Internet

 started in 1970 with 4 hosts, funded by the US Defence Department

 ARPANet → NSFNet → today’s Internet
 by the late 90s: US/non-US hosts split approximately 60/40
 remarkable sustained growth:
 annual growth rate for Internet hosts 16–118%, 1993–present (30% in 2006*)
 number of Internet hosts (Nov. 2006): 101,435,253*
 impossible to say how many Internet users there are!
 growth not just in size:
 everyone has http://www… on their products, ads, business cards
 .net and .com are the largest domains (together, 60% of all hosts)
 E-Commerce: annual growth rate approx. 190%; (£4 mln in UK every hour!^)
 Audio/video streaming: Voice over IP (VoIP), Video on Demand (VOD)
 specifications for multimedia communications
• Real-time Transport Protocol (RTP), Real-time Control Protocol (RTCP)
• Real Time Streaming Protocol (RTSP), Session Initiation Protocol (SIP)
 methods for supporting Quality of Service (QoS)
• Integrated Services (intserv), Differentiated Services (diffserv)
• Resource reSerVation Protocol (RSVP), MultiProtocol Label Switching (MPLS)
* Netcraft Web Server Survey - http://news.netcraft.com/archives/web_server_survey.html
^ E-commerce growth in UK - http://www.davechaffey.com/E-business/C1-Introduction/E-commerce-growth-rates-UK
Internet Web Sites - Evolution
Date Web sites
April 1997 1 million
February 2000 10 million
September 2000 20 million
July 2001 30 million
April 2003 40 million
May 2004 50 million
March 2005 60 million
August 2005 70 million
April 2006 80 million
August 2006 90 million
November 2006 100 million

- 3.5 million new sites last month only!

* Netcraft Web Server Survey - http://news.netcraft.com/archives/web_server_survey.html

Data Communications and Computer Networks

 Major features of the Internet
 Based on Internet Protocol (IP)
 uses datagram packet switching ⇒ connectionless

 open design, open implementations, open standardisation process

 independent of the physical medium

 scalable: as evidenced by its growth

 protocols have evolved over time, as problems arose and/or as application requirements
changed
 the desire to support QoS guarantees is adding connection-oriented concepts (such as
admission control and resource reservation) to the Internet service model

Flow of Information
SOURCE DESTINATION
Application Application
socket API TCP/UDP
NETWORK
TCP/UDP
socket API
ELEMENT
API = IP IP IP
Application H-to-N =
H-to-N H-to-N H-to-N
Programmin Host-to-
g Network
Interface
Data Communications and Computer Networks
IP offers a “best-effort” datagram service
 the Internet was originally intended for robust transfer of computer-to-
computer data over long distances
 best-effort service appropriate for data transfers
 no real-time requirements
 end-points can adapt to network conditions, if they want/need to
 connectionless packet-switching was preferred to circuit-switching
 no set-up delay
 no blocking (fire-and-forget, forward-if-possible, deliver-as-received)
however, there is no guarantee that any data reaches the destination
 flexibility in transmission bit-rates
in contrast with circuit-switching, which is usually tied to a few pre-determined
bit-rates
 no “path” ⇒ more reliable (route around problems)
 more efficient use of network resources when traffic is bursty
economics: transmission became more expensive than switching in late 1960s

Data Communications and Computer Networks

Internet Protocols

Datalink Frames
have a Trailer
also
(not shown here)

Data Communications and Computer Networks

Data Transmission in the OSI model

Data Communications and Computer Networks

 Data Transmission in the TCP/IP model

Internet Protocol
(IP)
Data Communications and Computer Networks
 Transmission Control Protocol/Internet Protocol (TCP/IP)

TCP/IP is industry standard transport protocol widely used in networking. It is

a standard protocol for Internet. TCP/IP is the default protocol used in
Windows 2000 and 2003 systems. TCP/IP is supported by variety of Operating
systems like MS-DOS, Windows, Unix, Linux, IBM mainframes, etc. Windows
TCP/IP suite includes various services like DHCP, DNS, WINS, etc.

Features of TCP/IP
1) Industry Standard Protocol.
2) Used by almost all operating systems
3) Internet standard protocol.
4) Used IP address for identifying host on network.
5) The current version is IPv4.
6) Robust and Routable protocol.
7) More Flexible and configuration then other protocol.
 8.3.1 Internet Protocol (IP) - IPv4 Header
all fragments of a Fragmentation of a datagram
given datagram is needed when an intermediate
have the same network has a max frame size
Identification value too small to carry the datagram

decreased by 1 padded to a multiple of 4 bytes (32 bits); can carry fields

at each “hop” that control routing, timing, management, security, ...
Data Communications and Computer Networks
 Internet Protocol Version 4 (IPv4)

IP Addressing
One of the most important topics in any discussion of TCP/IP is IP addressing. An IP address
is a numeric identifier assigned to each machine on an IP network. It designates the specific location of a
device on the network.
An IP address is a software address, not a hardware address. The latter is hard-coded on
a network interface card (NIC) and used for finding hosts on a local network. IP addressing was designed to
allow hosts on one network to communicate with a host on a different network regardless of the type of
LANs the hosts are participating in.
IP Terminology
In this section several important terms that are vital to your understanding of the Internet
Protocol. Here are a few to get you started:

Bit A bit is one digit, either a 1 or a 0.

