Computer Networks: A Systems Approach, 5e

Larry L. Peterson and Bruce S. Davie

Chapter 3
Internetworking

Copyright © 2010, Elsevier Inc. All rights Reserved 1

 Chapter 3
Internetworking
 What is internetwork
 An arbitrary collection of networks interconnected to provide
some sort of host-host to packet delivery service

A simple internetwork where H represents hosts and R represents routers

2
 Chapter 3
Internetworking
 What is IP
 IP stands for Internet Protocol
 Key tool used today to build scalable, heterogeneous
internetworks
 It runs on all the nodes in a collection of networks and defines the
infrastructure that allows these nodes and networks to function as
a single logical internetwork

A simple internetwork showing the protocol layers

3
 Chapter 3
IP Service Model
 Packet Delivery Model
 Connectionless model for data delivery
 Best-effort delivery (unreliable service)
 packets are lost
 packets are delivered out of order
 duplicate copies of a packet are delivered
 packets can be delayed for a long time
 Global Addressing Scheme
 Provides a way to identify all hosts in the network

4
 Chapter 3
Packet Format
 Version (4): currently 4
 Hlen (4): number of 32-bit words
in header
 TOS (8): type of service (not
widely used)
 Length (16): number of bytes in
this datagram
 Ident (16): used by fragmentation
 Flags/Offset (16): used by
fragmentation
 TTL (8): number of hops this
datagram has traveled
 Protocol (8): demux key (TCP=6,
UDP=17)
 Checksum (16): of the header
only
 DestAddr & SrcAddr (32)

5
 Chapter 3
IP Fragmentation and Reassembly
 Each network has some MTU (Maximum
Transmission Unit)
 Ethernet (1500 bytes), FDDI (4500 bytes), PPP (532
bytes)
 Strategy
 Fragmentation occurs in a router when it receives a
datagram that it wants to forward over a network
which has (MTU < datagram)
 Reassembly is done at the receiving host
 All the fragments carry the same identifier in the Ident
field
 Fragments are self-contained datagrams
 IP does not recover from missing fragments

6
 Chapter 3
IP Fragmentation and Reassembly

IP datagrams traversing the sequence of physical networks

7
 Chapter 3
IP Fragmentation and Reassembly

Header fields used in IP fragmentation. (a) Unfragmented packet; (b) fragmented packets.

8
 Chapter 3
Global Addresses
 Properties
 globally unique
 hierarchical: network + host
 4 Billion IP address, half are A type, ¼ is B type, and 1/8 is C type
 Format

 Dot notation
 10.3.2.4
 128.96.33.81
 192.12.69.77

9
 Chapter 3
IP Datagram Forwarding
 Strategy
 every datagram contains destination's address
 if directly connected to destination network, then forward to host
 if not directly connected to destination network, then forward to
some router
 forwarding table maps network number into next hop
 each host has a default router
 each router maintains a forwarding table
 Example (router R2)

10
 Chapter 3
IP Datagram Forwarding
 Algorithm
if (NetworkNum of destination = NetworkNum of one of my
interfaces) then
deliver packet to destination over that interface
else
if (NetworkNum of destination is in my forwarding table)
then
deliver packet to NextHop router
else
deliver packet to default router

For a host with only one interface and only a default router in its forwarding
table, this simplifies to

if (NetworkNum of destination = my NetworkNum)then

deliver packet to destination directly
else
deliver packet to default router

11
 Chapter 3
IP ADDRESSING
SUBNETTING & CIDR

12
 Chapter 3
Understanding IP Addresses
 An IP address is an address used in order to
uniquely identify a device on an IP network.
 The address is made up of 32 binary bits, which
can be divisible into a network portion and host
portion with the help of a subnet mask.
 The 32 binary bits are broken into four octets (1
octet = 8 bits).
 Each octet is converted to decimal and separated
by a period (dot). For this reason, an IP address is
said to be expressed in dotted decimal format (for
example, 172.16.81.100). The value in each octet
ranges from 0 to 255 decimal, or 00000000 -
11111111 binary.

