I was actually aware of myself asking multiple questions. I just thought it would be a good idea to elaborate on what I'm confused about pertaining subnetting, since I wasn't sure myself what the cause of my confusion was (unlike now, thanks to you! :).

Exactly what's the point of ip route outputting the classful boundary, when classful networking is obsolete? Do they just show this in case we're still using RIPv1 or IGRP? In other words, is there even a point in knowing the classful boundary?

Also, does show ip route always output the classful boundary of the subnets? Or will it differ on circumstances?

To your first comment: it's really just holdover from the classful routing days. They still show it because that's pretty much how it's always been done. Although I doubt it to be the reasoning for still displaying it this way, it wouldn't surprise me if there were deployments out there that still used classful addressing and routing. And yes, knowing the classful boundary can be useful in certain situations, although they are somewhat rare these days.

To your second comment: Yes. The major/classful network will always be shown as the "parent network route" for various subnets. Whether its natural (classful) mask is shown or not will depend on the length of its child subnets.

Also don't forget to mark the answer as accepted if it you feel it is such.

I'm confused about excercise 1 on 9tut.com/subnetting-tutorial/2. When I subnet I'm used to be given a network address like 4.0.0.0/24 and I would borrow host bits to create subnets. But in excercise 1 they figure out the major network boundary 4.0.0.0/8 and assume that all other class A networks are subnets to that major network boundary. Can you confirm this is correct? It confuses me because you told me that the classful boundary is rarely used these days, so I get the idea that this might be an old/obsolete way to think about subnetting.

To make my question a bit clearer, I'm going to give an example. Take a look back at the picture in my original question. The IP routing table says that 172.16.0.0/24 is subnetted and has 2 subnets. It seems the IP routing table also treats 172.16.1.0/24 and 172.16.2.0/24 as subnets because they are in the range of the major classful network 172.16.0.0/16 (i.e. 172.16.0.0 - 172.16.255.255). So this is probably how the routing table looks at subnets, correct?

One last thing: if I subnet a network address, I can always get this network back by supernetting the subnets, correct? This is then the answer to my last question my original post.

I've moved this to chat to avoid extended discussions in the comments.

I've also edited my answer to try to answer some of your questions

I just noticed

Let me read the edit!

Sure thing

What I meant by this question "One last thing: if I subnet a network address, I can always get this network back by supernetting the subnets, correct? This is then the answer to my last question my original post." was that, lets say I have a network 213.105.107.0/24 and I were to subnet this

into 213.105.107.0/25 and 213.105.107.128/25

if I now supernet these two subnets

will I always get the original network 213.105.107.0/24 back?

I tried it doing it just a sec ago, and it seemed so

but I can't prove if it works in all cases

so my question doesnt have anything to do with a classful network

All right, but that's a different question than the last question you asked in your original post

ah sorry

> Also, one thing I don't understand is how the router knows the original network (in this case 172.16.0.0/24) when only the subnets are given. How does it figure this out?

The thing that's throwing you off is how the router is presenting the "original network"

yeah the "(in this case 172.16.0.0)" should be omitted

that's what used to throw me off yeah

I understand the ip routing table shows the classful network now

and it will combine it with the CIDR of both the subnets given that they are equal

or otherwise it will show the natural mask

Before reading your answer I never knew that the ip routing tables show the classful networks, so I thought that the classful network they used to show was the "original network" that gets subnetted

Well, there's not really any "combining" - it's just showing you the prefix length that all subnets of the major network have. So say you're using 10.0.0.0/8 for a large network, and you have 300 subnets of 10.0.0.0/8, and all of them are /24's, the table would say "10.0.0.0/24 is subnetted, 300 subnets"

yeah exactly

I understand that now, thanks to you :)

Right... so as long as you combine enough, you should get the original network back

Subnetting is like dividing a pie up into equal parts

the key word being equal

That sounds like it isn't always the case?

that's all it is

I mean it's kind of arbitrary

So there's no way to get back the original network in all cases?

Once you get 213.105.107.0/24 and then combine it with 213.105.106.0/24 and then get 213.105.106.0/23

I thought supernetting was a way to get it back

it's just counting. i'm not sure where the desire to "get the original network back" comes from

I used to think it was possible because of my former confusion with the ip routing table, but now I know it's just the classful network

I was just wondering if it's possible to reverse engineer back

so that when you are given a network topology with different subnets

that u can figure out the original network

It's just out of curiousity now

if ur saying that's not important then I won't bother :)

I still find that excercise I was talking about in the comments confusing though.

