Dont understand subnetting

[gxsoul]

is this correct cause when i use a subnet calculator it gives a different answer

 

[coxhaus]

A class A IP address space is 16,000,000 not 65,000. A class B IP address space is around 64,000 maybe 65,000. I don't remember the exact number as it has been a lot of years.

All networks are N-2 for the number of hosts. The first IP address is the network and the last is the broadcast. You do the math for all that other crap. I am not.

 

[Tech Junky]

What is an RFC1918 Address?

An RFC1918 address is an IP address that is assigned by an enterprise to an internal host without requiring coordination with an Internet registry. Learn more about them in this blog post!

netbeez.net

CIDR Conversion Table

CIDR notation accomplishes the same task as network subnet masking. CIDR defines which IP addresses, and how many host IP addresses will be allowed on the network. The following table...

techlibrary.hpe.com

The only time IRL you do this by longhand is for a class or exam. Trust the calculator but verify it works on the equipment.

 

[sfx2000]

A class A IP address space is 16,000,000 not 65,000. A class B IP address space is around 64,000 maybe 65,000. I don't remember the exact number as it has been a lot of years.

His /16 is class A, but it's 65K hosts...

$ ipcalc 10.0.0.0/16
Address:   10.0.0.0             00001010.00000000. 00000000.00000000
Netmask:   255.255.0.0 = 16     11111111.11111111. 00000000.00000000
Wildcard:  0.0.255.255          00000000.00000000. 11111111.11111111
=>
Network:   10.0.0.0/16          00001010.00000000. 00000000.00000000
HostMin:   10.0.0.1             00001010.00000000. 00000000.00000001
HostMax:   10.0.255.254         00001010.00000000. 11111111.11111110
Broadcast: 10.0.255.255         00001010.00000000. 11111111.11111111
Hosts/Net: 65534                 Class A, Private Internet

It's still class A however...

Are you thinking of the /8?

$ ipcalc 10.0.0.0/8
Address:   10.0.0.0             00001010. 00000000.00000000.00000000
Netmask:   255.0.0.0 = 8        11111111. 00000000.00000000.00000000
Wildcard:  0.255.255.255        00000000. 11111111.11111111.11111111
=>
Network:   10.0.0.0/8           00001010. 00000000.00000000.00000000
HostMin:   10.0.0.1             00001010. 00000000.00000000.00000001
HostMax:   10.255.255.254       00001010. 11111111.11111111.11111110
Broadcast: 10.255.255.255       00001010. 11111111.11111111.11111111
Hosts/Net: 16777214              Class A, Private Internet

 

[coxhaus]

Yes. /8 is class A. /16 is class B. The mask defines the address space. He has an "A" for class but he is using /16.

If you have a class C which is part of a larger class A you don't call it a class A just because it's part of a larger class A.

And there is no problem having overlapping subnets. It is done all the time. I think they call it superscoping. You just have to be careful with broadcast IP addresses and not use them in another network. You can superscope over them but it is a good idea to avoid using them.

 

[dosborne]

How much do we get paid for doing someone's homework when google would have answered? LOL

 

[Tech Junky]

@coxhaus

The mask has nothing to do with rfc1918 classes it has to do with the amount of hosts you want in the address range. You can have a /30 in a class A for a point to point or a /24 for an office. Same goes for class B/C networks.

When doing this for an exam you get the basics not the IRL version that extends into real world scenarios.

 

[coxhaus]

Yes every network is a part of a class A. Private IP addresses are defined. Some are full class A and some are not.

 

[Tech Junky]

Yes every network is a part of a class A. Private IP addresses are defined. Some are full class A and some are not.

Do you have a point? Public / Private either way you can spilt the CIDR into other segments as you desire as long as there's no overlap.

I use /32's for my loopbacks / servers on my Linux box instead of dedicating physical interfaces. If you have the space in a subnet you can break it down extensively and with Cisco + subnet 0 you gain more options.

