Subnetting; lost address space??

We are creating some small local subnets from a local class-C address space; we really want the effect of a remote bridge, but this seems like the easiest way to do it. We use a pair of Linux boxes with PPP links connecting them, and have the local router proxy-ARP for the addresses on the small remote segment.

We allocated a small chunk of 8 addresses, (, with netmask of
The claim made was that the parent space loses

I have read a possible concern with this (Linux IP-subnet mini-HowTo), and want clarification.
It states: once subnetted, the smallest subnet granularity of a network segment will determine wastage over the entire class-C, because the first/last address of every possible subnet of that size is now unavailable in the parent C-space.
This seems wrong to me. The subnetting is transparent to the host address space.
If there was a subnet declared in the main router to the host space, then perhaps this would be true.

If the sub-netting was known at the router entry to the C-space, instead of via proxy-ARP which hides it, then this could be an issue, but even then is this FAQ correct?
-- creating a small subnet for 8 remote devices would create 32*2=64 address holes in the C-space?

For the sake of this example, let us assume that you have decided to subnetwork you C class IP network number
into 4 subnets (each of 62 usable interface/host IP numbers). However, two of these subnets are being combined into a larger single network, giving three physical networks.

Network         Broadcast       Netmask                 Hosts         62         62         124 (see note)

Note: the reason the last network has only 124 usable network addresses (not 126 as would be expected from the network mask) is that it is really a 'super net' of two subnetworks. Hosts on the other two networks will interpret as the network address of the 'non-existent' subnetwork. Similarly, they will interpret as the broadcast address of the 'non-existent' subnetwork."

Is this a correct analysis?
Who is Participating?
agolanConnect With a Mentor Commented:
I'll answer backwards.
The analysis is not correct, probably because the
"" is incorect as well, the valid mask should be
And then the and are parts of the
very-existent network
I don't beleive that a full lesson on IP and subneting could be held here, however, if you want to learn yourself follow those guidelines/rules:
1) While learning and during your first configurations ALLWAYS
translate the addresses to binary (Most Calcs will do it you).
2) The 32 bit IP address that results from a convertion to binary is built from a network part and a host part.
3) The tool used to select which part belongs to "network" and
which belongs to "host" is the network-mask or "Mask".
4) The most significant part of a Mask should be allways filled with binary 1's and point out the bits used for network in the 32 bit IP address.
5) The least significant part of a Mask should be allways filled
with binary 0's and point out which bits are used for the Hosts.
6) If all the bits in the host part are 0's this is the network address.
7) If all the bits in the host part are 1's this is this subnet broadcast address.
8) Under no condition should 2 defined networks have a shared part of each other IP 32 bits... very tough so let's explain it:
          IP = 00000001.00000001.000000001.00000001
MASK = 11111111.11111111.111111111.11111100
(means that this is a 2 bits network)
          IP = 00000001.00000001.000000001.00000101
MASK = 11111111.11111111.111111111.11111000

See we have a collision because:
is common to both networks.

Need more ?
guthrieAuthor Commented:

Yes, the quote from the HowTo is in error, I hadn't noticed the 126 -> 128 mistake.
Your 8th point is the relevant one. I agree, and this is what I proposed in my question.

isn't it also a correct summary (as before):
  The subnetting is transparent to the host address space.  If there was a subnet declared in the main router to the host space, then perhaps this would be true; i.e. this would be an impact of the rule that all routers on a network must agree in the subnet mask for a network, else this conflict could arise.

Well, it depends how you define "Transparent" If you mean that if you are not subneting a network, you don't have to reserve what would be the subnet's 0' and broadcast of networks you could subnet from this one. This is true.
Also, not only routers, but all the devices must agree on a subnet mask, this is very important, the devices (incl. routers) use the subnet mask to decide if they should find by broadcast another host that is supposedly on the same net (or subnet) and
if not, it should be sent sent to their "default-gateway", typically a router connected in the same network.
Eventually, there might be also a "route" defining that a given
targeted network (or subnet) is available via another gateway.
guthrieAuthor Commented:
Yes, thanks. I should have also repeated that we use ProxyARP on the router to the subnet; that is why is is "transparent", i.e. no other devices or routers on the main net, or anywhere need to see/know it.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.