Welcome to our community

I was looking through some Cisco training manuals today. Reading the section on subnetting they talk about the number of possible subnets available as 2^n where n is the number of bits borrowed from the host part of the IP address. But only 2^n-2 would be usable because of the "Not all 1's and All 0's" rule

Is this true today 2005 ????

Hi foobar
I was little puzzled by your post to start with, and then I became curious as to why you are asking this question.

Legacy software did not support subnets where all the bits in the subnet mask set to all 1's or all 0's. So as an example if you wanted to split up a class C network into 2 parts you would need to do something like this

xxxxxxxx.xxxxxxxx.xxxxxxxx.00nnnnnn - Cannot Use (00)
xxxxxxxx.xxxxxxxx.xxxxxxxx.01nnnnnn - OK (01)
xxxxxxxx.xxxxxxxx.xxxxxxxx.10nnnnnn - OK (10)
xxxxxxxx.xxxxxxxx.xxxxxxxx.11nnnnnn - Cannot Use (11)

So from your original 254 addresses you end up with 2 usable subnets of 62 addresses.

This is defined in RFC 950 (1991)

So unless you are working on very old systems you can ignore this problem. You can reference RFC 1878 (1995) for more information.

In summary modern subnetting uses 2^n (RFC 1878) and legacy stuff uses 2^n-2 (RFC 950). The big problem is when you are taking tests (MCSE, Cisco, CompTIA)

I did some browsing on the web and after looking at a load of the usual dross I found this link which may be worth a look.


here

Thanks for the reply Dogbert

So from what you have said and looking at your link. It would be true to say that:

Under RFC 950 you cannot use a mask of 255.255.255.128 to create two subnets from a class C network, but you could under RFC 1878.

To divide a class C into two equal sized subnets under RFC 950 you would need a mask of 255.255.255.192 and you loose more than half your available IP addresses.

The other problem I have is that the course materials uses the 2^n-2 (RFC 950), the modern world uses 2^n (RFC 1878). What will I be tested on??

Foobar
The link suggests that you will be tested on the 2^n method on their 2.0 track. You need to point your question at Cisco.

Hi all,

Apparently the "ip subnet-zero" command was introduced in IOS 10.0. This command enables the use of subnet 0 for interface addresses and routing updates. However, the Cisco document at the following link: Host and Subnet Quantities says, "Prior to Cisco IOS Software Release 12.0, the ip subnet-zero global configuration command was required to be able to configure subnet zero on an interface. In Cisco IOS 12.0, ip subnet-zero is enabled by default."


I've recently taken the CCNA exam after attending the Intro and ICND courses. Our instructors told us that the Cisco exam would use the 2^n method. As far as I can remember this was the case. The story seems to be that there will be no ambiguous questions where answers would be different by the 2^n and 2^n -2 methods.

regards

Simon