DHCP Help

[TimPNRHA]

Good Morning,

I need a little help redoing my DHCP Options as the current scope as run out address and still need to deploy an additional phones. I cannot just expand the subnet from 24 bit to 23 bit as it does not fall into the correct range for out network. Our current setup 10.11.11.0/24 if we go to a 23 bit that scheme would be 10.11.12.0, 10.11.13.0/23.

System IP Settings are:

10.11.11.40
255.255.255.0
10.11.11.1
Layer2 10.11.11.41
E2T 10.11.11.45/255.255.255.0

DHCP Options are:

3 Router 10.11.11.1 scope global
3 Router 10.11.59.129 scope subnet: 10_11_59_subnet (offsite with a dhcp helper)
125 id:ipphone.mitel.com;sw_tftp=10.11.11.40;call_srv=10.11.11.40;dscp=46v46s26
128 tftp server 10.11.11.40
129 Call Server 10.11.11.40
130 IP Phone Identifier MITEL IP PHONE
134 DSCP 46

This controller is part of 16 node cluster and is currently running on version 6.0 SP3.

My plan would be to leave the controller as 10.11.11.40 so I do not have any downtime, this is for a 24/7 healthcare facility. I would just create the following new:

DHCP Subnet
10_11_12_subnet
ip address 10.11.12.0
bit mask 255.255.254.0

DHCP IP Address Range
10_11_12_range
Start 10.11.12.50
End 10.11.13.254
Subnet 10_11_12_subnet(10.11.12.0)

What are your thoughts or how would you approach this?

[acejavelin]

Skipping over the fact this was a poorly designed implementation and bad choice of network scope from the implementation...

All facilities, even hospitals, can plan for downtime... I just did a 600+ phone hospital, including 4 remote clinics and also the ER/Trauma center for the local county, cut-over to a new system AND simultaneous telco switch, there was downtime, but it wasn't a problem because it was planned for properly.

Still here is one possibility, taking what you say to heart with zero downtime... build a new VLAN and deploy the new scope in that VLAN and change/implement your LLDP settings to put phones into that VLAN (remember this will only occur at boot of the phone, so existing phone will not be affected). Once it has been tested and verified functional there are many options to move all the phones to that VLAN, reset individually, reset switches, bounce ports in the switch, etc. Now you have a new VLAN for just voice (except the 5000, but properly implemented shouldn't be a problem), and you opened a whole bunch of your default data VLAN for other network devices.

In my opinion, the best practice here would be to plan a maintenance window and move the entire voice network into a new VLAN, including the 5000, for example into a 172.16.xx.xx subnet with ample room for growth (2-3 times projected 10 year growth), move the phones, 5000, and any other UC servers/services, into the new VLAN, and make a quick change to the other 16 nodes to talk to the new IP address. Again, with proper planning, testing, and implementation, you are still looking at minimal downtime at least for critical areas. You have also "future proofed" your VoIP network for unknown growth.

Also remember you are networked with 16 other sites, so it might be a good idea to get a little conference call going with system/network administrators at other sites to make sure just arbitrarily picking 172.16.16.0/22 for your VoIP network won't mess up some other implementation or plan for the future.

[TimPNRHA]

Thank you Ace,

I should have just used the actual ip settings sorry for the confusion,,, this is for the mitel 3300, current ip is 172.17.17.0/24 I cannot adjust this to a /23 as it falls to 172.17.16.0/23 which 172.17.16.0 is our dmz.

so next option is 172.17.18.0/23,, I am working on the maintenance window to see about the 20 minute outage for the reboot to to move my controller into 172.17.18.0/23 and create the new DHCP Scope options.. I will update NuPoint to see the new subnet 172.17.18.0 255.255.254.0 then all my other nodes IPXnet to the new system ip.. is this correct? is there anything that I need to worry about the Network Elements or with SDS?

[johnp]

Network elements should update automatically

[TimPNRHA]

Can someone please verify my process to complete this with minimal impact:

1. Update NuPoint subnet list to include 172.17.18.0 255.255.254.0 router 172.17.4.40

2.Create new DHCP Scope on the BUH Mitel Controller:
DHCP Subnet
172_17_18_subnet
ip address 172.17.18.0
bit mask 255.255.254.0

DHCP IP Address Range
172_17_18_range
Start 172.17.18.50
End 172.17.19.254
Subnet 172_17_18_subnet(172.17.18.0)

3 Router 172.17.18.1 scope global
3 Router 10.11.59.129 scope subnet: 10_11_59_subnet (offsite with a dhcp helper)
125 id:ipphone.mitel.com;sw_tftp=172.17.18.40;call_srv=172.17.18.40;dscp=46v46s26
128 tftp server 172.17.18.40
129 Call Server 172.17.18.40
130 IP Phone Identifier MITEL IP PHONE
134 DSCP 46

3.Adjust the BUH Mitel System IP Settings

System IP: 172.17.18.40
Subnet 255.255.254.0
Gateway: 172.17.18.1
Layer2 Switch: 172.17.18.41
E2T Card: 172.17.18.45

4. During reboot process log into all other controllers to update the IP/XNET Networking for BUH 200 from 172.17.17.40 to 172.17.18.40

Log into Nupoint, on the main Nupoint Configuration page:
-Go into the NuPoint Webconsole
-edit offline configuration *say yes to copy of active to offline*
-click network elements
-click on BUH200
-change IP Address to 172.17.18.40
-save
-commit changes and exit
- click on activate offline configuration

Verify network elements automatically updates to the new IP Address

Start with controllers in Battleford, then Lloyd Hospital, then Maidstone/Turtleford then remaining controllers to limit impact to NuPoint

[VinceWhirlwind]

You're probably much better off having a separate subnet per campus (small sites), per building (small buildings) or per floor.
 
The best design uses the following concept as a rule of thumb: the fewest numbers of subnets per switch, and the fewest switches per subnet.
 
This greatly facilitates operational management, change management and has the additional benefit of protecting different geographical areas from each other in case of network issues occurring at one site.
 
I always use 10. addresses to give myself maximum space. (I can't understand why anybody wants to use 192.168 or the 172.16 addresses).
 
I make the 2nd octet synonymous with a campus, then the 3rd octet synonymous with a building within that campus, or, for large buildings, a string of subnets running up the building.
For example, one hospital network I did had multiple 7--storey buildings on one campus, with 2 subnets per floor (East v. West)plus an old 1940s campus which was one huge sprawling single-storey building with about 32 subnets on it, one per wiring closet.
 
(I don't make the mistake of using the second octet to differentiate between services - that makes route summarisation impossible).
 
So I end up with a scheme like:
10.10.0.0/16 = Campus A
   10.10.10.0/24 = Building A, Ground, East
   10.10.11.0/24 = Building A, Ground, West
   10.10.12.0/24 = Building A, Lvl1, East
   .
   10.10.28.0/24 = Building B, Ground, East
   .
   10.10.45.0/24 = Building C, Ground
10.20.0.0/16 = Campus B
   10.20.10.0/24 = Porter
   10.20.11.0/24 = Store
   10.20.12.0/24 = Pathology
   10.20.13.0/24 = Renal
   10.20.14.0/24 = Oncology
   .
10.30.0.0/16 = Small Country Hospital A
   10.30.1.0/16 = Comms Room A
   10.30.2.0/16 = Comms Room B
10.40.0.0/16 = Clinic A
etc...
 
There really isn't any benefit to trying to cram as many geographically dispersed devices as possible onto the same subnet, and many drawbacks. There is no drawback to creating many multiple subnets each only using 20 devices.