alt.certification.mcse

Author

Subnetting

WD40

2002-10-28, 11:25 am

Hi, I have just started out doing a win2k MCSE, was wondering if anyone
knows of any good sites that explain subnetting, anything would be great
TIA. Tim

steve harris

2002-10-28, 11:25 am

WD40 wrote:

> Hi, I have just started out doing a win2k MCSE, was wondering if anyone
> knows of any good sites that explain subnetting, anything would be great
> TIA. Tim

www.learntosubnet.com

Todd Lammle's SYBEX CCNA book is great too.

I learned the binary method. No tricks or shortcuts.

pseudocyber

2002-10-28, 11:35 am

This is my post in another forum. I love talking subnetting. Let me knowif you have any questions!
(Quote

In a nut shell ...

There must be part of the address that is the network and part that is the host.

How does a computer know which is which? The subnet mask.

As Elf was alluding to - 255 in Dec is 11111111 in binary. So, when you have a subnet mask of 255.255.255.0 this in binary is: 11111111.11111111.11111111.00000000.

Without getting above MY head, the network portion of the subnet mask HAS to be all consecutive one's. This is where the magic numbers I mentioned come in. They are also one's from left to right. So, if you want to have a subnet mask of

255.255.255.128 then this would be:
11111111.11111111.11111111.10000000 get it?

There such things as "natural boundries" or Classful networks. These are referred to as Class A, Class B, ... Class E.

You'll usually only have to worry about A, B, and C.

An "A" has a 0 in the left most network bit - therefore any decimal number can only go up to 127. If it went to 128 or past it, the left most bit would be a 1. Therefore Class A networks are 1-126.

A "B" has a 1 in the left most network bit. Therefore, the number will be 128-191.

A "C" has two ones - 11. Therefore, it ranges from 192-223.

So, subnetmasks for these network are:
255.0.0.0 A
255.255.0.0 B
255.255.255.0 C

Get it?

If you get into the advanced stuff - "subnetting" then you start manipulating the "natural boundries" to be un-natural. You move them. If you move them to the "left" you get less networks, more hosts. If you move them to the "right" you get more networks, less hosts.

See that Class C subnet mask? There's 8 bits left for the host portion. How many hosts could you have?

The magic rule is 2^N - 2. So you could have 2^8-2 hosts = 254 hosts.

If you have a class C address from the internet, but want to have more than 1 network, you could "steal" some bits from the hosts and give it to the networks. So, your subnet mask could be: 255.255.255.128. This would give you 2 networks and you could have 126 hosts on each.

See how that works? If you needed more networks, move the bit more to the right. Be aware though, you are sacrificing the number of hosts you can have on each network.

So, there now! You're subnetting!

(If you have networks and move the network portion to the left, you're ... gasp .... supernetting! Also known as CIDR (Classless Interdomain Routing))

Check out extensive subnetting discussions here:

http://www.certifyexpress.com/forum...p?threadid=8918

http://www.certifyexpress.com/forum...ight=subnetting

http://www.certifyexpress.com/forum...ight=subnetting

Jeffrey L. Woods

2002-10-28, 1:27 pm

In article <pseudocyber.d8xgb@mail.examnotes.net>,
pseudocyber.d8xgb@mail.examnotes.net says...

> An "A" has a 0 in the left most network bit - therefore any decimal
> number can only go up to 127. If it went to 128 or past it, the left
> most bit would be a 1. Therefore Class A networks are 1-126.
>
> A "B" has a 1 in the left most network bit. Therefore, the number will
> be 128-191.
>
> A "C" has two ones - 11. Therefore, it ranges from 192-223.

Not to pick nits, but the above is NOT correct.

A B has "10" as the two leftmost bits. If you say that the ONLY
requirement is a binary 1 in the leftmost bit, then you've also included
class C networks, which you properly defined but which ALSO have a 1 in
the leftmost bit.

> (If you have networks and move the network portion to the left, you're
> .. gasp .... supernetting! Also known as CIDR (Classless Interdomain
> Routing))

CIDR is more than just supernetting. See the rfc for CIDR. Its main
benefit is route aggregation to prevent the route tables in the core
routers from growing too large, and that is accomplished through a sort
of supernetting, but that's not quite accurate.

pseudocyber

2002-10-28, 1:56 pm

Hmm. You are right. I should have said, "the left most significant bit".

CIDR & Supernetting ... I guess if you look at it that way, CIDR uses Supernetting, but supernetting doesn't use CIDR. However, I intended to mean that route summarization is the main use in supernetting.

Anna Annonymous

2002-10-29, 1:24 am

clear

Sponsored Links