13
 Chapter 3
Understanding IP Addresses
 Here is how binary octets convert to
decimal: The right most bit, or least
significant bit, of an octet holds a value of
20. The bit just to the left of that holds a
value of 21. This continues until the left-
most bit, or most significant bit, which
holds a value of 27. So if all binary bits are
a one, the decimal equivalent would be
255 as shown here:

14
 Chapter 3
Understanding IP Addresses
 1 1 1 11111
 128 64 32 16 8 4 2 1
(128+64+32+16+8+4+2+1=255)
 Here is a sample octet conversion when not all of
the bits are set to 1.
 0 1000001
 0 64 0 0 0 0 0 1 (0+64+0+0+0+0+0+1=65)
 And this is sample shows an IP address
represented in both binary and decimal.
 10. 1. 23. 19 (decimal)
 00001010.00000001.00010111.00010011 (binary)

15
 Chapter 3
Understanding IP Addresses
 These octets are broken down to provide
an addressing scheme that can
accommodate large and small networks.
There are five different classes of
networks, A to E. The focus is on
addressing classes A to C, since classes D
and E are reserved.

16
 Chapter 3
Understanding IP Addresses
 Given an IP address, its class can be
determined from the three high-order bits.

17
 Chapter 3
Understanding IP Addresses

18
 Chapter 3
Understanding IP Addresses
 In a Class A address, the first octet is the
network portion, so the Class A example in
figure has a major network address of
1.0.0.0 - 127.255.255.255. Octets 2, 3, and
4 (the next 24 bits) are for the network
manager to divide into subnets and hosts
as he/she sees fit. Class A addresses are
used for networks that have more than
65,536 hosts (actually, up to 16777214
hosts!).
19
 Chapter 3
Understanding IP Addresses
 In a Class B address, the first two octets
are the network portion, so the Class B
example in figure has a major network
address of 128.0.0.0 - 191.255.255.255.
Octets 3 and 4 (16 bits) are for local
subnets and hosts. Class B addresses are
used for networks that have between 256
and 65534 hosts.

20
 Chapter 3
Understanding IP Addresses
 In a Class C address, the first three octets
are the network portion. The Class C
example in figure has a major network
address of 192.0.0.0 - 233.255.255.255.
Octet 4 (8 bits) is for local subnets and
hosts - perfect for networks with less than
254 hosts.

21
 Chapter 3
Network Masks
 A network mask helps you know which
portion of the address identifies the
network and which portion of the address
identifies the node. Class A, B, and C
networks have default masks, also known
as natural masks, as shown here:
 Class A: 255.0.0.0
 Class B: 255.255.0.0
 Class C: 255.255.255.0

22
 Chapter 3
Network Masks
 An IP address on a Class A network that has
not been subnetted would have an
address/mask pair similar to: 8.20.15.1
255.0.0.0. To see how the mask helps you
identify the network and node parts of the
address, convert the address and mask to
binary numbers.
 8.20.15.1 =
00001000.00010100.00001111.00000001
 255.0.0.0 =
11111111.00000000.00000000.00000000

23
 Chapter 3
Network Masks
 Once you have the address and the mask
represented in binary, then identifying the
network and host ID is easier. Any address
bits which have corresponding mask bits
set to 1 represent the network ID. Any
address bits that have corresponding
mask bits set to 0 represent the node ID.

24
 Chapter 3
Network Masks

 8.20.15.1 =
00001000.00010100.00001111.00000001
 255.0.0.0 =
11111111.00000000.00000000.00000000
 -----------------------------------
 net id | host id
  
 netid = 00001000 = 8
 hostid = 00010100.00001111.00000001 = 20.15.1

25
 Chapter 3
Understanding Subnetting
 Subnetting allows you to create multiple
logical networks that exist within a single
Class A, B, or C network. If you do not
subnet, you are only able to use one
network from your Class A, B, or C
network, which is unrealistic.

26
 Chapter 3
Understanding Subnetting
 Each data link on a network must have a unique
network ID, with every node on that link being a
member of the same network.
 If you break a major network (Class A, B, or C) into
smaller subnetworks, it allows you to create a
network of interconnecting subnetworks.
 Each data link on this network would then have a
unique network/subnetwork ID.
 Any device, or gateway, connecting n
networks/subnetworks has n distinct IP addresses,
one for each network / subnetwork that it
interconnects.

