Subnet Zero....?

Fulcrum45

I've been dusting off my subnet-fu skills lately and I've run into something that has me a bit perplexed. One of the sites I use for subnetting practice questions says the following:

"Subnet zero is allowed as per Cisco standard practice"

I've googled this over and over but cant seem to find a description good enough for me to understand how this affects the # of possible subnets I can have in a given subnet.

If Subnet Zero is allowed does that mean I use {2n=} OR {2n-2=} when determining my number of subnets?

BobBobson

What it basically means is that with subnet zero enabled you can use the first and last subnet blocks in a set of subnet blocks.

For example if you had a 192.168.1.0 address, and subnetted it out with a /26 subnet mask, you would have the following subnet blocks:
192.168.1.0 - 63
192.168.1.64 - 127
192.168.1.128 - 191
192.168.1.192 - 255

Without subnet zero you would only be able to use the "64 through 127" and "128 through 191" subnets. WITH subnet zero you can use all 4 of them.

BobBobson

Are you certain about that Wolf? Cisco's documentation on this seems to claim that the "ip subnet-zero" command specifically allow the usage of the first and last subnet blocks, not the usage of the network and broadcast addresses.

The aforementioned documentation: Subnet Zero and the All-Ones Subnet - Cisco

Priston

Keith Barker describes it well I think

https://learningnetwork.cisco.com/thread/32136

Fulcrum45

Ok, Keith lays it down pretty good so that I get the concept anyway. So then why does 172.28.0.0 255.255.255.128 give me 512 Subnets & 126 Hosts instead of 2 Subnets and 126 hosts?

Am I just confusing Subnet Zero with something else?

BobBobson

Fulcrum45 wrote: »

Ok, Keith lays it down pretty good so that I get the concept anyway. So then why does 172.28.0.0 255.255.255.128 give me 512 Subnets & 126 Hosts instead of 2 Subnets and 126 hosts?

Am I just confusing Subnet Zero with something else?

That is just subnetting with a class B address, it has little to do with the ip subnet-zero command.

Remember, 172.28.0.0 is at its base a class B address, so the subnet mask it has before any subnetting is done is "255.255.0.0". Because of this the netmask given shows that a total of 9 bits for the network, rather than just one like you are thinking, and 9 network bits allows for 512 total subnetworks (2^9 = 512).

davenull

I'll put it this way:

You know how you have to subtract 2 when you calculate the total number of available ip addresses for hosts?

Without the ip subnet zero command you'd have to do the same when calculating the total number of subnets.

Fulcrum45

First off you guys are awesome- thank you!

Secondly, Bob- Even in the days of CIDR and classless networking we still need to be mindful of IP Classes and their corresponding subnet masks?
This is starting to make sense now....like a lightening bolt to the brain, but in a good way.

BobBobson

Fulcrum45 wrote: »

First off you guys are awesome- thank you!

Secondly, Bob- Even in the days of CIDR and classless networking we still need to be mindful of IP Classes and their corresponding subnet masks?
This is starting to make sense now....like a lightening bolt to the brain, but in a good way.

If you are subnetting from the private address ranges you definitely need to be aware of the base classful address that you are subnetting out, particularly its corresponding subnet mask. If for no other reason this is so you are aware of that base classful addresses corresponding subnet mask so you know how many bits you have to work with for that particular address when you are subnetting it out.

theodoxa

In the olden days (IOS 10 and IOS 11), Cisco and the IETF thought Network Engineers were idiots who would be easily "confused" by having subnets with the same network address or broadcast address as the network they were subnetted from. Cisco did not allow you to use the first or last subnet when you subnetted a network. The CCNA used to assume this behavior ("no ip subnet-zero"). I believe Cisco changed that either in the 2007 revision or the one before that and all Cisco exams since have assumed that the "Zero Subnet" is allowed unless it says otherwise (e.g. "the router is configured with the no ip subnet-zero command").

Fulcrum45

Appreciate the feedback- thanks Guys!