UEEN2013/UEEN2423

TCP/IP Network Fundamentals

(Topic 03)

Subnetting, VLSM and

Summarization
 IP Subnet Rule #1
• Every IP address within a “closed” network has to be unique.
– There can’t be two same IP addresses in the network
– Same applies to the Public IP addresses in the Internet

PC0 and PC2 can’t have the same IP in this network

 IP Subnet Rule #1
• The following networks are okay (for having same IP
addresses) since the networks are separated.
– There are two “closed” networks in the diagram.
 IP Subnet Rule #2
• The function of a subnet mask is to divide the IP address
into two parts: The Network ID and the Host ID.
• The process of getting the networking ID is to perform a
bitwise AND operation between the IP and subnet mask.
 IP Subnet Rule #2
• A simple example of getting the network ID and host ID:
– IP address = 128.143.137.144
– Subnet mask = 255.255.0.0
– Network ID = (128.143.137.144 & 255.255.0.0) => 128.143.0.0
– Host ID = 137.144 (of network 128.143.0.0)

• This is just a simplified demonstration. You need to convert the decimal

form to binary form for the ANDING operation:

Network ID Host ID

128.143 137.144
IP address
 IP Subnet Rule #2
• Another example to determine what the network ID is, for
any given IP address
– you have to convert both addresses into binary, and do a bitwise
AND operation.
• Example: IP address 156.154.81.56, network mask
255.255.255.240:

• This translates to a network ID of 156.154.81.48

• Sometimes, network ID is also called as network address
or subnet address (when you are referring to a subnet)
 IP Subnet Rule #3
• Network address within a “closed” network is
preferred to be unique.
Network ID=
Network ID= 192.168.2.0
192.168.4.0 Network ID= Network ID=
192.168.1.0 192.168.3.0
 IP Subnet Rule #4
• Examples of valid and invalid subnet mask
– 11111111.11110000.00000000.00000000 (valid)
– 11111111.11101101.00000000.00000000 (invalid)
• 1s staring from the left. No 0s in between 1s or otherwise
• There are only 32 valid subnet masks (theoretical
maximum).
– 128.0.0.0, 192.0.0.0, …, 255.192.0.0,…, 255.255.255.255
• Can be represented by ‘/’ notation
– e.g. /9, /24, etc
– /10 means ten ‘1’ from the left, and the remaining 22 bits are ‘0’
• 192.168.1.15/24 =>
– This interface has an IP = 192.168.1.15
– The subnet mask = 255.255.255.0
– Belongs to network 192.168.1.0
 Rule #4: Variable Network ID bits
• Since functional subnet masks can range from /8 to /30, which
means the network ID bits are also varies depending on the
subnet mask.
• This will cause the network ID to be the same for some subnet
masks.
– Be careful on your design.
Network ID Network ID
192.168.12.1/30 => 192.168.12.0 192.168.12.1/20 => 192.168.0.0
192.168.12.1/29 => 192.168.12.0 192.168.12.1/19 => 192.168.0.0
192.168.12.1/28 => 192.168.12.0 192.168.12.1/18 => 192.168.0.0
192.168.12.1/27 => 192.168.12.0 192.168.12.1/17 => 192.168.0.0
192.168.12.1/26 => 192.168.12.0 192.168.12.1/16 => 192.168.0.0
192.168.12.1/25 => 192.168.12.0 192.168.12.1/15 => 192.168.0.0
192.168.12.1/24 => 192.168.12.0 192.168.12.1/14 => 192.168.0.0
192.168.12.1/23 => 192.168.12.0 192.168.12.1/13 => 192.168.0.0
192.168.12.1/22 => 192.168.12.0 192.168.12.1/12 => 192.160.0.0
192.168.12.1/21 => 192.168.8.0 192.168.12.1/11 => 192.160.0.0
 Rule #4: Number of Host ID bits

• From the subnet mask, not only we know

the number of network ID bits, but also the
host ID bits.
– 32 bits = network ID bits + host ID bits