Byte A byte is 7 or 8 bits, depending on whether parity is used. For this course, always
assume that a byte is 8 bits.
Octet An octet, made up of 8 bits, is just an ordinary 8-bit binary number. In this course,
the terms byte and octet are completely interchangeable.
IP Addressing IP address is a 32-bit binary value divided into 4 octels of 8-bit
each. i.e 00000000.00000000.00000000.00000000. The range of IP address in
binary is 00000000.00000000.00000000.00000000 to
11111111.11111111.11111111.11111111. This is the Binary notation

We can also represent IP address in Decimal format (Dotted decimal). If we

convert binary to decimal the IP address can range from 0.0.0.0 to
255.255.255.255.
IP address is used to identify a host on a network. Each host should have a
unique IP address.

Examples of IP address
10.12.45.54
192.168.10.5
208.144.0.2

An IP address consist of two parts, one is Network ID and other is Host ID.
Network ID should be same for all the PCs in a same network segment and Host
ID should be unique for each PC. If two PC is having different network ID then
they will not be able to communicate with each other directly. So network ID
should be same for all the PC's in same segment.
 Classful Addressing
IPv4 addressing, at its inception, used the concept of classes. This architecture is called
classful addressing. Although this scheme is becoming obsolete, we briefly discuss it here to
show the rationale behind classless addressing. In classful addressing, the address space is
divided into five classes: A, B, C, D, and E. Each class occupies some part ofthe address space

IP address classes IP classes are used to create or allocate IP addressing in a proper

fashion. There are 5 different classes available. They are as under

Class A = N.H.H.H (1-126)

Class B = N.N.H.H (128-191)
Class C = N.N.N.H (192-223)
Class D = Multicast (224-239)
Class E = Research (240-255)
The designers of the Internet decided to create classes of networks based on network size.
The figure below summarizes the three classes of network

To ensure efficient routing, Internet designers defined a mandate for the leading-bits
section of the address for each different network class. For example, since a router
knows that a Class A network address always starts with a 0, the router might be able
to speed a packet on its
way after reading only the first bit of its address. This is where the address schemes define
the difference between a Class A, a Class B, and a Class C address.

Network Address Range: Class A

The first bit of the first byte in a Class A network address must always be off, or 0.
Consider the following network address:
0xxxxxxx
If we turn the other 7 bits all off and then turn them all on, we’ll find the Class A range of
network addresses:
00000000 = 0
01111111 = 127
This means a Class A address must be between 0 and 127 in the first byte, inclusive.
Network 127.0.0.1 is reserved for loopback tests. Designates the local node and allows that
node to send a test packet to itself without generating network traffic.
Network Address Range: Class B
The first bit of the first byte must always be turned on but the second bit must always be
turned off. If you turn the other 6 bits all off and then all on, you will find the range for a
Class B network:
10000000 = 128
10111111 = 191
This means a Class A address must be between from 128 to 191 in the first byte, inclusive.
Network Address Range: Class C
The first 2 bits of the first octet are always turned on, but the third bit can never be on.
Following the same process as the previous classes, convert from binary to decimal to find
the range. Here’s the range for a Class C network:
11000000 = 192
11011111 = 223
So, if you see an IP address that starts at 192 and goes to 223, you’ll know it is a Class C
IP address.
Network Address Ranges: Classes D and E
The addresses between 224 to 255 are reserved for Class D and E networks. Class D (224–
239) is used for multicast addresses and Class E (240–255) for scientific purposes.
For this course we don’t need to know them.
There are two types of IPv4 addresses:
 Private IP address
 Public IP address

Public IP Addresses
These are IP addresses that are routable on the internet.

Private IP Addresses
These are IP addresses that can be used on a private network, but they’re not routable through
the internet.
This is designed for the purpose of creating a measure of well-needed security, but it also
conveniently saves valuable IP address space. If every host on every network had to have real
routable IP addresses, we would have run out of IP addresses. But by using private IP
addresses ISPs and home users only need a relatively tiny group of bona fide IP addresses to
connect their networks to the Internet. This is economical because they can use private IP
addresses on their inside networks and get along just fine.

To accomplish this task, the ISP and the corporation, the end user, no matter who they are
need to use something called Network Address Translation (NAT), which basically takes a private IP address
and converts it for use on the Internet.
The reserved private addresses are listed in table below.
Private IP address Private IP address are called as non-routable IP
because they are not used in Internet. These IP address range are
used for Private network so as to avoid conflicts in Internet.

Address Class Reserved Address Range

Class A 10.0.0.0 through 10.255.255.255
Class B 172.16.0.0 through 172.31.255.255
Class C 192.168.0.0 through 192.168.255.255

Types of Address
Unicast = When a packet to send to a single host eg. host 10.0.0.1
is sending a packet to 10.0.0.2

Multicast = When a packet is send to group of host using a

multicast IP address. eg. Host 10.0.0.1 is sending a packet to
224.0.0.5

Broadcast = When a Packet is send to all host in the network. eg.

