Subnets, routing, more than one location

Hi all,
I think I get most of the routing / subnetting stuff as seperate topics. I just don't understand how they work together. If your subnets are at different geographical locations (NY and LA for example)and each connects to the internet, do all the routers on the internet need to know you have subnetted? How is traffic from the internet routed to the correct subnet? Can you subnet like this or are you required to have a common 'root' that connects to the outside world?

thanks
Dwight

Dwight,

Routing on the Internet is a big topic and hard to understand in a few paragraphs.

You can though see the big picture using a small and simple example, without getting into too much detail.

When you write a letter to your friend that's on the other side of the world, does your local post office know where exactly his house is ? No. They simply dont care cause thats someone's elses job when the letter gets to the destination country. From there its sent to the local post office of the city and finally the mailman will deliver it to your friend's home.

You can take this example and easily apply it to the way Internet routing happens and then you start to realise that you do not need any "common root" and understand that traffic is routed to the correct subnet by a local router that's aware of the subnet's existance.

I hope this helps!

Cheers,

Chis,
I'm sorry. I don't mean to be dense, but I still don't see it. May I rephrase my question and try again?

When I used the term 'root' , I was not talking about the internet as a whole. I was trying to refer to a common access point for my private network.

Using the mail analogy assume you have a company with the main office in NY and a branch office in LA. Someone in Chicago wants to send mail to the LA office. The post office will allow you to send mail west to California, then to LA, then to a ZIP code, and a specific address.

If you have subnetted your network all the routers on the Internet would need to know you have 2 physical locations in order to send email west to LA and east to NY.

You said "You can take this example and easily apply it to the way Internet routing happens and then you start to realise that you do not need any "common root" and understand that traffic is routed to the correct subnet by a local router that's aware of the subnet's existance."

How do you have a router that is "local" to both your NY and LA subnets?

thanks again
Dwight

Hello Dwight,

Maybe I can contribute some inputs here. But honestly I don't quite get your question and I also not very sure if I am right about my answer. Anyone if find mistakes or anything I missed after going through my answer below please correct me. This is how I can learn and I sincerely thank you all.

Imagine your have two big companies in LA and NY. Each company has its own local network and is assigned with an IP address.
Just for examples:
LA (Company A) - class C address 192.162.12.0
NY (Company - class C address 192.162.13.0

Each company uses subnet mask and further break down its own IP address to various smaller local subnets. Let's assume the company use local routers to segment its network into smaller subnets. Mind that these routers are locally owned by the company and they are not linked to the internet.

Now if I send a message to a host located in Company A (LA) that has a destination IP address 192.162.12.123. By default the internet can only see the address as a default class C address 192.162.12.0 (network ID) and doesn't care about the last octet. The packets move across the vast internet until it reaches the last final stop - the gateway (gateway router or sever) of the network of Company A (that is 192.162.12.0). Ok, now the packets have reached the network of Company A but not yet to the destination host/subnet. At this point, the gateway machine will apply its pre-configred subnet mask to the destination IP address to determine the correct local subnet and passes on the packets. When the local router of that particular subnet receives the packets, it will look into the complete address 198.162.12.123. The message is then sent to the destination host and that's how it is done.

So as you can see, subnetting is a local job. With subnet mask, you can simply subnet your assigned IP address anyway and it doesn't affect the operation of the internet.

Well, I hope I am right about this. Would be glad to hear feedback. Cheers.

Hi Steelben,

I'm beginning to belive the book I am reading is wrong and that is what has confused me. The book says "If an organization is large and has numerous computers, or if its computers are geographically dispersed, it makes good sense to divide a colossal network into smaller ones connected by routers." Your answer is correct, but for it to work company A and company B had 2 separate networks, each with a single gateway or "root" that connected to the internet. Then the 2 networks are subnetted. The book I am reading makes it sound like you could have 1 network and then subnet it to different geographical locations, like NY and LA. the only way I can see this working is if you had a private line between NY and LA.

thanks for all your time
Dwight

Dwight,

The idea of subnetting is to break up a single IP network number to smaller ones, so you will deal with less network traffic. From a geographical point of view this means your network should be small enough to handle this, like a school campus for example. The idea of having one physical network between LA and NY is quite impossible considering they're 3000 miles apart. The book you were reading was probably talking about a small area, perhaps within a couple of miles radius. I hope this clarifies it a bit more.

Maverick