Chapter 8
Advanced TCP/IP Network Design -
CLASSLESS ADDRESSING AND
VARIABLE-LENGTH SUBNET MASKS
Variable-Length Subnet Masks
Variable-length subnet masks specified how a single network ID
could have different subnet masks among its subnets.
Used correctly, VLSM could minimize the wasted IP addresses
Benefits
The major benefit of VLSM is that subnets can be defined to
different sizes as needed under a single Network ID, thereby minimizing, if not eliminating, wasted addresses.
As a result, an organization’s assigned IP address space is more
efficiently used.
Second, when correctly defined to match the physical topology
of the network, variable-length subnet masks can used to
permit router aggregation that minimizes the number of distinct routes that need to be advertised and processed by network
Implementation Requirements
In order for VLSM to be successfully implemented, the routers
on the network where VLSM is implemented must be able to share subnet masks and/or extended network prefixes along with each router advertisement
All routers supporting VLSM must support a longest match
routing algorithm.
This is particularly important in VLSM networks because
subnets can be embedded within subnets
Finally, the implemented network topology must match the
That is to say, the network designers must decide in advance
how many levels of subnets are required, and how many hosts per subnet must be supported at each level
Recursive Division of a Network Prefix
with VLSM
As previously described, VLSM allows an organization’s
assigned address space to be recursively divided into as many levels and sizes of subnets as required.
In order to better understand this process, we will first show
how the address space is divided and then show how the routes from that recursively divided address space can be aggregated to effectively reduce the amount of transmitted routing information.
In addition to reducing the amount of transmitted and stored
routing information, an added benefit is that the associated network topology and structure of one subnet is unknown to other subnets.
Figure 8-12 illustrates how a single network prefix can be recursively divided thanks to VLSM
Route Aggregation with VLSM
While the benefits of flexible subnet size definition is illustrated
in Figure 8-12, the route aggregation benefits of VLSM are illustrated in Figure 8-13.
Often, the terms summarization and aggregation are used
interchangeably to describe the process of reducing the number of routing advertisements between subnets by only advertising the common portion of subnet IDs.
Alternatively stated, summarization and aggregation mean that
subnet information is not shared between two networks when a router connects those networks
In some cases, however, a distinction is made between the two
terms.
In such cases, the term summarization is reserved to describe
those circumstances in which subnet addresses have been
rolled up all the way to the major network prefix as assigned by the Internet authorities.
In Figure 8-13, this would be the 121.0.0.0/8 major network
prefix.
On the other hand, the term aggregation is used to more
generally describe any circumstance when only the common portion of those addresses in a routing advertisement can represent a subnet’s entire address space
Notice in Figure 8-13, how each physical network that houses
multiple subnet IDs can have its routing information
summarized to a single route advertisement to the next higher layer of subnet.
Finally, the entire internetwork can be advertised to the
Internet routing tables by the single assigned network ID: 121.0.0.0.
Such route aggregation and the efficiencies gained therein, are
only possible if subnet masks are assigned in a planned manner so that subnet address assignment mirrors the actual topology of the network, as illustrated in Figure 8-13.
If assigned addresses are not organized to mirror the physical
topology of the network, then address aggregation
is not possible and the benefit of reduction of routing table size
Subnet Design Using VLSM
Subnet design with variable-length subnet masks is similar to
subnet design with fixed-length subnet masks, but the
decisions made regarding subnets for the entire network in the fixed-length subnet mask scenario are made independently at each level in the variable-length subnet mask scenario.
To elaborate, at each level (subnets, sub-subnets,
sub2-subnets, etc.), basically two questions must be answered:
1. How many subnets are required at this level, both now and in
the future?
2. What is the largest number of host required per subnet on
Defining Sub-Subnet Numbers with VLSM
Figure 8-14 provides an example of how subnet numbers are
defined in VLSM.
In this example, it was determined that six sub-subnets were
needed beneath the 121.253.0.0/16 subnet.
Since two subnets are reserved, we need to really be able to
define eight sub-subnets.
Two to the third power is eight, so it will take 3 additional bits
or /19 (/16+3 = /19) extended network prefix to provide the required six sub-subnets
Defining Sub2-Subnet Numbers with
VLSM
If it was then decided that the 121.253.160.0/19 sub-subnet
needed to be recursively divided into six sub2-subnets, so 3 additional bits of variable length subnet mask would be
required.
Defining Host Addresses for a Given
Subnet
With VLSM, defining host addresses involves the same process
for subnet, sub-subnets, or sub2-subnets.
Figure 8-16 illustrates the host definition process for
sub2-subnet 121.253.184.0/22 defined in Figure 8-15.
The extended network prefix of /22 tells us that 1022 host IDs
can be defined on this sub2-subnet. (32 bit address – 22
reserved bits = 10 bits available for host ID; two to the tenth power = 1,024 – 2 reserved host IDs = 1,022 available host IDs).
If 1,022 host IDs are way more than we could ever reasonably
use, we would probably want to consider defining another
subnet level so as not to strand or waste precious IP addresses.
Notice how the extended network prefix does not increase
when we define host IDs for a given subnet the way it did when we defined additional subnet levels to existing subnets
Notice how the third octet has changed from 184 to 187 on the
last few host IDs.
Does this mean that the subnet ID changed somehow? The
answer is no.
If you look in the extended network prefix column, you will see
that the subnet ID has not changed.
The reason the third octet changed is because the extended
network prefix was 22, leaving 2 bits of the third octet left over for use by the host ID.
Since the host IDs start using the rightmost bits first, it was
only when we got to the last few host IDs that we were forced to use the leftmost bits, which happened to belong in the third octet.
As a result, the third octet may have become 187, but the
Determining if VLSM IP Addresses Are
Part of the Same Subnet
Routers use the same algorithm to determine if IP addresses
are part of the same subnet, whether or not VLSM is used.
A router must somehow know the extended network prefix or
subnet mask, as well as the IP address.
In the case of fixed-length subnet masks, the router could use
its own interface’s subnet mask (since all subnet masks on a given network had to be the same), or it could assume the default subnet mask based on classful address class.
In the case of variable-length subnet masks, no such
assumptions can be made.
Extended network prefixes must accompany every advertised