Host 10.0.0.1 is sending a packet to 10.255.255.255 or
255.255.255.255
IPv4 Addresses
00000000 00000000 00000000 00000000 0.0.0.0
00000000 00000000 00000000 00000001 0.0.0.1
00000000 00000000 00000000 00000010 0.0.0.2
00000000 00000000 00000000 00000011 0.0.0.3

10001001 00101011 00000001 00110001 137.43.1.49

10001001 00101011 00000001 00110010 137.43.1.50
10001001 00101011 00000001 00110011 137.43.1.51

11111111 11111111 11111111 11111100 255.255.255.252

11111111 11111111 11111111 11111101 255.255.255.253
11111111 11111111 11111111 11111110 255.255.255.254
11111111 11111111 11111111 11111111 255.255.255.255

Data Communications and Computer Networks

IPv4 Address Format

Class Number of networks Number of hosts on each

of this Class network of this Class
all-0’s and all-1’s
A 126 224 − 2 (≈ 16 million) (in the Netid &/or
Hostid parts)
B 214 − 2 = 16,382 216 − 2 (≈ 65,000) have special
meanings
C 221 − 2 (≈ 2 million) 254

Data Communications and Computer Networks

Special IPv4 Addresses

All 0’s
This host

A host with this Hostid

Netid all 0’s Hostid
on the local network

Broadcast on the
All 1’s
local network

Broadcast on remote
Netid All 1’s
network specified by
Netid
refers to the network
Netid All 0’s
with specified Netid

127 (Anything) Loopback testing

IP packets with loopback addresses are processed by the machine which generated
them as if they were incoming packets (useful for debugging network software)

Data Communications and Computer Networks

Who assigns IPv4 Addresses?
IP address assignment managed by ICANN (Internet Corporation for
Assigned Names and Numbers)
ICANN is a non-profit corporation that was formed to assume responsibility
for the IP address space allocation, protocol parameter assignment, domain name
system management, and root server system management functions previously
performed under U.S. Government contract by IANA and other entities

ICANN assigns portions of the IP address space to other authorities.

Some of the authorities which assign addresses to ISPs:
APNIC: 61, 202, 203, 210, 211
Asia Pacific Network Information Centre
(www.apnic.net)

ARIN: 208, 209, 216

American Registry for Internet Numbers
(www.arin.net)
RIPE NCC: 193, 194, 195, 212, 62
Réseaux IP Européens Network Coordination Centre
(www.ripe.net)

EE210 Data Communications and Computer Networks

Names and IP addresses: the Domain Name System
There has to be a way to translate between “user-friendly” names and IP
addresses:
• “Name Server” is actually a hierarchy of servers End-user
called the Domain Name System (DNS) 2 1
cs.ariz [email protected]
ona.ed
• name-to-address mapping for one Name u Mail
or more root servers is “well-known” Server Program
and stored in the local name server’s
system configuration file
192.12.69.5 192.12.69.5 4
3
• hosts initialised with address of a local name server TCP
Example Hierarchy of Name Servers:
Root
name server 192.12.69.5 5
Arizona NASA
name server ..... name server IP

CS ECE if a name server cannot “resolve”

name server ..... name server the entire name, it returns an
answer for as much of it as it can
EE210 Data Communications and Computer Networks
 IP Router Architecture
IP packet arrives (in a datalink
frame):
1. line card applies datalink layer logic to ensure frame
I/O Line is
Port Card valid and packet successfully received;
s 2. validity check performed on IP header;
Line
Card 3. If destination address is a non-local host, routing table
Switching
lookup performed to determine how to forward the
Fabric packet;
Line 4. packets may have to be classified into predefined
Card service classes;

Line 5. TTL field decremented, new header checksum

Card computed, then packet sent to appropriate output
port;
6. datalink layer logic on output port’s line card inserts
Routing CPU datalink layer header and transmits the packet inside
a frame

[NOTE: some/all of steps 2-5 may be done in the input port,

not the routing CPU]

• if this process fails, ICMP error message sent to packet’s sender

• if packet arrival rate exceeds router’s forwarding capacity, packets must be buffered
• if buffers are full, packets must be discarded
EE210 Data Communications and Computer Networks
Forwarding IP packets
Example:

Problems forwarding IP packets ⇒ Internet Control Message Protocol (ICMP)

sends error message back to sending host. ICMP can also be used to tell hosts
about better routes, using ICMP–Redirect messages.
How does a sending host know the physical address corresponding to the IP address
of its intended destination? Answer: it uses the Address Resolution Protocol (ARP)

EE210 Data Communications and Computer Networks

ARP and Reverse ARP
 each host maintains an ARP cache of mappings between IP addresses and
physical addresses
 ARP cache entries time out after (typically) 15 minutes
 if destination’s IP address not in sending host’s ARP cache, sending host
broadcasts ARP query on local network
 ARP query asks for physical address for destination’s IP address
 ARP query includes sending host’s IP address and physical address (so
other hosts can enter this in their ARP caches)
 if destination is a host on local network, it sends its physical address to
sending host in an ARP reply
 if destination not on the local network, ARP server replies with its own
physical address: proxy ARP

 if a host doesn’t know its own IP address (e.g. diskless workstation being
booted), it uses Reverse ARP
 host broadcasts its physical address and asks for corresponding IP
address; RARP server replies

Data Communications and Computer Networks

Classless InterDomain Routing (CIDR)
 original class-based IP addressing scheme is inefficient
 by 1996, more than 50% of all Class B networks had < 50 hosts

 could give organisations which are large enough for a Class B address a
number of Class C addresses instead
 then every Internet backbone router would need an entry in its
routing table for each of these Class C network numbers
 CIDR: medium-large organisations get contiguous blocks of Class C
addresses (in powers of 2)
 example: company needs 8,000 IP addresses ⇒ gets 32 contiguous blocks of
Class C addresses ≡ 8,192 addresses
e.g. if given Class C network numbers 192.4.0 – 192.4.31, then first 19
bits of all addresses in company’s networks are the same (19-bit network
prefix):
 11000000 00000100 000 ⇒ first 19 bits of network mask are
1, rest
are 0
this company’s IP addresses are now written (e.g.) 192.4.19.252/19

 same idea is now applied to all addresses, not just Class C addresses
subnetting: share one network IP address among multiple physical networks
CIDR: collapse multiple adjacent IP addresses onto one network prefix
IP Subnetting
• IP address is composed of a Netid part and a Hostid part ⇒ 2-level hierarchy
• sometimes a 2-level hierarchy is insufficient for an organisation’s needs:

