Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

CCNA question...

Hi seniors, I have read so many articles on classful ip address vs classless ip address but could not get the idea. say there are two ip addresses as follows

1. 10.0.0.1 255.0.0.0

2. 10.0.0.1 255.255.0.0

Which address (above) is classful and which is classless. Lot of theory based on just classful and classless, so a clear concept is vital. Thanks for reply. Best regards. tahir

18 REPLIES
Hall of Fame Super Silver

Re: CCNA question...

Tahir

This is an important topic and is subtle to understand some of its aspects.

I would say that you can look at any address and say that it is claffful or it is classless. An address by itself is neither classful or classless. It is how the address is used that makes it classful or makes it classless.

10.0.0.1 255.0.0.0 is an address in a class A network that has not been subnetted.

10.0.0.1 255.255.0.0 is an address in a class A network in a subnet that is /16. If it is being used as classful all of the subnets will be /16. If it is being used as classless it is likely that subnets will be of different sizes.

Another aspect of classless vs classful is a consideration of contiguous or discontiguous subnets. In a classful environment the subnets will be contiguous while a classless environment will support discontiguous subnets. In a classful network summarization is automatic and is done at network boundaries. In a classless network summarization is not automatic and can be done at various places in the network.

HTH

Rick

Silver

Re: CCNA question...

Rick great response and very clear.

Rick

New Member

Re: CCNA question...

Thanks Rick. Lot of misconception is clear. say we have a network of 10.0.0.1/16 with 10 subnets 10.1.0.0/16 -- 10.10.0.0/16(no ip subnet zero) , it means it is classful network. Whereas if we have 5 networks of 10.0.0.1/16 as 10.1.0.0/16 and another 5 as 10.6.0.0/24 it means this network is classless. And all the ip addresses are either classful or classless. But we use different sizes networks in only VLSM ? You mean classful vs classless is only limited to VLSM ? Please make it clear further. Thanks. Tahir.

Super Bronze

Re: CCNA question...

Let me suggest a bit of history which might help clarify classful vs. classless.

During very early networking, computer resources were very, very expensive, which resulted in many clever approaches to minimize resource demands.

IP addresses are comprised of a network address and a host address. The question is what are the rules of this division and how do you convey it via a routing protocol.

Classful addressing splits the 32-bit IP address space into blocks where the address range itself also denotes the split between the network address and the host address. Having the IP address convey the split between network and host addresses saves dealing with a network mask; its implied. (Also the way the ranges were defined, looking at the first bit indicates a class "A" or not, and the second bit indicates a class "B" or not. Very advantageous when working with early computers where you might have to rotate bits through a sign or carry bit to determine whether it was a one or zero.)

With classless addresses, the division of network and host addresses aren't predefined, so we must communicate it along with the full IP address.

Given just the class "A" address 10.0.0.1, classful addressing rules tell us the mask is 255.0.0.0.

Given 10.0.0.1 and the mask 255.0.0.0, we have a classless /8 address that's the same as the class "A" address 10.0.0.1.

Given 10.0.0.1 and the mask 255.255.0.0, we have a classless /16 address. This address with mask can not be classful because its network/host division is different from a class "A".

Given 10.0.0.1 and the mask 255.255.255.0, we have a classless /24 address. This address with mask can not be classful because its network/host division is different from a class "A".

Given just the class "B" address 128.0.0.1, classful addressing rules tell us the mask is 255.255.0.0.

Given 128.0.0.1 and the mask 255.0.0.0, we have a classless /8 address. This address with mask can not be classful because its network/host division is different from a class "B".

Given 128.0.0.1 and the mask 255.255.0.0, we have a classless /16 address that's the same as the class "B" address 128.0.0.1.

Given 128.0.0.1 and the mask 255.255.255.0, we have a classless /24 address. This address with mask can not be classful because its network/host division is different from a class "B".