• For example, /25 =>

– 25 network ID bits with (32-25) = 9 host ID bits.
– Question: what is the available number of IP
address in this subnet?
 IP Subnet rule #5
• In order to communicate within the “LAN”, the IP address of all hosts and
gateways must have the same network ID.
– All IPs in the same subnet should be designed to the same network ID.
• That includes IPs for PC, server, gateway, etc.
• For the following example:
– If subnet mask = /16, all PC, server0, and gateway will be able to
communicate with each other
– If subnet mask = /24, gateway of Router0, and PC2 (172.16.26.X) will no
longer belong to the same network as the others.
 IP Subnet rule #5
PC0 can ping PC1

PC0 can’t ping PC3

Design guideline:
In order to communicate with
each other in a LAN, any IP
connecting to a switch should
share the same network
address (belong to the same
network)
 Application of Rule #5 in Gateway IP
• Remember that gateway IP has to be in the same
network (same network ID) with the rest of the
host IP.
– Otherwise, your subnet won’t work properly.
• So, please DON’T design your gateway IP as
This is okay.
below.

Don’t do this. Don’t do this.

 IP Subnet rule #6
• If the data’s destination IP does not have the same network ID as the
source IP, the data have to be sent to the default gateway (router).
– Using ARP, destination MAC address of default gateway will be obtained
for encapsulation purpose. Question: why use MAC add of default gateway?
• The router will then search in its routing table to find the destination
network. (will be discussed later)

If PC0 wants to:

ping 192.168.2.1
ping 180.7.4.3
ping 10.0.1.1

The data has to be sent to

the default gateway
 IP Subnet rule #6
• Example for rule #6:

• PC0 (192.168.1.1) ping 10.0.1.1

– Check for source IP network ID
• 192.168.1.1 & 255.255.255.0 = 192.168.1.0
– Check for destination IP network ID
• 10.0.1.1 & 255.255.255.0 = 10.0.1.0
• The source IP network and the destination IP network mismatch
• The host need to send the packet to the default gateway
– The subnet mask used in both cases, is the subnet
mask of source IP’s PC
 IP Subnet rule #6
• The algorithm for rule #6:

• Command destination_IP (e.g. ping 10.10.1.1)

• If (destination_IP & my_PC_subnet_mask == source_IP &

my_PC_subnet_mask)
– This means the destination IP is within same LAN,
– Check arp_table for this IP address for MAC address. (if not send out arp
to get this IP’s MAC address)
– Send the frame out to the IP address which is in the same LAN

• If (destination_IP & my_PC_subnet_mask != source_IP &

my_PC_subnet_mask)
– This means the destination IP is not in the same LAN
– Check OS for default_gateway_IP
– Check arp_table for the MAC address of default_gateway_IP
– Send the frame to the default gateway.
 IP Subnet rule #7
• Router (or any layer 3 devices) MUST be used for two hosts with
different network IDs to communicate as they reside in different
networks.
• Communication will not happen between hosts with different network
/subnet IDs that are connected to a switch
• A switch (layer 2) only provides communication for the PCs with the
same network ID
• Two or more different subnets have to be communicated via a router.
 Rules #7 in Router
• Routers are layer 3 devices which are suppose to link different
networks or subnets.

• The following example won’t work.

– Routers interfaces (ports) are in the same subnet as the PCs
 Example network of IP Subnet Rules #7
• The following example WORKS.
– Both PCs reside in different subnets and use their own gateway to
communicate with each other.
– The IP set for the interfaces won’t cause overlapping issue as they
belong to different subnets.
 IP Subnet Rule #8
• Two special cases on host ID bits which are all ‘0’s and all ‘1’s
– When Host ID bits are all ‘0’s, it is a network/subnet address.
– When Host ID is all ‘1’s, it is a broadcast address.

• These two addresses represent the “head” and the “tail” of the
given IP address block range.

• We CAN’T use these two addresses as host IP and gateway

IP.

• Example, IP address = 192.168.1.1, Subnet mask =

255.255.255.0
– Network address of the subnet = 192.168.1.0
– Broadcast address of the subnet = 192.168.1.255
– Question: What is the number of IP addresses that can be assigned to
hosts using this subset?
 Rule 8 in Router
• Router won’t accept network address and broadcast
address as its port IP address.

• The router will “complain” in the following cases and won’t

accept the addresses with the given subnet mask:
– IP = 192.168.1.0 subnet mask = 255.255.255.0
– IP = 192.168.1.255 subnet mask = 255.255.255.0
 What is a broadcast address?