Class B network

• a solution:
subnetting

the 3 subnetworks still

appear to the rest of the
Internet as a single
network
IP Subnetting (cont.)
 an IP packet from some other network destined for host 141.14.2.21
still reaches router R1, since the destination address is still a Class B
address with Netid 141.14 and Hostid 2.21 as far as the rest of the
Internet is concerned
 when the packet reaches router R1, the interpretation of the IP address
changes
 R1 knows that there are 3 levels of
hierarchy within the organisation, and
that in this case, the Netid is 141.14,
the Subnetid is 2, and the Hostid is 21

with subnetting, the Netid defines the site,

the Subnetid defines the physical network,
and the Hostid defines the actual machine

 how is this knowledge of the internal network hierarchy implemented in

the organisation’s routers?
 masking of IP addresses during the packet-forwarding process
 masking is done whether or not subnetting is being used
Data Communications and Computer Networks
IP Subnetting: Masking
 masking means
 taking bit-by-bit AND
 of IP address and mask
 (255 ≡ all 1’s, 0 ≡ all 0’s)
extracts network address
from an IP address

extracts subnetwork address

from an IP address

 mask address fields don’t have to be 255 or 0

 e.g. could have a mask = 255.255.192.0
Data Communications and Computer Networks
Subnetting Basics
At the introduction of the course, we were
defining only one network. What happens
if we wanted to take one network address
and create six networks from it? We
would have to do something called
subnetting, because that’s what allows you
to take one larger network and break it
Reduced network traffic
into a bunch of smaller networks.
We all appreciate less traffic of any
There are loads of reasons in favor of subnetting, including the following
kind. Networks are no different.
benefits:
Without trusty routers, packet traffic
could grind the entire network down to
a near standstill. With routers, most
traffic will stay on the local network;
only packets destined for other
Optimizedwill
networks network performance
pass through the router.
Routers
This create
is a result broadcast
of reduced domains.
network traffic.
The more broadcast domains you
create,
Simplified themanagement
smaller the broadcast
domains and the less network traffic
It’s easiernetwork
on each to identify and isolate network problems in a group of smaller connected
segment.
networks than within one gigantic network.
Facilitated spanning of large geographical distances

Because WAN links are considerably slower and more expensive than LAN links, a single large network that spans long distances can
create problems in every area previously listed. Connecting multiple smaller networks makes the system more efficient.

How to Create Subnets

To create subnetworks, you take bits from the host portion of the IP address and reserve them to define the subnet address.
To create a subnet follow these steps:
When you’ve chosen a possible subnet mask for your network and need to determine the

number of subnets, valid hosts, and broadcast addresses of a subnet that the mask provides,

all you need to do is answer five simple questions:

 How many subnets does the chosen subnet mask produce?

 How many valid hosts per subnet are available?

 What are the valid subnets?

 What’s the broadcast address of each subnet?

 What are the valid hosts in each subnet?

Here’s how you get the answers to those five big questions:

 How many subnets? 2^x = number of subnets. x is the number of masked bits, or the 1s. For example, in 11000000, the
number of 1s gives us 2^2 subnets. In this example, there are 4 subnets.
 How many hosts per subnet? 2^y – 2 = number of hosts per subnet. y is the number of unmasked bits, or the 0s. For example, in
11000000, the number of 0s gives us 2^6 – 2 hosts. In this example, there are 62 hosts per subnet. You need to subtract 2 for the

subnet address and the broadcast address, which are not valid hosts.
 What are the valid subnets? 256 – subnet mask = block size, or increment number. An example would be 256 – 192 = 64. The
block size of a 192 mask is always 64. Start
 counting at zero in blocks of 64 until you reach the subnet mask value and these are your subnets. 0, 64, 128, 192.

What’s the broadcast address for each subnet? Now here’s the really easy part. Since we counted our subnets in
the last section
 as 0, 64, 128, and 192, the broadcast address is always the number right before the next subnet.

For example, the 0 subnet has a broadcast address of 63 because the next subnet is 64. The 64 subnet has a
broadcast address of 127 because the next subnet is 128. And so on. And remember, the broadcast address of the
last subnet is always 255.

What are the valid

 hosts? Valid hosts are the numbers between the subnets, omitting the all 0s and all 1s. For
example, if 64 is the subnet number and 127 is the broadcast address, then 65–126 is the valid host range. It’s
always the numbers between the subnet address and the broadcast address.
Subnet Masks

For the subnet address scheme to work, every machine on the network must know which part of the host address
will be used as the subnet address. This is accomplished by assigning a subnet mask to each machine. A subnet
mask is a 32-bit value that allows the recipient of IP packets to distinguish the network ID portion of the IP address
from the host ID portion of the IP address.

The network administrator creates a 32-bit subnet mask composed of 1s and 0s. The 1s in the subnet mask represent
the positions that refer to the network or subnet addresses. Not all networks need subnets, meaning they use the
default subnet mask. This is basically the same as saying that a network doesn’t have a subnet address.The table
shows the default subnet masks for Classes A, B, and C. These default masks cannot change.
 In other words, you can’t make a Class B subnet mask read 255.0.0.0. If you try, the host will read that address as invalid and usually
won’t even let you type it in. For a Class A network, you can’t change the first byte in a subnet mask; it must read 255.0.0.0 at a
minimum. Similarly, you cannot assign 255.255.255.255, as this is all 1s, a broadcast address. A Class B address must start with
255.255.0.0, and a Class C has to start with 255.255.255.0.