Are there like, 2 ways of subnetting? A classful way and a classless way?

yeah it's not really that important

Subnetting originated with classful addressing and routing

because of CIDR it's no longer relevant

So in my response, what I meant was, the exercise is essentially forcing you to subnet 4.0.0.0/8

When in reality today, you could be allocated 4.0.0.0/20 and you would still be able to turn that into /24 subnets

there's not really such thing as an "original" network

it's either an aggregate or not

For example 192.168.1.0/24 would be an aggregate of 192.168.1.0/25 and 192.168.1.128/25

192.168.0.0/22 would be an aggregate of 192.168.0.0/24 - 192.168.1.3.0/24

Does that make sense?

I understand how to calculate supernets and subnets, but I don't understand how there's no original network

When you subnet a network, that network is the original network

In theory the "original network" is 0.0.0.0/0

that's my definition

all networks are technically subnets of that network

it's a hierarchy

I understand that yeah

If you're given an allocation, say 4.0.0.0/8 per the exercise, and you're using that as your starting point, then yes, theoretically you could supernet all the way back up your subnet chain to get that

if you can divide, then you can combine

but you're saying there's no point to do that?

in doing that*

why would there be? You're already going to know the allocation ahead of time...

Well you could be put into a network that someone else built

and he already subnetted the allocation

then it might be interesting to know the allocation

Okay, sure, but in basically all cases networks are addressed out of RFC1918 space. And the network ranges for those are already well-known in the industry

If you're referring to public allocations for things like internet access, those are trivial to obtain from either your router or documentation from the carrier

so if you were put into such a scenario, and there's no documentation, how would you figure out the initial allocation?

if we're still talking about publicly routable IP addresses, then my answer is trace where the connection to the outside world terminates, then "show run" on whatever piece of equipment that is

It means the question is forcing you to subnet based on "Class A"

ah that's clever :P

and you mean in a real life scenario you're never given a class a?

In a real life scenario, Classes aren't even used anymore

it's just allocations

Ah, so that's why u said the question was obsolete

Yes

Tell me how would you subnet 4.0.0.0/20 into /24's

well we're given 4 bits for subnetting

so there's 2^4 subnets

and 32 - 24 = 8 host bits

so 2^8 hosts

-2

so that would be

But I understand the analogy between this and that excercise now

Cool

I just find it weird that a 2014 post still explains about classful subnetting

:P

I want to thank you massively for your time dude

I hope I wasn't a burden :)

So back in the days the subnetting was always done with respect to the classful boundary?

Yep

Yeah I'm not sure why curriculums still insist on teaching classful things

Yeah it wasted alot of my time :/

quite annoying

It's quite a pain to figure out the difference between classful and classless subnetting unless you experienced how it evolved

CCNA is quite a pain itsself though lol

Well, once you get the hang of classless it will all click

I'm also a victim of confusion because of being forced to learn about classful first

so you're not alone :-)

I think most people are

untill cisco changes its curriculum

But I still wonder about how classful subnetting works

if I understand correctly

There's a major network/boundary and that is subnetted using the subnet mask, dividing it into smaller subnets?

I feel like I phrased that incorrectly

There's a major network/boundary like 4.0.0.0/8, and then it's divided into smaller subnets*

And then there's another one like 3.0.0.0/8

that is also divided into smaller subnets

And the smaller subnets are represented by a different cidr

Yeah, I think I get it now.

Well 4.0.0.0/8 is a major network boundary

3.0.0.0/8 would be distinct

Classful subnetting essentially means that there's no other way to subnet networks other than using a natural mask

so your options for classful subnetting are either /8, /16 or /24

They are both major network boundary's right?

yes, totally separate and distinct

All the way to 255

Correct

Taking this down many levels, say for example you're a company in a world where classful routing is still used

you've been allocated 192.168.0.0/24 through 192.168.0.3/24

First things first, you wouldn't be able to effectively turn those /24's into smaller subnets

Because there's not enough host bits or?

Right

But you could supernet them

Nope, couldn't do that either, because you're responsible for only those 4 networks

let's change it because it's RFC1918

so instead let's say you've been allocated 200.200.0.0/24 through 200.200.3.0/24

you have 254 * 4 usable host addresses right?

yh

since you can't make a /24 any smaller, you need to use RFC1918 space within your network to save your public IP addresses

Quite insane how you know the RFCs by heart

so on point to point serial links between your HQ and branch offices, you need to use a /24 for those

:P

lol i don't

RFC1918 is ubiquitous, everyone knows that one

What does it stand for?

I'd have to search it up

anyway do you see that using a whole /24 for a "network" that will only ever have two IP addresses is wasteful?

Yeah of course

VLSM isn't possible here?