27
 Chapter 3
Understanding Subnetting

 In order to subnet a network, extend the natural mask

using some of the bits from the host ID portion of the
address to create a subnetwork ID. For example, given a
Class C network of 204.17.5.0 which has a natural mask
of 255.255.255.0, you can create subnets in this manner:
204.17.5.0 - 11001100.00010001.00000101.00000000
255.255.255.224-1111111.11111111.11111111.11100000
 --------------------------|sub|----

28
 Chapter 3
Understanding Subnetting
 By extending the mask to be
255.255.255.224, you have taken three bits
(indicated by "sub") from the original host
portion of the address and used them to
make subnets.
 With these three bits, it is possible to create
eight subnets. With the remaining five host ID
bits, each subnet can have up to 32 host
addresses, 30 of which can actually be
assigned to a device.
 Since host ids of all zeros or all ones are not
allowed

29
 Chapter 3
Understanding Subnetting
 204.17.5.0 255.255.255.224 host address range 1 to 30
 204.17.5.32 255.255.255.224 host address range 33 to 62
 204.17.5.64 255.255.255.224 host address range 65 to 94
 204.17.5.96 255.255.255.224 host address range 97 to 126
 204.17.5.128 255.255.255.224 host address range 129 to
158
 204.17.5.160 255.255.255.224 host address range 161 to
190
 204.17.5.192 255.255.255.224 host address range 193 to
222
 204.17.5.224 255.255.255.224 host address range 225 to
254

30
 Chapter 3
Understanding Subnetting
 Note: There are two ways to denote these
masks. First, since you are using three bits
more than the "natural" Class C mask, you
can denote these addresses as having a 3-bit
subnet mask.
 Or, secondly, the mask of 255.255.255.224
can also be denoted as /27 as there are 27
bits that are set in the mask. This second
method is used with CIDR.
 One of these networks can be described with
the notation prefix/length. For example,
204.17.5.32/27 denotes the network
204.17.5.32 255.255.255.224.
31
 Chapter 3
Subnetting

32
 Chapter 3
Understanding Subnetting
 The network subnetting scheme in this
section allows for eight subnets, and the
network might appear as:

33
 Chapter 3
Understanding Subnetting
 Notice that each of the routers in figure is
attached to four subnetworks
 One subnetwork is common to both
routers.
 Also, each router has an IP address for
each subnetwork to which it is attached.
Each subnetwork could potentially support
up to 30 host addresses.

34
 Chapter 3
Understanding Subnetting
 This brings up an interesting point. The
more host bits you use for a subnet mask,
the more subnets you have available.
 However, the more subnets available, the
less host addresses available per subnet.
 For example, a Class C network of
204.17.5.0 and a mask of 255.255.255.224
(/27) allows you to have eight subnets,
each with 32 host addresses (30 of which
could be assigned to devices).

35
 Chapter 3
Understanding Subnetting

If you use a mask of 255.255.255.240 (/28), the break

down is:
204.17.5.0 - 11001100.00010001.00000101.00000000
255.255.255.240- 11111111.11111111.11111111.11110000
 --------------------------|sub |---
Since you now have four bits to make subnets with, you

only have four bits left for host addresses. So in this case
you can have up to 16 subnets, each of which can have up
to 16 host addresses (14 of which can be assigned to
devices).

36
 Chapter 3
Understanding Subnetting
 Take a look at how a Class B network
might be subnetted. If you have network
172.16.0.0 ,then you know that its natural
mask is 255.255.0.0 or 172.16.0.0/16.
 Extending the mask to anything beyond
255.255.0.0 means you are subnetting.
You can quickly see that you have the
ability to create a lot more subnets than
with the Class C network.

37
 Chapter 3
Understanding Subnetting
 If you use a mask of 255.255.248.0 (/21), how
many subnets and hosts per subnet does this
allow for?
 172.16.0.0 -
10101100.00010000.00000000.00000000
 255.255.248.0 -
11111111.11111111.11111000.00000000
 -----------------| sub |-----------
 You are using five bits from the original host bits
for subnets. This allows you to have 32 subnets
(25). After using the five bits for subnetting, you are
left with 11 bits for host addresses. This allows
each subnet so have 2048 host addresses (211),
2046 of which could be assigned to devices.
38
 Chapter 3
Exercise
 Given the Class C network of
204.15.5.0/24, subnet the network in order
to create the network in figure with the host
requirements shown.

