Solved: subnet zero and broadcast subnet

sarahr202 · ‎03-15-2009

Hi every body!

Few weeks ago, i posted a question on the forum " why waste two subnets( subnet zero and broadcast subnet)?" I find a very good material about it on new ccna book by Wendell Odom .

To make my point , let say we have class B network 172.172.0.0.. I subnetted it using mask 255.255.192.0

The all possible subntes are:

172.172.0.0(zero subnet).

172.172.64.0

172.172.128.0

172.172.192.0 ( broadcast subnet).

According to Wendell. the reason we avoid using 172.172.192.0 is ambiguity. For example, the major class B assigned to us , is 172.172.0.0. The broadcast address for this class b network is 172.172.255.255.

But the broadcast address for subnet 172.172.192.0 is also 172.172.255.255. When used, should this address be considered as broadcast address for subnet 172.172.192.0 or for class B network 172.172.0.0? For this reason, the broadcast subnet was avoided.

But i also read we can use this subnet now. Newer ios has " ip subnet zero" enabled by default. My question is when we do use this broadcast subnet, how do we handle the ambiguity mentioned above? Should we say this ambiguity is small price to pay rather than wasting the whole subnet after all this ambiguity will cause the whole network( i.e the all subnets in 172.172.0.0) to receive the broadcast traffic occasionally not all the time? ( the book did no go into how this ambiguity is overcome or not overcome at all)

Thanks a lot and have a nice weekend!

Richard Burts · ‎03-15-2009

First I would like to clear up something: the command ip subnet-zero does not have anything to do with the "broadcast" subnet. It is only involved with the "zero" subnet - the very first subnet in the network (and technically it is the subnet where the subnetting bits are all zero). In very early Routing RFCs this subnet was restricted and in early versions of IOS the subnet was restricted. The ip subnet-zero command was used to override that restriction. For a long time that restriction has been lifted.

I would then observe that while abstractly there is ambiguity about the broadcast subnet, that there is no practical impact of it. The only time that the entire network 172.172.0.0 has a network broadcast address of 172.172.255.255 is if the network is not subnetted. In this case a packet addressed to 172.172.255.255 will go to every device in the network. As soon as the network is subnetted you can no longer send a single broadcast that will go to the entire network. As soon as the network is subnetted the address of 172.172.255.255 is uniquely associated with the last subnet of the network.

HTH

Rick

HTH

Rick

View solution in original post

Richard Burts · ‎03-15-2009

I have a lot of respect for Wendell Odom and the folks at Cisco Press who produce these books and they produce books whose accuracy is pretty good. But in this case they are absolutely incorrect. IOS has NEVER prevented use of the broadcast subnet.

I have worked with IOS since release 9.14 where "no ip subnet-zero" was the default. With no ip subnet-zero if you attempted to configure an interface with the zero subnet you received an error message. But if you configured the broadcast subnet it was accepted and worked just fine.

There may have been debate in the earlier releases about whether you SHOULD use the broadcast subnet or not. And lots of sources said that you should not use either of the subnet zero or the broadcast subnet (and the discussion tended to lump the subnets togeter in discussion). But IOS has never prevented the use of the broadcast subnet the way that it did prevent use of subnet zero.

HTH

Rick

HTH

Rick

View solution in original post

Richard Burts · ‎03-15-2009

As I think about this point I believe that there is another factor which has tended to confuse the discussion about possible use of the broadcast subnet. In looking at the number of usable addresses in a subnet we lose the zero address and the broadcast address. I believe that many people carry that concept over to the discussion about usable subnets and conclude that you can not use zero and broadcast subnets.

Since there used to be recommendations that you SHOULD not use the broadcast subnet (see for example RFC 950) many people carry the restriction that you CAN NOT use the broadcast from the question of addresses (where it is correct) to the question of subnets (where it is not correct).

HTH

Rick

HTH

Rick

View solution in original post

Richard Burts · ‎03-15-2009

First I would like to clear up something: the command ip subnet-zero does not have anything to do with the "broadcast" subnet. It is only involved with the "zero" subnet - the very first subnet in the network (and technically it is the subnet where the subnetting bits are all zero). In very early Routing RFCs this subnet was restricted and in early versions of IOS the subnet was restricted. The ip subnet-zero command was used to override that restriction. For a long time that restriction has been lifted.

I would then observe that while abstractly there is ambiguity about the broadcast subnet, that there is no practical impact of it. The only time that the entire network 172.172.0.0 has a network broadcast address of 172.172.255.255 is if the network is not subnetted. In this case a packet addressed to 172.172.255.255 will go to every device in the network. As soon as the network is subnetted you can no longer send a single broadcast that will go to the entire network. As soon as the network is subnetted the address of 172.172.255.255 is uniquely associated with the last subnet of the network.

HTH

Rick

HTH

Rick

sarahr202 · ‎03-15-2009

Thanks Rick!

here is quote from my book;

Isbn 10 1-58720-182-8

title: CCENT/CCNA icnd1 official guide second edition

author: Wendell Odom.

======================

" The third factor that defines whether the two special subnets should be used is based on a global configuration command " ip subnet zero". If the ip subnet zero is configured or not listed, then zero subnet and the other special subnet the broadcast subnet are both allowed"

chater # 12,pg# 361, first paragraph.

=====================================

Let me quote you for easy reference;

"As soon as the network is subnetted you can no longer send a single broadcast that will go to the entire network. As soon as the network is subnetted the address of 172.172.255.255 is uniquely associated with the last subnet of the network."

If this was always the case then what prevented the use of broadcast subnet in the past?

Thanks a lot and have a nice weekend!

Richard Burts · ‎03-15-2009

I have a lot of respect for Wendell Odom and the folks at Cisco Press who produce these books and they produce books whose accuracy is pretty good. But in this case they are absolutely incorrect. IOS has NEVER prevented use of the broadcast subnet.

I have worked with IOS since release 9.14 where "no ip subnet-zero" was the default. With no ip subnet-zero if you attempted to configure an interface with the zero subnet you received an error message. But if you configured the broadcast subnet it was accepted and worked just fine.

There may have been debate in the earlier releases about whether you SHOULD use the broadcast subnet or not. And lots of sources said that you should not use either of the subnet zero or the broadcast subnet (and the discussion tended to lump the subnets togeter in discussion). But IOS has never prevented the use of the broadcast subnet the way that it did prevent use of subnet zero.

HTH

Rick

HTH

Rick

sarahr202 · ‎03-15-2009

Thanks a lot Rick!

Richard Burts · ‎03-15-2009

As I think about this point I believe that there is another factor which has tended to confuse the discussion about possible use of the broadcast subnet. In looking at the number of usable addresses in a subnet we lose the zero address and the broadcast address. I believe that many people carry that concept over to the discussion about usable subnets and conclude that you can not use zero and broadcast subnets.

Since there used to be recommendations that you SHOULD not use the broadcast subnet (see for example RFC 950) many people carry the restriction that you CAN NOT use the broadcast from the question of addresses (where it is correct) to the question of subnets (where it is not correct).

HTH

Rick

HTH

Rick

sarahr202 · ‎03-18-2009

Thanks Rick!

it was just the wording issue then.

Rfc 950 says broadcast subnet should not be used , which does not mean it can not be used.

Richard Burts · ‎03-18-2009

That is exactly right. It is a subtle but important distinction between should not be used and can not be used.

HTH

Rick

HTH

Rick