NAT confusion

Bearded Steve

I'm getting mixed up with NAT.

Static map i understand but Dynamic NAT and Overload is confusing me.

As i understand it, to represent a range of Inside local addresses to a host on an outside network, you must define a pool of Inside Global addresses and associate that pool to your range of Inside local addresses.

eg

i have a range of addresses assigned to me by my ISP 10.100.100.64 - 10.100.100.66 / 26

I have a private addressing scheme on my local network in the range 192.168.1.33 - 192.168.1.45 /27

ip nat pool mypool 10.100.100.33 - 10.100.100.35 netmask 255.255.255.224
ip nat inside source list 1 pool mypool overload
access-list 1 permit 192.168.1.32 0.0.0.31

Is the keyword "overload" needed in this configuration? Some sources i have found use it and some don't

When would you use the following statement instead?

ip nat inside source list 1 interface serial0 overload

Bearded Steve

Any takers?

johnnynodough

When you see overload, think Port Address Translation.

Static NAT - One inside local IP address maps to one inside global address. One inside local maps to one global inside at a time. Configured manually. Nat Table is based on IP address only.

Dynamic NAT - One inside local IP address maps to oneinside global address, dynamically, on a need as basis. Used commonly when you have mutiple outside address and mutiple inside address. Only one inside local can map to any one inside global address at a time. If you have 2 inside global addresses, only two inside local computers can access the internet at a given time. Same access as above. NAT table translation is based on IP address only.

Dynamic NAT with PAT (Overload)- With one inside global address, you can have 65K+ connected to the internet at one time, the NAT table will store and translate based on IP address and port number.


At home, when you use your broadband router, that is performing NAT overload, which is the most common method of NAT, as it is the only method that preserves IP addresses on the net.

SV

Bearded Steve wrote:

Is the keyword "overload" needed in this configuration? Some sources i have found use it and some don't

I agree with Steve. Even I have noticed it... but do not know the reason. Could any one having real devices (lab) please try configuring NAT overload with these commands:

1) Router_1(config)# ip nat inside source list 2 pool mypool overload
2) Router_1(config)# ip nat inside source list 2 pool mypool

(Is the word overload really required?)

I just have a simulator.... will try it tonight.

johnnynodough

Look at the addressing scheme -

Public IPS -
10.100.100.64 - 10.100.100.66 / 26 - 2 available addresses
Private IPS -
192.168.1.33 - 192.168.1.45 /27 - 13 internal addresses

The only way the can all be translated at the same time is with overload (NAT W/ PAT) , so yes, overload is needed in this scenario, otherwise only 2 pc's at a time will have remote communication abilities.