Although classful addresses were nice in that they implied their mask, one might want to treat their network addresses differently, such as splitting the larger class A or B network addresses into multiple networks; subnetting. Or you might want to aggregate several true class C contiguous network addresses into a supernet. I.e. supernetting or subnetting was using IP addresses in a way that didn't match classful rules while using classful addresses.

New Member

Re: CCNA question...

Thanks rburts & josephdoherty for good explaination.

Tahir

Hall of Fame Super Silver

Re: CCNA question...

The explanation given by Joseph is interesting but in some parts it is inaccurate. It discusses the classful address as if it were defined by the natural mask (A=8, B=16, C=24) and implies that classful networks do not support subnetting. This is not true. With Classful networks you can subnet but all subnets within that network must be the same size.

Then Joseph says:

Given 10.0.0.1 and the mask 255.255.0.0, we have a classless /16 address. This address with mask can not be classful because its network/host division is different from a class "A".

This also represents a mis-statement. It is quite possible to build a very classful network with 10.0.0.1 255.255.0.0

I will say again: an IP address (especially an address taken by itself) is neither classful or classless. What determines the classful nature or the classless nature is the context within which it is used.

To go back to a question that Tahir raised: VLSM is one of the characteristics that differentiates classful and classless (VLSM is used in classless and classful can not vary the size of the subnet). But VLSM is not the only differentiator. Other aspects that determine classful or classless include contiguous/discontiguous subnets (classful requires contiguous while classless supports discontiguous) and summarization (classful does summarization automatically at network boundaries while classless may summarize manually and may summarize at points other than network boundaries).

HTH

Rick

Super Bronze

Re: CCNA question...

Rick, I've carefully reviewed your comments and I'm unable to see the inaccuracies, in this instance, although I'm always ready to concede its not the best possible explanation.

You begin with:

"The explanation given by Joseph is interesting but in some parts it is inaccurate. It discusses the classful address as if it were defined by the natural mask (A=8, B=16, C=24) and implies that classful networks do not support subnetting. This is not true."

I do assume when we're discussing classful IP addresses, they're as defined by the "Addressing" section of RFC791. RFC defines classes A/B/C, as in:

High Order Bits Format Class

--------------- ------------------------------- -----

0 7 bits of net, 24 bits of host a

10 14 bits of net, 16 bits of host b

110 21 bits of net, 8 bits of host c

my third and fourth paragraphs seem to agree with the RFC. Could you elaborate why this is not true?

As to ". . . implies that classful networks do not support subnetting." above, my last paragraph explicitly mentions you can subnet classful addresses. Unclear how I implied otherwise.

You continue with:

"Then Joseph says:

Given 10.0.0.1 and the mask 255.255.0.0, we have a classless /16 address. This address with mask can not be classful because its network/host division is different from a class "A".

This also represents a mis-statement. It is quite possible to build a very classful network with 10.0.0.1 255.255.0.0"

I didn't write you're unable to build a network using 10.0.0.1 255.255.0.0 or use it within a classful network. What I was trying to make clear is 10.0.0.1 with the mask 255.255.0.0 could not be represented as the classful address 10.0.0.1 which implies the mask 255.0.0.0. I had hoped the series of examples would help make clear where a classful address with its implied mask corresponds with a classless address and its mask, or not.

You write "I will say again: an IP address (especially an address taken by itself) is neither classful or classless." That's true, it's just a 32 bit number, but only when there's no context at all. When you add the context of classful interpretation, the IP address implies a network/host division. Knowing this, from the original post:

"say there are two ip addresses as follows

1. 10.0.0.1 255.0.0.0

2. 10.0.0.1 255.255.0.0

Which address (above) is classful and which is classless"

Address 1 with its mask can be represented as classful, address 2 with its mask can not.

Hall of Fame Super Silver

Re: CCNA question...

Joseph

Can you explain what you meant when you said this:

Given 10.0.0.1 and the mask 255.255.0.0, we have a classless /16 address. This address with mask can not be classful because its network/host division is different from a class "A".