Subnetting Class C Addresses

In a Class C address, only 8 bits are available for defining the hosts. Remember that subnet bits start at the left and go to the right,
without skipping bits. This means that the only Class C subnet masks can be the following:

We can’t use a /31 or /32 because we have to have at least 2 host bits for assigning IP addresses to hosts.
Example #1: 255.255.255.128 (/25)
Since 128 is 10000000 in binary, there is only 1 bit for subnetting and 7 bits for hosts. We’re going to subnet the Class C network
address 192.168.10.0.
192.168.10.0 = Network address
255.255.255.128
 = Subnet mask
Now, let’s answer the big five:


How many subnets? Since 128 is 1 bit on (10000000), the answer would be 21 = 2.

How many hosts per subnet? We have 7 host bits off (10000000), so the equation would be 27 – 2 = 126 hosts.


What are the valid subnets? 256 – 128 = 128. Remember, we’ll start at zero and count in our block size, so our subnets are 0, 128.

What’s the broadcast address for each subnet? The number right before the value ofthe next subnet is all host bits turned on and

equals the broadcast address. For the zero subnet, the next subnet is 128, so the broadcast of the 0 subnet is 127.

What are the valid hosts? These are the numbers between the subnet and broadcast address. The easiest way to find the hosts is
to write out the subnet address and the broadcast address. This way, the valid hosts are obvious. The following table shows the
0 and 128 subnets, the valid host ranges of each, and the broadcast address of both subnets:
Example #2: 255.255.255.192 (/26)
In this second example, we’re going to subnet the network address 192.168.10.0 using the
subnet mask 255.255.255.192.
192.168.10.0 = Network address
255.255.255.192 = Subnet mask
Now, let’s answer the big five:
 How many subnets? Since 192 is 2 bits on (11000000), the answer would be 22 = 4
subnets.
 How many hosts per subnet? We have 6 host bits off (11000000), so the equation
would be 26 – 2 = 62 hosts.
 What are the valid subnets? 256 – 192 = 64. Remember, we start at zero and count in
our block size, so our subnets are 0, 64, 128, and 192.
 What’s the broadcast address for each subnet? The number right before the value of
the next subnet is all host bits turned on and equals the broadcast address. For the zero subnet, the next
subnet is 64, so the broadcast address for the zero subnet is 63.

What are the valid hosts? These are the numbers between the subnet and broadcast
address. The easiest way to find the hosts is to write out the subnet address and the
broadcast address. This way, the valid hosts are obvious. The following table shows
the 0, 64, 128, and 192 subnets, the valid host ranges of each, and the broadcast
address of each subnet:
 Subnetting Class B Addresses
With Class B addresses, we have a lot more possible subnet masks than we do with a Class C
network address:

We know the Class B network address has 16 bits available for host addressing. This means we can use up to 14 bits for
subnetting (because we have to leave at least 2 bits for host addressing). Using a /16 means you are not subnetting with
class B, but it is a mask you can use.
This section will give you an opportunity to practice subnetting Class B addresses. Again, I
have to mention that this is the same as subnetting with Class C, except we start in the third octet with the exact same
numbers.
Example #1: 255.255.128.0 (/17)
172.16.0.0 = Network address
255.255.128.0 = Subnet mask
 Subnets? 21 = 2 (same as Class C).
 Hosts? 215 – 2 = 32,766 (7 bits in the third octet, and 8 in the fourth).

 Valid subnets? 256 – 128 = 128. 0, 128. Remember that subnetting is performed in
the third octet, so the subnet numbers are really 0.0 and 128.0, as shown in the next
table. These are the exact numbers we used with Class C; we use them in the third
octet and add a 0 in the fourth octet for the network address.
 Broadcast address for each subnet?

 Valid hosts?
The following table shows the two subnets available, the valid host range, and the broadcast
address of each:

Example #2: 255.255.192.0 (/18)

172.16.0.0 = Network address
255.255.192.0 = Subnet mask
 Subnets? 22 = 4.
 Hosts? 214 – 2 = 16,382 (6 bits in the third octet, and 8 in the fourth).
 Valid subnets? 256 – 192 = 64. 0, 64, 128, 192. Remember that the subnetting is
performed in the third octet, so the subnet numbers are really 0.0, 64.0, 128.0, and
192.0, as shown in the next table.
 Broadcast address for each subnet?