The OP question / exam capture is more related to the fundamentals and how they're dumbed down to the extreme basics to get used to the math associated with counting by 8's. When you start out as a n00b you don't get into the gritty and messy production tricks you use to make things work.

Simply they put it into:
255.0.0.0 = A 10.x.x.x /8
255.255.0.0 = B 172.16-31.x.x /12
255.255.255.0 = C 192.168.0.0/16

In reality though you're not confined to those at all within the specified CIDR other than you can't have a bigger than assigned scope. If you have a /16 you can't provision a /15 to double your network hosts. However you can split a /24 into several /30's or /29 or /28 and so on.

With the likes of Google / Level3 / Facebook / ATT / VZW etc. having the bigger scopes that could get them into large swaths of IP's to use there's options to use A class routing w/o worry of exhaustion. With some tricks like MPLS you can setup things that are more like a VLAN but, routable / tagged / private from internet snooping. For instance one network that's popular uses a 5.x.x.x MPLS scheme for nationwide routing. When you go to lookup the IP though you get some little ISP in Turkey. DOD uses some MPLS trickery as well to route things in a similar fashion. Throwing things into a tagged MPLS environment allows you to use IP's more liberally because they typically only need a couple of IP's on either end that are public to exit the network to the outside world. ATT business routing does things with 10.x.x.x or 12.x.x.x

 

[sfx2000]

Yes. /8 is class A. /16 is class B. The mask defines the address space. He has an "A" for class but he is using /16.

He has a class A network not because it's a /8 or a /16 - that has no bearing - one can have a class A network with a /24 which is only 254 hosts..

It's relevant as he's working with private 10dot's - and the RFC1918 suggests as much.

ipcalc 10.0.0.0/24
Address:   10.0.0.0             00001010.00000000.00000000. 00000000
Netmask:   255.255.255.0 = 24   11111111.11111111.11111111. 00000000
Wildcard:  0.0.0.255            00000000.00000000.00000000. 11111111
=>
Network:   10.0.0.0/24          00001010.00000000.00000000. 00000000
HostMin:   10.0.0.1             00001010.00000000.00000000. 00000001
HostMax:   10.0.0.254           00001010.00000000.00000000. 11111110
Broadcast: 10.0.0.255           00001010.00000000.00000000. 11111111
Hosts/Net: 254                   Class A, Private Internet

He has a class A network because the leading bit is "0" - review the bitwise representation I show above.

In IPV4 classful networks, the leading bit/bits define what class it is...

Class A - 0
Class B - 10
Class C - 110
Class D - 1110 (which is multicast)

@Tech Junky and @coxhaus - check and review your previous statements.

Putting my developer hat back on the shelf now...

 

[sfx2000]

With the likes of Google / Level3 / Facebook / ATT / VZW etc. having the bigger scopes that could get them into large swaths of IP's to use there's options to use A class routing w/o worry of exhaustion. With some tricks like MPLS you can setup things that are more like a VLAN but, routable / tagged / private from internet snooping. For instance one network that's popular uses a 5.x.x.x MPLS scheme for nationwide routing. When you go to lookup the IP though you get some little ISP in Turkey. DOD uses some MPLS trickery as well to route things in a similar fashion. Throwing things into a tagged MPLS environment allows you to use IP's more liberally because they typically only need a couple of IP's on either end that are public to exit the network to the outside world. ATT business routing does things with 10.x.x.x or 12.x.x.x

MPLS doesn't use addresses, it uses labels - MPLS is essentially Layer 2 with some Layer 3 added...

Best way to consider MPLS is VPN on steroids - that's all most people need to know - it's a virtual wire

Telco carriers - their internal networking is largely abstracted from the public, but internally, they use a lot of RFC1918 ranges internally - and depending on last mile, for example, in wireless operators, one is usually nat'ed in a 10dot range, because (a) it doesn't matter, and (b) easier to manage with the P-GW/SGSN-GGSN networks across MME's

Carriers do use MPLS internally across markets and sometimes from the core out to the edge, or from core to core, esp with the majors where thru mergers/acquisitions, can and often due have multiple cores.

