Can setting up a VLAN give us more addresses?
We have a 192.168.0.X (255.255.255.0) class network and are running out of addresses. We have about 100 thin clients that only need to access the Servers that host them. If I set up the thin clients to use one VLAN, their servers (which are their gateway to the rest of the network) to use both that VLAN and a primary VLAN, and everything else to just use the primary VLAN, would that free up 100 addresses so we wouldn't have to change all our nodes to use a different class of addresses? Would it work if the thin clients used a 192.168.1.X (255.255.255.0) address scheme and the primary VLAN used a 192.168.0.X address scheme? What type of address would the thin client servers use to be able to access both VLANs, the 192.168.0.X or the 192.168.1.X? Would the thin client servers need to be assigned 2 addresses?
ASKER
We are using DHCP, but there are a lot of static addresses on the network, some from third party companies (security, phone, etc) which we don't know how to change, some from legacy devices that you can't easily change the address of, and a lot of computers reference these static addresses (ex. printing to a TCP/IP port), so if the address changes, the computers need to be changed to reference the new address. I was hoping that the VLAN solution would be an easier transition for us.
Can a server be assigned two addresses? What's wrong with doing that? I thought that is common practice for a computer to be able to access more than one VLAN? I thought that it was just a matter of configuring the port of the switch it's plugged into to "Trunk" mode instead of "Access" mode (in Cisco terms)?
Can a server be assigned two addresses? What's wrong with doing that? I thought that is common practice for a computer to be able to access more than one VLAN? I thought that it was just a matter of configuring the port of the switch it's plugged into to "Trunk" mode instead of "Access" mode (in Cisco terms)?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
You could use just use 2 networks or better still if it's just that your running out of address change your network mask from 255.255.255.0 to 255.255.254.0 which will mean that 192.168.0.1 and 192.168.1.1 are on the same network and can see each other. The range would actually be 192.168.0.1 to 192.168.1.25, which gives you twice as many addresses. 255.255.252.0 will give you 4 times as many etc. This would do away with the need to setup VLANS across your network switches as well as setting up any trunk ports etc and also any routing required to route the 2 networks together.
Changing your network mask on static devices might be a bit of a pain but you can change it easily in your DHCP server scope.
Changing your network mask on static devices might be a bit of a pain but you can change it easily in your DHCP server scope.
ASKER
Why won't a computer with IP 192.168.0.1 and subnet 255.255.255.0 see a computer with IP 192.168.1.1 and subnet 255.255.254.0? Shouldn't both computers be on the same network? The following is my understanding ...
Consider 3rd octet ...
00000000 - IP #1 (3rd octet)
11111111 - SubNet #1 (all bits define the Network)
00000000 - Network
00000001 - IP#2 (3rd octet)
11111110 - SubNet#2 (only 7 bits define the Network)
00000000 - Network (Network comes out the same because the last bit defines the Node. Also, the 1st, 2nd, and 4th octets are the same for both IP and subnet)
Consider 3rd octet ...
00000000 - IP #1 (3rd octet)
11111111 - SubNet #1 (all bits define the Network)
00000000 - Network
00000001 - IP#2 (3rd octet)
11111110 - SubNet#2 (only 7 bits define the Network)
00000000 - Network (Network comes out the same because the last bit defines the Node. Also, the 1st, 2nd, and 4th octets are the same for both IP and subnet)
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
To add to my above comment, If your other machine has subnet 255.255.255.0 then it's binary would be
11111111 11111111 11111111 00000000
I've bolded the above to show where our other subnet went up top, This now means that your 2 address ranges cannot see each other because when broken down into Binary the change happens before the last 0, in other words they are still on 2 different networks. If that makes sense.
11111111 11111111 11111111 00000000
I've bolded the above to show where our other subnet went up top, This now means that your 2 address ranges cannot see each other because when broken down into Binary the change happens before the last 0, in other words they are still on 2 different networks. If that makes sense.
ASKER
I think I found a way to get more nodes on our network without changing the class of addresses we are using (i.e. without changing the subnet mask of all of the devices on our network). The thin clients and their servers will have a subnet mask of 255.255.254.0, and everything else will stay at subnet mask 255.255.255.0. The thin clients will have an address of 192.168.1.X, and everything else, including the servers of the thin clients, will still have an address of 192.168.0.X. The thin clients will be able to communicate with their servers and only with their servers (which is fine), servers will still be able to communicate with the rest of the devices on the network because then though they have a different subnet mask, the first 3 octets of the IP will be the same as the rest of the devices on the network (except the thin clients).
Does anyone have any reasons why this approach is not a good idea? I tried this and it worked. However, at the time I tried this we got network disconnect problems, and all the thin clients hosted by one of our servers were locked up (frozen). Someone restarted that server. to remedy this. I don't know if it is a coincidence that this happen when I changed the TCP/IP settings of a couple of computers. One of the computers I changed the address of already got an address by DHCP that was logged in WINS. I didn't delete the WINS entry when I changed its address to a static one. Would this cause a conflict?
Does anyone have any reasons why this approach is not a good idea? I tried this and it worked. However, at the time I tried this we got network disconnect problems, and all the thin clients hosted by one of our servers were locked up (frozen). Someone restarted that server. to remedy this. I don't know if it is a coincidence that this happen when I changed the TCP/IP settings of a couple of computers. One of the computers I changed the address of already got an address by DHCP that was logged in WINS. I didn't delete the WINS entry when I changed its address to a static one. Would this cause a conflict?
ASKER
ad4m1 - Thank you for going through all of that with me. I understand subnet masks much better now.
cdusio - I ended up using two IP addresses on the thin client's servers. Thanks for the idea.
cdusio - I ended up using two IP addresses on the thin client's servers. Thanks for the idea.
otherwise, I *think* your solution is sound, but I am a bit concerned about the servers using both VLAN addresses.