 Valid hosts?
The following table shows the four subnets available, the valid host range, and the broadcast
address of each:

Subnetting Class A Addresses

Class A subnetting is not performed any differently than Classes B and C, but there are 24 bits to play with
instead of the 16 in a Class B address and the 8 in a Class C address.
Let’s start by listing all the Class A masks:

Example #1: 255.255.0.0 (/16)

Class A addresses use a default mask of 255.0.0.0, which leaves 22 bits for subnetting since
you must leave 2 bits for host addressing. The 255.255.0.0 mask with a Class A address is
using 8 subnet bits.
 Subnets? 28 = 256.
 Hosts? 216 – 2 = 65,534.
 Valid subnets? What is the interesting octet? 256 – 255 = 1. 0, 1, 2, 3, etc. (all in the
second octet). The subnets would be 10.0.0.0, 10.1.0.0, 10.2.0.0, 10.3.0.0, etc., up to
10.255.0.0.
 Broadcast address for each subnet?
 Valid hosts?
The following table shows the first two and last two subnets, valid host range, and broadcast
addresses for the private Class A 10.0.0.0 network:
Example #2: 255.255.240.0 (/20)

255.255.240.0 gives us 12 bits of subnetting and leaves us 12 bits for host addressing.
 Subnets? 212 = 4096.
 Hosts? 212 – 2 = 4094.
 Valid subnets? What is your interesting octet? 256 – 240 = 16. The subnets in the
second octet are a block size of 1 and the subnets in the third octet are 0, 16, 32, etc.
 Broadcast address for each subnet?
 Valid hosts?

The following table shows some examples of the host ranges the first three and the last subnets:
8.3.2 IPv6: Next Generation IP
 in early 90s, IETF held an open design process to design
the next version of IP: IPv6

 motivated by problems with IPv4 (the current version of IP)

 still have limited address space, even with subnetting & CIDR
 didn’t handle real-time “flows”: routers ignored the IPv4 Type Of
Service field
 at the time, IPv4 security features were not widely used

 IPv6 is not compatible with IPv4, but it is compatible with

TCP/UDP/OSPF/BGP/DNS…

 IPv4 and IPv6 will co-exist for a (possibly long) time

 IPv6 implementations exist, but not many products so
far
Data Communications and Computer Networks
IPv6: Next Generation IP
 major features of IPv6:
 address space problems taken care of by using 128-bit
addresses
 there are a maximum 2128 IPv6 addresses (approximately 3.4×1038)
of
 even the most pessimistic expectations of IPv6 addressing
efficiency predict 16,000 IPv6 addresses per square metre of the
earth’s surface…
 simplified packet header functions
 fragmentation at the source only; no header checksum
 better support for options
 e.g. hop-by-hop options, routing options, fragmentation options
 extension headers are present when corresponding options used
 support for per-flow handling and traffic classes
 flow designated by source & destination addresses and flow number
 support for authentication and security

Data Communications and Computer Networks

Datalink Layer in Local Area Networks

Most LANs use broadcast technology (i.e. multiple

nodes share a channel). It must be determined which
node can use the shared channel at any given time

• when two or more nodes want to transmit at the same time,

they are said to contend for the channel

• if two or more nodes transmit at the same time, usually

assumption is that all the transmissions collide ⇒ no
transmission is successful

Data Communications and Computer Networks

Overview: Issues for Broadcast LANs
 Overview of Local Area Networks?
 Basic topologies

 Broadcast Networks
 MAC and LLC
Sublayers
 The 802.X
standard
Ethernet
Token Ring
Token Bus
FDDI
Wireless LANs

Data Communications and Computer Networks

LAN Overview: What Are LANs?

 LANs are used to provide inexpensive and fast

interconnection of distributed communities of DTEs
(minicomputers, PCs, printers, telephones)

 Usually localised within a single or a group of buildings

 Since LANs are typically owned and maintained by one

organisation they are also referred to as Private Data
Networks
 Because of the relatively short distances involved LANs can
provide very high data transmission rates, even up to
hundreds of Mbps, between networked DTEs

Data Communications and Computer Networks

LAN Overview: Combined Topologies

 In large computer networks, such as in an University,

often a number of LAN topology types co-exist

Gateway INTERNET
School of Nursing

Backbone Network

VAX

School of Electronic Eng.

Data Communications and Computer Networks

Datalink Layer: Broadcast networks - Contention resolution
• 3 basic strategies:
• divide the channel into independent sub-channels, one for
each
transmission (e.g. using TDM or FDM) – turns the channel into a set of point-
to-point links
• collision resolution: allow nodes to start transmitting whenever they like,
but nodes check for collisions; if a collision is detected, each node involved
waits a random amount of time (why?), and starts transmitting again
• improvement – active nodes listen before/while transmitting; if
collision detected, stop transmitting, then wait random amount of time…
• reservations: a node must obtain the token (giving permission to transmit)
before transmitting; when finished, node releases the token to its neighbour
• modification – token doesn’t have to be passed from neighbour to
neighbour, as long as all nodes get (maybe unequal) chances to transmit

Data Communications and Computer Networks

Datalink layer: Broadcast networks – LLC and MAC

• HDLC not suitable for broadcast (or multi-access) channels

• in 1985, the IEEE started Project 802 to define standards for broadcast
LANs to enable communications between equipment from different vendors
• LLC: Logical Link Control sublayer (non-architecture-specific, common
to all broadcast LANs; concerned with logical addresses, control
information, and data). Standardised by 802.2 protocol.
• MAC: Medium Access Control sublayer
• MAC protocol is specific to the LAN technology being used
• e.g. 802.3 (CSMA/CD), 802.4 (Token Bus), 802.5 (Token Ring), …
• resolves contention for the channel - specifies flags, error control
schemes, physical addresses
802.2 Logical Link Control LLC
Layer 2
802.3 802.4 802.5 802.6 802.7 802.8 802.9 MAC
csma-cd tok. bus tok. ring DQDB broadb. fiber integr. Phy. Layer 1

Data Communications and Computer Networks

Datalink layer: Broadcast networks – LLC and MAC
• data unit at LLC sublayer is called a PDU (Protocol Data Unit)

logical addresses

PDU control field

identical to HDLC
control field

• PDU has no flag fields, no CRC/checksum, and no physical (machine)

addresses - these are supplied by the MAC sublayer

Data Communications and Computer Networks

Datalink layer: Broadcast networks – New Standards
• Recently, three more standards have been added:

