Logical subnets on the same physical segment

Dracula28

Say you have two logical subnets on the same physical segment:

Subnet A 192.168.0.0 /24
Subnet B 192.168.1.0 /24

Would the two subnets need a router to communicate with each other?

The reason why I'm asking this is, because I'm reading about DHCP superscopes again, and I never really seem to fully understand that. In the training kit, there is an example, where a DHCP superscope is leasing out ip addresses to several logical subnets, and there is no DHCP relay agent involved (page 7-29, figure 7-7). Isn't a DHCP server able to only lease out addresses to computers on the same subnet as itself is on (unless there is a relay agent or 1542-compliant router involved?)

And in superscopes, how can the clients determine which addresses they are supposed to lease?

dynamik

Yes, a router is needed for the hosts to communicate with each other since they are on different subnets. However, a DHCP relay agent or 1542 compliant router is not needed because they are all on the same physical network segment. The reason you need either of those is because routers do not (typically) forward DHCP broadcasts to other networks. This really wasn't explained well, but I believe that is correct. It didn't really click for me either until I just typed it out now (I'm literally a few pages ahead of you in my review).

You might be able to use something like options classes or reservations to specify addresses to specific clients, but it shouldn't matter which addresses they get.

Dracula28

Yeah that made sense. If they are on the same physical segment, I guess it would not matter what address they got. Because they would just use the default gateway to communicate with the other logical subnet. But then I guess the training kits example on figure 7-9 is wrong, as it would not matter what address they got. Or am I wrong?

dynamik

I hate 7-9 and 7-10

To be honest, I really couldn't figure out what they were saying. I understand that the other DHCP server would send a NACK because the client was trying to renew an address that wasn't in it's scope. It also makes sense that this wouldn't happen if they both used a multinet and excluded each other's subnet. However, I don't see why it would matter if a client got an address on the other subnet. All the machines are connected to a hub, which is connected to a router, so why would it matter? They also never go into how to restrict the clients to a single DHCP. They're just saying that the client will continue to renew the address at the DHCP from which is was leased. Which is by no means a guarantee that it will continue to do so in the future. The server could fail, or a variety of other things could happen that would prevent it from renewing at that server. I think that was just a poorly written section.

snadam

dynamik wrote:

I hate 7-9 and 7-10

To be honest, I really couldn't figure out what they were saying. I understand that the other DHCP server would send a NACK because the client was trying to renew an address that wasn't in it's scope. It also makes sense that this wouldn't happen if they both used a multinet and excluded each other's subnet. However, I don't see why it would matter if a client got an address on the other subnet. All the machines are connected to a hub, which is connected to a router, so why would it matter? They also never go into how to restrict the clients to a single DHCP. They're just saying that the client will continue to renew the address at the DHCP from which is was leased. Which is by no means a guarantee that it will continue to do so in the future. The server could fail, or a variety of other things could happen that would prevent it from renewing at that server. I think that was just a poorly written section.

HA! THAT MAKES 3 OF US!!!

Other than the 'rebinding state' section, it really doesnt go into detail how to restrict/renew leases if something happens to a DHCP server.


EDIT: MY bad, the rebinding state (and lease renewal process for that matter) is described in chapter 8...This book has a tendency to throw words, and protocols, and other practices that are "described later in this book"..

sprkymrk

I'll have to check out that book and see what is confusing everyone. For superscopes, royal and I had a good discussion here:

http://www.techexams.net/forums/viewtopic.php?t=24392

royal

And an article that explains the scope vs superscope and when the addresses are on the same physical segment Vs across different physical segments with a router in between.
http://www.shudnow.net/2007/11/20/dhcp-scope-vs-superscope/

snadam

royal wrote:

And an article that explains the scope vs superscope and when the addresses are on the same physical segment Vs across different physical segments with a router in between.
http://www.shudnow.net/2007/11/20/dhcp-scope-vs-superscope/

would that be on the 291 sticky?

Dracula28

Thanks for the feedback guys.

royal

snadam wrote:

royal wrote:

And an article that explains the scope vs superscope and when the addresses are on the same physical segment Vs across different physical segments with a router in between.
http://www.shudnow.net/2007/11/20/dhcp-scope-vs-superscope/

would that be on the 291 sticky?

It is now.

snadam

royal wrote:

snadam wrote:

royal wrote:

And an article that explains the scope vs superscope and when the addresses are on the same physical segment Vs across different physical segments with a router in between.
http://www.shudnow.net/2007/11/20/dhcp-scope-vs-superscope/

would that be on the 291 sticky?

It is now.

many thanks royal!