Classful VS Classless protocols, help?

[KraftDinner]

Alright just as the title suggests, could someone give me a definition and difference between classful and classless protocols, preferably something that's easy to understand? Some examples would be nice too! 

 

This forum has been a huge help for all my networking questions so far, so thanks in advance!

[Wombo]

Alright just as the title suggests, could someone give me a definition and difference between classful and classless protocols, preferably something that's easy to understand? Some examples would be nice too! 

 

This forum has been a huge help for all my networking questions so far, so thanks in advance!

Classfull protocols only know the class boundaries. This often involves them not even sending subnet mask information at all. Rip v1 for example does not send subnet mask information as it is strictly a classfull routing protocol.

 

Classless protocols do send subnet mask information and fully support VLSM.

 

For example a classfull protocol would summarize 10.10.10.0/24 to 10.0.0.0 and send that on its way. The recieving end would then look at the 10.0.0.0 address and apply a /8 to this as it can tell it is class A. A classless protocol would keep the mask, /24 (would be sent as 255.255.255.0) and send this with the information for 10.10.10.0. This allows for much more complex routing and a better use of the available address space. While classfull protocols functioned well for their time the simplicity just doesn't work for us anymore. Classfull protocols are entirely legacy. They are useful to understand however as it lends itself well to learning the functions of auto-summarization in protocols such as BGP and EIGRP.

[KraftDinner]

Classfull protocols only know the class boundaries. This often involves them not even sending subnet mask information at all. Rip v1 for example does not send subnet mask information as it is strictly a classfull routing protocol.

 

Classless protocols do send subnet mask information and fully support VLSM.

Ok, this makes senses, but what kind of subnet mask information are they sending?

[Wombo]

Ok, this makes senses, but what kind of subnet mask information are they sending?

With classfull protocols they don't send any subnet mask information. The receiving end looks at the address and determines what class it belongs to and applies the correct mask to that.

 

For example let's say we are using RIP v1 and we have an interface 10.10.10.1/24. We want to send this to a neighboring device.

 

RIP v1 would look at the address and determine what class it belongs to, in this case class A, and summarize it to the class boundary. This gives us 10.0.0.0/8. RIP v1 will then remove the subnet mask and just send information for 10.0.0.0. The receiving end will look at the address and identify which class each address belongs to. In this case the receiving end will add the /8 back on and place the route into the routing table.

 

As you can see if we only wanted to send information for the 10.10.10.0/24 network we really can't with a classfull protocol. This is why they are entirely legacy now as we either have to send the entire classfull range, or nothing at all.

[KraftDinner]

With classfull protocols they don't send any subnet mask information. The receiving end looks at the address and determines what class it belongs to and applies the correct mask to that.

 

For example let's say we are using RIP v1 and we have an interface 10.10.10.1/24. We want to send this to a neighboring device.

 

RIP v1 would look at the address and determine what class it belongs to, in this case class A, and summarize it to the class boundary. This gives us 10.0.0.0/8. RIP v1 will then remove the subnet mask and just send information for 10.0.0.0. The receiving end will look at the address and identify which class each address belongs to. In this case the receiving end will add the /8 back on and place the route into the routing table.

 

As you can see if we only wanted to send information for the 10.10.10.0/24 network we really can't with a classfull protocol. This is why they are entirely legacy now as we either have to send the entire classfull range, or nothing at all.

So basically with RIPV1 we would have to send packets to the entire network, and not a specific subnet? 

[Wombo]

So basically with RIPV1 we would have to send packets to the entire network, and not a specific subnet? 

Well not exactly. As RIPv1 doesn't allow us to use a mask when adding networks with the network command we are limited to enabling RIPv1 for an entire classfull range. We can then use the passive interface command to limit where the RIPv1 broadcast messages are sent. The fact still remains however that with classfull protocols we either send the entire subnet range, or nothing at all.

[KraftDinner]

Well not exactly. As RIPv1 doesn't allow us to use a mask when adding networks with the network command we are limited to enabling RIPv1 for an entire classfull range. However we can use the passive interface command to limit where the RIPv1 broadcast messages are sent. The fact still remains however that with classfull protocols we either send the entire subnet range, or nothing at all.

I see, and with classless protocols we can send a certain subnet range as VLSM is enabled and allows us to divide subnets based on the number of hosts?

[Wombo]

I see, and with classless protocols we can send a certain subnet range as VLSM is enabled and allows us to divide subnets based on the number of hosts?

Yup. We can also do custom sumarizations to optimize routing tables.

 

Using specific masks also lets us chose exactly the subnets we want to advertise. Sometimes it may be the cast that you want to advertise the 10.10.10.0/24 network, but not the 10.11.10.0/24 network. With a classfull protocol this is pretty much impossible. However with a classless protocol we can enable routing only on the specific interfaces we want.

 

Classless is the way everything is now. Make sure you understand the differences between the two and how to identify the classes of addresses, but don't worry to much about the clasful protocols themselves.

[KraftDinner]

Yup. We can also do custom sumarizations to optimize routing tables.

 

Using specific masks also lets us chose exactly the subnets we want to advertise. Sometimes it you may want to advertise the 10.10.10.0/24 network, but not the 10.11.10.0/24 network. With a classfull protocol this is pretty much impossible. However with a classless protocol we can enable routing only on the specific interfaces we want.

 

Classless is the way everything is now. Make sure you understand the differences between the two and how to identify the classless of addresses.

Ok this is all starting to make sense, so just to clarify, classful routing associates 1 subnet mask to every network connected to a specific router? And classless routing supports VLSM and allows us to send a certain subnet range associated with each network connected to a specific router?

[Wombo]

Ok this is all starting to make sense, so just to clarify, classful routing associates 1 subnet mask to every network connected to a specific router? And classless routing supports VLSM and allows us to send a certain subnet range associated with each network connected to a specific router?

Well a classful protocol can use more than one subnet mask per router. All it is doing is summarizing to the classful boundaries (A,B, or C).

 

For example say we have 3 networks on a router.

 

10.10.10.0/24

172.16.10.0/24

192.0.0.0/8

 

If we added all these to the classful routing protocol it would summarize these routes to their classful boundaries. In this case 10.0.0.0/8, 172.16.0.0/16 and 192.0.0.0/24.

 

It is important the know all the classful boundaries and how to identify what class an address belongs to. In this case I gave examples of a class A, B and C.

 

As for classless you're pretty much right. Classless means we can use any mask and we are not tied to any classful boundaries, but we can always use them if we want.

 

Classful - obeys the class boundaries

Classless - does not