•802.11 is the Wireless LAN standard

•802.11a (up to 54 Mbps in the 5 GHz band)
•802.11b (up to 11 Mbps in the 2.4 GHz band)
•802.11g (up to 54 Mbps in the 2.4 GHz
band)
•…
•802.16 is the a Wireless Point-to-multipoint standard

•802.15 is a Personal Area Network. This is similar to Bluetooth

Datalink layer: Ethernet
• mid-70s: Xerox PARC developed first Ethernet to connect 100 computers on
a 1 km cable; subsequently developed by Xerox, DEC, and Intel
• Ethernet is slightly different to IEEE 802.3
• channel access method: CSMA/CD (Carrier Sense Multiple Access with
Collision Detection) - try to reduce the likelihood and effects of a
collision
• CSMA: a node wishing to transmit must first listen to the channel (e.g. by
measuring the channel’s voltage level). If the channel is busy, some other
node is transmitting and our node must wait until it detects that the channel
is idle. When the channel is determined to be idle, our node can transmit.
• CD: during transmission, our node listens to the channel and if another
transmission is detected (e.g. higher voltage level than expected for one
transmission), all nodes involved in the collision stop transmitting
immediately. Each node then computes a randomly-sized time interval,
waits for that amount of time, and begins the transmission attempt again.
• basic problem with CSMA/CD is that, theoretically, a node wishing to
transmit may never be able to (Maximum Medium Access Time-MMAT =
∞): even with random waiting times, the node’s transmission attempts may
Datalink layer: collisions in CSMA/CD / Ethernet
• even with Carrier Sensing before transmission, collisions can still occur
because it takes non-zero time for signal to propagate along the channel
• if the propagation time along the length of the channel is denoted by τ,
can show that the worst-case collision detection time is (approximately)
2τ

noise burst generated by node B to warn all other nodes about collision
• this shows why there is a minimum frame length in CSMA/CD systems: all
frames must take more than 2τ to transmit (“pad” information field, if
necessary)
Data Communications and Computer Networks
 Datalink layer: CSMA/CD / Ethernet Frame

physical addresses 802.2 PDU

Preamble is to allow Receiver to synchronise with incoming transmission.

SFD is the start flag in Ethernet.
Node physical address: 6 bytes, encoded on node’s Network Interface
Card (NIC). Each NIC ever made has a unique 6-byte address.

Maximum length of Data field = 1500 bytes.

Note: There is no provision for ACK/NAK
This must be implemented at a higher layer
⇒ Ethernet is an unreliable medium
 Datalink layer: Ethernet Implementations (1)
10Base5: 10 Mbps, Baseband (digital encoding),
maximum channel length = 500 metres
(also called “Thick Ethernet”)

Data Communications and Computer Networks

 Datalink layer: Ethernet Implementations (2)
10Base2: 10 Mbps, Baseband (digital encoding),
maximum channel length = 185 metres
(also called “Thin Ethernet” or “cheapnet”)

transceiver circuitry moved into NIC; no need for AUI cables

Data Communications and Computer Networks
 Datalink layer: Ethernet Implementations (3)
10BaseT: 10 Mbps, Baseband (digital encoding), twisted pair
wiring, maximum length hub-to-node = 100 metres

logically still a bus topology, although physical topology is a star

Data Communications and Computer Networks
 Datalink layer: Ethernet Implementations (4)
Switched Ethernet: replace hub in 10BaseT with a switch

line
card

typically: 4-32 line cards

connected by a high-
speed (e.g. 1 Gbps)
backplane inside the
switch
10BaseT: when a node sends a Frame to the hub, it is sent out on all
other hub interfaces ⇒ only one node can transmit at a time
Switched Ethernet: when a node sends a Frame to the switch, it is sent out only
on the destination node interface ⇒ switch can forward more Frames at a
time Data Communications and Computer Networks
 Datalink layer: Ethernet Implementations (5)
Fast Ethernet: increase channel bandwidth to 100 Mbps

to achieve this 10-fold speedup,

the maximum channel length
must be reduced by a factor of
10:
2500 metres (with repeaters)
reduced to 250 metres

Gigabit Ethernet: increase channel bandwidth to 1000 Mbps = 1 Gbps

• mainly intended for use with optical fibres
• maximum channel length reduced to 25 metres
• usually for backbone connection between Fast Ethernet
networks
Data Communications and Computer Networks
Datalink layer: CSMA/CD Advantages & Disadvantages
+ low average access time under light loading conditions
+ easy to move/upgrade/change nodes, due to distributed channel access protocol
+ graceful performance degradation as number of nodes increases
+ the most popular LAN technology, with a huge installed base
– substantial analogue component: collision detection circuitry
– minimum valid frame length is inefficient if data is short (e.g. 1
– character) priorities not supported ⇒ difficult to support (e.g.) real-time
– traffic channel access time unpredictable
– as load increases: access time increases, and is theoretically unbounded
• the potential problems with Ethernet led to the development of some alternative
technologies in the early 80s
• IBM chose a ring topology for office automation applications – led to the
Token Ring (802.5) standard
• General Motors, and others interested in factory automation, chose a
bus
topology as a good match to layout of assembly lines – led to the Token Bus
(802.4) standard
• key point: the Maximum Medium Access Time (MMAT) is bounded, assuming
the network is working correctly
Data Communications and Computer Networks
Datalink layer: Fibre Distributed Data Interface (FDDI)
• standardised by ANSI and ITU-T, originally as a high-speed alternative
to
Ethernet and Token Ring LANs since FDDI runs at 100 Mbps
• copper-wire
• access method is version now available:
token-passing, CDDI
limited by time
• supports 2 kinds of traffic: real-time and non-real-
time
• now• real-time traffic
mostly used as asent first when
backbone token received
to connect other LANs of various types
can be up to 200 km,
with 1000 nodes