• A broadcast address
– is a network address that allows information to
be sent to all nodes on a LAN, rather than to a
specific network host.
• Broadcast is used in ARP and DHCP.
• Router will stop broadcast from reaching
other subnets.
• Thus, an IP network/subnet is also called a
broadcast domain.
 IP subnet rule #9
• The first usable IP and the last usable IP.
– Usable IP addresses mean they can be used in hosts, PCs, printers,
gateways, and servers.
• The first usable IP = network address + 1
– More precisely, host ID = 1
– If network adress = 192.168.3.0, first usable IP = 192.168.3.1
• The last usable IP = broadcast address – 1
– More precisely, host ID = All ‘1’s – 1
– If broadcast address = 192.168.3.255, last usable IP = 192.168.3.254
• Cisco guidelines (not rules):
– Last usable IP is preferred to be
• router IP address = gateway IP
– First usable IP is preferred to be
• server, printer or important machines
• any host that requires a static IP
 LAN IP design guideline
• Given a range of IP address:
– First few IP addresses are
preferred to be used for setting First IP
static IP for servers and printers e.g. 192.168.1.1 Static IP
• Starting from the first IP and
counting down
• e.g: 192.168.1.1 for DHCP server;
192.168.1.2 for printer Static IP
– Last IP address is preferred to be
used for setting router IP (gateway Mid-range IP
IP)
• e.g. 192.168.1.254 for gateway Dynamic IP
– Mid-range IP addresses are
preferred to be set as the DHCP
range for the PCs
• e.g 192.168.1.10 to 192.168.1.253
e.g. 192.168.1.254 Static IP
• In this range, we reserve first 10 IP
for servers and printer and the last Last IP
IP for gateway
 Why Last Usable IP for Gateway?
• Case study:

• If the gateway IP is set at the beginning of the range (e.g.

192.168.1.3), chances are very high that you set an host IP
which is the same as the default gateway IP.
 IP Subnet Rules #10
• When Internet addresses were
standardized (in early 1980s), the IP
addresses were divided up into 5
classes:
• Class A:
– Network prefix is 8 bits long.
– Default mask: 255.0.0.0. or /8
• Class B:
– Network prefix is 16 bits long.
– Default mask: 255.255.0.0 or /16
• Class C:
– Network prefix is 24 bits long.
– Default mask: 255.255.255.0 or /24
• Class D:
– contains multicast addresses
• Class E:
– Experimental
 IP Subnet Rules #10
• We can only use class A, B, and C
for host IP address.
– Class A, B, C IP addresses are
called unicast IP addresses
• We CAN’T use class D and E IP
address for “normal” IP address.
– Class D IP addresses are called
multicast IP addresses
 Broadcast, Multicast, Unicast
• Broadcast
– One to all
– Used in DHCP, ARP
• Multicast
– One to many
– Class D IP
– Normally used in streaming IP
TV or streaming radio
• Unicast
– One to one
– Class A, B, C IP
– What we have been using for
devices
 IP Subnet Rules #10
• Unicast IP = 1 device 1 IP
• Multicast IP = all devices that “tune into that channel” share the IP
• Unicast IP and multicast IP can be used at the same time in 1 host.
• Example in the following diagram:
– Two streaming servers with multicast IP (224.0.0.112 and 224.0.0.113)
• For example streaming web cam videos.
– All PCs and servers still require their unique IP address
– If all PCs “tune in” the “channel” of 224.0.0.112, all PCs and Streaming
Server 1 will share the same multicast IP (224.0.0.112) on top of their own
unicast IP address.
 IP Subnet Rules #11
• The range of private IP addresses which is NOT used in public IP address
for global Internet.
• The Internet Assigned Numbers Authority (IANA) has reserved the
following three IP address blocks for private use (local networks):
– Class A: 10.0.0.0 - 10.255.255.255
– Class B: 172.16.0.0 - 172.31.255.255
– Class C: 192.168.0.0 - 192.168.255.255
• IANA has reserved the following IP addresses as Automatic Private IP
Addressing (APIPA) for Windows platforms (except NT).
– 169.254.0.0 - 169.254.255.255
• APIPA is used in Windows where the IP addresses (169.254.x.x) are
assigned automatically by the OS when the DHCP service is not available
(and the option of “Obtain an IP address automatically” is on).
– If the DHCP service is working fine in your LAN and you still get the IP
169.254.x.x for your Windows, it means your cable, connection or NIC is
facing problem.
 Private and Public IP Address
• Private IP addresses are the IP addresses that are used at home, office,
etc. as your source IP with the purpose of saving money.
• You need to pay money for an Public IP address (monthly subscription).
– And the domain name too (e.g. www.utar.edu.my, www.intel.com)
• Public IP is global, and you need to apply and obtain it through IANA (via
ISP)
• Private IP is not unique.
– There are thousands of 192.168.1.1 in the world
• Special IP = 127.0.0.1, is reserved to be the loopback address (localhost)
– To test the network software
– The domain name localhost = 127.0.0.1
– Ex: ping locahost
 IP Subnet Rule #12
• A host can have different IP addresses according to the
number of network interface cards installed.
• Examples:
– A PC with 1 NICs requires 1 IP address
– A PC with 3 NICs requires 3 IP addresses
• One IP for each NIC
– A router with 2 Fast Ethernet ports and 2 T1 serial ports need 4 IP
addresses
• One IP for each ports
• A DNS server can have 2 NICs with 2 different IP
addresses
 IP Rule #13
• Number of host ID bits as the LAN design parameter.

