Nerds, subnetting?

128 | 64 | 32 | 16 | 8 | 4 | 2 | 1

10.22.183.14/16 (CIDR)

10.22.183.14 / 255.255.0.0 (MASK)

to binary:

00001010.00010110.10110111.00001110 (ANDing)
11111111.11111111.00000000.00000000

00001010.00010110.00000000.00000000

to decimal:

10.22.0.0 = Subnet ID

16 bit subnet = 128ths place?

So, is this correct?

10.22.0.0 - 10.22.0.127 (N = 0 / H = 1-126 / B = 127)
10.22.0.128 - 10.22.0.255 (N = 128 / H = 129-254 / B = 255)
10.22.1.0 - 10.22.1.127 (N = 0 / H = 1-126 / B = 127)
10.22.1.128 - 10.22.1.255 (N = 128 / H = 129-254 / B = 255)
10.22.2.0 - 10.22.2.127 (N = 0 / H = 1-126 / B = 127)
10.22.2.128 - 10.22.2.128 (N = 128 / H = 129-254 / B = 255)

and so on

Did I goof? If so, where?

Are you trying to subnet the larger /16 into smaller subnets that have 126 usables per subnet?

10.22.183.14/16 means that the network portion of the IP is 10.22 and any IP in the 10.22.0.0 - 10.22.255.255 range is considered to be in that subnet.

If you want to break it down farther you will have to increase the size of the mask.

Source: I'm a CCIE

LonelyInAtl is a much better source than I am, but I briefly looked over your issue as well, and the way you're doing the subnetting suggests that you're actually calculating with a /25 mask rather than a /16 mask.

/16 would have the first subnet be...

10.22.0.0-10.22.255.255, and with one bit taken away for the network and one for broadcast, leaves you with 65534 useable hosts. Your next subnet will begin at 10.23.0.0...

A /25 would be pretty much exactly what you want to create the configuration you have, with the binary mask being 11111111.11111111.11111111.10000000.

So your first subnet would be 10.22.0.0 to 10.22.0.128, with 126 useable hosts.

BTW, is your question that you have a 10.22.183.14/16, and want to find out the number of useable hosts, first and last host, network bit & broadcast address?

source: CCNA only, alas.

LonelyInAtl said:

Are you trying to subnet the larger /16 into smaller subnets that have 126 usables per subnet?

10.22.183.14/16 means that the network portion of the IP is 10.22 and any IP in the 10.22.0.0 - 10.22.255.255 range is considered to be in that subnet.

If you want to break it down farther you will have to increase the size of the mask.

Source: I'm a CCIE

Click to expand...

OK, thank you. That makes sense.

---

IgnoredOne said:

LonelyInAtl is a much better source than I am, but I briefly looked over your issue as well, and the way you're doing the subnetting suggests that you're actually calculating with a /25 mask rather than a /16 mask.

/16 would have the first subnet be...

10.22.0.0-10.22.255.255, and with one bit taken away for the network and one for broadcast, leaves you with 65534 useable hosts. Your next subnet will begin at 10.23.0.0...

A /25 would be pretty much exactly what you want to create the configuration you have, with the binary mask being 11111111.11111111.11111111.10000000.

So your first subnet would be 10.22.0.0 to 10.22.0.128, with 126 useable hosts.

BTW, is your question that you have a 10.22.183.14/16, and want to find out the number of useable hosts, first and last host, network bit & broadcast address?

source: CCNA only, alas.

Click to expand...

This was helpful as well. Actually I didn't have a particular question that needed to be solved. I was trying to recall the process of subnetting and apparently added in steps that weren't there.

So a more appropriate question may be, you are given an IP address of 10.22.183/16 and required to provide at least (picks random number) 20 subnets, how do you proceed?

You have a class A address there, so your ocet of change would be 0-255, you should have 256 total possible subnets, each with 65534(256 * 256 -2) possible hosts with a /16 mask. You can create 20 easily from that 256 range;

10.22.xx
10.23.xx
10.21.xx
etc...

Class A subnetting:

nnnnnnnn.oooooooo.hhhhhhhh.hhhhhhhh

Where n = network address(fixed at 10 here), o = ocet of change(0-255), and h = available ocets for host addresses.

I think you might have a VLSM question there? I'm not as familiar with that.