a device which can connect

2 different networks at the
Datalink layer
Data Communications and Computer Networks
 Datalink layer: FDDI (cont.)
• FDDI implemented as a dual ring
• usually, all transmissions on primary ring; other (secondary) ring used
only for backup when some part of the primary ring fails

difference from Token

Ring: in FDDI, node
releases token
immediately after
Frame transmission
normal ring
operatio failur
n e
• Frame format similar
to IEEE 802.5:

number of bytes
Data Communications and Computer Networks
Data Link Layer: Wireless LANs

 Wireless LAN’s are relatively recent

 Advantages
Relatively little infrastructure is required
Users can move around (laptops)

 Disadvantages
Equipment still quite expensive
Security
Errors will be more frequent
Lower Bit Rates

Data Communications and Computer Networks

Data Link Layer: Wireless LANs
 What MAC is appropriate for Wireless LAN
 Try CSMA
 Interference at the Sender is not the same as the receiver
 Difficult to listen and transmit
 C cannot hear A, but will ruin A->B Frame

A B C D

 Beware ‘Hidden’ and ‘Exposed’ Station Problems

 Hidden Station:
A is transmitting to B
C senses medium, and concludes it can transmit
This wipes out frame at A
 Exposed Station
B is transmitting to A
C cannot transmit as medium is busy
But in reality, it could transmit to D
Data Communications and Computer Networks
Data Link Layer: Wireless LANs

 Multiple Access with Collision Avoidance

 Works with a CTS, RTS scheme

Range of A

RTS
C A CTS B D

E
Range of B

Data Communications and Computer Networks

Data Link Layer: Wireless LANs

 A wishes to transmit to B
 C, B and E hear the RTS from A
 A, E and D hear the CTS from
B
 A starts to transmit
 D knows not to, even though it cannot hear
A
 Actually, C can transmit, as B will not hear it. Must only wait for
CTS
 Collisions can occur between RTS frames, but not as
serious
 MACAW (MACA for Wireless adds
Acknowledgements)
 CSMA is used to reduce RTS collisions

Data Communications and Computer Networks

Data Link Layer: Wireless LANs

Upper Layers

Logical Link Control

MAC
sublay
er
802.11 802.11 802.11 802.11a 802.11b 802.11g
Infrared FHSS DSSS OFDM HR-DSSS OFDM

Data Communications and Computer Networks

3.4.1 Guided Media: Two-Wire Open Lines

Signal Wire
Reference Wire

 Used mainly for directly connecting devices

 over small distances (< 50 m) and
 at moderate bit rates (< 19.2 bps)

 Signals get distorted due to:

 Crosstalk between the two signals
 Susceptibility to Noise Signals from other
(external)
electrical sources
Data Communications and Computer Networks
Guided Media: Twisted Pair Lines

 The proximity of signal and reference lines means noise

signals are picked up by both wires
 Shielded twisted pairs offer better noise immunity
 skin effect, which increases attenuation as the bit rate of
the transmitted signal increases.

 Offer higher bit rates

 1.5 Mbps up to 5 km
 51 Mbps less than 300
m

 E.g. high speed data to home (ADSL) and LANs

Data Communications and Computer Networks

Guided media: Coaxial Cable

 This is made up of an inner conductor surrounded by an

insulator and that surrounded again by the outer
conductor
 Use of the dielectric material and outer conductor
effectively isolate the core conductor from external noise
interference.

 Coaxial cables can be used at rates from over 10Mbps,

over distances of several hundred meters.

 E.g. Cable TV and Cable Modem

Data Communications and Computer Networks

Guided Media: Optical Fibre

 Optic fibre does not use electrical signals to transmit the

data, rather it uses light. It transmits these signals
through thin glass fibres.
 Light beams are also immune to electromagnetic
interference and crosstalk

 Currently up to 40 Gbps with repeaters every 50 km.

Theoretically, up to 50 Tbps = 50,000 Gbps.

 E.g. access and backbone networks, very fast LANs

Data Communications and Computer Networks

3.4.2 Wireless: Terrestrial Microwave/Radio
 Provides omnidirectional or unidirectional signalling
depending on transmitter and antenna.
 Suffer from factors such as bad weather conditions
and obstruction by man-made objects.

 Use microwaves for large distances, radio waves for

shorter distances. Limited by the curvature of the Earth.

 E.g. AM and FM radio, TV, Cellular telephony, wireless

LANs

Data Communications and Computer Networks

3.4.3 Satellite
 The first satellites were launched in the 1970’s and now the most
common are the Geostationary ones at 36,000 km above the Earth.
 Satellites are using direct line of sight between the transmitters
and receivers.
 Data transmitted using electromagnetic (radio) waves propagating
through the atmosphere.
 Typically many signals will be multiplexed onto a single satellite
channel utilising a high bit rate.
 Applications:
 Areas with little wired infrastructure
 Mobile communication
 Broadcast communication
 Rapid deployment (military)

Data Communications and Computer Networks