39
 Chapter 3
Examples
 Looking at the network shown in figure,
you can see that you are required to create
five subnets. The largest subnet must
support 28 host addresses. Is this possible
with a Class C network? and if so, then
how?
 You can start by looking at the subnet
requirement. In order to create the five
needed subnets you would need to use
three bits from the Class C host bits. Two
bits would only allow you four subnets (22).
40
 Chapter 3
Examples
 Since you need three subnet bits, that
leaves you with five bits for the host
portion of the address. How many hosts
does this support? 25 = 32 (30 usable).
This meets the requirement.
 Therefore you have determined that it is
possible to create this network with a
Class C network. An example of how you
might assign the subnetworks is:

41
 Chapter 3
Examples

 netA: 204.15.5.0/27 host address range 1 to 30

 netB: 204.15.5.32/27 host address range 33 to
62
 netC: 204.15.5.64/27 host address range 65 to
94
 netD: 204.15.5.96/27 host address range 97 to
126
 netE: 204.15.5.128/27 host address range 129
to 158

42
 Chapter 3
VLSM Example
 In all of the previous examples of subnetting,
notice that the same subnet mask was applied for
all the subnets. This means that each subnet has
the same number of available host addresses.
 You can need this in some cases, but, in most
cases, having the same subnet mask for all
subnets ends up wasting address space.
 For example, in the above exercise, a class C
network was split into eight equal-size subnets;
however, each subnet did not utilize all available
host addresses, which results in wasted address
space. Figure illustrates this wasted address
space.
43
 Chapter 3
VLSM Example

44
 Chapter 3
VLSM Example
 Figure illustrates that of the subnets that are
being used, NetA, NetC, and NetD have a lot
of unused host address space.
 It is possible that this was a deliberate design
accounting for future growth, but in many
cases this is just wasted address space due
to the fact that the same subnet mask is
being used for all the subnets.
 Variable Length Subnet Masks (VLSM) allows
you to use different masks for each subnet,
thereby using address space efficiently.

45
 Chapter 3
VLSM Example
 Given the same network and requirements
as in Exercise develop a subnetting
scheme with the use of VLSM, given:
 netA: must support 14 hosts
 netB: must support 28 hosts
 netC: must support 2 hosts
 netD: must support 7 hosts
 netE: must support 28 host

46
 Chapter 3
VLSM Example
 Determine what mask allows the required number of hosts.
 netA: requires a /28 (255.255.255.240) mask to support 14
hosts
 netB: requires a /27 (255.255.255.224) mask to support 28
hosts
 netC: requires a /30 (255.255.255.252) mask to support 2
hosts
 netD*: requires a /28 (255.255.255.240) mask to support 7
hosts
 netE: requires a /27 (255.255.255.224) mask to support 28
hosts
  
 * a /29 (255.255.255.248) would only allow 6 usable host
addresses
 therefore netD requires a /28 mask.

47
 Chapter 3
VLSM Example
 The easiest way to assign the subnets is to assign
the largest first. For example, you can assign in
this manner:
 netB: 204.15.5.0/27 host address range 1 to 30
 netE: 204.15.5.32/27 host address range 33 to 62
 netA: 204.15.5.64/28 host address range 65 to 78
 netD: 204.15.5.80/28 host address range 81 to 94
 netC: 204.15.5.96/30 host address range 97 to 98
 This can be graphically represented as shown in
Figure 5:

48
 Chapter 3
VLSM Example

49
 Chapter 3
CIDR
 Classless Interdomain Routing (CIDR) was
introduced to improve both address space
utilization and routing scalability in the
Internet. It was needed because of the
rapid growth of the Internet and growth of
the IP routing tables held in the Internet
routers.

50
 Chapter 3
CIDR
 CIDR moves away from the traditional IP
classes (Class A, Class B, Class C, and so
on).
 In CIDR , an IP network is represented by
a prefix, which is an IP address and some
indication of the length of the mask.
 Length means the number of left-most
contiguous mask bits that are set to one.
 So network 172.16.0.0 255.255.0.0 can be
represented as 172.16.0.0/16.