• For example: For a subnet with a mask of 255.255.255.128, how many

useable IP that we can allocate in that subnet?
– 255.255.255.128 => /25 => 32-25 = 7 host bits
– 27 => 128 host ID => theoretically 128 IP address.
– Actual allocation of host IP = 128 – 1 – 1 = 126 useable IP addresses
– -2 because you can’t use network address and broadcast address.
– The remaining 126 can be used for hosts, i.e., laptops, servers, printers or
default gateway, IP for managing a switch, etc.

• If a LAN is desired to have just 27 PCs (assume no gateway, IP for

switch, etc.), what should be the subnet mask?
– Formula: 2H ≥ “number of hosts” + 2; H = host ID bits
– 2 = 1 network address + 1 broadcast address
– 2H ≥ 27 + 2. So, it is suffice to have H = 5.
– H = 5; Subnet mask = /(32-5) = /27 => 255.255.255.224 (VLSM, SM is no
longer fix to the masks of class A, B and C)
 IP Subnet Rules #14
• The principle of IP design in subnetting and supernetting
• Subnetting
– 1 big network is splitted into a few smaller networks.
• Supernetting
– 2 or more smaller networks are joined to form a larger network.

subnetting
supernetting
 IP Subnet Rules #14
• Split the host number portion of an IP address into a subnet
number and a (smaller) host number (borrowing host bits).
• Results in a few small blocks of IP addresses.
• The subnet mask will be extended with more ‘1’

network prefix host number

network prefix subnet number host number

extended network prefix

 IP Subnetting example (1)
• Perform subnetting for the following LAN.
– 172.16.0.0/16
– Question: How many IP addresses under this network?
– Question: What is the range of IP address of this network?
• Now, lets do subnetting.
– Let’s borrow 8 bits from the host bits.
 IP Subnetting example (2)
• Original subnet mask = /16
• New extended subnet mask = /(16 + 8) = /24
16 bits 16 bits

172.16 0.0

8 bits 8 bits
network prefix Subnet # hosts

extended network prefix (24 bits)

8 bits
New network prefix hosts
• Questions: (i) How many subnets you can create? (ii) How
many IP addresses per subnet?
 IP Subnetting example (3)
• Implementation
– Subnet 1 network into 3 LANs
– 1 network ID (172.16.0.0/16) becomes 3 subnet IDs (172.16.1.0/24,
172.16.2.0/24, 172.16.3.0/24)
– All subnet masks have been changed from /16 to /24
– 1 gateway -> 3 gateways
 Types of Subnetting
• Traditional classful subnetting
– Design with an equal size subnet mask across
“the network”. Ex: class A /8, class B /16, and
class C /24.
– Waste a lot of IP addresses, particularly for
class A and B.
• VLSM (Variable Length Subnet Mask)
– Creates different sizes of subnets using
different lengths of subnet mask.
– More efficient use of the IP addresses.
 Traditional classful Subnetting
• First, determine which class does the IP
belongs to
– Class A (24 host bits), B (16 host bits) or C (8 host
bits)

• Second, determine how many subnet bits you

want to allocate.

