Static IP Address Question

[Robert Tickle]

If I tell my router's DHCP server to start assigning IP addresses at 192.168.1.100, and if I configure static IP addresses in my devices to be below 192.168.1.100, do I still need to reserve those static IP addresses in the DHCP server?

 

[ColinTaylor]

No. However you may still want to. Not for the IP address (which is set on the client) but to create a corresponding host name entry in DNS for the client. You just have to remember to update the router's IP entry should you change it on the client.

But then it begs to question, if you're creating entries in the router's DHCP server anyway, why not set to clients to use DHCP and administer everything centrally? (The answer to that is that you might want some devices to work even if the router was offline, e.g. security cameras.)

 

[Jeffrey Young]

To further what Colin said re the DNS, if you do need to stay with a static IP for reasons, you can always add your machines FQDN to the /configs/hosts.add file.

 

[ApexRon]

For my DHCP server, I chose a Class C address range and divided it into 4 subnets. I used the second subnet for my DHCP server (192.168.x.65 to 192.168.x.126). I use the third subnet to configure some devices for static IP address. There are some devices that use DHCP for which I manually set their IP address from the DHCP range. Works well.

 

[ColinTaylor]

For my DHCP server, I chose a Class C address range and divided it into 4 subnets. I used the second subnet for my DHCP server (192.168.x.65 to 192.168.x.126). I use the third subnet to configure some devices for static IP address. There are some devices that use DHCP for which I manually set their IP address from the DHCP range. Works well.

Just to be pedantic you don't actually have 4 subnets (unless you have 4 routers), you have a single subnet with a DHCP pool from .65 to .126. You could equally have chosen 50 to 100, 100 to 200 or any other arbitrary range within 192.168.x.0/24.

 

[ApexRon]

Just to be pedantic you don't actually have 4 subnets (unless you have 4 routers), you have a single subnet with a DHCP pool from .65 to .126. You could equally have chosen 50 to 100, 100 to 200 or any other arbitrary range within 192.168.x.0/24.

Ya, I skimmed over that. To restate, I divided the address range of a Class C address into 4 sub-nets and used the address range of the second sub-net as the address range for the DHCP server. It makes my network addresses spread sheet easier. Let's see if I can illustrate:
1st Sub-Net: Unused | 2nd Sub-Net: DHCP | 3rd Sub-Net: Static | 4th Sub-Net: Unused
The router LAN interface uses the entire Class C.

I want my MacBook to use a specific address at home so I had been defining a static IP on the Mac. When I travel, I would have to have another network profile on my Mac to flip to DHCP. So I now leave my Mac with a DHCP profile and have a specific IP address from DHCP pool assigned on my router. When I travel, all I have to do is hook up to a different SSID.

 

[drinkingbird]

Ya, I skimmed over that. To restate, I divided the address range of a Class C address into 4 sub-nets and used the address range of the second sub-net as the address range for the DHCP server. It makes my network addresses spread sheet easier. Let's see if I can illustrate:
1st Sub-Net: Unused | 2nd Sub-Net: DHCP | 3rd Sub-Net: Static | 4th Sub-Net: Unused
The router LAN interface uses the entire Class C.

I want my MacBook to use a specific address at home so I had been defining a static IP on the Mac. When I travel, I would have to have another network profile on my Mac to flip to DHCP. So I now leave my Mac with a DHCP profile and have a specific IP address from DHCP pool assigned on my router. When I travel, all I have to do is hook up to a different SSID.

I think the point of clarification he was trying to make is you only have one subnet. Just because you are assigning IPs along subnet boundaries doesn't make them subnets. I'd call them "ranges". To make it easier on yourself, why not just use round numbers like 0, 50, 100, etc?

Personally I'd use DHCP reservations and change all your machines to DHCP. Centrally managed, helps client list work better, lots less work if you ever want to change DNS, etc.

 

[ApexRon]

I think the point of clarification he was trying to make is you only have one subnet. Just because you are assigning IPs along subnet boundaries doesn't make them subnets. I'd call them "ranges". To make it easier on yourself, why not just use round numbers like 0, 50, 100, etc?

Personally I'd use DHCP reservations and change all your machines to DHCP. Centrally managed, helps client list work better, lots less work if you ever want to change DNS, etc.

Retired Network Design Engineers think in sub-nets

 

[Robert Tickle]

No. However you may still want to. Not for the IP address (which is set on the client) but to create a corresponding host name entry in DNS for the client. You just have to remember to update the router's IP entry should you change it on the client.

But then it begs to question, if you're creating entries in the router's DHCP server anyway, why not set to clients to use DHCP and administer everything centrally? (The answer to that is that you might want some devices to work even if the router was offline, e.g. security cameras.)

Thank you for this reply. It was very helpful. My understanding of the static IP address table in a router was incomplete. I think I understand better now, but let me ask a followup question to check. When a wifi client requests a connection to a network, the client presents both the stored password and its MAC address. Assuming the password is correct, the DHCP server looks to see if the presented MAC address is in the static IP address table. If the MAC address is there, the DHCP server assigns the client the IP address stored with the MAC address. If the presented MAC address is not in the table, the DHCP server assigns the client the next available IP address in the pool? Am I close?

 

[Jeffrey Young]

Thank you for this reply. It was very helpful. My understanding of the static IP address table in a router was incomplete. I think I understand better now, but let me ask a followup question to check. When a wifi client requests a connection to a network, the client presents both the stored password and its MAC address. Assuming the password is correct, the DHCP server looks to see if the presented MAC address is in the static IP address table. If the MAC address is there, the DHCP server assigns the client the IP address stored with the MAC address. If the presented MAC address is not in the table, the DHCP server assigns the client the next available IP address in the pool? Am I close?

Yes Sir. That is right.

 

[drinkingbird]

Retired Network Design Engineers think in sub-nets

Eh, different schools of thoughts, I've been a network engineer and architect for almost 25 years and my home network runs separate /24s to keep it easy and within that I use even number boundaries for division. With private IPs, really the only time I feel it necessary to subnet is /31s for point to point links to make it obvious that it is a transit link, and eliminate unnecessary broadcasts.

 

[ColinTaylor]

Am I close?

Yes, pretty much.

Side note: To avoid confusing other people I suggest you don't use the term "static address" when referring to the router's DHCP entries. Asus calls them "manually assigned", most other people would call them "DHCP reservations". This is to avoid confusing them with a "static IP address" which typically refers to an IP address that is manually set in a client's network interface settings.

 

[drinkingbird]

Yes, pretty much.

Side note: To avoid confusing other people I suggest you don't use the term "static address" when referring to the router's DHCP entries. Asus calls them "manually assigned", most other people would call them "DHCP reservations". This is to avoid confusing them with a "static IP address" which typically refers to an IP address that is manually set in a client's network interface settings.

Eh manually assigned still kinda sounds like static..... DHCP reservation is better.

 

[ApexRon]

Devices in your network will have an IP address either by using DHCP or by manually configuring on the device itself.

The router’s DHCP server will assign the device’s IP address either ‘randomly’ or a DHCP reserved IP address based on device MAC address.

A device with a manually configured IP address could use an IP address that was assigned by the router to another device, duplicating the IP address on the network. Additionally, multiple devices could be manually configured with the same IP address, duplicating the IP address on the network. Duplicate IP addresses are not a good thing. Bottom line is that if you manually configure device IP addresses you need to take whatever measures necessary to prevent duplication.