Internal Clients Only
Dedicated Line

This class of site either has an IP Network with clients and possibly some hosts, or is preparing to install one, but in either case has no intention of providing services to the outside world. There may be Internet-style services available internally, but the desire is to spend no time thinking about security from prying eyes on the outside. For this site, the steps to an Internet connection are straightforward.
This is how your network appears in this configuration. A router (and modem or DSU) connects you to your Internet Services Provider. You may wish to lease this router from your ISP for convenience. The NAT will be your security, so there's not much router setup required.

The LAN that is shared by the router and your NAT is "public." You must assume that someone with ill intent can reach this LAN. Host computers must never be placed on a public LAN like this.

The NAT is our NetNAT. Its job is to let your internal clients access the Internet, while turning aside all service requests from the outside.

The "Private Net" LAN segment is within your facility. This segment is your world, with your internal clients and hosts.

To the outside world, you appear to have a single IP Address: the IP Address of the NAT's Ethernet port on the public segment.


If your Internet Service Provider uses RFC 1490 compliant Frame Relay for a dedicated link, you can eliminate the extra router and LAN hardware, and use our NetNAT for the connection. We offer Frame Relay adapters in two versions, one with a 56K speed limit and one with a T-1 speed limit. Both feature built-in CSU/DSU, with a smart T-1 CSU on the T-1 model.

The NAT is our NetNAT. Its job is to let your internal clients access the Internet, while turning aside all service requests from the outside.

The "Private Net" LAN segment is within your facility. This segment is your world, with your internal clients and hosts.

To the outside world, you appear to have a single IP Address: the IP Address of the NAT on the Frame Relay link.


How Does This Work?

Your world's workstations and hosts each have a unique IP Address, that is used in communication with others. When a client application wishes to connect to a server, it performs the following steps:
  1. creates a "socket" for its use in the connection-to-be,
  2. puts its IP Address into the "source IP Address" fields of the socket structure,
  3. makes up a number in the range 1024 through 65534,
  4. put this number in the "source Port Number" fields of the structure,
  5. places the server's IP Address in the "destination" part of the structure,
  6. puts the service's Port Number in the "destination Port Number" part, and
  7. opens the connection to the server.
In TCP and UDP, the primary IP protocols in use, a connection is defined by the unique combination of the four numbers we just saw put into the socket structure. These were: Each end now has a "socket," that may be used to communicate with the other end. People called it a "socket" to help us visuallize its nature. Imagine an old-time telephone switchboard, with thousands of sockets and a bunch of patch cables. Each socket corresponded to a destination, and by plugging the correct cable into the correct socket, you could have two-way communication with that destination.

If the client and server are on your private LAN, the communication happens directly between the two participants. If the server is not on your LAN, the client will send the connect request to the NAT instead. The request is still addressed to the ultimate server, but the message goes to the NAT as your network's default gateway to the outside.

The NetNAT Magic

As the message passes through the NetNAT, the source IP Address is saved and replaced with the IP Address of the NetNAT's PPP interface. Because the source port may already be in use by another client, the NetNAT saves the original one and replaces it with a freshly-allocated one from the NetNAT's pool. Thus, the outside world sees a different IP Address and (probably) a different port number than were used by the client application.

The opposite occurs on a response. The original IP Address and port are restored to the destination part of the message header before the message is sent to the client device. The NetNAT maintains a context block for each active connection. No message from the outside is permitted in unless it completely matches the characteristics of the connection: the four numbers described above.

The end result is that the Internet sees a single IP Address at your site. You may have thousands of devices that access the Internet, but nobody will know that but you.


Special Support

There is always a chance that, by changing the IP Address and/or Port Number, you will no longer be able to use certain services. We have already built in support for the two most common of these services: As we find others, and if they seem destined for greatness, we will try to add support.
This page was last modified on 30 September, 1995.

This information is proprietary to Network Safety. Network Safety, WebElite, DialNAT and NetNAT are trademarks of Network Safety. For information on our products and services, please contact our sales department.

This page was prepared using WebElite, our professional editor for the Web.