Great. So, now going back to your public allocations. When you're announcing these allocations to your internet provider, you are NOT able to aggregate ("supernet") them into 200.200.0.0/22

Ah right

so you'd have to announce each individual /24 to the provider separately

So basically back in the days you could only have /24 /16 and /8

but it was possible to subnet a class A into /24s?

VLSM is not possible because VLSM means "variable length subnet mask"

Right, that's all you had to work with

ahh

It was possible to subnet a Class A into /24's

and into /16s rightr?

Sure

I see

But if I understand correctly, back then there was no CIDR

Right

so how could you give 4.0.0.0/24 to someone

when it all depends on the first octet

4.0.0.0/24 is still a subnet of 4.0.0.0/8

Yeah, but I mean, how could you differ between the two

when you don't have a subnet mask

The class of an address is what depends on the first octet

not the subnet mask

Well lets say you divide 4.0.0.0/8 into 4.0.0.0/24, 4.0.1.0/24 etc

Right

When you look at the first octet of 4.0.1.0 it's still the same

as 4.0.0.0

so they are both class A

so how would you know they're different /24s

or do we assume that all class A are divided into /16s and then the /16s are divided into /24s?

It's up to the engineer on how they want to divide it up. Just because an address is class A does not mean it must be a /8

I don't get that

hmm

the natural mask is 8

so if the engineer decides he wants a 4.0.1.0/24

how would he do that without a subnet mask

4.0.0.0/8 would be the major network boundary; 4.0.1.0/24 would be a subnet of 4.0.0.0/8

Yeah I understand that

but say ur a company, and you're allocated 4.0.1.0/24

Right

how can u be given that allocation without a CIDR

connected routers would just see 4.0.1.0 right?

in the routing table

hmm I think I get it

Nearby routers don't need to know the subnet mask

They wouldn't have had to, that was one of the key points of classful routing

They would just automatically assume that 4.0.1.0 - 4.0.1.255 would be under that network

the network and host portions were determined by looking at the address alone

not the subnet mask

yeah

now, you can still subnet classfully, but classless routing happened when routing protocols were developed to send the subnet mask along with the IP address in routing updates

they also didn't have NAT back then right?

I doubt it

And yes, you're right

the routers make assumptions about what other routers know

ahh

hence why you can still subnet 4.0.1.0/24 and 4.0.2.0/24 on one router, but other adjacent routers will believe that the 4.x.x.x router knows about 4.0.0.0/8

does that make sense?

yeah cause if nearby routers don't know the subnet mask of the router, how would they know that 4.0.1.1 would be inside 4.0.1.0/24

Right. It would be confusing for a router to be "in between" two other routers, and those two routers on either side are advertising subnets of 10.0.1.0/24 and 10.0.0.0/24

The router "in the middle" would believe that either router knows about 10.0.0.0/8

ahh yeah that's what was confusing me

I also read that in that scenario load balancing occurs

So basically it just sends all packets destined for a 10.0.0.1 address to any nearby 10.x.x.x?

Not sure, would be neat to find out though

by load balancing I mean that it would send half to one router and the other half of the data to the other router

it splits it up because it doesn't know what to do with it

this is straight from my CCNA course: "Classful routing protocols also create problems in discontiguous networks. A discontiguous network is when subnets from the same classful major network address are separated by a different classful network address. "

and then it illustrates this with some pictures

Right

it just tells to use classless routing protocols from that point

to avoid these problems

because it doesn't know what to do with the packets

Makes sense

Yep

Hence the issues with classful addressing and routing :-)

Once you complete any tests or homework that forces you to remember anything classful, do yourself a favor and forget it

Yeah but the dumb thing IMO is that they teach 50% of it

and then expect me to understand why classful routing protocols create problems in discontiguous networks

Right

even though they never really explained classful subnetting etc properly

but then when u ignore those things u find urself very confused later on when they continue to dive deeper into such concepts

Sure

the same thing happened with the question I originally asked

they never said anything about the major network boundary or classful boundary

but they do tell me to use the ip routing table

and then I just wonder myself how other people in my course are going to learn all this

I guess most would just ignore it and continue

but I feel like that's not very clever since the basics are essential before you can move on

indeed

Anyways, thanks for your time dude! I really appreciate it

I hope I'll reach your level of proficiency one day :)

I can't think of any more questions so I feel like I understand it now

Sure. My email address is [email protected] if you can think if any more

you prob never use skype? :P

just for chatting I mean

not with webcam and everything

I do but I'm rarely on. jajensen83

ahh

I'll be glad to respond though if you catch me while I am on :-)

I will add you anyways :) I'll definitely be sending you questions every now and then

Sounds good!

in case I get confused

I've got to go now mate!

Cool, take care

talk soon

Cya

yea :D