- Posts: 77
- Thank you received: 0
subnet masks
19 years 2 months ago #10365
by mew
Replied by mew on topic Re: subnet masks
I have to admit that it has been awhile since I read this sites section on subnetting. Since I have now encouraged others to read it I decided to read it again. I remember telling Chris that I thought his definition for classless IP address was in contrast to every other author I have read on the subject. This is the source of much confusion. If you ignore Chris’s insistence of calling any IP address with something other than a default subnet mask a classless IP address I can still say Chris does an excellent job of explaining how to subnet and how subnets work.
IP Addresses by themselves are not classful or classless no matter what subnet mask you use. When discussing IP addresses and their use with classless routing protocols the term class is eliminated all together. You no longer have class A, B, C, etc. You do not even have subnet masks. You have a Prefix and a Prefix Mask thus the term classless.
On the other hand every device I configure static addresses on still ask’s for “IP address” and “subnet mask”. This would seem to make the use of a Prefix and Prefix Mask appear more theoretical than practical. This is the cost of rapid changes in technology and having to deal with legacy processes as well as legacy hardware.
You need to understand that the classful routing protocols (RIP v1 and IGRP) do not exchange subnet mask information. However, you can subnet class A, B, or C networks in a classful manner which means that throughout the network you use the same subnet mask. These routing protocols will report that the networks have been subnetted based on the mask used on it’s interfaces. If you use the same subnetmask throughout the network it will function properly.
Using a different subnet mask is allowed with classless routing protocols and is referred to as VLSM. For example if router 1 (R1) is a Cisco 2501 and ethernet 0 and has an address of 192.168.0.1 with a subnet mask of 255.255.255.192 and it’s Serial 0 has an address of 192.168.0.65 with a subnet mask of 255.255.255.252 and is connected to router 2’s (R2) Serial 0 which is a Cisco 2501 that has an address of 192.168.0.66 with a subnet mask of 255.255.255.252 and has an ethernet 0 with an address of 192.168.0.97 with a subnet mask of 255.255.255.224.
Using RIPv1 router 1 will not accept the route to the 192.168.0.96 network because it will not know that the subnet’s address doesn’t over lap with the subnets it is already connected to. The same is true for Router 1 and the 192.168.0.0 network. However changing the routing protocol to RIP v2 (Note: this is the classless form of RIP) will fix this problem because classless routing protocols exchange subnet mask information.
On the other hand to illustrate a classless addressing scheme and having it not work with a classful routing protocol then fixing it with classless routing protocol is not proof enough.
Therefore, if you have a subnetwork on R1 of 192.168.0.32/27 where you assign 192.168.0.33 with a subnet mask of 255.255.255.224 to e0 on R1. And you have a subnetwork of 192.168.0.64/27 where you assign an address of 192.168.0.65 with a subnet mask of 255.255.255.224 to the s0 interface which is connected to the s0 interface of R2 that has the address 192.168.0.66 with a subnet mask of 255.255.255.224. And you have a subnetwork of 192.168.0.96/27 with and IP address of 192.168.0.97 with a subnet mask of 255.255.255.224 assigned to R2’s e0. This will result in RIPv1 advertising the subnets to each other (Note: this is classful subnetting).
I have used two Cisco 2501’s to prove this. I also turned off IP classless (on by default due to the IOS used) to make sure that nothing classless was used. RIPv1 on R1 and RIPv1 on R2 advertise the subnetworks 192.168.0.32/27 and 192.168.0.96/27 respectively. “Show IP Route” produces the results that indicate that with a classful routing protocol and with “IP Classless” turned off that RIP routes are advertised using classful subnetting rules. Note that the subnet mask used on all interfaces was 255.255.255.224. The very same subnet mask referred to in “Subnet Masks & Their Effect” on this web page.
IP Addresses by themselves are not classful or classless no matter what subnet mask you use. When discussing IP addresses and their use with classless routing protocols the term class is eliminated all together. You no longer have class A, B, C, etc. You do not even have subnet masks. You have a Prefix and a Prefix Mask thus the term classless.
On the other hand every device I configure static addresses on still ask’s for “IP address” and “subnet mask”. This would seem to make the use of a Prefix and Prefix Mask appear more theoretical than practical. This is the cost of rapid changes in technology and having to deal with legacy processes as well as legacy hardware.
You need to understand that the classful routing protocols (RIP v1 and IGRP) do not exchange subnet mask information. However, you can subnet class A, B, or C networks in a classful manner which means that throughout the network you use the same subnet mask. These routing protocols will report that the networks have been subnetted based on the mask used on it’s interfaces. If you use the same subnetmask throughout the network it will function properly.
Using a different subnet mask is allowed with classless routing protocols and is referred to as VLSM. For example if router 1 (R1) is a Cisco 2501 and ethernet 0 and has an address of 192.168.0.1 with a subnet mask of 255.255.255.192 and it’s Serial 0 has an address of 192.168.0.65 with a subnet mask of 255.255.255.252 and is connected to router 2’s (R2) Serial 0 which is a Cisco 2501 that has an address of 192.168.0.66 with a subnet mask of 255.255.255.252 and has an ethernet 0 with an address of 192.168.0.97 with a subnet mask of 255.255.255.224.
Using RIPv1 router 1 will not accept the route to the 192.168.0.96 network because it will not know that the subnet’s address doesn’t over lap with the subnets it is already connected to. The same is true for Router 1 and the 192.168.0.0 network. However changing the routing protocol to RIP v2 (Note: this is the classless form of RIP) will fix this problem because classless routing protocols exchange subnet mask information.
On the other hand to illustrate a classless addressing scheme and having it not work with a classful routing protocol then fixing it with classless routing protocol is not proof enough.
Therefore, if you have a subnetwork on R1 of 192.168.0.32/27 where you assign 192.168.0.33 with a subnet mask of 255.255.255.224 to e0 on R1. And you have a subnetwork of 192.168.0.64/27 where you assign an address of 192.168.0.65 with a subnet mask of 255.255.255.224 to the s0 interface which is connected to the s0 interface of R2 that has the address 192.168.0.66 with a subnet mask of 255.255.255.224. And you have a subnetwork of 192.168.0.96/27 with and IP address of 192.168.0.97 with a subnet mask of 255.255.255.224 assigned to R2’s e0. This will result in RIPv1 advertising the subnets to each other (Note: this is classful subnetting).
I have used two Cisco 2501’s to prove this. I also turned off IP classless (on by default due to the IOS used) to make sure that nothing classless was used. RIPv1 on R1 and RIPv1 on R2 advertise the subnetworks 192.168.0.32/27 and 192.168.0.96/27 respectively. “Show IP Route” produces the results that indicate that with a classful routing protocol and with “IP Classless” turned off that RIP routes are advertised using classful subnetting rules. Note that the subnet mask used on all interfaces was 255.255.255.224. The very same subnet mask referred to in “Subnet Masks & Their Effect” on this web page.
19 years 2 months ago #10370
by nske
Replied by nske on topic Re: subnet masks
Great post on a tricky subject! 
Ouch my head hurts :!:
Time to create page: 0.123 seconds