9

Other than not being able to claim any more IP addresses in a subnet, what other bad things happen when a subnet reaches 100% utilization, if anything?

Ron Maupin
  • 98,218
  • 26
  • 115
  • 191
UnworthyToast
  • 201
  • 1
  • 5
  • 5
    Total protonic reversal! – AbraCadaver May 25 '17 at 19:53
  • IP doesn't care whether an address is used or not - as far as the computer is concerned, subnet utilization might as well be a made up concept. – user253751 May 25 '17 at 21:32
  • Whatever you do, just don't put a hub on that subnet. Not that hubs are really used anymore. But maybe you work for the Government, then it's possible--one see's the craziest stuff when working for the Government :) – Jeff.Clark May 26 '17 at 08:51
  • 1
    Is there a particular reason for all the sarcastic/humouristic comments? The guy asks a perfectly sensible question. Sure, the answer might be obvious to us, but not to everybody. Is it so hard to say "except for no more addresses being available, nothing bad happens" and leave it at that? – AnoE May 26 '17 at 11:21

1 Answers1

19

A subnet (network) is really just a collection of contiguous addresses within a binary mask. It is simply a logical way to divide address block. If you run out of addresses in a network (subnet), then you are simply out of addresses in that network. Adding any more hosts would require reclaiming unused addresses, expanding the network (may not be possible), or adding a new network (requires a router to communicate between networks).

Ron Maupin
  • 98,218
  • 26
  • 115
  • 191
  • 1
    Okay. That's what I thought, but since I'm relatively new to the networking arena I wanted to make sure there weren't other consequences that could happen that I was unaware of. Thanks! – UnworthyToast May 25 '17 at 19:12
  • Think about it this way: just configure your hosts to use one more bit. Boom, your subnet is at 50% capacity again. But … the network didn't change at all! So, clearly, there cannot be a difference between a half-empty subnet and a full subnet. – Jörg W Mittag May 25 '17 at 23:50
  • 1
    What happens if the subnet which fills up is `0.0.0.0/0` or `::/0` ? ;-) – TessellatingHeckler May 26 '17 at 02:18
  • That doesn't make sense because you would be trying to use invalid IP addresses. No router will let you assign a `0.0.0.0/0` network to an interface. – Ron Maupin May 26 '17 at 02:24
  • 3
    @TessellatingHeckler Then you invent IPv6 – JBentley May 26 '17 at 10:04
  • @RonMaupin So the problem already occurs when you fill up a `/1` net ... – Hagen von Eitzen May 26 '17 at 11:02
  • 2
    @HagenvonEitzen, from a a practical standpoint, you would not be able to have a single network (LAN) with the over two billion hosts hosts on it that would be required to fill a `/1` network. Aside from the cost, there is the matter of space required, and the fact that a broadcast network would collapse long before reaching that many hosts. Even with matchbox-sized (two cubic inches) devices, you are looking at over 70,000 cubic meters of devices, which must have power and cooling that requires more space, and most LAN protocols will not work over those types of distances. – Ron Maupin May 26 '17 at 14:40
  • 2
    FYI: Most dhcp servers in routers handle this by ignoring dhcp requests from new devices – Ferrybig May 30 '17 at 15:06