3

Is it easier to find subnet overlaps in VLSM and FLSM. I really don't see the difference. The algorithm to find overlaps in VLSM and FLSM is not different right? Both of them involve finding the broadcast addresses and checking if they are different/same?

Danis Fermi
  • 323
  • 1
  • 3
  • 16
  • The NESE community is pretty stringent on answering questions sourced from training materials or schoolwork or educational classes. I recommend editing your question to remove the phrase My professor asked me a question. – Eddie Sep 05 '16 at 17:35

1 Answers1

0

There is no possibility of an overlap with FLSM. There can only be an overlap with VLSM. To that end, it will be much easier to "find" an overlap in FLSM, since there never is one. Whereas "finding" a overlap in VLSM will involve some binary math. Let me explain.


FLSM, or fixed-length subnet mask implies every mask is the same across multiple subnets.

For example, breaking up a Class C network (aka, /24) into FLSM sub-networks of /27 would yield ONLY these results:

  • x.x.x.0/27
  • x.x.x.32/27
  • x.x.x.64/27
  • x.x.x.96/27
  • x.x.x.128/27
  • x.x.x.160/27
  • x.x.x.192/27
  • x.x.x.224/27

These are the only 8 possible /27's within a /24. Every NetID listed above represents a subnetwork that includes addresses of the NetID itself, and the following 31 addresses.

There is no such thing as a x.x.x.150/27 address, the NetID and mask do not pair, so there is no way to have a /27 range represent 150-181 (and would therefore introduce an overlap possibility).

As a result, in FLSM, there is no possibility of an overlap if you are using NetID's and Masks that pair properly.


VLSM, or variable-length subnet mask implies every mask is not necessarily the same across multiple subnets.

For example, there are countless ways to break up a /24 into multiple subnetworks, one of which might be...

  • x.x.x.0/27
  • x.x.x.32/28
  • x.x.x.48/29
  • x.x.x.56/29
  • x.x.x.60/30
  • x.x.x.64/27
  • x.x.x.96/28
  • x.x.x.112/28
  • x.x.x.128/26
  • x.x.x.192/28
  • x.x.x.208/28
  • x.x.x.224/27

To find overlaps, you would have to calculate the IP range for each subnet, and ensure that no IP addresses are "counted" twice. That involves doing the traditional subnetting math.

Since the masks differ for each subnetwork, in VLSM, there is absolutely a possibility that an overlap exists.


To close, I pose the question to you. Every NetworkID and SubnetMask above pair properly, which is to say they represent a valid subnetwork and subnet mask combination.

As we learned, with properly paired FLSM, there is no overlap. With properly paired VLSM, there might be an overlap.

In the Ssb-network list above, there is an overlap in the VLSM range above. Can you find which it is? How easy was it, comparatively?

Eddie
  • 15,026
  • 6
  • 44
  • 84
  • x.x.x.56/29 doesn't fit. It overlaps the next network: x.x.x.60/30. One could argue that any /27, even if they are all the same is using VLSM since the mask varies from the original masks defined by the IPv4 classes. That is basically the definition of VLSM. – Ron Maupin Sep 05 '16 at 18:32
  • @RonMaupin winner! Although, I was hoping someone who could use the subnetting practice would have found it. ;) Also,your statement about the /27's and FLSM vs VLSM isn't entirely correct. I also once thought that FLSM meant classful boundaries, but this isn't true. Lab out two routers, configure each of them with 3 /27 loopbacks in the 10.0.0.0/8 network, and a /27 interlink, then enable RIPv1 (which does not support VLSM) and you'll see the FLSM RIPv1 shares /27 routes. – Eddie Sep 05 '16 at 19:01
  • Well, the point of VLSM was that it freed us from the classful masks. When it was introduced, people were glad that they could create eight /27 subnets from one Class C network. It didn't matter that all the subnets were /27, they got to use masks different from the class mask. That is VLSM. It seems that people have forgotten history. VLSM predates RIPv2. Don't confuse VLSM and CIDR. RIPv1 can use VLSM, but not CIDR. – Ron Maupin Sep 05 '16 at 19:06
  • RIPv1 can use VLSM, but not CIDR -- How so? The lab I am looking at right at the moment says otherwise. But we're off tangent, shall we defer to chat? – Eddie Sep 05 '16 at 19:10
  • OK, we can chat. As you say, RIPv1 can advertise prefixes with other than the classful mask (VLSM), but it simply does not use CIDR since that would confuse RIPv1 if two routers advertise subnets of the same classful network. – Ron Maupin Sep 05 '16 at 19:13
  • 1
    @Eddie I got your point. Binary math for VLSM is much harder. That was a great explanation by the way. – Danis Fermi Sep 05 '16 at 19:38