAT would refer to the IP address 192.168.1.10 as the inside local address and its mapped IP address 203.0.113.10 as the inside global address.