51
 Chapter 3
CIDR
 CIDR also depicts a more hierarchical
Internet architecture, where each domain
takes its IP addresses from a higher level.
 This allows for the summarization of the
domains to be done at the higher level.

52
 Chapter 3
Classless Addressing
 Classless Inter-Domain Routing
 A technique that addresses two scaling concerns in
the Internet
 The growth of backbone routing table as more and more
network numbers need to be stored in them
 Potential exhaustion of the 32-bit address space
 Address assignment efficiency
 Arises because of the IP address structure with class A, B,
and C addresses
 Forces us to hand out network address space in fixed-size
chunks of three very different sizes
 A network with two hosts needs a class C address
 Address assignment efficiency = 2/255 = 0.0078=.78%
 A network with 256 hosts needs a class B address
 Address assignment efficiency = 256/65535 = 0.0039= .39%

53
 Chapter 3
Classless Addressing
 Exhaustion of IP address space centers on exhaustion of
the class B network numbers
 Solution
 Say “NO” to any Autonomous System (AS) that requests a class
B address unless they can show a need for something close to
64K addresses
 Instead give them an appropriate number of class C addresses

 What is the problem with this solution?

54
 Chapter 3
Classless Addressing
 Problem with this solution
 Excessive storage requirement at the routers.
 If a single AS has, say 16 class C network
numbers assigned to it,
 Every Internet backbone router needs 16 entries in its
routing tables for that AS
 This is true, even if the path to every one of these
networks is the same
 If we had assigned a class B address to the AS
 The same routing information can be stored in one
entry
 Efficiency = 16 × 255 / 65, 536 = 6.2%

55
 Chapter 3
Classless Addressing
 CIDR tries to balance the desire to minimize the
number of routes that a router needs to know
against the need to hand out addresses
efficiently.

 CIDR uses aggregate routes

 Uses a single entry in the forwarding table to tell the
router how to reach a lot of different networks
 Breaks the rigid boundaries between address classes

56
 Chapter 3
Classless Addressing
 Consider an AS with 16 class C network numbers.
 Instead of handing out 16 addresses at random, hand
out a block of contiguous class C addresses
 Suppose we assign the class C network numbers from
192.4.16 through 192.4.31
 Observe that top 20 bits of all the addresses in this range
are the same (11000000 00000100 0001)
 We have created a 20-bit network number (which is in between
class B network number and class C number)

57
 Chapter 3
Classless Addressing
 Requires to hand out blocks of class C addresses that
share a common prefix
 The convention is to place a /X after the prefix where X is
the prefix length in bits
 For example, the 20-bit prefix for all the networks
192.4.16 through 192.4.31 is represented as 192.4.16/20

 By contrast, if we wanted to represent a single class C

network number, which is 24 bits long, we would write it
192.4.16/24

58
 Chapter 3
Classless Addressing
 How do the routing protocols handle this
classless addresses
 It must understand that the network number may be of
any length
 Note that representing a network address
in this way is similar to the <mask, value>
approach used in subnetting, as long as
masks consist of contiguous bits starting
from the most significant bit.
 All routers must understand CIDR addressing

59
 Chapter 3
Classless Addressing

Route aggregation with CIDR

60
 Chapter 3
IP Forwarding Revisited
 IP forwarding mechanism assumes that it can
find the network number in a packet and then
look up that number in the forwarding table

 We need to change this assumption in case of

CIDR

 CIDR means that prefixes may be of any length,

from 2 to 32 bits

61
 Chapter 3
IP Forwarding Revisited
 It is also possible to have prefixes in the forwarding
tables that overlap
 Some addresses may match more than one prefix

 For example, we might find both 171.69 (a 16 bit prefix)

and 171.69.10 (a 24 bit prefix) in the forwarding table of
a single router

 A packet destined to 171.69.10.5 clearly matches both

prefixes.
 The rule is based on the principle of “longest match”
 171.69.10 in this case
 A packet destined to 171.69.20.5 would match 171.69
and not 171.69.10

62