It seems pretty clear to me that you explicitly say that 10.0.0.1 255.255.0.0 can not be classful. I maintain that that address can indeed be used as a classful address.

You seem to equate classful with using the natural mask. I do not agree that the criteria for classful equates to using the natural mask.

HTH

Rick

Super Bronze

Re: CCNA question...

Rick,

You write:

"Can you explain what you meant when you said this:

Given 10.0.0.1 and the mask 255.255.0.0, we have a classless /16 address. This address with mask can not be classful because its network/host division is different from a class "A". "

I thought I had explained in my prior post.

You write:

"It seems pretty clear to me that you explicitly say that 10.0.0.1 255.255.0.0 can not be classful. I maintain that address can indeed be used as a classful address."

Again, I thought my prior post explained this, but when you emphasis just one statement, taken out of context, one might think that, as you write, "10.0.0.1 255.255.0.0 can not be classful" could be taken to mean such an IP address and mask can not be used within a classful network. Such an interpretation was not intended, and I had hoped, within full context of my post, would not have been. Again, I agree that 10.0.0.1 255.255.0.0 can be used within a classful network.

You also write:

"You seem to equate classful with using the natural mask. I do not agree that the criteria for classful equates to using the natural mask."

Returning to the original posted question:

"1. 10.0.0.1 255.0.0.0

2. 10.0.0.1 255.255.0.0

Which address (above) is classful and which is classless."

I contend since #1 can be provided by the IP address 10.0.0.1, alone, using classful rules, while #2 can not, one could call #1 classful and #2 classless. Here, it seems, we differ. I'm assuming you believe the correct answer is none of the above.

Perhaps the confusion is based on I was trying to simply show whether certain networks could be represented with just an IP address using classful rules, which they can. I did not touch upon all the implications of using classful vs. classless network addressing, such as all the issues of discontinuous subnets that you broach. Of course, discontinuous subnets are the way they are because of classful network addressing which explains why a subnetted class A like 10.0.0.1 255.255.0.0 is different from a class B 128.0.0.1 255.255.0.0, yet are similar when both are subnetted like 10.0.0.1 255.255.255.0 and 128.0.0.1 255.255.255.0 but different from the class C 192.0.0.1 255.255.255.0.

Hall of Fame Super Blue

Re: CCNA question...

Hi Rick / Joseph

Very interesting conversation, hope you don't mind me joining in.

"It seems pretty clear to me that you explicitly say that 10.0.0.1 255.255.0.0 can not be classful. I maintain that that address can indeed be used as a classful address. "

I would agree and disagree with this and myabe that's part of the problem.

As an IP address/subnet mask i can't see how 10.0.0.1 255.255.0.0 can be called classful. A classful 10.x.x.x network must have a subnet mask of 255.0.0.0 because the leading bit in the first octet is 0 and this makes it a class A address.

However a classful routing protocol simply means the routing protocol cannot handle different subnet masks for the same major network ie. it applies the same subnet mask to all versions of the major net so

10.5.1.0 255.255.0.0

10.5.6.0 255.255.0.0

10.5.7.0 255.255.0.0

are all 10.5.0.0/16 to a classful routing protocol.

So i would say that the criteria for classful with IP addressing does equate to the natural mask, but the criteria for classful with a routing protocol does not.

Apologies if i have stepped on any one's toes, just thought it was an interesting thread

Jon

Super Bronze

Re: CCNA question...

Jon,

Nice post!

Perhaps a better explanation of what I was trying to convey.

Thanks.

Hall of Fame Super Silver

Re: CCNA question...

If you equate classful as using the "natural" subnet mask then Joseph is correct that 10.0.0.1 255.0.0.0 is classful and 10.0.0.1 255.255.0.0 is not. I do not find that to be the way that most people use the terms classful and classless. And it leads to statements that say that 10.0.0.1 255.255.0.0 is a classless address which can be used in a classful network. Perhaps we should ask Tahir to clarify the context of his question.

HTH

Rick

Hall of Fame Super Blue

Re: CCNA question...