• Third, compute the new subnet mask.

• Fourth, plan the IP range of each subnet.

 Traditional Classful Subnetting
• There are two ways to determine the subnet bits
(you can choose either one the following):
– The number of subnets that you want to have in the
network that you are going to “split”
• For example, you want 10 subnets in the network, then you need
S subnet bits. (2S >= 10)
– The maximum number of hosts that you want to allocate
• For example, if you have a class B IP address, and you need
maximum1000 hosts in any one of the subnet, then you need N
subnet bits where ( 32 = 16 + S + H bits) 2H >= 1000 => H = 10
bits. Then N = 32 – 16 – 10 = 4 bits.
• 16 is the original host bits for class B IP address.
 Steps to Subnetting (1)
• Case Study:
– You have a Class C address of 192.168.100.0
/24. You need nine subnets.
– What is the IP plan of network numbers,
broadcast numbers, and valid host numbers?
– What is the subnet mask needed for this plan?
 Steps to Subnetting (2)
• Since the subnet mask is /24, we only need to focus on the
4th octet.
– The 4 octet is the host bits, represented by ‘H’
– The ‘N’ bits is the new “subNet” bits.
 Steps to Subnetting (3)

*using ip subnet zero can solve the problem

Steps to Subnetting (4)
Steps to Subnetting (5)
Steps to Subnetting (6)
Steps to Subnetting (7)
Steps to Subnetting (8)
Steps to Subnetting (8)

Continue
next page
Steps to Subnetting (9)
Steps to Subnetting (10)
 VLSM (1)
• Variable-length subnet masking (VLSM) is a
more realistic way of subnetting to efficiently
use of all of the bits.

• VLSM is the process of “subnetting a subnet”

and using different subnet masks for different
subnets in your IP plan.

• What you have to remember is that you need

to make sure that there is no overlapping in
any of the addresses.
 VLSM (2)
• Remember that when you perform classful (or
what I sometimes call classical) subnetting, all
subnets have the same number of hosts because
they all use the same subnet mask.
• This leads to inefficiencies.
– For example, if you borrow 4 bits on a Class C network,
you end up with 14 valid subnets of 14 valid hosts.
(instead of 16 subnets)
– A serial link to another router only needs 2 hosts, but
with classical subnetting, you end up wasting 12 of
those hosts.
 VLSM Example #1 (1)
• Given the following network 192.168.1.0/24, how is it
going to be subnetted?
– We only need to take care of the 10 hosts subnet, 11 hosts
subnet, and 30 hosts subnet.
– Assume that there won’t be expansion on the number of host
per subnet.
 VLSM Example #1 (2)
• Using the traditional subnetting methodology.
– There are 3 subnets and the maximum # host in a
subnet is 30.
– Either we have 2 subnet bits or 5 host bits.
– If we choose 2 subnets bit design (6 bits host),
there won’t be much expansion on subnets but
growth on hosts.
– In this case, we choose 5 bits host design (3
subnet bits), as specified in the question.
 VLSM Example #1 (3)
• With traditional classful subnetting
25 = 32 ip addresses. So,
192.168.1.64/27 S0: 192.168.1.0 – 192.168.1.31
S1: 192.168.1.32 – 192.168.1.63
S2: 192.168.1.64 – 192.168.1.127

192.168.1.32/27 192.168.1.0/27
 VLSM Example #1 (4)

• However, the traditional classful subnetting in

this case, is not “IP address efficient”.
• Given that there won’t be expansion on the
number of host per subnet.
• The 10 hosts subnet and 11 hosts subnet will
have around 20 IP addresses (per subnet) that
are not utilized.
 VLSM Example #1 (5)
• If we use VLSM scheme, this is how we are
going to do it.
• 1st we choose the largest network 30 hosts
– Host bits is still 5
– We can still use the S0 192.168.1.0/27, useable
IP range from 192.168.1.1 to 192.168.1.30
• 2nd we “group” the 10 hosts subnet and 11
hosts subnet into one “30 hosts” subnet.
– We can use the S1 192.168.1.32/27, useable IP
range from 192.168.1.33 to 192.168.1.62
 VLSM Example #1 (6)
• Now, we perform subnetting the subnet.
• We subnet the 192.168.1.32/27 into two
subnets.
• New SM:
11111111.11111111.11111111.11110000
• With new SM, we can get 2 subnets, each 16
IP addresses.
• We will have:
– 192.168.1.32/28
• Usable IP range 192.168.1.33 to 192.168.1.46
– 192.168.1.48/28
• Usable IP range 192.168.1.49 to 192.168.1.62
– How is this possible? Next slide will explain
 VLSM Example #1 (6) in detail…
• IP: 11000000.10101000.00000001.0010#0000 192.168.1.32
SM: 11111111.11111111.11111111.1111#0000 /28 Borrow 1 bit

