DHCP Split Scopes

cbriant

How likely is it that i'll get questions on DHCP split scopes? Can anyone who has done the 293 exam let me know if there were any questions on this?

Chris.

dynamik

Possibly. What troubles you about split scopes? You basically just put the same scope on your DHCP servers, and exclude ranges so none of the servers will lease the same address. This is typically (or at least suggested by MS) done in an 80/20 ratio. Also, you'll need to setup identical reservations on both the servers.

royal

One thing I'd like to add. There is another way to do split scopes.

Method #1 - The way dynamik stated. Add the same scope on both servers and set up different exclusions on both servers.
Method #2 - Don't add exclusions, but only add the scope using the ranges each DHCP server will hand out.

Here is an example that may help you out in better understanding:
192.168.1.1 not part of your scope and reserved for router
192.168.1.2 through 192.168.1.20 reserved for servers (this can be a part of your scope and reserved or just not added to the scope at all)

Method #1
DHCP Server #1 (80%)
Scope - 192.168.1.21 - 192.168.1.254
Exclusion- 192.168.1.221 - 192.168.1.254
Result - Only hands out 192.168.1.21 - 192.168.1.220

DHCP Server #2 (20%)
Scope - 192.168.1.21 - 192.168.1.254
Exclusion - 192.168.1.21 - 192.168.1.220
Result - Only hands out 192.168.1.221 - 192.168.1.254

Method #2
DHCP Server #1 Scope (80%) - 192.168.1.21 - 192.168.1.220
DHCP Server #2 Scope (20%) - 192.168.221 - 192.168.1.254

As you can see with both methods, one server hands out 80% of the addresses for the same range while the other server hands out 20%. The difference here, is Method #1 uses the same scope between the two servers but use a different set of exclusions whereas Method #2 uses a different scope altogether thus negating the need to use exclusions.

Using exclusions vs not using exclusions is merely just a preference. And btw, 80/20 is the book information. It's better to do 50/50 or 60/40 in real-world due to 1 server going down, the other can pick up a lot of the slack instead of only having a few addresses left. Regardless, it's easy to bring a new DHCP server up anyways. Just set up a new server and enable conflict detection so you don't start having ip conflicts.

hetty

The only thing I would add is, after going in to a small network where no useful network documentation existed I could tell straight away that there was another DHCP server on the network after seeing the reservations entered in method 1. With method 2 you don’t have that extra little bit of information to hand if you are going in blind, it must be documented or you need to track it down.

donald7862003

royal wrote:

One thing I'd like to add. There is another way to do split scopes.

Method #1 - The way dynamik stated. Add the same scope on both servers and set up different exclusions on both servers.
Method #2 - Don't add exclusions, but only add the scope using the ranges each DHCP server will hand out.

Here is an example that may help you out in better understanding:
192.168.1.1 not part of your scope and reserved for router
192.168.1.2 through 192.168.1.20 reserved for servers (this can be a part of your scope and reserved or just not added to the scope at all)

Method #1
DHCP Server #1 (80%)
Scope - 192.168.1.21 - 192.168.1.254
Exclusion- 192.168.1.221 - 192.168.1.254
Result - Only hands out 192.168.1.21 - 192.168.1.220

DHCP Server #2 (20%)
Scope - 192.168.1.21 - 192.168.1.254
Exclusion - 192.168.1.21 - 192.168.1.220
Result - Only hands out 192.168.1.221 - 192.168.1.254

Method #2
DHCP Server #1 Scope (80%) - 192.168.1.21 - 192.168.1.220
DHCP Server #2 Scope (20%) - 192.168.221 - 192.168.1.254

As you can see with both methods, one server hands out 80% of the addresses for the same range while the other server hands out 20%. The difference here, is Method #1 uses the same scope between the two servers but use a different set of exclusions whereas Method #2 uses a different scope altogether thus negating the need to use exclusions.

Using exclusions vs not using exclusions is merely just a preference. And btw, 80/20 is the book information. It's better to do 50/50 or 60/40 in real-world due to 1 server going down, the other can pick up a lot of the slack instead of only having a few addresses left. Regardless, it's easy to bring a new DHCP server up anyways. Just set up a new server and enable conflict detection so you don't start having ip conflicts.

Microsoft will say that it is better to exclude the addresses in DHCP splitt scope rather than just not including them in the scope

1MeanAdmin

I thought the "Microsoft way" is to use superscopes when implementing 80/20 rule. I may be wrong though.

hetty

midiman wrote:

I thought the "Microsoft way" is to use superscopes when implementing 80/20 rule. I may be wrong though.

They are not the same thing. A split-scope is the implementation of the 80/20 rule for fault tollerance. Whereas a superscope is the grouping of multiple scopes on the DHCP server.

Split-scope vs Superscope

royal

hetty wrote:

midiman wrote:

I thought the "Microsoft way" is to use superscopes when implementing 80/20 rule. I may be wrong though.

They are not the same thing. A split-scope is the implementation of the 80/20 rule for fault tollerance. Whereas a superscope is the grouping of multiple scopes on the DHCP server.

Split-scope vs Superscope

Also: http://www.shudnow.net/2007/11/20/dhcp-scope-vs-superscope/

hetty

royal wrote:

Also: http://www.shudnow.net/2007/11/20/dhcp-scope-vs-superscope/