Rick

Still think that the terms are confusing ie.

"10.0.0.1 255.255.0.0 is a classless address which can be used in a classful network"

Again i think this is mixing up 2 things.

If i was told that a network was using classful addressing i would take that to mean it was using

Class A addressing with a 255.0.0.0 subnet mask

Class B addressing with a 255.255.0.0 subnet

mask

Class C addressing with a 255.255.255.0 subnet.

So when we talk about classful networks it is very important to specify whether we mean

1) classful in terms of addressing

2) classful in terms of the routing protocol.

If we are referring to the routing protocl then this implies nothing about the length of the subnet mask in relation to Class A, B, C but it does say that the same network mask will be used for the same major network. So any combination of major net/subnet mask can be used in a classful network as far as the routing protocol is concerned.

I guess it depends on experience but most people i have worked with understand classful addressing to be Class A, B, C with their natural masks.

HTH

Jon

Hall of Fame Super Blue

Re: CCNA question...

Tahir

Apologies if i have hijacked your thread, hopefully you find this just as interesting as we seem to.

Jon

New Member

Re: CCNA question...

Hi Jon, No apologies please, everyone has the right to contribut into the forum, a very informative discussion is here that we new network persons cannot find elsewhere. I thanks for everyone who contributed into and hopefully will continue.

Regards,Tahir

New Member

Re: CCNA question...

Hi seniors, uptill yet i concluded from the discussion that 1.10.0.0.1 255.0.0.0 is classful whereas 2.10.0.0.1 255.255.0.0 is classless. I got this question during my study for ccna, and writer say second address is classless very reson that it is not on natural boundary as IOS will think. But my seniors say "it depends on how that address is being used, I mean then how IOS will know what situation is there, if anythink is defined in IOS then I am not aware of. Please guide me. Thnaks again to all seniors who participated in it and hopefully will further clear it.

Hall of Fame Super Blue

Re: CCNA question...

Tahir

" But my seniors say "it depends on how that address is being used"

This is the point we have been discussing really and it is obvious that different people have different views so perhaps there is no definitive answer.

I would argue that when we talk about how the address is being used then we are really talking about classful vs classless routing protocols ie. RIP/IGRP = classful, OSPF/EIGRP/RIPv2 = classless.

But i still fundamentally agree with Joseph in that the address itself can be talked about as to whether it is classless or classful based on whether it uses the natural mask ie.

Class A = 255.0.0.0

Class B = 255.255.0.0

Class C = 255.255.255.0

HTH

Jon

Super Bronze

Re: CCNA question...

Hope this doesn't muddy the waters further, but since you also asked about concepts, wanted to touch on a few points.

For "book" answers, as you initially posted, they're likely looking for whether the given address and mask can be represented by a classful A/B/C IP address. Classful rules define certain masks for certain address ranges and are known as "natural" masks; natural because they are implied.

As Jon notes in his latest reply, some routing protocol are classful, e.g. RIPv1/IGRP, and some are classless, e.g. RIPv2/OSPF/EIGRP. What makes them different is the former only has passes along addresses, the latter passes along addresses and their masks. Because of this difference, how network topologies are designed for classful vs. classless addressing may differ. Classless can mimic a classful design, but can do more. The rules for using "natural" IP addresses differ from using "non-natural" IP addresses within a classful network. Both, though, can be used and usually were. (E.g. point-to-point links with /30 addressing.)

When thinking of classful vs. classless IP addressing then, we're really thinking within the framework of network topology and specific routing protocols. (As also noted by Jon.)

Beyond dealing with classful vs classless, and dealing with just IP addresses and masks, they don't have to conform to either. Two examples come to mind, first a host's perspective with a interface IP address and mask. Host's perspective isn't classful or classless, it's just and IP address and mask. (Perhaps one of the things Rick had in mind.) Second, we haven't touched upon static routing. Here too, you just see IP address and mask, it's left to you how to utilize them.

353
Views
23
Helpful
18
Replies