VLSM Subnetting Question

[TeuffelHunden]

I'm in a course right now that requires the use of a lot of VLSM subnetting.

I've been given some conflicting information about the use of the last subnet (Class C). Some people say you can use it, others say you can't because it contains the broadcast ID for the entire Net ID. But if that's the case, then why can you use the first subnet, which contains the Net ID for the entire net?

SO the question is...can you use the last group of IP addresses (eg 192.168.1.252-192.168.1.255) as a subnet? Please provide reason and/or reference for your response, since I will need to back up anything I say in class.

Thanks!

[Thespis377]

Well, let's see if a Network Admin can help you out here. In your example, 192.168.1.252-192.168.1.255, you forgot an IP. You forgot 192.168.1.0. This is known as the Network IP. And the last IP in the block is the Broadcast. You will sometimes see networks expressed ast 192.168.1.0/24. That gives you 254 usable addresses, b/c the first and last one cannot be used. The gateway is usually the 1st. For example 10.0.0.0/23 will give you 511 usable addresses, 10.0.0.1-10.0.1.255. traditionally your gateway will be 10.0.0.1. Again, this is only by tradition. You can use any IP in the block you want. If you would like, I'll even show you how to do subnet math. Again, it's not that hard once you figure out that it's just simple binary math.

[Ddruid_SMP]

SO the question is...can you use the last group of IP addresses (eg 192.168.1.252-192.168.1.255) as a subnet? Please provide reason and/or reference for your response, since I will need to back up anything I say in class.

Thanks!

192.168.1.252/30 (255.255.255.252)
.252 - Network
.253 - Valid IP
.254 - Valid IP
.255 - Broadcast

Cheat Sheet (http://www.cisco.com/univercd/cc/td/doc/product/iaabu/pix/pix_sw/v_60/config/subnets.pdf)

[Spion]

^ correct ^ go with a /29 mask if you need 4 valid addresses

[Thespis377]

As a Network Engineer, I failed reading comprehension. DUH. Sorry for the long winded dribble that didn't answer your question. Ddruid_SMP is correct. Remember, any network will consist of the Network IP and the Broadcast IP. Spion is correct. If you need at least 4 addresses to use, you would need a 192.168.1.248/29 network. This gives you the following:

192.168.1.248: Network
.249 Valid IP
.250 "
.251 "
.252 "
.253 "
.254 "
.255 Broadcast

[ErikD]

Yep basically as long as you get at least two valid IP addresses you are fine. Good to use on point to point links. No reason it is any less true with a class C address than with any other.

[TeuffelHunden]

Thanks for all the replies everyone. So what everyone's agreeing on is that I CAN use the 192.168.1.252/30 subnet. That's what I thought also.

What prompted the question was that I was told I could not because the .255 must be reserved for the broadcast ID for the /24 net.

So does anyone know where it says that? I'd like something to reference when I argue my point :)

[TeuffelHunden]

1
Cheat Sheet (http://www.cisco.com/univercd/cc/td/doc/product/iaabu/pix/pix_sw/v_60/config/subnets.pdf)


Nevermind...this is EXACTLY what I was looking for!! Thanks Ddruid_SMP!!

Thanks to everyone for their help!

[NWEng]

Damn, I missed a pretty good subnetting conversation. It's good to find other engineers on the forums.
What prompted the question was that I was told I could not because the .255 must be reserved for the broadcast ID for the /24 net.
It still is the broadcast address for any device the /24 network using a /24 mask for that address range, and it's also the broadcast for that /30 network for those devices in that range using a /30 mask. However, I wouldn't recommend using the last /30 within an active /24 due to the conflicting broadcast addresses. Either carve up the /24 or leave it alone; but, don't "nest" subnets. Bad times.

[TeuffelHunden]

So if I sent a broadcast message to .255 would the entire /24 get it, or just the /30 subnet? And is there a way to differentiate when sending the message?

[Cheator]

you wont send a broadcast message. It is done through ARP. So you wont be using that address at all, unless you are calculating subnets. And as for where it goes to, the broadcast address for that subnet is used for that subnet only. So if your broadcast address is 192.168.1.255, then it broadcasts to the 192.168.1.0 network only.

Finally my CCNA program is useful!

[NWEng]

This is all generalization and theory, really. "You" won't be sending anything via the broadcast address. However, one of the more real world use of the broadcast address is used within Windows for WINS and SMB/Netbios, but that's down the road for you and irrelevant at this time.

The broadcast address for the 192.168.1.0 /24 and 192.168.1.252 /30 (or the 192.168.1.248 /29, etc.,) is the same address, 192.168.1.255. If some some reason both subnets are on the same wire, (collision domain,) probably due to a misconfiguration, they are essentially in the same broadcast domain.


It is done through ARP.
I'm not sure what you're saying, what is done through ARP? ARP is broadcast via an Ethernet broadcast address, not IP. Remember, ARP is requesting the MAC address of a particular IP and it's only is carried on the requestors collision domain.

[Cheator]

I'm not sure what you're saying, what is done through ARP? ARP is broadcast via an Ethernet broadcast address, not IP. Remember, ARP is requesting the MAC address of a particular IP and it's only is carried on the requestors collision domain.
Yeah but arp is layer 3 as well. It can request the MAC address of a certain IP. It wasn't initially designed for layer 3 but it functions on it as well and I belive uses the broadcast address. Then again, I could be wrong :). That was level 1. But, reading up on it, it looks like ARP requests are sent out as packets, and would probably, therefore, be layer 3.

All that to say, you might be right. Sorry if I caused some confusion. But one thing is for sure and a constant in the universe: network addresses and broadcast addresses are unusable IPs.

[ErikD]

The bottom line is that yes, 192.168.1.0/24 and 192.168.1.252/30 do share the same broadcast address. However both subnets shouldn't be getting used in the same collasion domain anyway. You should be designing the network to use all available IP addresses with no overlap. Those subnets on the same collision domain would overlap for the whole address space of the /30.

So it individually they work fine and are valid subnets. Together they are not. Remebember that each of the /30 subnets would be an address space for the /24, so technically all addresses of each /30 subnet could be used by the /24. That is why you can't have them working together.

The whole thing with subnetting is you need to subnet a portion that won't be getting used by IPs for devices. So 192.168.1.0/28 and 192.168.1.252/30 would play nicely together.

[TeuffelHunden]

Thanks for all the input everyone!!