SS0: 11000000.10101000.00000001.0010#0000 192.168.1.32 Subnet ID

11000000.10101000.00000001.0010#0001 192.168.1.33
...
11000000.10101000.00000001.0010#1110 192.168.1.46
11000000.10101000.00000001.0010#1111 192.168.1.47 Broadcast

SS1: 11000000.10101000.00000001.0011#0000 192.168.1.48 Subnet ID

11000000.10101000.00000001.0011#0001 192.168.1.49
...
11000000.10101000.00000001.0011#1110 192.168.1.62
Broadcast
11000000.10101000.00000001.0011#1111 192.168.1.63
 VLSM Example #1 (7)
• With VLSM
192.168.1.32/28

192.168.1.48/28 192.168.1.0/27
 VLSM Example #1 (8)
192.168.1.32/28

I need two IPs for this

Network. 

192.168.1.48/28 192.168.1.0/27

• Can you help?

 VLSM Conclusion

• Looking at the plan, you can see that no number

is used twice.

• You have now created an IP plan for the

network and have made the plan as efficient as
possible, wasting no addresses in the serial
links and leaving room for future growth.

• This is the power of VLSM.

 Route Summarization
• Route summarization is needed to reduce the
number of routes that a router advertises to its
neighbor.
– Router summarization make the “combined” subnets
appear to be a “supernet”
• For every route you advertise, the size of the
update and routing table grows.
• Route summarization greatly reduces the size of
the routing table
– It has been said that if there were no route
summarization, the Internet backbone would have
collapsed from the sheer size of its own routing tables
back in 1997.
 CIDR
• CIDR (Classless Inter Domain Routing) is also known as
supernetting as it effectively allows multiple subnets to
be grouped together for network routing.

• CIDR was adopted to help ease the load imposed on

internet and large network backbone routers by the
increasing size of routing tables.
– Large routing tables have several adverse effects:
• Routers require more memory in order to store and manipulate their
routing tables which increases operation costs.
• Routing latency is increased due to the large amount of data contained
in the routing tables.
• Network bandwidth usage is increased by routing updates when
routers exchange their routing tables.
 CIDR
• CIDR (Classless Inter-Domain Routing)
encompasses:
– the VLSM technique of specifying arbitrary-length
prefixes.
• An address in CIDR notation is written with a suffix
indicating the number of bits in the prefix, such as
192.168.0.0/16, where /16 is the suffix, and 192.168.0.0 is
the prefix.
– the aggregation of multiple contiguous prefixes into
supernets, thus reducing the number of entries in
the global routing table.
• Aggregation hides multiple levels of subnetting from the
Internet routing table, and reverses the process of
subnetting with VLSM.
Route Summarization Example (1)
Route Summarization Example (2)
Route Summarization Example (3)
Route Summarization Example (4)
Route Summarization Example (5)
Route Summarization Example (6)
Supernetting & Route Summarization
• Supernet and route summarization use the
same IP address computation that combines
a few IP of a few subnets into one.
• The difference is that supernetting will
physically combine a few subnets together
whereas route summarization only compute
the supernet network ID (in router) without
changing the network.
Supernetting & Route Summarization
• Both route summarization &
supernetting produce the
same summarized IP
network ID.
– However supernetting
changes the network
Original topology.

Route
supernetting Summarization
172.16.128.0/22
 Pros and Cons of subnetting

• Pros:
– Better security and management
– More host IP being utilized within the subnet.
• Cons:
– Waste some host IP in the form of network
address, broadcast address, and gateway IPs
– Network design is more complex in order to
perform the IP planning.