Should also mention that MPLS is used with partners (for example ATT roaming into Rogers - that's over MPLS links)

(note - edited for clarity)

 

[Ranger802004]

MPLS doesn't use addresses, it uses labels - MPLS is essentially Layer 2 with some Layer 3 added...

Best way to consider MPLS is VPN on steroids - that's all most people need to know - it's a virtual wire

Telco carriers - their internal networking is largely abstracted from the public, but internally, they use a lot of RFC1918 ranges internally - and depending on last mile, for example, in wireless operators, one is usually nat'ed in a 10dot range, because (a) it doesn't matter, and (b) easier to manage with the P-GW/SGSN-GGSN networks across MME's

Carriers do use MPLS internally across markets and sometimes from the core out to the edge, or from core to core, esp with the majors where thru mergers/acquisitions, can and often due have multiple cores.

Should also mention that MPLS is used with partners (for example ATT roaming into Rogers - that's over MPLS links)

(note - edited for clarity)

When I was Network Engineer for an Enterprise environment, we had nothing but MPLS at all of our sites and Datacenters...definitely worth the money and a lot less packet overhead compared to VPN Tunnels.

 

[sfx2000]

definitely worth the money and a lot less packet overhead compared to VPN Tunnels.

The downside of MPLS is that one pays for the bandwidth whether it is used or not - it's a high quality link with defined class/quality of service, but if sized incorrectly, it can be very expensive indeed...

That's where SD-WAN can come into play - there's tradeoffs of course, but properly architected, one can use both - MPLS as primary, and SD-WAN for yield on demand.

 

[Ranger802004]

The downside of MPLS is that one pays for the bandwidth whether it is used or not - it's a high quality link with defined class/quality of service, but if sized incorrectly, it can be very expensive indeed...

That's where SD-WAN can come into play - there's tradeoffs of course, but properly architected, one can use both - MPLS as primary, and SD-WAN for yield on demand.

That is true, we definitely used ours with backend email flow, call routing, DC to DC traffic, etc.

 

[coxhaus]

The downside of MPLS is that one pays for the bandwidth whether it is used or not - it's a high quality link with defined class/quality of service, but if sized incorrectly, it can be very expensive indeed...

That's where SD-WAN can come into play - there's tradeoffs of course, but properly architected, one can use both - MPLS as primary, and SD-WAN for yield on demand.

When I retired we were heading to MPLS. We had the Cisco hardware in place. It was a matter of a transition plan from A to B on how we were going to get there. I left it for the new guy taking my place. I am thinking the video conferencing would also have been converted over from ATM using our Cisco LS1 switch to ethernet.

 

[sfx2000]

With the telco's pushing to kill off TDM lines, MPLS has shown to be the replacement for ATM in many cases... in those use cases, MPLS has all the properties of both internet and ATM.

Oddly enough, one can run MPLS over ATM...

 

[coxhaus]

With the telco's pushing to kill off TDM lines, MPLS has shown to be the replacement for ATM in many cases... in those use cases, MPLS has all the properties of both internet and ATM.

Oddly enough, one can run MPLS over ATM...

ATM is ok if you can live in those small cell sizes. It is kind of old technology. ATM has been around a long time. It was what you used a long time ago for video conferencing.

 

[GHammer]

So, past the MPLS/ATM/CDMA/HALO/FM/AM/BBC/IPLS, what's the answer to the original question?

 

[ColinTaylor]

So, past the MPLS/ATM/CDMA/HALO/FM/AM/BBC/IPLS, what's the answer to the original question?

The answer is..

a) The OP doesn't care because after making the post he never returned.
b) Yes, the "answers" written in the OP are correct.

Post #10 is probably the best explanation as to why the answer as written is correct.