[Pepsione1]

What happens when you have 0.0.0.0 subnet mask on a machine? My understanding on this topic is the subnet mask is a used as the second operand of a AND operation. The first operand would be the IP address. But what does that really mean to network traffic and how packets will be routed from this machine and from a router to this machine.

 

[HFat]

You should only use even bit values to keep jitter under control. I use 255.255.255.255 myself because it gives a warmer soundstage. Nice synergy with my class A switch too.

 

[EvilGnome6]

A subnet mask of 0.0.0.0 would mean that all IP addresses are local to the machine. Since no address would be calculated as being in a remote subnet, it would broadcast all traffic to the local subnet and never send it to the router.

In short, the machine would only be to access hosts on the local subnet and never be able to communicate with hosts outside the subnet.

 

[Pepsione1]

I don't know how I got this post in this sub-forum. Please move it.

So what's consider in the same subnet? aaa.bbb.ccc.ddd.eee The aaa.bbb. are the network ID and ccc. is the subnet ID and eee is the node ID.

Is that correct? um.. I am still unclear about this.

 

[EvilGnome6]

If you do things in binary, it's a little easier:

Let's say your IP address is 192.168.0.1 and your subnet mask is 255.255.255.0. If you convert these to binary, you get:

11000000.10101000.00000000.00000001
11111111.11111111.11111111.00000000

The 1's indicate which portion of the address is the subnet (In this case, 192.168.0) and the 0's indicate which portion of the address is the host address (In this case 1).

If you were try to send a message to 192.168.0.2, the computer looks at its own address and subnet mask determine whether or not the address is in the same subnet (in this case it is). When it's in the same subnet, the packet is broadcast on the local subnet.

Now, let's say you're sending a message to 10.0.0.1. the computer looks at it and says, "hey, 10.0.0 is a different subnet than I have" and forwards the packet to the router instead of broadcasting it.

If your subnet mask was set to 0.0.0.0, no portion of the address is allocated as the network portion and every other address will be considered as "local". The computer will never forward a packet to a router and be unable to communicate with anything outside of the local subnet.

 

[EvilGnome6]

Quote:

Originally Posted by Pepsione1 /img/forum/go_quote.gif So what's consider in the same subnet? aaa.bbb.ccc.ddd.eee The aaa.bbb. are the network ID and ccc. is the subnet ID and eee is the node ID.

Actually, the whole "network id" versus "subnet id" is a legacy "classful" convention that nobody really bothers with anymore. These days everything is "classless".

 

[Pepsione1]

Quote:

Originally Posted by EvilGnome6 /img/forum/go_quote.gif If your subnet mask was set to 0.0.0.0, no portion of the address is allocated as the network portion and every other address will be considered as "local". The computer will never forward a packet to a router and be unable to communicate with anything outside of the local subnet.

Computer A has address 192.168.0.1 subnet mask 255.255.255.0
Computer B has address 10.10.10.10 subnet mask 0.0.0.0

When A sends a packet to B, it would not be able to obviously (unless a router is in a place) because their network ID are different

When B sends a packet to A, would the packet arrive at A? It should right because B wouldn't realize that it doesn't have network ID because of the subnet mask. Therefore it wouldn't know 192.168.0. isn't on the same network as itself.


Another example
Computer A has address 192.168.0.1 subnet mask 255.255.255.0
Computer B has address 192.168.5.1 subnet mask 255.255.0.0

A wouldn't be able to communicate with B however B would be able to do so with A. Am I correct?

These examples are based on the fact computer A & B are on the same physical network connected through a switch.

 

[EvilGnome6]

Quote:

Originally Posted by Pepsione1 /img/forum/go_quote.gif A wouldn't be able to communicate with B however B would be able to do so with A. Am I correct?

Yes, both of those examples are correct.

The other point to consider is that if you're using TCP, the two computers will perform a three way handshake before they actually transfer any data. Kind of an "are you there?" - "yes I am" - "here's some data for you".

Since two-way communications between A and B isn't working, TCP won't work.

 

[Pepsione1]

Is it possible to have a network with 0.0.0.0 subnet mask on every node? I have seen this configuration and all the nodes are connected together in the same network physically. I wonder if there are any drawbacks with this configuration? Besides the fact that this network will never be able to communicate with another because it's treating all traffic as local traffic.

 

[EvilGnome6]

Quote:

Originally Posted by Pepsione1 /img/forum/go_quote.gif I wonder if there are any drawbacks with this configuration? Besides the fact that this network will never be able to communicate with another because it's treating all traffic as local traffic.

No real drawbacks unless you have many, many hosts. You will be in effect creating a single broadcast domain where all traffic is sent to every host all the time. This is why we have routers to separate networks to keep local (broadcast) traffic and remote traffic separate.

With less than a few hundred computers, this is not too much of an issue. If you have thousands of computers, your network will start to get bogged down with broadcast traffic.