IgnoredOne said:

You have a class A address there, so your ocet of change would be 0-255, you should have 256 total possible subnets, each with 65534(256 * 256 -2) possible hosts with a /16 mask. You can create 20 easily from that 256 range;

10.22.xx
10.23.xx
10.21.xx
etc...

Class A subnetting:

nnnnnnnn.oooooooo.hhhhhhhh.hhhhhhhh

Where n = network address(fixed at 10 here), o = ocet of change(0-255), and h = available ocets for host addresses.

I think you might have a VLSM question there? I'm not as familiar with that.

Click to expand...

Thanks for the feedback. I'm just randomly plucking this stuff out of my psyche in an attempt to refresh myself. The thing about subnetting that always confuses me is that it is three-dimensional and can be approached from several different angles. Prepping for CompTIA Network + exam to be specific.

What book are you using? And enable PMs if you can.

One good rule of thumb using VLSM is that you always subnet the largest size network first and work your way down from there.

IgnoredOne said:

What book are you using? And enable PMs if you can.

Click to expand...

Truth be told, I haven't committed to a book purchase yet. I'm studying through online resources and manufacturer literature at this point.

---

LonelyInAtl said:

One good rule of thumb using VLSM is that you always subnet the largest size network first and work your way down from there.

Click to expand...

I need to read on VLSM. A quick google search seemed to reference it as something similar to CIDR? I do appreciate the feedback, was just curious who around these parts worked in IT. I'm going to head back to Cisco's site for a while and study.

Basically, CIDR is supernetting while VLSM allows you to use a different mask for each subnet. VLSM allows for more efficient use of IP space whereas CIDR is geared more towards reducing the size of routing tables by allowing aggregation of networks into one route advertisement.

After reading that, no wonder I can't get a date. I'm too much of a geek!

Nah, I don't think its geeky to know subnetting. Pretty sure that's an essential networking skill; geeky would be knowing how to configure BGP!

IgnoredOne said:

Nah, I don't think its geeky to know subnetting. Pretty sure that's an essential networking skill; geeky would be knowing how to configure BGP!

Click to expand...

Guilty as charged. BGP with route maps, metrics, prepend, loopbacks, etc. Also iBGP, eBGP, EIGRP, RIP, OSPF, ad nauseum.

Basic BGP is really easy. Just configure your local AS, remote AS, neighbor, and the networks you want to redistribute.

But it sure felt good getting that dual track CCIE!

LonelyInAtl said:

Basically, CIDR is supernetting while VLSM allows you to use a different mask for each subnet. VLSM allows for more efficient use of IP space whereas CIDR is geared more towards reducing the size of routing tables by allowing aggregation of networks into one route advertisement.

After reading that, no wonder I can't get a date. I'm too much of a geek!

Click to expand...

Hehe! Not many girls find this stuff interesting. That is for sure.

LonelyInAtl said:

But it sure felt good getting that dual track CCIE!

Click to expand...

Envy.

also: bitches don't know about my subnetting

Alright, I'm back! Feeling frisky with my new subnetting skills! (or complete lack thereof)

So how's about this:

10.220.17.55 /8 or /16 or /24 = Classful, so it would always have 254 usable hosts per network. Correct?

But, if I change it up, like this:

10.220.17.55 /27 = Classless (like the town drunk, lol), so now I'm gotta do a little more 'rithmetic.

128 / 64 / 32 / 16 / 8 / 4 / 2 / 1

10.220.17.55 = 00001010 . 11011100 . 00010001 . 00110111

/27 ............ = 11111111 . 11111111 . 11111111 . 11100000 (or, Mask: 255.255.255.224)

"AND" it out!!!

00001010 . 11011100 . 00010001 . 00100000

otherwise known as:

10.220.17.32 (Network Address?)

Since we "stole" three bits from the last octet for the network address, that is the "32s" place, so our Magic Number is 32, meaning that each SubNetwork will have 32 addresses (1 subnet address, 30 usable hosts, 1 broadcast)

So, the subnetworks are:

10.220.17.0 (0 subnetwork address, 1 - 30 hosts, 31 broadcast)
10.220.17.32 (32 subnetwork address, 33 - 62 hosts, 63 broadcast) Our original address is in this subnet?
10.220.17.64 (64 subnetwork address, 65-94 hosts, 95 broadcast)
10.220.17.96 (96 subnetwork address, 97 - 126 hosts, 127 broadcast)
10.220.17.128 (128 subnetwork address, 129 - 158 hosts, 159 broadcast)
10.220.17.160 (160 subnetwork address, 161 - 190 hosts, 191 broadcast)
10.220.17.192 (192 subnetwork address, 193 - 222 hosts, 223 broadcast)
10.220.17.224 (224 subnetwork address, 225 - 254 hosts, 255 broadcast)

This freaking concept may just kill me yet.

