IP subnet zero theory?

Answered Question
Mar 25th, 2004

Hi all,

Today I was asked an odd question about the use of IP subnet zero in theory. Since I am not the expert on the material and haven't memorized the RFC I figured I would ask everyone.

Is this ever possible? as a host address

Please explain?



I have this problem too.
0 votes
Correct Answer by omal about 11 years 8 months ago

Yes that should be possible. This is a class 'A'address. Remember class is between 0 - 127. But we don't take because we use that address to give a default route. 127 network also we don't use because 127 range is for testing, eg: loopback address.

Therefore theoratically this should work.


  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 3 (1 ratings)
Correct Answer
omal Fri, 03/26/2004 - 13:38

Yes that should be possible. This is a class 'A'address. Remember class is between 0 - 127. But we don't take because we use that address to give a default route. 127 network also we don't use because 127 range is for testing, eg: loopback address.

Therefore theoratically this should work.


Anonymous (not verified) Fri, 03/26/2004 - 19:53


The IP address is not a valid one.The valid network address for class A can be between 1-126. It cannot be 0 (zero). You cannot say your network address is 0(zero).

The very first valid network address in class A scheme is . Any network address less than is considered invalid.

IP subnet zero allows to use the major network address as the first subnet,when subnetting the major network. If IP subnet zero is not configured or not supported, the major network address cannot be used for the first subnet.

For example, consider the following IP address

IP address:

This is class A address. The network mask (default) for this IP address is If you want to subnet this major network in to four sub networks, you need to borrow 2 bits from the host portion. That gives you the subnet mask as :00000001.11000000.00000000.00000000

It is always better to convert the decimal numbers to binary to reach the correct result.

1. 2.3.4 :00000001.00000010.00000011.00000100

255. 0.0.0 :11111111.00000000.00000000.00000000 :00000001.11000000.00000000.00000000

Here the network portion is 1. x.x.x

the host portion is x. 2.3.4

the subnet portion is x.192.x.x

since the assignment of host address is controlled by the Network Administrator, the valid host range is to In binary

1. 0. 0. 1 :00000001.00000000.00000000.00000001 :00000001.11111111.11111111.11111110 1.192. 0. 0 :00000001.11000000.00000000.00000000

And hence

The first subnetwork address (or subnet zero) :00000001.00000000.00000000.00000000

The first host in the first subnet

1. 0. 0. 1 :00000001.00000000.00000000.00000001

The last host in the first subnet

The broadcast address in the first subnet

The second subnetwork address

1. 64. 0. 0 :00000001.01000000.00000000.00000000

The first host in the second subnet

1. 64. 0. 1 :00000001.01000000.00000000.00000001

The last host in the second subnet

The broadcast address in the second subnet

The third subnetwork address

1.128. 0. 0 :00000001.10000000.00000000.00000000

The first host in the third subnet

1.128. 0. 1 :00000001.10000000.00000000.00000001

The last host in the third subnet

The broadcast address in the third subnet

The fourth(last) subnetwork address

1.192. 0. 0 :00000001.11000000.00000000.00000000

The first host in the fourth subnet

1.192. 0. 1 :00000001.11000000.00000000.00000001

The last host in the fourth subnet

The broadcast address in the fourth subnet

If subnet zero is neither configured nor supported, that particular router does not know the existance of the first and the fourth subnetworks and those two subnetworks are not forwarded to other routers in the network.

I hope I have provided enough information to clear your doubt.

Thanks and Regards


Network Administrator

BCT Networks,

blazesod Sat, 03/27/2004 - 02:20

Hi CSCO~004,

I understand the point you made. Thanks for explaining it in such detail. My reasoning behind the question is based on the formula for calculating # of hosts and # of subnets. The hosts portion is 2^n-2 where you lose 1 for broadcast and 1 for the network. This week I was informed # of subnets is no longer configured the same way. Now the formula (unless I am misinformed) is just 2^n=subnets. A simple calculation will create the possibility of using the 0.x.x.x network.

if n=0 then 2^n(0) =1

Historically (1997) I assigned a host the IP address of (OS= nt3.51) and had it talk on the network. I think it forced the card into "sniffer" mode. Obviously this is not possible in NT4.0 and higher but on a router you don't have those configuration limitations. Finally, I would like to know if wildcard masks have been used as a subnetmask alternative.

Since I am getting mixed info, "any TAC techies out there want to put an idea through the think tank?"

Anonymous (not verified) Sat, 03/27/2004 - 12:05

Hi blazesod,

The first paragraph in your message is exactly right, except the last sentence in that paragraph. What you need to understand is the difference between the network and the subnetwork.

The network address is assigned by the Service provider. You have no control over that. And no private addressing is suggested to start with the network address as zero. No service provider will assign the network address to start with 0(zero). The subnetworks are designed from the network address by the network Designer of the Organisation based on the internal requirements. So there is no possibility of having any network to start with zero as the first octet.

Simple theory will not work in real time scenario. There are so many faces for any statement and everything cannot be discussed in detail in this forum.

All your doubts related to IP Addressing would be cleared if you could visit the following webpage:


If you can't or don't want to spend time in reading the link or in general if you don't have a reading habit to search the abundant and easily browsable technical information available with Cisco Connection Online, and depending on others to clear your silly doubts, there is a possibility that you will be carried away with false knowledge.

The real techies will ignore the questions like the one posted by you. Because they don't want to waste the time on answering simple basic questions. They would love to reply only for the questions genuinly difficult.

I understood from your second response that you are not interested to visit any web page that explains the IP Addressing Scheme or any RFC related to IP Addressing. I am sorry to say that this is not the proper way to gain the knowledge.

Please take some pain in referring the resources available related to IP Addressing. The doubt you have raised is really very basic. Cisco Connection Online is wonderful resource for any doubts related Cisco Products and technical information.

Wish you all the best.

Thanks and Regards,


Network Administrator

BCT Networks

blazesod Sun, 03/28/2004 - 02:49


I believe you may have misunderstood me. Please accept my apology if you are offended by my post.



protocl Sat, 03/27/2004 - 21:17

in order to uitilize '0', first configure the router to use '0' the rest is up to the math....subnetting.

jacquesd@isserv... Sat, 03/27/2004 - 21:47

The issue here is not about subnetting. It is what the legal addresses in a class A network is. Even if subnet zero is configured on a router, still cannot be used.

Simply put, the first and last networks at a classful NETWORK boundary cannot be used.

Class A: 1-126 (0 is first and 127 is last - also used for diags)

Class B: 128.1 - 191.254

Class C: 192.0.1 - 223.255.254


blazesod Sun, 03/28/2004 - 02:33

Hi all,

Thanks everyone for posting. In the interest of not causing an arguement I am going to consider the conversation solved. I was not asking what the valid ranges for IP classes, private (reserved), RFC #'s or NIC organizational addresses are. I was simply relating an improbable possibility question that I was asked last week.

omal Sun, 04/04/2004 - 07:36

Hi Dave,

I can understand your situation.

Well, it's like this. Generally it says Class A range is 0 - 127. BUT network has been reserved for testing perpurse. is DEFINITELY INVALID as it is been taken to define the default route (gateway of the last resort). Considering these things, I don't see any problem in taking the IP address you have given as a valid host. Theoritically it should be a valid host address.

But the best thing for you to do is, if you have a router, just configure that and see, (without ip subnet-zero command) if the router accepts it. If you have to give the command "ip subnet-zero" to configure that ip address, then it is not valid, at least according to Cisco.

All the best!



This Discussion