Multiple Network Address Translation 3-3
Dynamic mapping
Dynamic mapping, often referred to as many-to-few, offers an extension to the advantages provided by static mapping. Instead of requiring a one-to-one association of public addresses and private addresses, as is required in static mapping, dynamic mapping uses a group of public IP addresses to dynamically allocate static mappings to private hosts that are communicating with the public network. If a host on the private network initiates a connection to the Internet, for example, the Netopia Router automatically sets up a one-to-one mapping of that host’s private IP address to one of the public IP addresses allocated to be used for Dynamic NAT. As long as this host is communicating with the Internet, it will be able to use that address. When traffic from that host ceases, and no traffic is passed from that host for five minutes, the public address is made available again for other private hosts to use as necessary.
When addresses are returned to the group of available addresses, they are returned to the head of the group, being the most recently used. If that same host requests a connection an hour later, and the same public address is still available, then it will be mapped to the same private host. If a new host, which has not previously requested a connection, initiates a connection it is allocated the last, or oldest, public address available.
Dynamic NAT is a way of sharing a range of public, or exterior, NAT addresses among one or more groups of private, or interior, hosts. This is intended to provide superior support for applications that traditionally have difficulty communicating through NAT. Dynamic NAT is intended to provide functionality beyond many-to-one and one-to-one translation. Netopia’s NAT implementation makes it possible to have a static mapping of one public address to one private address, thus allowing applications such as NetMeeting to work by assuring that any traffic sent back to the source IP address is forwarded through to the internal machine.
Static one-to-one mapping works well if you have enough IP addresses for all the workstations on your LAN. If you do not, Dynamic NAT allows machines to make full use of the publicly routable IP addresses provided by the ISP as necessary, on demand. When these public IP addresses are no longer being used by a particular workstation, they are returned to a pool of available addresses for other workstations to use.
A common example is a DSL customer’s application. Most DSL ISPs only provide customers with a few IP addresses for use on their network. For networks with more than four or five machines it is usually mandatory to use NAT. A customer may have 15 workstations on the LAN, all of which need Internet access. The customer is only provided five IP addresses by their ISP. The customer has eight hosts, which only need to use email and have Web access, but another seven hosts, which use NetMeeting to communicate with clients once or twice a day. NetMeeting will not work unless a static one-to-one mapping exists for the machine running NetMeeting to use for communication. The customer does not have enough IP addresses to create a one-to-one mapping for each of the seven users. This is where dynamic NAT applies.
The customer can configure four of these addresses to be used for Dynamic NAT. The fifth address is then used for the eight other machines that do not need one-to-one mappings. As each machine configured to use addresses from the dynamic pool tries to connect to the Internet it is allocated a public IP address to use temporarily. Once the communication has been terminated, that IP address is freed for one of the other six hosts to use.
