Subnetting Questions

MasterZorlak · ‎12-17-2007

Hey there,

I was wondering (more or less hoping) someone could help me with the basics to subnetting. I am seriously banging my head against the wall. Don't get me wrong, I have been reading a lot online and I am slowly understanding (I think), but some things just aren't clicking.

Mostly, I don't understand how PC's on the LAN are able to communicate between subnets... (Does that question make any sense?)

In short, I have basically been given the task of setting up a new site in our company. I want to be able to scale this LAN to handle somewhere around 16,000 hosts. I was thinking of using the subnet of 255.255.192.0 to accomplish this. What I don't understand is how 255.255.192.0 = 16,xxx potential hosts... Will the PC's on the LAN get different subnet addresses like 255.255.192.0 and 255.255.192.1? Can a PC with the .1 subnet communicate back to a PC with the .0 subnet correctly? Where does the broadcast address come in?

I know these must sound like silly questions, but I've yet to find a good online resource to explain them. Any basic help would be so greatly appreciated, or even any helpful links.

Thanks!

nordick26 · ‎12-17-2007

Hi,

reffer to:

http://www.cisco.com/en/US/tech/tk365/technologies_tech_note09186a00800a67f5.shtml

Say if this doesn't help

Regards

Ivan

Richard Burts · ‎12-17-2007

Perhaps review of some basics will help you with this. Many networks (especially class A and class B networks) are too large to be used as a single entity. So a network can be subdivided to create subnets that are a more useable size. The number of network bits in the address and the number of host bits in the address are indicated in the subnet mask. In the subnet mask a binary 1 indicates a network/subnet bit and a binary 0 indicates a host bit.

So if you want to scale a LAN to handle 16,000 hosts then you need a subnet mask with 14 binary 0s in the mask. This would be a mask of 255.255.192.0.

I wonder if that is really what you want to do (or actually if it is what you SHOULD want to do). A LAN with 16,000 hosts is very large. The amount of broadcast traffic generated in a LAN that large is probably going to interfere with performance of that LAN. If the site needs to scale to 16,000 that is one thing and can be done by assigning multiple LANs to the site.

To answer the other part of your question, if a PC in one subnet wants to communicate with a PC in a different subnet, they need some device capable of routing between subnets. Each subnet needs at least one layer 3 device capable of routing between subnets. These devices are configured as the default gateway for each subnet to which they connect. So PC1 sends to its default gateway and the default gateway forwards to the layer 3 device/router of the other subnet, and that device forwards to PC2. That is how PCs in different subnets communicate with each other.

HTH

Rick

HTH

Rick

MasterZorlak · ‎12-17-2007

Ahh OK, that is now starting to click! I didn't understand how the 0's and the 1's were equal to different bits.

In all actuality our LAN does not need to support 16,000 hosts. However, the nature of our company falls into the financial sector. We just had a large security / technology audit and evidently, with bringing this new site online, the technology auditor felt that our network needed that kind of scalability to be in proper protocol. Naturally management is now up in arms about that. In reality, scalability for 2000 hosts would be adequate.

So now I am not entirely sure how to approach the situation.

This does raise an interesting question though. What does a typical network administrator do when a small or medium sized business spills over their Class-C address space? I imagine they can easily go over 255 hosts.

sandeep1310 · ‎12-17-2007

Supernetting is the answer for that...

Rgrds,

SK