Rather than explain it, I found a good table in this article about classful IP...

http://en.wikipedia.org/wiki/Classful_network#Introduction_of_address_classes


Your VLSM appears to be correct.

LonelyInAtl said:

Rather than explain it, I found a good table in this article about classful IP...

http://en.wikipedia.org/wiki/Classful_network#Introduction_of_address_classes


Your VLSM appears to be correct.

Click to expand...

Lol, thanks! Reading it now. :club:

LonelyInAtl said:

Rather than explain it, I found a good table in this article about classful IP...

http://en.wikipedia.org/wiki/Classful_network#Introduction_of_address_classes


Your VLSM appears to be correct.

Click to expand...

But that's not actually VLSM, is it? Since there's only one mask being used, the /27 mask. Would be Classless, but not quite Variable.

VLSM would be more akin to a /25 mask for the x.x.x.128 range, with enough ips left for two /26 masks to cover the 129...193, 194...258 range.

How that would actually be implemented, on the other hand beyond the theoretical, I suppose, boggles me a little(along with OSPF configuration). I imagine that one could setup a router to have interface fa0/0 to the /25 interface for a particularly large host range, for example, and have the other two interfaces fa0/1 and fa0/2 to the smaller host ranges in the /26 range, limiting broadcast traffic.

My company right now simply uses the private C class range with a /24 mask, so that each subnet is assigned to a branch. This would permit 256 branch locations with 256 hosts in each; quite possibly not the best configuration, although given that we only seem to have around 10 to 75 computers at each site, not much of an issue yet.

IgnoredOne said:

LonelyInAtl said:

Rather than explain it, I found a good table in this article about classful IP...

http://en.wikipedia.org/wiki/Classful_network#Introduction_of_address_classes


Your VLSM appears to be correct.

Click to expand...

But that's not actually VLSM, is it? Since there's only one mask being used, the /27 mask. Would be Classless, but not quite Variable.

VLSM would be more akin to a /25 mask for the x.x.x.128 range, with enough ips left for two /26 masks to cover the 129...193, 194...258 range.

How that would actually be implemented, on the other hand beyond the theoretical, I suppose, boggles me a little(along with OSPF configuration). I imagine that one could setup a router to have interface fa0/0 to the /25 interface for a particularly large host range, for example, and have the other two interfaces fa0/1 and fa0/2 to the smaller host ranges in the /26 range, limiting broadcast traffic.

My company right now simply uses the private C class range with a /24 mask, so that each subnet is assigned to a branch. This would permit 256 branch locations with 256 hosts in each; quite possibly not the best configuration, although given that we only seem to have around 10 to 75 computers at each site, not much of an issue yet.

Click to expand...

Technically true. I pretty much use VLSM to mean any network broken into smaller networks. In the real world you really won't ever use the term. I can't remember the last time I used it outside of a classroom.

And remember, there are only 254 USABLE IPs when you take into account the network and broadcast. The gateway router generally takes up one IP, so you really could only have 253 hosts in the /24.

You are correct that you would need a router between subnets. However, you can also configure a "router on a stick" using a single interface to route between the disparate subnets using VLANs. All the traffic would be on the same wire but broadcasts could be an issue in busy networks. However, they would be filtered from the physical interfaces on the switch based on Port VLAN membership or Dot1Q VLAN tagging.

In the "real world" most places do allocate much larger networks for sites than they need to, and that's perfectly acceptable. We have a customer to allocates a /15 to every site using a private addressing scheme and they have no issues whatsoever. It's better to have too many available IPs than to have to re-IP and deal with the routing issues involved.

LonelyInAtl said:

You are correct that you would need a router between subnets. However, you can also configure a "router on a stick" using a single interface to route between the disparate subnets using VLANs. All the traffic would be on the same wire but broadcasts could be an issue in busy networks. However, they would be filtered from the physical interfaces on the switch based on Port VLAN membership or Dot1Q VLAN tagging.

Click to expand...

That is genuinely interesting; I think I'll be poking you with a lot of questions as I focus on improving my networking knowledge. Almost completely virtual interfaces are kind of old technology at this point, I admit, but I still know very little about them and its **** cool.